diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000000000000000000000000000000000000..53b7f797bcca64ce5e42272fa234cf9e8a410f3f --- /dev/null +++ b/.gitattributes @@ -0,0 +1,73 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.avro filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.lz4 filter=lfs diff=lfs merge=lfs -text +*.mds filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +# Audio files - uncompressed +*.pcm filter=lfs diff=lfs merge=lfs -text +*.sam filter=lfs diff=lfs merge=lfs -text +*.raw filter=lfs diff=lfs merge=lfs -text +# Audio files - compressed +*.aac filter=lfs diff=lfs merge=lfs -text +*.flac filter=lfs diff=lfs merge=lfs -text +*.mp3 filter=lfs diff=lfs merge=lfs -text +*.ogg filter=lfs diff=lfs merge=lfs -text +*.wav filter=lfs diff=lfs merge=lfs -text +# Image files - uncompressed +*.bmp filter=lfs diff=lfs merge=lfs -text +*.gif filter=lfs diff=lfs merge=lfs -text +*.png filter=lfs diff=lfs merge=lfs -text +*.tiff filter=lfs diff=lfs merge=lfs -text +# Image files - compressed +*.jpg filter=lfs diff=lfs merge=lfs -text +*.jpeg filter=lfs diff=lfs merge=lfs -text +*.webp filter=lfs diff=lfs merge=lfs -text +# Video files - compressed +*.mp4 filter=lfs diff=lfs merge=lfs -text +*.webm filter=lfs diff=lfs merge=lfs -text +*.tsv.gz filter=lfs diff=lfs merge=lfs -text +*.idx filter=lfs diff=lfs merge=lfs -text + +# Binary/large deposit formats that the HuggingFace default set above does not cover. Without these +# a future .xlsx supplement or an uncompressed .fasta lands as a PLAIN BLOB in a dataset repo -- +# the mirror-image of the bug that once let a shipped .fasta escape a *.fasta ignore rule by its +# extension. Nothing tracked today is affected: 0 .xlsx and 0 bare .fasta are tracked. +# *.tsv and *.csv are deliberately NOT here: five small tables are tracked as plain blobs on +# purpose (context_synonyms.tsv, toil_floors{,_mmu}.tsv, harmonized/manifest.tsv, summary.tsv), +# and an LFS pattern would convert a 402-byte manifest for nothing. Compressed *.tsv.gz is already +# covered above. +*.xlsx filter=lfs diff=lfs merge=lfs -text +*.fasta filter=lfs diff=lfs merge=lfs -text diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000000000000000000000000000000000000..6b797664f735fdd7571ec4015eb160b5932a9473 --- /dev/null +++ b/.gitignore @@ -0,0 +1,51 @@ +dump/ +src/ +.claude/ +.DS_Store +**/.DS_Store + +# raw downloads (never pushed) +raw/ + +# MixMHCpred3 comparison material (never pushed). Two of the four files are publisher +# supplementary .xlsx — copyrighted, not ours to redistribute — and the directory's only +# provenance is three bare URLs in benchmark/mixmhcpred3_ref.md, which does not meet the +# SOURCES.md standard the rest of this repo now holds. Kept on disk, off the Hub. +benchmark/ + +private/ +# NB: tcga/ is PUBLIC (neoantigen predictions derived from open-access TCGA somatic + +# expression data; per-barcode HLA/neoantigens as published by TCIA / Thorsson 2018). + +# internal working table, not published +summary_all_datasets.tsv + +# whole-proteome window indexes: derived, rebuilt on demand by mhcmatch, never published. +# Eight (species, length) cells were 17 GB -- 41% of this dataset -- against a 55 MB FASTA that +# reproduces any one of them in seconds, and fetching one measured SLOWER than building it +# (mouse L=9: 1.49 GB in 47.0 s unauthenticated, against 25.0 s to build). Withdrawn 2026-09-06. +proteome_index/ + +# ICB cohort assembly, staged 2026-09-10. Our harmonised tables under icb_cohorts/harmonized/ +# are published (typing.parquet and manifest.tsv -- our own six-allele genotypes over public +# typings, and a source->count manifest). The publisher supplementary workbooks under +# icb_cohorts/raw/ are copyrighted and not ours to redistribute; they stay ignored permanently. +icb_cohorts/raw/ + +# NB: hla_loh/ is PUBLIC and tracked from 2026-09-16. It holds per-allele HLA-I loss-of- +# heterozygosity over TCGA/PCAWG and Hartwig, derived from open-access data exactly as tcga/ is: +# ICGC donor ids and TCGA barcodes are public pseudonymous research identifiers, and PCAWG's +# `donor_wgs_exclusion_white_gray` is its published QC whitelist flag. An earlier rule here called +# the directory a re-identification risk; that was wrong and is withdrawn. What genuinely may not +# ship is the three copyrighted Springer Nature workbooks, so they -- and the third-party PCAWG +# sample sheet -- live in `raw/hla_loh/` under the `raw/` rule above, and `hla_loh/harmonize.py` +# rebuilds the published parquets from them. Facts are not copyrightable; a publisher's workbook is. + +# The mimicry k-mer cache. mhcmatch WRITES this itself (`mimicry.py` under +# $MHCMATCH_CALIBRATION_CACHE), keyed by a blake2b of the schema epoch, class, species, lengths and +# each deposit's name and size -- so it is derived data that any consumer regenerates on demand, and +# `docs/cli/environment.rst` says in as many words that it is safe to delete. 2.07 GB across 50 +# files (two cache cells, human and mouse), tracked and pushed with no SOURCES.md entry. Same call as the 17 GB `proteome_index/` withdrawal in 2f76f91, for the same +# reason. NOTE: this rule stops it being re-added; the 50 files already tracked need a +# `git rm --cached`, which is a push to the Hub and so is the author's to make. +mimicry_refs/ diff --git a/README.md b/README.md new file mode 100644 index 0000000000000000000000000000000000000000..2d922a074e9cd758d62791dd83d548c5dc5c786e --- /dev/null +++ b/README.md @@ -0,0 +1,120 @@ +--- +pretty_name: pMHC Data Compendium +license: other +license_name: mixed-upstream +license_link: https://huggingface.co/datasets/isalgo/pmhc_data#licensing +language: + - en +tags: + - immunology + - immunopeptidomics + - neoantigen + - mhc + - hla + - t-cell + - cancer-vaccine + - proteomics +size_categories: + - 10M Vivian J, Rao AA, Nothaft FA, Ketchum C, Armstrong J, Novak A, Pfeil J, Narkizian J, Deran AD, +> Musselman-Brown A, Schmidt H, Amstutz P, Craft B, Goldman M, Rosenbloom K, Cline M, O'Connor B, +> Hanna M, Birger C, Kent WJ, Patterson DA, Joseph AD, Zhu J, Zaranek S, Getz G, Haussler D, +> Paten B. **Toil enables reproducible, open source, big biomedical data analyses.** +> *Nat Biotechnol* 2017;35(4):314-316. +> PMID [28398314](https://pubmed.ncbi.nlm.nih.gov/28398314/) · +> PMC5546205 · doi:[10.1038/nbt.3772](https://doi.org/10.1038/nbt.3772) + +> Goldman MJ, Craft B, Hastie M, Repecka K, McDade F, Kamath A, Banerjee A, Luo Y, Rogers D, +> Brooks AN, Zhu J, Haussler D. **Visualizing and interpreting cancer genomics data via the Xena +> platform.** *Nat Biotechnol* 2020;38(6):675-678. +> PMID [32444850](https://pubmed.ncbi.nlm.nih.gov/32444850/) · +> PMC7386072 · doi:[10.1038/s41587-020-0546-8](https://doi.org/10.1038/s41587-020-0546-8) + +**The TCGA study code is joined through the barcode, never through the disease name.** Xena labels a +TCGA sample with a disease name (`Uveal Melanoma`); the rest of this compendium keys on the study +abbreviation (`UVM`). A hand-written name->code crosswalk can be silently wrong, so the code is taken +from PanCanAtlas `merged_sample_quality_annotations.tsv` on the 15-character sample barcode, which is +exact. **17,474 of 19,131 matrix columns are used**: 9,616 TCGA tumour aliquots and 7,862 GTEx +samples. TARGET (734 columns) is dropped -- paediatric, and none of its diseases has a matched normal +here -- as are TCGA adjacent normals and aliquots the source flags `Do_not_use`. + +**Upstream values are `log2(TPM + 0.001)` and are inverted on build**, so what is stored is TPM. + +**Caveats.** +- **The GTEx contexts are Toil-era, not v11.** `reference_expression.tsv.gz` is GTEx v11; this is the + earlier release Toil reprocessed, so the two GTEx halves are neither the same samples nor the same + depth. They are also not spelled the same: Toil writes `Skin - Sun Exposed (Lower Leg)` and GTEx + v11 writes `Skin - Sun Exposed (Lower leg)`. **A context string cannot be carried from one table + to the other without checking.** +- **Three tissues are thin and are kept with their `n` rather than dropped**: Bladder (9), + Cervix - Ectocervix (6), Fallopian Tube (5). They are the matched normals of BLCA, OV and CESC, so + dropping them would remove the safety read for three tumour types. `Cervix - Endocervix` has fewer + than five samples and is absent; CESC still resolves through Ectocervix. Every TCGA study clears 36. +- **353 of 58,581 symbols carry more than one Ensembl gene id** (0.6 %). The row kept per + (symbol, context) is the one with the **highest median**, not a sum across loci: a lookup asks + whether a gene is on in this context, and the dominant locus answers it without inventing a total + across paralogues. 164,862 duplicate rows were collapsed this way. +- `n` is a **sample** count here, as in the GTEx half, and not the transcript count the mouse table's + `n` is. + +**What a floor built from it looks like**, as the 25th percentile of non-zero gene medians: +TCGA runs 0.1200 (KICH) to 0.3700 (LAML) TPM, a **3.08x** spread over 33 studies; GTEx runs 0.1000 +(Whole Blood) to 0.4000 (Testis), **4.00x** over 53 tissues. Pooled, TCGA is 0.1800 TPM over 882,959 +non-zero gene rows and GTEx 0.2700 over 1,402,705. **A tumour's own floor is consistently below its +matched normal's** -- SKCM 0.1600 against skin 0.3050 (0.52x), BLCA 0.1700 against bladder 0.3600 +(0.47x), LUAD 0.2000 against lung 0.3500 (0.57x), BRCA 0.1900 against breast 0.3000 (0.63x). + +**Format: parquet, and the record is not in the bootstrap set.** As gzipped TSV this table was +70.5 MB and 0.5 s to read with polars (5.2 s with the standard library); dictionary-encoded and +zstd-compressed it is **38.6 MB and 0.04 s**. On a link measured here at 154 kB/s that is three and +a half minutes of download saved, which is the difference between a bootstrap that finishes and one +that looks like it has hung. `mhcmatch bootstrap --reference` stages only the three companions +below — 6.6 MB between them, carrying the same numbers — and never this file. + +**Label arrays in the `.npz` are fixed-width unicode, not `dtype=object`.** An object array is +pickled, and `np.load` needs `allow_pickle=True` to read it back; a pickled array in a file fetched +over the network executes on load. The deposit is written so that flag is never needed. + +### Its three companions, and why scoring never reads the table itself + +Added 2026-08-26 beside it. `reference_expression_toil.tsv.gz` is the record; these are what a +program actually opens. + +| file | size | what it is | +|---|--:|---| +| `toil_matrix.npz` | 6,632,665 B | the whole table as a dense **58,581 gene x 86 context float32 matrix**, contexts keyed `"|"` | +| `toil_floors.tsv` | 8,365 B | 88 rows -- each context's abundance floor at q05, q10, q25, plus a pooled row per cohort | +| `context_synonyms.tsv` | 29,334 B | 448 rows, 184 aliases -- how a free-text origin becomes a context | + +**The matrix exists so a lookup does not cost a whole table.** Answering *what is this gene in this +tumour type* from the TSV means building a five-million-entry dictionary of dictionaries: measured, +**5.20 s and 3,168 MB resident**. The matrix is **0.05 s and 29 MB** -- 100x the speed on 1/109 of +the memory -- and every lookup a resolution chain makes is one array index. Both return the same +floor to four decimals, which is how the matrix is checked against the table it came from. A gene absent from a context is `0.0` and not `nan`: every gene here was +measured in every context that cleared the sample threshold, so absence is silence, and silence is a +measurement. Whether a *gene* is in the reference at all is a separate question and is reported +separately -- silence and ignorance must not collapse into one number. + +**`context_synonyms.tsv` is derived, not hand-written.** Long format, one target per row, so it +sorts and joins like data rather than holding lists in cells: + +| alias_kind | where the aliases come from | rows | +|---|---|--:| +| `code` | the 33 TCGA studies in the matrix | 83 | +| `site` | Xena `_primary_site`, joined to the PanCanAtlas study code on the 15-character barcode | 218 | +| `disease` | Xena `detailed_category`, joined the same way | 83 | +| `gtex` | the 53 GTEx contexts in the matrix | 64 | + +**One alias resolves to two things and each is its own row**, which is why the alias count and the +row count differ: 150 rows carry `target_kind = tcga_code` and 298 carry `gtex_context`, so +`hepatocellular` names both `LIHC` and `Liver`. 152 of the 448 rows are `derived:*` and 296 are +`curated:*`, and the `provenance` column says which on every row. The curated majority is the +matched-normal half and the adjectival forms a free-text origin actually arrives in +(`cervical`, `pancreatic`, `uterine corpus`); the derived half is what the matrix and the barcode +join supply on their own. Hand-writing this would have been wrong in the places nobody checks: **an organ is +more than one study more often than not** -- `Lung` is `LUAD` *and* `LUSC`, `Kidney` is three -- and a +crosswalk that picks one is silently wrong for the rest. The tumour-to-matched-normal half is carried +from `mhcmatch.expression.TUMOR_TISSUE` against the spellings this matrix uses, because Toil writes +`Brain - Frontal Cortex (Ba9)` where GTEx v11 writes `(BA9)` and an exact match drops it. + +An alias that does not resolve **raises**. Falling back to the pooled reference would return a +plausible number computed from the wrong distribution with nothing downstream able to tell. + +**A note on what these floors are for.** `c` is a property of a transcriptome, and the numerator is +one candidate's abundance. Where the numerator is not in reference TPM the two stop cancelling, and +the obvious repair -- a median of ratios against the reference -- **must not be estimated from a +candidate list**. A mutation reaches such a list only if it was called, and only where the gene was +seen in RNA, so the ratio measures that conditioning instead of the library: on three screens whose +columns are all deposited as TPM it returns 1.78, 2.18 and 3.15, never below 1. `mhcmatch` gates it +on coverage of the reference transcriptome for that reason, and a candidate list cannot clear it. + +--- + +## Mouse — `reference_expression_mmu.tsv.gz` and `protein_abundance_mmu.tsv.gz` + +Added 2026-08-21. The compendium's normal-tissue safety read was GTEx, and GTEx is human, so a mouse +epitope had no reference to be scored against. These two files are the mouse counterpart. + +**They are two files, not one, and that is deliberate.** The human table's value column is literally +`median_tpm`. RNA and protein are different measurements in different units, and writing a SILAC +abundance into a `_tpm` column would be a units lie no caller could detect. + +### `reference_expression_mmu.tsv.gz` — the GTEx analogue + +| | | +|---|---| +| rows | **659,050** | +| schema | `key`, `key_type`, `source`, `context`, `median_tpm`, `q25_tpm`, `q75_tpm`, `n` — **column for column identical to the human file** | +| keys | 18,830 gene symbols (`key_type = gene`) | +| contexts | **35 adult tissues** | +| `source` | `fantom5_mouse` | +| provenance | **experimental** — CAGE tag counts; the per-gene median across transcripts and the quantile summary are **computed** | + +**Origin.** EBI Expression Atlas [E-MTAB-3579](https://www.ebi.ac.uk/gxa/experiments/E-MTAB-3579), +the RIKEN FANTOM5 mouse CAGE atlas. `q25_tpm` / `median_tpm` / `q75_tpm` are taken from the Atlas' +own five-number summary per gene × group; they are not re-derived here. + +> Lizio M, Harshbarger J, Shimoji H, Severin J, Kasukawa T, Sahin S, Abugessaisa I, Fukuda S, +> Hori F, Ishikawa-Kato S, Mungall CJ, Arner E, Baillie JK, Bertin N, Bono H, de Hoon M, Diehl AD, +> Dimont E, Freeman TC, Fujieda K, Hide W, Kaliyaperumal R, Katayama T, Lassmann T, Meehan TF, +> Nishikata K, Ono H, Rehli M, Sandelin A, Schultes EA, 't Hoen PAC, Tatum Z, Thompson M, +> Toyoda T, Wright DW, Daub CO, Itoh M, Carninci P, Hayashizaki Y, Forrest ARR, Kawaji H. +> **Gateways to the FANTOM5 promoter level mammalian expression atlas.** +> *Genome Biol* 2015;16:22. +> PMID [25723102](https://pubmed.ncbi.nlm.nih.gov/25723102/) · +> doi:[10.1186/s13059-014-0560-6](https://doi.org/10.1186/s13059-014-0560-6) + +**Only plain `adult` groups are kept** — 35 of the deposit's 170. GTEx is adult normal tissue; +folding an E14 embryo, a pregnancy timepoint or a lactating mammary gland into a "is this gene on in +normal tissue" safety read answers a different question. The other stages are in the source and are +recoverable by re-running without the filter. + +**A parsing trap worth recording.** Atlas group labels are two `;`-separated parts and **the order +is not fixed**: 136 of the 170 are stage-first (`adult; thymus`) and 32 are tissue-first +(`adrenal gland; adult`). Taking either position blindly mislabels a third of the table — the first +build of this file produced contexts named `adult` and `whole body` that way. The stage is matched by +vocabulary wherever it sits, and the tissue is whatever remains. + +**Checked against known biology, not assumed.** Of 35 tissues, thymus ranks **first** for `Prss16` +(31 TPM), `Psmb11` (88), `Dntt` (961), `Rag1`, `Cd8a` (131) and `Lck` (12) — thymic epithelium and +thymocyte genes. As a negative control `Cela1` peaks in pancreas at 225,010 TPM against 0.3 in +thymus. + +**Caveats.** +- **`n` is 1 for every row.** FANTOM5 provides one library per adult tissue, so `q25`/`q75` describe + the spread across transcripts of a gene, **not across animals**. GTEx `n` runs to several hundred + donors. Do not read the two `n` columns as the same quantity. +- **`Aire` is 0 TPM in thymus here, and that is a detection floor, not biology.** Aire is expressed + in a rare mTEC subset; bulk CAGE at this depth cannot see it. Any promiscuous-expression question + needs sorted mTEC data, not this file. +- CAGE measures 5′-capped tag density. It is reported as TPM and behaves like expression, but it is + not RNA-seq TPM and the two are not numerically interchangeable across sources. + +### `tumor_expression_mmu.tsv.gz` — the mouse TUMOUR read, and why it is a separate file + +Added 2026-09-04. `reference_expression_mmu.tsv.gz` answers the *safety* half of the expression +question -- is this gene on in normal tissue -- and there was no counterpart for the *ranking* half. +Mouse has no TCGA, so a mouse neoantigen's own abundance fell back to the gene's normal-tissue +median, which is the number `expr_norm` already carries. Measured on the 923 class-I rows of the +IEDB mouse neoantigen deposit, `expr_lvl` and `expr_norm` came out **identical on 100 %** of the +rows lacking a deposited abundance: two fitted terms on one column. + +| | | +|---|---| +| rows | **149,640** | +| schema | `key`, `key_type`, `source`, `context`, `median_tpm`, `q25_tpm`, `q75_tpm`, `n` — **column for column identical to the two files above** | +| keys | 24,940 gene symbols (`key_type = gene`) | +| contexts | **6 syngeneic tumour models**: B16F10, CT26, E0771, LLC, MC38, Panc02 | +| `source` | `gse245293_syngeneic` | +| `n` | **3 on every row** — three biological replicates per model, so `q25`/`q75` are a spread across *animals*, unlike the FANTOM5 file's spread across transcripts | +| provenance | **experimental** — RNA-seq counts; the length normalisation, the per-model quantiles and the gene-level aggregation are **computed** | + +**Origin.** NCBI GEO [GSE245293](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE245293), +supplementary file `GSE245293_Normalized_count_table_tumor.txt.gz` (2.9 MB), staged at +`raw/expression_mouse/`. 23 samples in the series; the 18 `*_tumor_rep*` columns are used and the +8 CAF (cancer-associated fibroblast) columns are **not** -- those are stroma, not the tumour +transcriptome, and folding them in would answer a different question. + + curl -O https://ftp.ncbi.nlm.nih.gov/geo/series/GSE245nnn/GSE245293/suppl/GSE245293_Normalized_count_table_tumor.txt.gz + python src/build_tumor_expression_mmu.py # in ~/vcs/projects/public_expression_peptidomes + +> Comparison of syngeneic tumor models reveals different tumor microenvironments with distinct +> subsets of cancer-associated fibroblasts polarized by infiltrating T cells. GEO GSE245293, +> deposited 2023-10-13, released 2024-01-29. *Mus musculus*, expression profiling by high +> throughput sequencing. + +**Units are computed and the file says so.** The deposit is **normalised counts** with a `Length` +column, not TPM. Length-normalising each sample and rescaling to a million gives a TPM-like +quantity comparable *within* this table; it is **not** interchangeable with FANTOM5 CAGE TPM or +with GTEx RSEM TPM. `mhcmatch.rank.expr_level` is unit-free when its floor is a quantile of the +same column, which is exactly why a floor for these contexts must be taken from this file. + +**Checked against known biology, not assumed.** Three independent melanocytic markers separate the +melanoma line by three orders of magnitude: `Tyr` **398.9**, `Dct` **4,611.8** and `Pmel` +**4,636.7** TPM in B16F10 against ≤ 4.3 in all five other models. They would all move together if +the sample-to-model mapping were wrong. + +**Colon markers were tried and do not work here; do not re-try them.** `Muc2` reads ~0 in every +model -- CT26 is an undifferentiated colon carcinoma and makes no goblet-cell mucin -- and +`Cdx1`/`Cdx2` do not separate CT26 and MC38 from the rest (Cdx2 is highest in Panc02, 8.8 against +CT26's 1.4). The expectations were wrong, not the data. + +**Coverage of the IEDB mouse neoantigen deposit**, joining on model × gene symbol: **347 of 968 +class-I rows (36 %)** and **304 of 522 class-II rows (58 %)**. The class-II figure is the reason +this file exists -- that arm previously had a deposited abundance on 13 % of its rows. + +**Two models it does not carry** that the neoantigen deposit names often: `P815` (89 rows) and the +`T3`/`CMS5`/`Meth A` fibrosarcomas (150 rows between them). Those still take the pooled floor. + +### `PRJNA1271699` — located, not fetched + +[PRJNA1271699](https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1271699), *"Mouse syngeneic model +RNAseq sequencing"* (CrownBio), submitted 2025-06-04: **120 runs over 12 models × 10 replicates** +-- A20, B16BL6, B16F10, CT26, EMT6, H22, Hepa16, LL2, MC38, Pan02, RM1, Renca -- paired-end +Illumina NovaSeq 6000, **418 GB** of raw reads. It would cover Renca and EMT6, which nothing here +does. + +**NCBI carries raw reads only; there is no processed matrix.** The BioProject abstract advertises +an interactive resource at `https://muscreen.crownbio.com/syngeneic_omics`, and **as of 2026-09-04 +that URL is dead** -- `/syngeneic_omics` returns 404 and the host root returns a bare Apache +default page (403). So the processed tables the abstract promises are not currently retrievable, +and using this project means downloading 418 GB and running an alignment pipeline. Recorded here +so the cost is known rather than rediscovered. + +**This is the source to re-process if these models are wanted here** -- it is the one that covers +P815, Renca and EMT6, which `tumor_expression_mmu.tsv.gz` does not. + + # metadata only, no reads: + curl -sS "https://trace.ncbi.nlm.nih.gov/Traces/sra-db-be/runinfo?acc=PRJNA1271699" + +### `toil_matrix_mmu.npz` + `toil_floors_mmu.tsv` — the harmonised mouse compendium + +Added 2026-09-04. The mouse counterpart of `toil_matrix.npz` / `toil_floors.tsv`, same array names +and the same `source|context` key shape, so code written against the human pair reads this one. + +| | | +|---|---| +| shape | **26,737 genes x 68 contexts**, dense float32 | +| `genes` | official NCBI symbol, synonyms resolved via Ensembl release-100 (= GENCODE M25) and NCBI Gene | +| `contexts` | `fantom5_mouse\|` x 35 normal, `gse245293\|` x 6 tumour, `prjna1271699+\|` x 27 tumour | +| `values` | linear TPM-like, **harmonised across sources**; `NaN` = the gene is absent from that source, which is not the same as measured at zero | +| `n_samples` | 3 for GSE245293 (replicate **mean**), 1 for FANTOM5, 0 where the source reports none | +| `toil_floors_mmu.tsv` | 71 rows -- each context's floor at q05/q10/q25, plus a pooled row per source, same seven columns as the human file | + +**Sources.** FANTOM5 mouse CAGE (EBI E-MTAB-3579); GEO **GSE245293** (6 syngeneic models x 3 +replicates); and the CrownBio syngeneic panel, labelled `prjna1271699+` for NCBI BioProject +**PRJNA1271699** plus the further models that panel carries beyond the 12 the accession lists. +GSE245293 wins the five models both cover. The 27 filled models include **P815, Renca and EMT6**, +which no other deposit here reaches. + +**Harmonised, and here is what that means.** The three sources sat on three scales -- pooled mean +`log2(1 + TPM)` of 1.721, 2.513 and 2.059. Each source's non-zero values are quantile-mapped onto +the mean of the three sources' quantile functions; **per source, not per context**, so +between-context structure survives. Zero maps to zero. The map is monotone and its grid spans each +source's support, so **every within-context rank is preserved exactly** -- which is what +`mhcmatch.rank.expr_level` consumes. Non-zero spread falls 0.577 -> 0.087 log2 units, and the +pooled q25 floor agrees to 0.714--0.717 across all three. + +Two things it does not do. It cannot correct a **gene-specific** platform effect: `Actb` reads 3,032 +in B16F10 and 10.4 in FANTOM5 thymus after harmonisation, because CAGE at that promoter genuinely +ranks it mid-distribution. And it cannot correct the **detection-rate** difference (65.2 / 78.1 / +65.0 % non-zero), because a monotone map that respects zero cannot move it. + +**Absolute values are no longer any one source's own number.** That is the trade: comparability +across all 68 contexts, at the cost of each source's internal scale. The unharmonised per-source +tables are not deposited; `tumor_expression_mmu.tsv.gz` and `reference_expression_mmu.tsv.gz` above +remain the unharmonised reads of two of the three sources. + +Built by `src/build_deposit_mmu.py` in `~/vcs/projects/public_expression_peptidomes`; the protocol +and its numbers are `results/expression_harmonization.md` there. + +**For comparison, the human pair needed none of this.** Same measurement on `toil_matrix.npz`: +`toil_gtex` and `toil_tcga` pooled means differ by **0.011** log2 units, against 0.565 for the +mouse CAGE-vs-RNA-seq seam. Both human halves went through one Toil RSEM recompute, and it worked. + +### `protein_abundance_mmu.tsv.gz` — the protein-level read + +| | | +|---|---| +| rows | **133,848** | +| schema | `key`, `key_type`, `source`, `context`, `abundance`, `unit`, `n` | +| keys | 5,148 gene symbols | +| contexts | **26 tissues**, including thymus | +| `source` / `unit` | `geiger_silac_mouse` / `ppb_ibaq` | +| provenance | **experimental** — SILAC-normalised MS abundance | + +**Origin.** EBI Expression Atlas [E-PROT-11](https://www.ebi.ac.uk/gxa/experiments/E-PROT-11). + +> Geiger T, Velic A, Macek B, Lundberg E, Kampf C, Nagaraj N, Uhlen M, Cox J, Mann M. +> **Initial quantitative proteomic map of 28 mouse tissues using the SILAC mouse.** +> *Mol Cell Proteomics* 2013;12(6):1709–1722. +> PMID [23436904](https://pubmed.ncbi.nlm.nih.gov/23436904/) · +> doi:[10.1074/mcp.M112.024919](https://doi.org/10.1074/mcp.M112.024919) + +**Why keep a protein table at all.** A peptide can only be presented if the protein was translated, +so for immunopeptidome work protein abundance is the more direct evidence and RNA is a proxy. It is +much shallower — 5,148 genes against 18,830 — so it complements the RNA table rather than replacing +it. + +### Why not E-PROT-13 + +[E-PROT-13](https://www.ebi.ac.uk/gxa/experiments/E-PROT-13) (Huttlin's mouse phosphorylation and +expression atlas) was the first candidate considered. It covers **9 tissues and does not include +thymus**, so it cannot serve either purpose here. Of the **65 mouse baseline experiments in +Expression Atlas, only six contain thymus**; E-MTAB-3579 (35 adult tissues) and E-PROT-11 (26) are +the only two of those with broad tissue coverage. + +### Regenerate + + cd ~/vcs/projects/public_expression_peptidomes && ./src/build_expression_mmu.py + +--- + +## Human — the `hpa_consensus` arm, and why GTEx alone was not enough + +Added 2026-08-21, in the same pass that built the mouse references. + +**GTEx has no thymus.** Its 53 contexts contain no thymus, no lymph node and no bone marrow — adult +thymus involutes and is not routinely sampled, so the normal-tissue safety read was blind to the one +tissue central tolerance actually happens in. Expression Atlas is no better: of its **123 human +baseline experiments only three carry any thymus group**, and all three are compromised for this +purpose — E-MTAB-3871 is fetal tissue from fetuses with congenital defects, E-MTAB-3827 is sorted +thymocyte subsets rather than tissue, and E-MTAB-9641 is thymic epithelium cultured on a 3T3-J2 +feeder layer. + +HPA samples thymus directly, so its consensus tissue set is what fills the hole. + +| | | +|---|---| +| rows | **1,025,751** | +| keys | 20,151 gene symbols (`key_type = gene`) | +| contexts | **51 tissues, including thymus** | +| `source` | `hpa_consensus` | +| provenance | **experimental** — RNA-seq; the HPA consensus normalisation is **computed** | + +**Origin.** [`rna_tissue_consensus.tsv.zip`](https://www.proteinatlas.org/download/tsv/rna_tissue_consensus.tsv.zip) +from the Human Protein Atlas. + +> Uhlén M, Fagerberg L, Hallström BM, Lindskog C, Oksvold P, Mardinoglu A, Sivertsson Å, Kampf C, +> Sjöstedt E, Asplund A, Olsson I, Edlund K, Lundberg E, Navani S, Szigyarto CA, Odeberg J, +> Djureinovic D, Takanen JO, Hober S, Alm T, Edqvist PH, Berling H, Tegel H, Mulder J, Rockberg J, +> Nilsson P, Schwenk JM, Hamsten M, von Feilitzen K, Forsberg M, Persson L, Johansson F, Zwahlen M, +> von Heijne G, Nielsen J, Pontén F. +> **Tissue-based map of the human proteome.** +> *Science* 2015;347(6220):1260419. +> PMID [25613900](https://pubmed.ncbi.nlm.nih.gov/25613900/) · +> doi:[10.1126/science.1260419](https://doi.org/10.1126/science.1260419) + +**It is added as a third `source`, not merged into GTEx and not replacing it.** The schema was +designed with a `source` column for exactly this; every existing `gtex` and `tcga` row is untouched. + +**Checked against known biology.** Of 51 tissues, thymus ranks **first** for `PRSS16` (770 nTPM), +`PSMB11` (28.5), `DNTT` (306), `CD1A` (296), `RAG1` (368), `CCL25` (718), `CD3E` (596) and `LCK` +(326). Negative controls behave: `ALB` peaks in liver at 198,524 nTPM against 0.3 in thymus, `INS` in +pancreas, `MYH7` in skeletal muscle. + +**Caveats.** +- **nTPM is not GTEx TPM.** Both are TPM-family units but the normalisation differs, so compare + *within* a `source`, not across. This is what the `source` column is for. +- `n` is 1 — HPA consensus publishes one value per gene × tissue, so `q25_tpm` and `q75_tpm` equal + `median_tpm` and carry no spread. GTEx `n` runs to several hundred donors; the two `n` columns are + not the same quantity. +- **`AIRE` is detectable here (1.2 nTPM, rank 2 of 51) but reads 0 in the mouse CAGE atlas.** That is + a difference in assay depth, not in biology — do not read it as a species difference. diff --git a/expression/context_synonyms.tsv b/expression/context_synonyms.tsv new file mode 100644 index 0000000000000000000000000000000000000000..ebb923abeef07a20fb226d4d8e3134647c302b33 --- /dev/null +++ b/expression/context_synonyms.tsv @@ -0,0 +1,449 @@ +alias alias_kind target_kind target provenance +ACC code gtex_context Adrenal Gland curated:TUMOR_TISSUE +ACC code tcga_code ACC derived:toil_matrix +Acute Myeloid Leukemia disease gtex_context Whole Blood curated:TUMOR_TISSUE +Acute Myeloid Leukemia disease tcga_code LAML derived:xena_detailed_category +Adipose - Subcutaneous gtex gtex_context Adipose - Subcutaneous derived:toil_matrix +Adipose - Visceral (Omentum) gtex gtex_context Adipose - Visceral (Omentum) derived:toil_matrix +Adrenal Gland gtex gtex_context Adrenal Gland derived:toil_matrix +Adrenal gland site gtex_context Adrenal Gland curated:TUMOR_TISSUE +Adrenal gland site tcga_code ACC derived:xena_primary_site +Adrenocortical Cancer disease gtex_context Adrenal Gland curated:TUMOR_TISSUE +Adrenocortical Cancer disease tcga_code ACC derived:xena_detailed_category +Artery - Aorta gtex gtex_context Artery - Aorta derived:toil_matrix +Artery - Coronary gtex gtex_context Artery - Coronary derived:toil_matrix +Artery - Tibial gtex gtex_context Artery - Tibial derived:toil_matrix +BLCA code gtex_context Bladder curated:TUMOR_TISSUE +BLCA code tcga_code BLCA derived:toil_matrix +BRCA code gtex_context Breast - Mammary Tissue curated:TUMOR_TISSUE +BRCA code tcga_code BRCA derived:toil_matrix +Bile duct site gtex_context Liver curated:TUMOR_TISSUE +Bile duct site tcga_code CHOL derived:xena_primary_site +Bladder gtex gtex_context Bladder derived:toil_matrix +Bladder site gtex_context Bladder curated:TUMOR_TISSUE +Bladder site tcga_code BLCA derived:xena_primary_site +Bladder Urothelial Carcinoma disease gtex_context Bladder curated:TUMOR_TISSUE +Bladder Urothelial Carcinoma disease tcga_code BLCA derived:xena_detailed_category +Brain site gtex_context Brain - Amygdala curated:TUMOR_TISSUE +Brain site gtex_context Brain - Anterior Cingulate Cortex (Ba24) curated:TUMOR_TISSUE +Brain site gtex_context Brain - Caudate (Basal Ganglia) curated:TUMOR_TISSUE +Brain site gtex_context Brain - Cerebellar Hemisphere curated:TUMOR_TISSUE +Brain site gtex_context Brain - Cerebellum curated:TUMOR_TISSUE +Brain site gtex_context Brain - Cortex curated:TUMOR_TISSUE +Brain site gtex_context Brain - Frontal Cortex (Ba9) curated:TUMOR_TISSUE +Brain site gtex_context Brain - Hippocampus curated:TUMOR_TISSUE +Brain site gtex_context Brain - Hypothalamus curated:TUMOR_TISSUE +Brain site gtex_context Brain - Nucleus Accumbens (Basal Ganglia) curated:TUMOR_TISSUE +Brain site gtex_context Brain - Putamen (Basal Ganglia) curated:TUMOR_TISSUE +Brain site gtex_context Brain - Spinal Cord (Cervical C-1) curated:TUMOR_TISSUE +Brain site gtex_context Brain - Substantia Nigra curated:TUMOR_TISSUE +Brain site tcga_code GBM derived:xena_primary_site +Brain site tcga_code LGG derived:xena_primary_site +Brain - Amygdala gtex gtex_context Brain - Amygdala derived:toil_matrix +Brain - Anterior Cingulate Cortex (Ba24) gtex gtex_context Brain - Anterior Cingulate Cortex (Ba24) derived:toil_matrix +Brain - Caudate (Basal Ganglia) gtex gtex_context Brain - Caudate (Basal Ganglia) derived:toil_matrix +Brain - Cerebellar Hemisphere gtex gtex_context Brain - Cerebellar Hemisphere derived:toil_matrix +Brain - Cerebellum gtex gtex_context Brain - Cerebellum derived:toil_matrix +Brain - Cortex gtex gtex_context Brain - Cortex derived:toil_matrix +Brain - Frontal Cortex (Ba9) gtex gtex_context Brain - Frontal Cortex (Ba9) derived:toil_matrix +Brain - Hippocampus gtex gtex_context Brain - Hippocampus derived:toil_matrix +Brain - Hypothalamus gtex gtex_context Brain - Hypothalamus derived:toil_matrix +Brain - Nucleus Accumbens (Basal Ganglia) gtex gtex_context Brain - Nucleus Accumbens (Basal Ganglia) derived:toil_matrix +Brain - Putamen (Basal Ganglia) gtex gtex_context Brain - Putamen (Basal Ganglia) derived:toil_matrix +Brain - Spinal Cord (Cervical C-1) gtex gtex_context Brain - Spinal Cord (Cervical C-1) derived:toil_matrix +Brain - Substantia Nigra gtex gtex_context Brain - Substantia Nigra derived:toil_matrix +Brain Lower Grade Glioma disease gtex_context Brain - Amygdala curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Anterior Cingulate Cortex (Ba24) curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Caudate (Basal Ganglia) curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Cerebellar Hemisphere curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Cerebellum curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Cortex curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Frontal Cortex (Ba9) curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Hippocampus curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Hypothalamus curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Nucleus Accumbens (Basal Ganglia) curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Putamen (Basal Ganglia) curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Spinal Cord (Cervical C-1) curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease gtex_context Brain - Substantia Nigra curated:TUMOR_TISSUE +Brain Lower Grade Glioma disease tcga_code LGG derived:xena_detailed_category +Breast site gtex_context Breast - Mammary Tissue curated:TUMOR_TISSUE +Breast site tcga_code BRCA derived:xena_primary_site +Breast - Mammary Tissue gtex gtex_context Breast - Mammary Tissue derived:toil_matrix +Breast Invasive Carcinoma disease gtex_context Breast - Mammary Tissue curated:TUMOR_TISSUE +Breast Invasive Carcinoma disease tcga_code BRCA derived:xena_detailed_category +CESC code gtex_context Cervix - Ectocervix curated:TUMOR_TISSUE +CESC code tcga_code CESC derived:toil_matrix +CHOL code gtex_context Liver curated:TUMOR_TISSUE +CHOL code tcga_code CHOL derived:toil_matrix +COAD code gtex_context Colon - Sigmoid curated:TUMOR_TISSUE +COAD code gtex_context Colon - Transverse curated:TUMOR_TISSUE +COAD code tcga_code COAD derived:toil_matrix +Cells - Ebv-Transformed Lymphocytes gtex gtex_context Cells - Ebv-Transformed Lymphocytes derived:toil_matrix +Cells - Leukemia Cell Line (Cml) gtex gtex_context Cells - Leukemia Cell Line (Cml) derived:toil_matrix +Cells - Transformed Fibroblasts gtex gtex_context Cells - Transformed Fibroblasts derived:toil_matrix +Cervical & Endocervical Cancer disease gtex_context Cervix - Ectocervix curated:TUMOR_TISSUE +Cervical & Endocervical Cancer disease tcga_code CESC derived:xena_detailed_category +Cervix site gtex_context Cervix - Ectocervix curated:TUMOR_TISSUE +Cervix site tcga_code CESC derived:xena_primary_site +Cervix - Ectocervix gtex gtex_context Cervix - Ectocervix derived:toil_matrix +Cholangiocarcinoma disease gtex_context Liver curated:TUMOR_TISSUE +Cholangiocarcinoma disease tcga_code CHOL derived:xena_detailed_category +Colon site gtex_context Colon - Sigmoid curated:TUMOR_TISSUE +Colon site gtex_context Colon - Transverse curated:TUMOR_TISSUE +Colon site tcga_code COAD derived:xena_primary_site +Colon - Sigmoid gtex gtex_context Colon - Sigmoid derived:toil_matrix +Colon - Transverse gtex gtex_context Colon - Transverse derived:toil_matrix +Colon Adenocarcinoma disease gtex_context Colon - Sigmoid curated:TUMOR_TISSUE +Colon Adenocarcinoma disease gtex_context Colon - Transverse curated:TUMOR_TISSUE +Colon Adenocarcinoma disease tcga_code COAD derived:xena_detailed_category +DLBC code gtex_context Spleen curated:TUMOR_TISSUE +DLBC code gtex_context Whole Blood curated:TUMOR_TISSUE +DLBC code tcga_code DLBC derived:toil_matrix +Diffuse Large B-Cell Lymphoma disease gtex_context Spleen curated:TUMOR_TISSUE +Diffuse Large B-Cell Lymphoma disease gtex_context Whole Blood curated:TUMOR_TISSUE +Diffuse Large B-Cell Lymphoma disease tcga_code DLBC derived:xena_detailed_category +ESCA code gtex_context Esophagus - Gastroesophageal Junction curated:TUMOR_TISSUE +ESCA code gtex_context Esophagus - Mucosa curated:TUMOR_TISSUE +ESCA code gtex_context Esophagus - Muscularis curated:TUMOR_TISSUE +ESCA code tcga_code ESCA derived:toil_matrix +Endometrium site gtex_context Uterus curated:TUMOR_TISSUE +Endometrium site tcga_code UCEC derived:xena_primary_site +Esophageal Carcinoma disease gtex_context Esophagus - Gastroesophageal Junction curated:TUMOR_TISSUE +Esophageal Carcinoma disease gtex_context Esophagus - Mucosa curated:TUMOR_TISSUE +Esophageal Carcinoma disease gtex_context Esophagus - Muscularis curated:TUMOR_TISSUE +Esophageal Carcinoma disease tcga_code ESCA derived:xena_detailed_category +Esophagus site gtex_context Esophagus - Gastroesophageal Junction curated:TUMOR_TISSUE +Esophagus site gtex_context Esophagus - Mucosa curated:TUMOR_TISSUE +Esophagus site gtex_context Esophagus - Muscularis curated:TUMOR_TISSUE +Esophagus site tcga_code ESCA derived:xena_primary_site +Esophagus - Gastroesophageal Junction gtex gtex_context Esophagus - Gastroesophageal Junction derived:toil_matrix +Esophagus - Mucosa gtex gtex_context Esophagus - Mucosa derived:toil_matrix +Esophagus - Muscularis gtex gtex_context Esophagus - Muscularis derived:toil_matrix +Eye site tcga_code UVM derived:xena_primary_site +Fallopian Tube gtex gtex_context Fallopian Tube derived:toil_matrix +GBM code gtex_context Brain - Cortex curated:TUMOR_TISSUE +GBM code gtex_context Brain - Frontal Cortex (Ba9) curated:TUMOR_TISSUE +GBM code tcga_code GBM derived:toil_matrix +Glioblastoma Multiforme disease gtex_context Brain - Cortex curated:TUMOR_TISSUE +Glioblastoma Multiforme disease gtex_context Brain - Frontal Cortex (Ba9) curated:TUMOR_TISSUE +Glioblastoma Multiforme disease tcga_code GBM derived:xena_detailed_category +HNSC code gtex_context Esophagus - Mucosa curated:TUMOR_TISSUE +HNSC code gtex_context Minor Salivary Gland curated:TUMOR_TISSUE +HNSC code tcga_code HNSC derived:toil_matrix +Head & Neck Squamous Cell Carcinoma disease gtex_context Esophagus - Mucosa curated:TUMOR_TISSUE +Head & Neck Squamous Cell Carcinoma disease gtex_context Minor Salivary Gland curated:TUMOR_TISSUE +Head & Neck Squamous Cell Carcinoma disease tcga_code HNSC derived:xena_detailed_category +Head and Neck region site gtex_context Esophagus - Mucosa curated:TUMOR_TISSUE +Head and Neck region site gtex_context Minor Salivary Gland curated:TUMOR_TISSUE +Head and Neck region site tcga_code HNSC derived:xena_primary_site +Heart - Atrial Appendage gtex gtex_context Heart - Atrial Appendage derived:toil_matrix +Heart - Left Ventricle gtex gtex_context Heart - Left Ventricle derived:toil_matrix +KICH code gtex_context Kidney - Cortex curated:TUMOR_TISSUE +KICH code tcga_code KICH derived:toil_matrix +KIRC code gtex_context Kidney - Cortex curated:TUMOR_TISSUE +KIRC code tcga_code KIRC derived:toil_matrix +KIRP code gtex_context Kidney - Cortex curated:TUMOR_TISSUE +KIRP code tcga_code KIRP derived:toil_matrix +Kidney site gtex_context Kidney - Cortex curated:TUMOR_TISSUE +Kidney site tcga_code KICH derived:xena_primary_site +Kidney site tcga_code KIRC derived:xena_primary_site +Kidney site tcga_code KIRP derived:xena_primary_site +Kidney - Cortex gtex gtex_context Kidney - Cortex derived:toil_matrix +Kidney Chromophobe disease gtex_context Kidney - Cortex curated:TUMOR_TISSUE +Kidney Chromophobe disease tcga_code KICH derived:xena_detailed_category +Kidney Clear Cell Carcinoma disease gtex_context Kidney - Cortex curated:TUMOR_TISSUE +Kidney Clear Cell Carcinoma disease tcga_code KIRC derived:xena_detailed_category +Kidney Papillary Cell Carcinoma disease gtex_context Kidney - Cortex curated:TUMOR_TISSUE +Kidney Papillary Cell Carcinoma disease tcga_code KIRP derived:xena_detailed_category +LAML code gtex_context Whole Blood curated:TUMOR_TISSUE +LAML code tcga_code LAML derived:toil_matrix +LGG code gtex_context Brain - Amygdala curated:TUMOR_TISSUE +LGG code gtex_context Brain - Anterior Cingulate Cortex (Ba24) curated:TUMOR_TISSUE +LGG code gtex_context Brain - Caudate (Basal Ganglia) curated:TUMOR_TISSUE +LGG code gtex_context Brain - Cerebellar Hemisphere curated:TUMOR_TISSUE +LGG code gtex_context Brain - Cerebellum curated:TUMOR_TISSUE +LGG code gtex_context Brain - Cortex curated:TUMOR_TISSUE +LGG code gtex_context Brain - Frontal Cortex (Ba9) curated:TUMOR_TISSUE +LGG code gtex_context Brain - Hippocampus curated:TUMOR_TISSUE +LGG code gtex_context Brain - Hypothalamus curated:TUMOR_TISSUE +LGG code gtex_context Brain - Nucleus Accumbens (Basal Ganglia) curated:TUMOR_TISSUE +LGG code gtex_context Brain - Putamen (Basal Ganglia) curated:TUMOR_TISSUE +LGG code gtex_context Brain - Spinal Cord (Cervical C-1) curated:TUMOR_TISSUE +LGG code gtex_context Brain - Substantia Nigra curated:TUMOR_TISSUE +LGG code tcga_code LGG derived:toil_matrix +LIHC code gtex_context Liver curated:TUMOR_TISSUE +LIHC code tcga_code LIHC derived:toil_matrix +LUAD code gtex_context Lung curated:TUMOR_TISSUE +LUAD code tcga_code LUAD derived:toil_matrix +LUSC code gtex_context Lung curated:TUMOR_TISSUE +LUSC code tcga_code LUSC derived:toil_matrix +Lining of body cavities site tcga_code MESO derived:xena_primary_site +Liver gtex gtex_context Liver derived:toil_matrix +Liver site gtex_context Liver curated:TUMOR_TISSUE +Liver site tcga_code LIHC derived:xena_primary_site +Liver Hepatocellular Carcinoma disease gtex_context Liver curated:TUMOR_TISSUE +Liver Hepatocellular Carcinoma disease tcga_code LIHC derived:xena_detailed_category +Lung gtex gtex_context Lung derived:toil_matrix +Lung site gtex_context Lung curated:TUMOR_TISSUE +Lung site tcga_code LUAD derived:xena_primary_site +Lung site tcga_code LUSC derived:xena_primary_site +Lung Adenocarcinoma disease gtex_context Lung curated:TUMOR_TISSUE +Lung Adenocarcinoma disease tcga_code LUAD derived:xena_detailed_category +Lung Squamous Cell Carcinoma disease gtex_context Lung curated:TUMOR_TISSUE +Lung Squamous Cell Carcinoma disease tcga_code LUSC derived:xena_detailed_category +Lymphatic tissue site gtex_context Spleen curated:TUMOR_TISSUE +Lymphatic tissue site gtex_context Whole Blood curated:TUMOR_TISSUE +Lymphatic tissue site tcga_code DLBC derived:xena_primary_site +MESO code tcga_code MESO derived:toil_matrix +Mesothelioma disease tcga_code MESO derived:xena_detailed_category +Minor Salivary Gland gtex gtex_context Minor Salivary Gland derived:toil_matrix +Muscle - Skeletal gtex gtex_context Muscle - Skeletal derived:toil_matrix +Nerve - Tibial gtex gtex_context Nerve - Tibial derived:toil_matrix +OV code gtex_context Fallopian Tube curated:TUMOR_TISSUE +OV code gtex_context Ovary curated:TUMOR_TISSUE +OV code tcga_code OV derived:toil_matrix +Ovarian Serous Cystadenocarcinoma disease gtex_context Fallopian Tube curated:TUMOR_TISSUE +Ovarian Serous Cystadenocarcinoma disease gtex_context Ovary curated:TUMOR_TISSUE +Ovarian Serous Cystadenocarcinoma disease tcga_code OV derived:xena_detailed_category +Ovary gtex gtex_context Ovary derived:toil_matrix +Ovary site gtex_context Fallopian Tube curated:TUMOR_TISSUE +Ovary site gtex_context Ovary curated:TUMOR_TISSUE +Ovary site tcga_code OV derived:xena_primary_site +PAAD code gtex_context Pancreas curated:TUMOR_TISSUE +PAAD code tcga_code PAAD derived:toil_matrix +PCPG code gtex_context Adrenal Gland curated:TUMOR_TISSUE +PCPG code tcga_code PCPG derived:toil_matrix +PRAD code gtex_context Prostate curated:TUMOR_TISSUE +PRAD code tcga_code PRAD derived:toil_matrix +Pancreas gtex gtex_context Pancreas derived:toil_matrix +Pancreas site gtex_context Pancreas curated:TUMOR_TISSUE +Pancreas site tcga_code PAAD derived:xena_primary_site +Pancreatic Adenocarcinoma disease gtex_context Pancreas curated:TUMOR_TISSUE +Pancreatic Adenocarcinoma disease tcga_code PAAD derived:xena_detailed_category +Paraganglia site gtex_context Adrenal Gland curated:TUMOR_TISSUE +Paraganglia site tcga_code PCPG derived:xena_primary_site +Pheochromocytoma & Paraganglioma disease gtex_context Adrenal Gland curated:TUMOR_TISSUE +Pheochromocytoma & Paraganglioma disease tcga_code PCPG derived:xena_detailed_category +Pituitary gtex gtex_context Pituitary derived:toil_matrix +Prostate gtex gtex_context Prostate derived:toil_matrix +Prostate site gtex_context Prostate curated:TUMOR_TISSUE +Prostate site tcga_code PRAD derived:xena_primary_site +Prostate Adenocarcinoma disease gtex_context Prostate curated:TUMOR_TISSUE +Prostate Adenocarcinoma disease tcga_code PRAD derived:xena_detailed_category +READ code gtex_context Colon - Sigmoid curated:TUMOR_TISSUE +READ code gtex_context Colon - Transverse curated:TUMOR_TISSUE +READ code tcga_code READ derived:toil_matrix +Rectum site gtex_context Colon - Sigmoid curated:TUMOR_TISSUE +Rectum site gtex_context Colon - Transverse curated:TUMOR_TISSUE +Rectum site tcga_code READ derived:xena_primary_site +Rectum Adenocarcinoma disease gtex_context Colon - Sigmoid curated:TUMOR_TISSUE +Rectum Adenocarcinoma disease gtex_context Colon - Transverse curated:TUMOR_TISSUE +Rectum Adenocarcinoma disease tcga_code READ derived:xena_detailed_category +SARC code tcga_code SARC derived:toil_matrix +SKCM code gtex_context Skin - Not Sun Exposed (Suprapubic) curated:TUMOR_TISSUE +SKCM code gtex_context Skin - Sun Exposed (Lower Leg) curated:TUMOR_TISSUE +SKCM code tcga_code SKCM derived:toil_matrix +STAD code gtex_context Stomach curated:TUMOR_TISSUE +STAD code tcga_code STAD derived:toil_matrix +Sarcoma disease tcga_code SARC derived:xena_detailed_category +Skin site gtex_context Skin - Not Sun Exposed (Suprapubic) curated:TUMOR_TISSUE +Skin site gtex_context Skin - Sun Exposed (Lower Leg) curated:TUMOR_TISSUE +Skin site tcga_code SKCM derived:xena_primary_site +Skin - Not Sun Exposed (Suprapubic) gtex gtex_context Skin - Not Sun Exposed (Suprapubic) derived:toil_matrix +Skin - Sun Exposed (Lower Leg) gtex gtex_context Skin - Sun Exposed (Lower Leg) derived:toil_matrix +Skin Cutaneous Melanoma disease gtex_context Skin - Not Sun Exposed (Suprapubic) curated:TUMOR_TISSUE +Skin Cutaneous Melanoma disease gtex_context Skin - Sun Exposed (Lower Leg) curated:TUMOR_TISSUE +Skin Cutaneous Melanoma disease tcga_code SKCM derived:xena_detailed_category +Small Intestine - Terminal Ileum gtex gtex_context Small Intestine - Terminal Ileum derived:toil_matrix +Soft tissue,Bone site tcga_code SARC derived:xena_primary_site +Spleen gtex gtex_context Spleen derived:toil_matrix +Stomach gtex gtex_context Stomach derived:toil_matrix +Stomach site gtex_context Stomach curated:TUMOR_TISSUE +Stomach site tcga_code STAD derived:xena_primary_site +Stomach Adenocarcinoma disease gtex_context Stomach curated:TUMOR_TISSUE +Stomach Adenocarcinoma disease tcga_code STAD derived:xena_detailed_category +TGCT code gtex_context Testis curated:TUMOR_TISSUE +TGCT code tcga_code TGCT derived:toil_matrix +THCA code gtex_context Thyroid curated:TUMOR_TISSUE +THCA code tcga_code THCA derived:toil_matrix +THYM code tcga_code THYM derived:toil_matrix +Testicular Germ Cell Tumor disease gtex_context Testis curated:TUMOR_TISSUE +Testicular Germ Cell Tumor disease tcga_code TGCT derived:xena_detailed_category +Testis gtex gtex_context Testis derived:toil_matrix +Testis site gtex_context Testis curated:TUMOR_TISSUE +Testis site tcga_code TGCT derived:xena_primary_site +Thymoma disease tcga_code THYM derived:xena_detailed_category +Thymus site tcga_code THYM derived:xena_primary_site +Thyroid gtex gtex_context Thyroid derived:toil_matrix +Thyroid Carcinoma disease gtex_context Thyroid curated:TUMOR_TISSUE +Thyroid Carcinoma disease tcga_code THCA derived:xena_detailed_category +Thyroid Gland site gtex_context Thyroid curated:TUMOR_TISSUE +Thyroid Gland site tcga_code THCA derived:xena_primary_site +UCEC code gtex_context Uterus curated:TUMOR_TISSUE +UCEC code tcga_code UCEC derived:toil_matrix +UCS code gtex_context Uterus curated:TUMOR_TISSUE +UCS code tcga_code UCS derived:toil_matrix +UVM code tcga_code UVM derived:toil_matrix +Uterine Carcinosarcoma disease gtex_context Uterus curated:TUMOR_TISSUE +Uterine Carcinosarcoma disease tcga_code UCS derived:xena_detailed_category +Uterine Corpus Endometrioid Carcinoma disease gtex_context Uterus curated:TUMOR_TISSUE +Uterine Corpus Endometrioid Carcinoma disease tcga_code UCEC derived:xena_detailed_category +Uterus gtex gtex_context Uterus derived:toil_matrix +Uterus site gtex_context Uterus curated:TUMOR_TISSUE +Uterus site tcga_code UCS derived:xena_primary_site +Uveal Melanoma disease tcga_code UVM derived:xena_detailed_category +Vagina gtex gtex_context Vagina derived:toil_matrix +White blood cell site gtex_context Whole Blood curated:TUMOR_TISSUE +White blood cell site tcga_code LAML derived:xena_primary_site +Whole Blood gtex gtex_context Whole Blood derived:toil_matrix +adrenal gtex gtex_context Adrenal Gland curated:spelling +adrenal site gtex_context Adrenal Gland curated:spelling +adrenal site tcga_code ACC curated:spelling +biliary site gtex_context Liver curated:spelling +biliary site tcga_code CHOL curated:spelling +blood site gtex_context Whole Blood curated:spelling +blood site tcga_code LAML curated:spelling +bone site tcga_code SARC curated:spelling +cerebral site gtex_context Brain - Amygdala curated:spelling +cerebral site gtex_context Brain - Anterior Cingulate Cortex (Ba24) curated:spelling +cerebral site gtex_context Brain - Caudate (Basal Ganglia) curated:spelling +cerebral site gtex_context Brain - Cerebellar Hemisphere curated:spelling +cerebral site gtex_context Brain - Cerebellum curated:spelling +cerebral site gtex_context Brain - Cortex curated:spelling +cerebral site gtex_context Brain - Frontal Cortex (Ba9) curated:spelling +cerebral site gtex_context Brain - Hippocampus curated:spelling +cerebral site gtex_context Brain - Hypothalamus curated:spelling +cerebral site gtex_context Brain - Nucleus Accumbens (Basal Ganglia) curated:spelling +cerebral site gtex_context Brain - Putamen (Basal Ganglia) curated:spelling +cerebral site gtex_context Brain - Spinal Cord (Cervical C-1) curated:spelling +cerebral site gtex_context Brain - Substantia Nigra curated:spelling +cerebral site tcga_code GBM curated:spelling +cerebral site tcga_code LGG curated:spelling +cervical site gtex_context Cervix - Ectocervix curated:spelling +cervical site tcga_code CESC curated:spelling +cholangiocarcinoma site gtex_context Liver curated:spelling +cholangiocarcinoma site tcga_code CHOL curated:spelling +colonic site gtex_context Colon - Sigmoid curated:spelling +colonic site gtex_context Colon - Transverse curated:spelling +colonic site tcga_code COAD curated:spelling +colorectal site gtex_context Colon - Sigmoid curated:spelling +colorectal site gtex_context Colon - Transverse curated:spelling +colorectal site tcga_code COAD curated:spelling +colorectum site gtex_context Colon - Sigmoid curated:spelling +colorectum site gtex_context Colon - Transverse curated:spelling +colorectum site tcga_code COAD curated:spelling +cutaneous site gtex_context Skin - Not Sun Exposed (Suprapubic) curated:spelling +cutaneous site gtex_context Skin - Sun Exposed (Lower Leg) curated:spelling +cutaneous site tcga_code SKCM curated:spelling +endometrial site gtex_context Uterus curated:spelling +endometrial site tcga_code UCEC curated:spelling +esophageal site gtex_context Esophagus - Gastroesophageal Junction curated:spelling +esophageal site gtex_context Esophagus - Mucosa curated:spelling +esophageal site gtex_context Esophagus - Muscularis curated:spelling +esophageal site tcga_code ESCA curated:spelling +gastric gtex gtex_context Stomach curated:spelling +gastric site gtex_context Stomach curated:spelling +gastric site tcga_code STAD curated:spelling +glioblastoma site gtex_context Brain - Amygdala curated:spelling +glioblastoma site gtex_context Brain - Anterior Cingulate Cortex (Ba24) curated:spelling +glioblastoma site gtex_context Brain - Caudate (Basal Ganglia) curated:spelling +glioblastoma site gtex_context Brain - Cerebellar Hemisphere curated:spelling +glioblastoma site gtex_context Brain - Cerebellum curated:spelling +glioblastoma site gtex_context Brain - Cortex curated:spelling +glioblastoma site gtex_context Brain - Frontal Cortex (Ba9) curated:spelling +glioblastoma site gtex_context Brain - Hippocampus curated:spelling +glioblastoma site gtex_context Brain - Hypothalamus curated:spelling +glioblastoma site gtex_context Brain - Nucleus Accumbens (Basal Ganglia) curated:spelling +glioblastoma site gtex_context Brain - Putamen (Basal Ganglia) curated:spelling +glioblastoma site gtex_context Brain - Spinal Cord (Cervical C-1) curated:spelling +glioblastoma site gtex_context Brain - Substantia Nigra curated:spelling +glioblastoma site tcga_code GBM curated:spelling +glioblastoma site tcga_code LGG curated:spelling +glioma site gtex_context Brain - Amygdala curated:spelling +glioma site gtex_context Brain - Anterior Cingulate Cortex (Ba24) curated:spelling +glioma site gtex_context Brain - Caudate (Basal Ganglia) curated:spelling +glioma site gtex_context Brain - Cerebellar Hemisphere curated:spelling +glioma site gtex_context Brain - Cerebellum curated:spelling +glioma site gtex_context Brain - Cortex curated:spelling +glioma site gtex_context Brain - Frontal Cortex (Ba9) curated:spelling +glioma site gtex_context Brain - Hippocampus curated:spelling +glioma site gtex_context Brain - Hypothalamus curated:spelling +glioma site gtex_context Brain - Nucleus Accumbens (Basal Ganglia) curated:spelling +glioma site gtex_context Brain - Putamen (Basal Ganglia) curated:spelling +glioma site gtex_context Brain - Spinal Cord (Cervical C-1) curated:spelling +glioma site gtex_context Brain - Substantia Nigra curated:spelling +glioma site tcga_code GBM curated:spelling +glioma site tcga_code LGG curated:spelling +head and neck site gtex_context Esophagus - Mucosa curated:spelling +head and neck site gtex_context Minor Salivary Gland curated:spelling +head and neck site tcga_code HNSC curated:spelling +hepatic gtex gtex_context Liver curated:spelling +hepatic site gtex_context Liver curated:spelling +hepatic site tcga_code LIHC curated:spelling +hepatocellular gtex gtex_context Liver curated:spelling +hepatocellular site gtex_context Liver curated:spelling +hepatocellular site tcga_code LIHC curated:spelling +leukaemia site gtex_context Whole Blood curated:spelling +leukaemia site tcga_code LAML curated:spelling +leukemia site gtex_context Whole Blood curated:spelling +leukemia site tcga_code LAML curated:spelling +lymphoma site gtex_context Spleen curated:spelling +lymphoma site gtex_context Whole Blood curated:spelling +lymphoma site tcga_code DLBC curated:spelling +mammary site gtex_context Breast - Mammary Tissue curated:spelling +mammary site tcga_code BRCA curated:spelling +melanoma site gtex_context Skin - Not Sun Exposed (Suprapubic) curated:spelling +melanoma site gtex_context Skin - Sun Exposed (Lower Leg) curated:spelling +melanoma site tcga_code SKCM curated:spelling +mesothelioma site tcga_code MESO curated:spelling +ocular site tcga_code UVM curated:spelling +oesophageal site gtex_context Esophagus - Gastroesophageal Junction curated:spelling +oesophageal site gtex_context Esophagus - Mucosa curated:spelling +oesophageal site gtex_context Esophagus - Muscularis curated:spelling +oesophageal site tcga_code ESCA curated:spelling +oesophagus site gtex_context Esophagus - Gastroesophageal Junction curated:spelling +oesophagus site gtex_context Esophagus - Mucosa curated:spelling +oesophagus site gtex_context Esophagus - Muscularis curated:spelling +oesophagus site tcga_code ESCA curated:spelling +ovarian gtex gtex_context Ovary curated:spelling +ovarian site gtex_context Fallopian Tube curated:spelling +ovarian site gtex_context Ovary curated:spelling +ovarian site tcga_code OV curated:spelling +pancreatic gtex gtex_context Pancreas curated:spelling +pancreatic site gtex_context Pancreas curated:spelling +pancreatic site tcga_code PAAD curated:spelling +paraganglioma site gtex_context Adrenal Gland curated:spelling +paraganglioma site tcga_code PCPG curated:spelling +pheochromocytoma site gtex_context Adrenal Gland curated:spelling +pheochromocytoma site tcga_code PCPG curated:spelling +prostatic gtex gtex_context Prostate curated:spelling +prostatic site gtex_context Prostate curated:spelling +prostatic site tcga_code PRAD curated:spelling +pulmonary gtex gtex_context Lung curated:spelling +pulmonary site gtex_context Lung curated:spelling +pulmonary site tcga_code LUAD curated:spelling +pulmonary site tcga_code LUSC curated:spelling +rectal site gtex_context Colon - Sigmoid curated:spelling +rectal site gtex_context Colon - Transverse curated:spelling +rectal site tcga_code READ curated:spelling +renal site gtex_context Kidney - Cortex curated:spelling +renal site tcga_code KICH curated:spelling +renal site tcga_code KIRC curated:spelling +renal site tcga_code KIRP curated:spelling +sarcoma site tcga_code SARC curated:spelling +soft tissue site tcga_code SARC curated:spelling +testicular gtex gtex_context Testis curated:spelling +testicular site gtex_context Testis curated:spelling +testicular site tcga_code TGCT curated:spelling +thymic site tcga_code THYM curated:spelling +thyroid site gtex_context Thyroid curated:spelling +thyroid site tcga_code THCA curated:spelling +urothelial gtex gtex_context Bladder curated:spelling +urothelial site gtex_context Bladder curated:spelling +urothelial site tcga_code BLCA curated:spelling +uterine gtex gtex_context Uterus curated:spelling +uterine site gtex_context Uterus curated:spelling +uterine site tcga_code UCS curated:spelling +uterine corpus site gtex_context Uterus curated:spelling +uterine corpus site tcga_code UCEC curated:spelling +uveal site tcga_code UVM curated:spelling diff --git a/expression/protein_abundance_mmu.tsv.gz b/expression/protein_abundance_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..c49fe56edfd0cdb9123c9f6c3d0d63d51d39ab72 --- /dev/null +++ b/expression/protein_abundance_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e8ca44c2f01f9bab28f0f41ae66bb70a189f9f61d81925a645af3c39c9907cd +size 1206843 diff --git a/expression/reference_expression.tsv.gz b/expression/reference_expression.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..60b1c3ff24ac646c6dc4ce1a1c32a74e86ba78b5 --- /dev/null +++ b/expression/reference_expression.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b66301491e4fb0945eebfafa267dc287c8d4e9ffe2e81d0c8f351972f509605e +size 114886559 diff --git a/expression/reference_expression_mmu.tsv.gz b/expression/reference_expression_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..29e00c3a293b54754f4129c3c0f0089d680d6c25 --- /dev/null +++ b/expression/reference_expression_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:02cf907bb4cb2bb2565401d72fc230b24c1f7e7b8847da2c714d2054f9c18bdd +size 3116340 diff --git a/expression/reference_expression_toil.parquet b/expression/reference_expression_toil.parquet new file mode 100644 index 0000000000000000000000000000000000000000..0832b3615bad45ee543bcc5f5b115698c5fcfc2c --- /dev/null +++ b/expression/reference_expression_toil.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a461178a37bddb67283213fc6c05282eed7638f8278bff91baf94c4d85bb0aa8 +size 38553668 diff --git a/expression/toil_floors.tsv b/expression/toil_floors.tsv new file mode 100644 index 0000000000000000000000000000000000000000..1ed8cff7277b02d900662bab1202abccf7089c66 --- /dev/null +++ b/expression/toil_floors.tsv @@ -0,0 +1,89 @@ +source context q05 q10 q25 n_genes n_samples +toil_gtex Adipose - Subcutaneous 0.040000916 0.08000192 0.2900018 26040 318 +toil_gtex Adipose - Visceral (Omentum) 0.040000916 0.070001826 0.2599961 26119 197 +toil_gtex Adrenal Gland 0.02999974 0.06000199 0.23999915 26015 128 +toil_gtex Artery - Aorta 0.040000916 0.08999935 0.31998995 25524 207 +toil_gtex Artery - Coronary 0.040000916 0.08000192 0.31000295 25751 118 +toil_gtex Artery - Tibial 0.040000916 0.08000192 0.3299993 24589 281 +toil_gtex Bladder 0.049998555 0.099998094 0.36000732 26736 9 +toil_gtex Brain - Amygdala 0.02999974 0.06000199 0.23999915 26046 69 +toil_gtex Brain - Anterior Cingulate Cortex (Ba24) 0.02999974 0.06000199 0.2500071 26594 83 +toil_gtex Brain - Caudate (Basal Ganglia) 0.02999974 0.06000199 0.23999915 27075 109 +toil_gtex Brain - Cerebellar Hemisphere 0.049998555 0.08999935 0.34998956 27122 97 +toil_gtex Brain - Cerebellum 0.049998555 0.099998094 0.3700023 27425 119 +toil_gtex Brain - Cortex 0.040000916 0.08000192 0.2799918 27159 105 +toil_gtex Brain - Frontal Cortex (Ba9) 0.040000916 0.070001826 0.2749992 26770 102 +toil_gtex Brain - Hippocampus 0.02999974 0.06000199 0.23999915 26358 84 +toil_gtex Brain - Hypothalamus 0.030000996 0.070001826 0.2599961 27235 82 +toil_gtex Brain - Nucleus Accumbens (Basal Ganglia) 0.02999974 0.06000199 0.24500312 27068 104 +toil_gtex Brain - Putamen (Basal Ganglia) 0.02999974 0.06000199 0.23000562 25961 81 +toil_gtex Brain - Spinal Cord (Cervical C-1) 0.040000916 0.07500187 0.2849968 26134 60 +toil_gtex Brain - Substantia Nigra 0.02999974 0.070001826 0.2599961 26187 57 +toil_gtex Breast - Mammary Tissue 0.040000916 0.08000192 0.29999262 27044 179 +toil_gtex Cells - Ebv-Transformed Lymphocytes 0.019999515 0.040000916 0.15999635 25259 107 +toil_gtex Cells - Leukemia Cell Line (Cml) 0.019999515 0.044999734 0.17000377 24871 70 +toil_gtex Cells - Transformed Fibroblasts 0.02999974 0.06000199 0.23000562 24328 256 +toil_gtex Cervix - Ectocervix 0.030000215 0.06000088 0.23999858 28094 6 +toil_gtex Colon - Sigmoid 0.040000916 0.08000192 0.29999262 25948 141 +toil_gtex Colon - Transverse 0.040000916 0.08000192 0.27000663 26613 167 +toil_gtex Esophagus - Gastroesophageal Junction 0.040000916 0.08000192 0.2900018 25786 137 +toil_gtex Esophagus - Mucosa 0.040000916 0.070001826 0.27000663 25938 273 +toil_gtex Esophagus - Muscularis 0.040000916 0.08000192 0.2900018 25631 245 +toil_gtex Fallopian Tube 0.06000199 0.10999756 0.38000983 28082 5 +toil_gtex Heart - Atrial Appendage 0.02999974 0.06000199 0.20000455 25363 175 +toil_gtex Heart - Left Ventricle 0.019999515 0.040000916 0.15999635 23986 202 +toil_gtex Kidney - Cortex 0.024999628 0.050001454 0.20999782 26516 28 +toil_gtex Liver 0.02999974 0.049998555 0.18000482 24649 110 +toil_gtex Lung 0.049998555 0.08999935 0.34998956 27206 288 +toil_gtex Minor Salivary Gland 0.040000916 0.08000192 0.27000663 27258 55 +toil_gtex Muscle - Skeletal 0.02999974 0.049998555 0.18000482 23401 396 +toil_gtex Nerve - Tibial 0.049998555 0.099998094 0.3700023 27034 278 +toil_gtex Ovary 0.040000916 0.07500187 0.29499722 26838 88 +toil_gtex Pancreas 0.02999974 0.049998555 0.18000482 25047 167 +toil_gtex Pituitary 0.049998555 0.099998094 0.36000732 28477 107 +toil_gtex Prostate 0.044999734 0.08999935 0.3499985 27692 100 +toil_gtex Skin - Not Sun Exposed (Suprapubic) 0.040000916 0.08000192 0.2799918 26707 233 +toil_gtex Skin - Sun Exposed (Lower Leg) 0.040000916 0.08000192 0.3049978 26572 324 +toil_gtex Small Intestine - Terminal Ileum 0.040000916 0.08000192 0.2749992 27637 92 +toil_gtex Spleen 0.040000916 0.08000192 0.29999262 27157 100 +toil_gtex Stomach 0.02999974 0.06000199 0.23000562 25735 175 +toil_gtex Testis 0.070001826 0.12000176 0.4000072 36855 165 +toil_gtex Thyroid 0.049998555 0.099998094 0.3700023 27300 279 +toil_gtex Uterus 0.040000916 0.085000634 0.3299993 26839 78 +toil_gtex Vagina 0.040000916 0.08999935 0.3299993 27096 85 +toil_gtex Whole Blood 0.019999515 0.02999974 0.099998094 21838 337 +toil_gtex __pooled__ 0.040000916 0.070001826 0.27000663 1402705 396 +toil_tcga ACC 0.019999515 0.040000916 0.17000377 24378 76 +toil_tcga BLCA 0.019999515 0.040000916 0.17000377 26115 403 +toil_tcga BRCA 0.019999515 0.049998555 0.1899941 27126 1087 +toil_tcga CESC 0.019999515 0.040000916 0.17000377 26331 301 +toil_tcga CHOL 0.024999628 0.049999047 0.20999782 26374 36 +toil_tcga COAD 0.019999515 0.049998555 0.1750043 26169 282 +toil_tcga DLBC 0.019999515 0.040000916 0.14999785 24919 47 +toil_tcga ESCA 0.019999515 0.040000916 0.14000322 30561 180 +toil_tcga GBM 0.02999974 0.070001826 0.2500071 27300 153 +toil_tcga HNSC 0.019999515 0.040000916 0.15999635 25899 514 +toil_tcga KICH 0.019999515 0.02999974 0.12000176 25726 65 +toil_tcga KIRC 0.019999515 0.040000916 0.15999635 27722 512 +toil_tcga KIRP 0.019999515 0.049998555 0.1899941 26360 283 +toil_tcga LAML 0.049998555 0.099998094 0.3700023 28056 173 +toil_tcga LGG 0.02999974 0.06000199 0.21999718 27931 507 +toil_tcga LIHC 0.019999515 0.040000916 0.14999785 24179 367 +toil_tcga LUAD 0.02999974 0.049998555 0.20000455 27131 508 +toil_tcga LUSC 0.019999515 0.049998555 0.18000482 27614 483 +toil_tcga MESO 0.019999515 0.049998555 0.18000482 26673 87 +toil_tcga OV 0.019999515 0.040000916 0.14000322 28858 418 +toil_tcga PAAD 0.02999974 0.049998555 0.21999718 27483 157 +toil_tcga PCPG 0.024999628 0.049998555 0.20000455 26205 178 +toil_tcga PRAD 0.019999515 0.049998555 0.1899941 27144 492 +toil_tcga READ 0.019999515 0.049998555 0.18000482 26295 89 +toil_tcga SARC 0.02999974 0.049998555 0.1899941 25531 254 +toil_tcga SKCM 0.019999515 0.040000916 0.15999635 25675 468 +toil_tcga STAD 0.019999515 0.040000916 0.14000322 29404 410 +toil_tcga TGCT 0.02999974 0.049998555 0.18000482 29059 147 +toil_tcga THCA 0.019999515 0.040000916 0.1899941 26518 507 +toil_tcga THYM 0.019999515 0.049998555 0.20000455 27330 119 +toil_tcga UCEC 0.019999515 0.040000916 0.18000482 26404 177 +toil_tcga UCS 0.02999974 0.049998555 0.20000455 26981 57 +toil_tcga UVM 0.019999515 0.040000916 0.14999785 23508 79 +toil_tcga __pooled__ 0.019999515 0.049998555 0.18000482 882959 1087 diff --git a/expression/toil_floors_mmu.tsv b/expression/toil_floors_mmu.tsv new file mode 100644 index 0000000000000000000000000000000000000000..a16401a68503546acc30343c58d6703c024de9b1 --- /dev/null +++ b/expression/toil_floors_mmu.tsv @@ -0,0 +1,72 @@ +source context q05 q10 q25 n_genes n_samples +fantom5_mouse __pooled__ 0.0990793183 0.166985497 0.717383385 429465 1 +fantom5_mouse accessory axillary lymph node 0.166985497 0.166985497 1.38365448 12020 1 +fantom5_mouse adrenal gland 0.0459894836 0.0459894836 0.405304521 13385 1 +fantom5_mouse aorta 0.0459894836 0.0459894836 0.405304521 13162 1 +fantom5_mouse bone (os femoris) 0.717383385 0.717383385 3.26015902 10340 1 +fantom5_mouse cecum 0.717383385 3.26015902 3.26015902 9173 1 +fantom5_mouse cerebellum 0.166985497 0.166985497 1.38365448 11958 1 +fantom5_mouse colon 0.166985497 0.166985497 0.577963889 12659 1 +fantom5_mouse corpora quadrigemina 0.0459894836 0.166985497 0.717383385 13084 1 +fantom5_mouse corpus striatum 0.309359103 0.309359103 1.38365448 11807 1 +fantom5_mouse cortex 0.0990793183 0.166985497 0.717383385 12867 1 +fantom5_mouse diencephalon 0.0990793183 0.0990793183 1.38365448 12892 1 +fantom5_mouse epididymis 0.166985497 0.166985497 0.577963889 13016 1 +fantom5_mouse eyeball 0.0990793183 0.0990793183 0.405304521 13595 1 +fantom5_mouse hippocampus 0.0990793183 0.0990793183 0.717383385 12710 1 +fantom5_mouse intestinal mucosa 0.166985497 0.166985497 0.577963889 11107 1 +fantom5_mouse intestine 0.166985497 0.166985497 0.577963889 11967 1 +fantom5_mouse lung 0.0990793183 0.166985497 0.717383385 13370 1 +fantom5_mouse medulla oblongata 0.0990793183 0.0990793183 0.577963889 13242 1 +fantom5_mouse olfactory brain 0.166985497 0.166985497 1.38365448 12296 1 +fantom5_mouse ovary 0.0459894836 0.0459894836 0.405304521 14064 1 +fantom5_mouse pancreas 0.493148953 1.38365448 3.26015902 9097 1 +fantom5_mouse pituitary gland 0.166985497 0.166985497 1.38365448 12222 1 +fantom5_mouse prostate 0.0990793183 0.0990793183 0.493148953 12308 1 +fantom5_mouse skin 0.0990793183 0.0990793183 0.577963889 13526 1 +fantom5_mouse spinal cord 0.0990793183 0.0990793183 0.717383385 13303 1 +fantom5_mouse spleen 0.0990793183 0.0990793183 0.493148953 12228 1 +fantom5_mouse stomach 0.309359103 1.38365448 3.26015902 10780 1 +fantom5_mouse submandibular gland 0.405304521 0.405304521 1.38365448 10990 1 +fantom5_mouse testis 0.405304521 0.405304521 3.26015902 11113 1 +fantom5_mouse thymus 0.166985497 0.166985497 0.717383385 11856 1 +fantom5_mouse tongue 0.0990793183 0.0990793183 0.717383385 12850 1 +fantom5_mouse urinary bladder 0.0990793183 0.0990793183 0.577963889 13257 1 +fantom5_mouse uterus 0.0990793183 0.0990793183 0.493148953 13252 1 +fantom5_mouse vagina 0.0990793183 0.0990793183 0.577963889 12979 1 +fantom5_mouse vesicular gland 0.309359103 0.309359103 1.38365448 10990 1 +gse245293 B16F10 0.0863542557 0.147608891 0.709316909 19074 3 +gse245293 CT26 0.0808949843 0.133418322 0.602678776 19182 3 +gse245293 E0771 0.0942154527 0.164973333 0.778372049 19101 3 +gse245293 LLC 0.0898912326 0.147942439 0.675284326 19700 3 +gse245293 MC38 0.0903859511 0.157022417 0.767261088 19890 3 +gse245293 Panc02 0.094235152 0.165870145 0.870251417 19764 3 +gse245293 __pooled__ 0.0882425308 0.152038917 0.714352965 116711 3 +prjna1271699+ 4T1 0.0907283276 0.201277331 0.991839886 16827 0 +prjna1271699+ 4T1orth 0.129452452 0.210650012 0.996328354 16699 0 +prjna1271699+ A20 0.0949191302 0.157220826 0.662428319 16502 0 +prjna1271699+ B16BL6 0.129452452 0.205734506 0.889154255 15882 0 +prjna1271699+ B16F0 0.076880008 0.129452452 0.528137326 17105 0 +prjna1271699+ B16F1 0.076880008 0.129452452 0.539604843 16719 0 +prjna1271699+ C1498 0.0816102177 0.138391808 0.556235015 15929 0 +prjna1271699+ CloneM3 0.129452452 0.205734506 0.910723627 15760 0 +prjna1271699+ Colon26 0.0862378627 0.157220826 0.844071329 16671 0 +prjna1271699+ EL4 0.0949191302 0.166985497 0.701014161 17045 0 +prjna1271699+ EMT6 0.0949191302 0.166985497 0.894779742 16177 0 +prjna1271699+ GL261 0.129452452 0.205734506 0.98558861 17720 0 +prjna1271699+ H22 0.0907283276 0.157220826 0.695897818 16640 0 +prjna1271699+ Hepa1-6 0.0907283276 0.157220826 0.814445496 17151 0 +prjna1271699+ J558 0.0862378627 0.147964045 0.652128041 16481 0 +prjna1271699+ JC 0.0816102177 0.143396497 0.711802483 17128 0 +prjna1271699+ KLN205 0.0862378627 0.152635559 0.844071329 18160 0 +prjna1271699+ L1210 0.0862378627 0.138391808 0.528137326 16918 0 +prjna1271699+ L5178R 0.0816102177 0.138391808 0.556235015 16175 0 +prjna1271699+ MB49 0.0816102177 0.129452452 0.539604843 16623 0 +prjna1271699+ MBT2 0.0862378627 0.143396497 0.657023311 17120 0 +prjna1271699+ MPC11 0.201277331 0.269641459 0.867251277 15036 0 +prjna1271699+ Neuro2a 0.0816102177 0.133858606 0.647043347 17522 0 +prjna1271699+ P815 0.0907283276 0.152635559 0.766629577 15932 0 +prjna1271699+ RM1 0.0862378627 0.143396497 0.68285197 16868 0 +prjna1271699+ Renca 0.0907283276 0.157220826 0.820747018 16407 0 +prjna1271699+ WEHI164 0.0862378627 0.143396497 0.67239815 17406 0 +prjna1271699+ __pooled__ 0.0898167714 0.152635559 0.717383385 450603 0 diff --git a/expression/toil_matrix.npz b/expression/toil_matrix.npz new file mode 100644 index 0000000000000000000000000000000000000000..a53497c8e3459c8157f5abce8fc8ea847e463b45 --- /dev/null +++ b/expression/toil_matrix.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:005df372d40ce4ea1f8d56e1f8e842550d4edebcc81a67ad26a7c20e1612a875 +size 6632665 diff --git a/expression/toil_matrix_mmu.npz b/expression/toil_matrix_mmu.npz new file mode 100644 index 0000000000000000000000000000000000000000..6c306cb2be871de3ed4afdceffa04ae6845d75f4 --- /dev/null +++ b/expression/toil_matrix_mmu.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8ad1e180053facb0e823e83865a60cd77acb5cf3cdaf30ad4887dee2d6c5100f +size 2664374 diff --git a/expression/tumor_expression_mmu.tsv.gz b/expression/tumor_expression_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6386c239ae37b4ca0be59caa6548a805909d191f --- /dev/null +++ b/expression/tumor_expression_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6c2abec625d3e440878e722fe561266cbd0ae5914640d5f5ecfe37d357f343b9 +size 1999293 diff --git a/hla_loh/SOURCES.md b/hla_loh/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..50cc7aca6fe22601a1c05382bd8dfe42d3b7e67b --- /dev/null +++ b/hla_loh/SOURCES.md @@ -0,0 +1,134 @@ +# `hla_loh/` — sources + +Per-allele HLA-I **loss of heterozygosity** and genetic immune escape, over 6,319 tumours +(4,439 Hartwig metastatic + 1,880 PCAWG primary), plus the join that carries the PCAWG arm onto +TCGA barcodes so it meets `tcga/tcga_neoantigens.tsv.gz`. + +The question this directory exists to answer: **when a tumour loses one HLA allele, is the lost one +the allele that better presented its neoantigens?** That is the escape mechanism a polyepitope +cassette has to be designed against, and `bench/cassette/tcga_loh_perallele.py` in the benchmark +repo is the consumer. + +## Provenance in one line + +Everything here is **derived** — LILAC allele-specific copy-number calls on whole-genome sequencing, +published as supplementary data — except `do_to_tcga_barcode.tsv`, which is a lookup we performed +against the GDC API. Nothing in this directory is a measurement we made. + +**Primary source.** Martínez-Jiménez F, Priestley P, Shale C, Baber J, Rozemuller E, Cuppen E. +*Genetic immune escape landscape in primary and metastatic cancer.* **Nature Genetics 55(5):820–831 +(2023)**, [doi:10.1038/s41588-023-01367-1](https://doi.org/10.1038/s41588-023-01367-1). + +## What ships, and what deliberately does not + +**The publisher's workbooks are not redistributable; the facts in them are.** The three +`41588_2023_1367_MOESM*.xlsx` are copyrighted Springer Nature supplementary material. They are kept +in `raw/hla_loh/` (gitignored) and `harmonize.py` rebuilds neutral long-format parquet from them. +This is the same call already taken for `icb_cohorts/`: harmonised output published, publisher +workbook held back. An `.xlsx` is a layout and a copyrightable compilation; a table of allele copy +numbers is neither. + +### Published (this directory) + +| file | rows × cols | what it is | provenance | +|---|---|---|---| +| `gie_allele.parquet` | 37,914 × 9 | one row per **(sample, allele slot)** — 6,319 samples × the six slots `A1 A2 B1 B2 C1 C2`. Carries `allele`, `copy_number`, `n_mutations`, plus `cohort`/`cancer_type`. This is Supplementary Data 2 unpivoted | derived (LILAC on WGS) | +| `gie_sample.parquet` | 6,319 × 32 | one row per **sample**: the escape read-outs — `loh_lilac`, `copy_number_neutral_loh_hla`, `loh_focal`/`loh_nonfocal`/`loh_hfocal`, `mut_hla_lilac`, the five pathway flags with their detail strings, `genetic_immune_escape`, per-locus `ploidy_minor`/`ploidy_major`/`del_HLA-*` | derived | +| `hla_typing_callers.parquet` | 226,008 × 8 | one row per **(sample, caller, tissue, slot)** — 6,278 samples × 3 callers (`LILAC`, `xHLA`, `polysolver`) × normal/tumour × 6 slots. Supplementary Data 1, unpivoted. The caller-concordance reference; **not** used in the join | derived | +| `sample_metadata.parquet` | 7,619 × 40 | per-sample cohort, cancer type, biopsy site, treatment history, MSI/HR status and mutation load (`sbs`/`dbs`/`indel`/`sv`, clonal and subclonal). Supplementary Table 2 | derived | +| `escape_genes.parquet` | 21 × 6 | the immune-escape gene panel the paper's calls are made over, with pathway assignment and admissible mutation types. Supplementary Table 1 | curated gene list | +| `perallele_tcga_loh.parquet` | 2,747 × 3 | **the deliverable the benchmark consumes.** `donor` (TCGA barcode), `hla_allele`, `allele_lost`. 481 TCGA donors overlapping `bench/cassette/scored.parquet`; 238 loss calls | derived (join, 2026-09-11) | +| `do_to_tcga_barcode.tsv` | 711 × 4 | `icgc_donor_id` → `case_uuid` → `tcga_barcode` → `gdc_project`, for the PCAWG donors that are TCGA cases | derived (GDC `/cases` lookup) | +| `harmonize.py` | — | rebuilds the five parquets above from `raw/hla_loh/` | — | + +### Held back in `raw/hla_loh/` (gitignored) + +| file | why it is not published | +|---|---| +| `41588_2023_1367_MOESM4_ESM.xlsx` | copyrighted Springer Nature supplement (Supplementary Tables 1–2) | +| `41588_2023_1367_MOESM5_ESM.xlsx` | copyrighted (Supplementary Data 1) | +| `41588_2023_1367_MOESM6_ESM.xlsx` | copyrighted (Supplementary Data 2) | +| `pcawg_sample_sheet.2016-10-18.tsv` | third-party PCAWG release file, re-fetchable (command below); an input, not our output | + +**The Hartwig arm of the source study needs its own access application** for the underlying +sequencing. Nothing here is that sequencing — these are the published per-sample call tables. + +## Identifiers, and why they are public + +- **PCAWG samples are keyed `DO#####`** (`icgc_donor_id`), Hartwig samples `HMF#####A`. Both are + pseudonymous research identifiers assigned by the consortium, as published. +- **TCGA barcodes** (`TCGA-XX-XXXX`) are likewise public — `tcga/` in this compendium already ships + per-barcode HLA and neoantigen calls on the same basis (open-access somatic + expression, as in + TCIA / Thorsson 2018). +- PCAWG's `donor_wgs_exclusion_white_gray` is the consortium's **published QC whitelist/greylist** + flag, not a consent tier. An earlier `.gitignore` rule in this repo described this directory as a + re-identification risk; that reading was wrong and has been withdrawn. + +## Three things a consumer has to know + +**1. `n_mutations` is a count, not a flag.** `*_MUT` takes `0.0, 0.5, 1.0, 1.5, 2.0, 3.0, 4.0` +across the six slots (37,707 zero, 207 non-zero, 42 of them non-integral). Reading it as a boolean +loses every fractional and multi-hit call. + +**2. Four columns are comma-separated locus lists, not booleans.** +`copy_number_neutral_loh_hla`, `loh_focal`, `loh_nonfocal`, `loh_hfocal` carry values like +`HLA-A,HLA-C,HLA-B`. They sit beside ten genuine booleans in the same sheet and look identical at a +glance; `copy_number_neutral_loh_hla` has 784 non-null values that a bool cast would have silently +nulled. They are kept verbatim as the source wrote them. + +**3. The allele vocabulary is this repo's, not the workbook's.** The workbooks write `A*02:01`; +`gie_allele.parquet` and `hla_typing_callers.parquet` write **`HLA-A*02:01`**, matching +`perallele_tcga_loh.parquet`, `tcga/`, and `mhcmatch`. Two spellings for one molecule is the +recurring failure mode in this compendium, and a lookup that silently matches neither returns the +same shape as "nothing was presented". + +## The join + +``` +gie_sample / gie_allele (cohort = PCAWG, sample_id = DO#####) + │ icgc_donor_id + ▼ +do_to_tcga_barcode.tsv 711 PCAWG donors that are TCGA cases + │ tcga_barcode + ▼ +perallele_tcga_loh.parquet ──► bench/cassette/scored.parquet (481 donors overlap) +``` + +All 711 bridge rows join to a PCAWG sample in `gie_sample.parquet`, and all 481 donors in +`perallele_tcga_loh.parquet` are present in the bridge. The Hartwig arm has no TCGA counterpart and +does not enter the join. + +## Regenerate + +```zsh +# the five harmonised parquets, from the workbooks in raw/hla_loh/ +python hla_loh/harmonize.py + +# the PCAWG sample sheet (the ICGC DCC portal is retired; this is the operations repo mirror) +gh api repos/ICGC-TCGA-PanCancer/pcawg-operations/contents/lists/sample_sheet/pcawg_sample_sheet.2016-10-18.tsv \ + -H "Accept: application/vnd.github.raw+json" > raw/hla_loh/pcawg_sample_sheet.2016-10-18.tsv + +# the workbooks +curl -O https://static-content.springer.com/esm/art%3A10.1038%2Fs41588-023-01367-1/MediaObjects/41588_2023_1367_MOESM4_ESM.xlsx +curl -O https://static-content.springer.com/esm/art%3A10.1038%2Fs41588-023-01367-1/MediaObjects/41588_2023_1367_MOESM5_ESM.xlsx +curl -O https://static-content.springer.com/esm/art%3A10.1038%2Fs41588-023-01367-1/MediaObjects/41588_2023_1367_MOESM6_ESM.xlsx +``` + +`do_to_tcga_barcode.tsv` and `perallele_tcga_loh.parquet` were built on 2026-09-11 and are **not +scripted**. Per-allele lost = `loh_lilac` true at sample level **and**, at each heterozygous locus, +the allele whose tumour copy number is the minor one (`min(CN) < 0.5 <= max(CN)`); a homozygous +locus is never called lost, because the allele name is retained. The GDC step POSTs +`{"filters":{"op":"in","content":{"field":"case_id","value":[...]}},"fields":"case_id,submitter_id,project.project_id"}` +to `https://api.gdc.cancer.gov/cases` in batches, resolving the sample sheet's `submitter_donor_id` +(a TCGA **case UUID** for `*-US` projects, not a barcode) to the barcode. + +## Not harmonised, and why + +Four sheets of Supplementary Data 1 (`Platinum-Yoruban trio`, `TRACERx hla-I typing`, +`TRACERx LOH of HLA`, and the three `GenDX validation` sheets) are the paper's **caller-validation** +tables — 54 to 600 rows each, against orthogonal typing. Nine sheets of Supplementary Data 2 +(`metastatic GIE`, `primary GIE`, the per-gene and mutual-exclusivity breakdowns, `GIE primary vs +metastatic`) are the paper's **aggregate result** tables, every one of them a group-by over +`gie_sample.parquet`. Neither set is an input to anything here, and both are re-derivable or +re-fetchable from the source. `Supp. Table 2 - cancer type data` (58 rows) is a value-count over +`sample_metadata.parquet` and is dropped for the same reason. diff --git a/hla_loh/do_to_tcga_barcode.tsv b/hla_loh/do_to_tcga_barcode.tsv new file mode 100644 index 0000000000000000000000000000000000000000..35729e7cc1a691f0b8787f565daaf63265473a8d --- /dev/null +++ b/hla_loh/do_to_tcga_barcode.tsv @@ -0,0 +1,712 @@ +icgc_donor_id case_uuid tcga_barcode gdc_project +DO10172 58facedb-fcb8-4ecf-8338-2bfa4947acef TCGA-A6-6141 TCGA-COAD +DO10486 ddf69ed9-4c53-48b8-9f01-495bd60c3425 TCGA-AA-3514 TCGA-COAD +DO10631 f87c6dcc-1a9e-4df5-962c-1f946e2611f8 TCGA-AA-A01T TCGA-COAD +DO10908 737b35e1-d668-4fce-9b6e-76946c7952b6 TCGA-27-2523 TCGA-GBM +DO10960 9853c6bb-a42c-4698-abe5-f3c897ebc8f6 TCGA-06-0214 TCGA-GBM +DO10975 74139255-a635-4c87-814d-3dd04ed630a8 TCGA-06-0190 TCGA-GBM +DO11046 e17a6048-7a72-42c0-ad3f-97cbff02bc9f TCGA-19-2629 TCGA-GBM +DO11148 0078b0c4-68a9-483b-9aab-61156d263213 TCGA-14-1034 TCGA-GBM +DO11202 b1c4579c-72fa-4974-830c-b6a81ace45f4 TCGA-14-2554 TCGA-GBM +DO11382 01a92062-967a-4900-8dc7-a5ecd3b3f8e2 TCGA-06-0171 TCGA-GBM +DO11501 c184c3ca-7ad3-4202-b108-cb9fd5f5d947 TCGA-41-5651 TCGA-GBM +DO11782 be3a7ef3-34ed-40e1-9d9c-187940596b26 TCGA-27-2528 TCGA-GBM +DO11794 5c984433-33cf-42fc-b3ba-511efcdcab19 TCGA-06-0211 TCGA-GBM +DO11854 bf339349-062f-4ea9-a0b2-d87d3a21099e TCGA-14-1402 TCGA-GBM +DO11860 5613f5f6-086d-470b-8f77-6dcb7f8625b7 TCGA-06-0744 TCGA-GBM +DO12034 386b629e-fab1-4033-b088-45d6eeb4a13e TCGA-06-0745 TCGA-GBM +DO12070 a515cf2d-e918-4958-9bf6-e611b425a97e TCGA-14-1823 TCGA-GBM +DO12352 a2ac9937-f351-4d78-9261-264bf6c21e0c TCGA-02-2483 TCGA-GBM +DO12370 93ed7a2b-b0cb-4a84-871f-5c34a0b6a640 TCGA-19-5960 TCGA-GBM +DO12382 309005a2-93a8-4566-b8d3-6b9310144266 TCGA-19-1389 TCGA-GBM +DO12454 713190ed-c6c1-4695-814b-85ca9b95e6a0 TCGA-06-5415 TCGA-GBM +DO12484 3f960d3b-a58c-43d0-a8a4-f3555b399c9d TCGA-06-5411 TCGA-GBM +DO1249 0dca98b0-f43e-45b6-9a02-00092c78678c TCGA-D8-A27H TCGA-BRCA +DO1257 0bc5744c-5fa3-45bb-87d0-70a02068b392 TCGA-BH-A0DG TCGA-BRCA +DO1274 1549dc64-3dab-43fc-96e9-b07d520957e1 TCGA-D8-A27F TCGA-BRCA +DO12742 521ea765-1bd1-423d-a75d-091243df37a9 TCGA-02-2485 TCGA-GBM +DO1281 1b703058-e596-45bc-80fe-8b98d545c2e2 TCGA-AR-A2LK TCGA-BRCA +DO12814 ea3b5da2-6a12-400c-bf0f-e442f5ec1132 TCGA-06-0157 TCGA-GBM +DO1287 b094e8b2-8ece-4b36-8025-18073a8b873c TCGA-AO-A124 TCGA-BRCA +DO1290 b2aac45b-2073-4c7a-adb9-769a4fdcc111 TCGA-E9-A1NH TCGA-BRCA +DO1291 9fefbe7c-f66a-4940-843e-285cb7b392c1 TCGA-AR-A24Z TCGA-BRCA +DO12952 352768f9-3ce1-419c-beef-6515c78f5d7a TCGA-19-2620 TCGA-GBM +DO1301 18eb4dfc-556f-4bf3-a411-4780209ed1e0 TCGA-BH-A0B3 TCGA-BRCA +DO13132 f45f4a30-6b4c-4f15-9140-959d6a25a45f TCGA-27-1831 TCGA-GBM +DO13144 6c5154d2-af36-492f-b520-d925528824e4 TCGA-32-1970 TCGA-GBM +DO13192 3258cb3b-f63b-463c-b2e4-d638149157c5 TCGA-06-0145 TCGA-GBM +DO1328 6fa2a667-9c36-4526-8a58-1975e863a806 TCGA-AC-A2BK TCGA-BRCA +DO13318 6c28f086-6a25-40b6-93eb-bba0014acda6 TCGA-06-2557 TCGA-GBM +DO13348 c931f3bd-74c5-4ebd-bc0f-c7c6becd25ab TCGA-06-0686 TCGA-GBM +DO13474 2c5fa2d4-8e35-42e4-8bca-9fb3371a19c8 TCGA-26-5132 TCGA-GBM +DO13510 8da3103e-3e6c-4176-a583-d5fe5e60601e TCGA-06-0125 TCGA-GBM +DO13575 b37d944c-c65c-46d6-bb8b-3eb37cb85b68 TCGA-06-2570 TCGA-GBM +DO1384 20e8106b-1290-4735-abe4-7621e08e3dc8 TCGA-BH-A0E0 TCGA-BRCA +DO13869 de7b7cac-f094-4d59-8651-e991e34ea093 TCGA-06-0210 TCGA-GBM +DO1392 8785012f-f73e-4d68-87cf-1d804af32782 TCGA-A7-A13D TCGA-BRCA +DO13971 054f472f-98cb-4559-b2e2-b5f800fc8eef TCGA-06-0221 TCGA-GBM +DO14049 2f4a127f-101a-4192-b3e8-f9be2c8648bc TCGA-19-2624 TCGA-GBM +DO14171 f030e410-2585-461e-ad23-9fdf026ac06c TCGA-CR-7382 TCGA-HNSC +DO14192 5fce954c-11b9-41f4-9471-4b5bdc0efa7f TCGA-CR-6491 TCGA-HNSC +DO14204 d5eedfb5-4fa3-4a1e-ad0d-ea550b22a5ba TCGA-CV-6433 TCGA-HNSC +DO14232 70741b73-9683-42bd-87fc-6a021a70103b TCGA-BA-5556 TCGA-HNSC +DO14236 3e5f451a-5882-4914-ae1a-95c898c2bcca TCGA-BA-6872 TCGA-HNSC +DO14238 05fbd98c-28bd-41bc-ae60-16ac0c16c723 TCGA-CV-7090 TCGA-HNSC +DO14242 bf6b931e-cb2e-4cc0-b603-e9fc71fdc509 TCGA-BA-6873 TCGA-HNSC +DO14250 fafd6f5b-1d76-4537-bd1c-e0bd7b4e2166 TCGA-BA-6869 TCGA-HNSC +DO14264 c1f50a22-38df-41cc-a1f4-f7985504a7ac TCGA-BB-4225 TCGA-HNSC +DO14282 9f89510c-ed07-471f-b35e-7c87c237b9fe TCGA-CR-7385 TCGA-HNSC +DO14296 ba9d4568-71a8-41fc-a2be-0bf784e36d6a TCGA-CR-6480 TCGA-HNSC +DO14343 d89b1fd6-bef4-4803-8ed3-3b442be600b6 TCGA-CV-5432 TCGA-HNSC +DO14420 d189a6c6-210f-4957-b500-457d5ce3f867 TCGA-CR-7404 TCGA-HNSC +DO14566 4075717b-2240-4842-a00a-0c4dbcb910c4 TCGA-CV-6961 TCGA-HNSC +DO14582 c7df3466-b9a7-4818-883b-d0cd08483570 TCGA-CR-5249 TCGA-HNSC +DO14614 e8ac39e6-377a-412c-b18d-f6c4fc654ec6 TCGA-HD-7753 TCGA-HNSC +DO14678 3c975b68-4a6e-4c97-a8d1-b9a427d4957c TCGA-CV-5442 TCGA-HNSC +DO14686 cd032ddb-55f2-4c77-8dcf-e4e630f7de6f TCGA-CR-6470 TCGA-HNSC +DO14718 534f85f2-b9eb-45be-87d5-c3dffd13d541 TCGA-CV-7255 TCGA-HNSC +DO14742 d2ca9a4d-c14c-40c0-871c-88794a02e4aa TCGA-CV-7180 TCGA-HNSC +DO14750 e303b8c4-6428-4134-bb9d-a95dc1d8e35e TCGA-CQ-6228 TCGA-HNSC +DO14814 f33aa93f-b32d-4c4f-9a71-1e5ab69c6899 TCGA-CN-6989 TCGA-HNSC +DO14862 cb99fb14-641c-4fb3-bf1d-17efb8cd982a TCGA-BA-5149 TCGA-HNSC +DO15014 8fc1f1be-d2d5-4b3a-9973-f4d964018beb TCGA-CV-5973 TCGA-HNSC +DO15022 4cffea0b-90a7-4c86-a73f-bb8feca3ada7 TCGA-CN-5365 TCGA-HNSC +DO15046 ad6c9d09-2c03-4786-a72d-dd2aa5f603d4 TCGA-CX-7086 TCGA-HNSC +DO15110 1f744ada-6de1-446d-b62c-76a7fc8b40af TCGA-CN-4737 TCGA-HNSC +DO1537 a482fc4c-ff1d-46a0-a968-7df06422cc4b TCGA-B6-A0WX TCGA-BRCA +DO15398 6909302d-358a-46cd-8052-eedb09c83c83 TCGA-CN-6994 TCGA-HNSC +DO15414 53d9a130-4a3a-4491-85cd-9a44415a1632 TCGA-DQ-5625 TCGA-HNSC +DO15430 d7641ca1-a062-4dd2-9414-447af7311a84 TCGA-CN-5374 TCGA-HNSC +DO15503 90a8d2cb-1f31-45f9-8575-eed3cb9a7798 TCGA-CV-7100 TCGA-HNSC +DO15591 64bb5550-2735-4401-a0db-58ec1020a32d TCGA-CV-5431 TCGA-HNSC +DO15791 e7d49212-e9cb-4cd7-923a-bdfc82a4a150 TCGA-CN-6011 TCGA-HNSC +DO15815 8c238d30-df8e-4e6b-98fc-21696269a294 TCGA-CV-6956 TCGA-HNSC +DO15870 0956bc67-dd55-4c54-a56a-f22f2856662a TCGA-CR-6472 TCGA-HNSC +DO15911 180036a2-3b56-405e-a1fe-d5932517b6c7 TCGA-CV-5443 TCGA-HNSC +DO15943 468649e7-1525-4b8a-8ab5-ddb27db5f022 TCGA-CR-6482 TCGA-HNSC +DO16015 9e178d80-fe90-4998-9dae-726c4beceef7 TCGA-CR-5250 TCGA-HNSC +DO16136 9da334ed-314c-4445-ba29-ffd9bb42403c TCGA-AK-3428 TCGA-KIRC +DO16233 900b5f21-e8fb-4a3f-b8e3-ef098716e508 TCGA-CJ-4899 TCGA-KIRC +DO16248 3f93d1cf-a307-4e29-9e6f-239e9ffaa179 TCGA-CW-6093 TCGA-KIRC +DO16269 1e417b93-837d-4110-9090-239eb82c3e20 TCGA-BP-4968 TCGA-KIRC +DO1663 c364e81c-eb1e-4870-ab37-9c661f5f2e3d TCGA-AR-A0TX TCGA-BRCA +DO16769 8aaa4e25-5c12-4ace-96dc-91aaa0c4457c TCGA-B0-5094 TCGA-KIRC +DO16999 9b84d013-713b-4571-888c-12863fdd4a52 TCGA-BP-4327 TCGA-KIRC +DO17054 fb9bafa5-7133-4955-8156-4eb6763dc8e1 TCGA-B0-5695 TCGA-KIRC +DO17074 33cf0893-0411-4758-aa90-602bfedf0850 TCGA-BP-4807 TCGA-KIRC +DO17085 f2801b21-5444-4cc3-a642-c60c6d82cd3d TCGA-CZ-5454 TCGA-KIRC +DO17373 2939c03a-6f3f-4f7b-b246-34361baadeb9 TCGA-CJ-5681 TCGA-KIRC +DO17751 9f5c0a43-d222-484a-9249-d288387dd7a1 TCGA-BP-5010 TCGA-KIRC +DO17848 cc353686-e2b6-46a1-8095-48e2bc44b11f TCGA-BP-4977 TCGA-KIRC +DO18016 577847b9-f9b6-4954-868f-3d5ab2b4f694 TCGA-CZ-5453 TCGA-KIRC +DO18056 4c325ee1-9b8e-45d4-a0a7-e47ec1eac402 TCGA-CJ-5682 TCGA-KIRC +DO18164 9ecdde9c-2179-4830-bd10-c4fa40f26066 TCGA-BP-4756 TCGA-KIRC +DO18572 2ed3296a-5b10-40d8-8af4-90dc031657cd TCGA-AK-3455 TCGA-KIRC +DO18651 4b2ec8aa-8460-440e-b75b-6b92ae7a3ffc TCGA-DV-5566 TCGA-KIRC +DO18671 781acafa-a58b-4180-bbca-c5b675d51b17 TCGA-BP-5168 TCGA-KIRC +DO18701 f5759059-c0e3-4f1a-af96-5c7197d3c33c TCGA-CZ-5987 TCGA-KIRC +DO18839 bbbce1ba-c739-43ba-b9cf-a4f746491ae3 TCGA-B2-4102 TCGA-KIRC +DO18988 1adb05d4-e1ca-4ede-81c7-ff24a243abae TCGA-B2-4099 TCGA-KIRC +DO19028 659294b9-ded9-498a-bc00-c1d4d456aa4a TCGA-A3-3363 TCGA-KIRC +DO19380 78048432-2de1-498c-9cf1-9c5bd43daa39 TCGA-CJ-4870 TCGA-KIRC +DO19733 0487fc41-386c-4d76-9084-daf959bf5e98 TCGA-CW-5585 TCGA-KIRC +DO20082 bbdaa931-e922-49be-bfbf-fa0c2ae27d7a TCGA-CW-6087 TCGA-KIRC +DO20416 7664c241-b369-4d76-9cd1-7f2bdc1d267f TCGA-BP-4326 TCGA-KIRC +DO2055 441919eb-8c9f-478c-bda3-8de4f8295e8a TCGA-C8-A12L TCGA-BRCA +DO20604 7ca06dbe-f073-4aee-86ff-a545fad5ea93 TCGA-CJ-4878 TCGA-KIRC +DO20768 a121406a-30b2-4717-ae96-edcb17d911e2 TCGA-B9-4115 TCGA-KIRP +DO20770 71b35cd8-ecee-4599-bb24-b1ef78118423 TCGA-B9-4117 TCGA-KIRP +DO2078 f6e1ec78-5ad9-4879-9b7e-262d17b166ad TCGA-BH-A18R TCGA-BRCA +DO20780 5c9a6c40-982e-45c7-98d3-aeb6cad16986 TCGA-B9-4113 TCGA-KIRP +DO20839 2272c592-b9b6-4483-9f7f-6e185699b900 TCGA-AL-3468 TCGA-KIRP +DO2084 aeb6a3a6-b57a-4c53-90e6-e833777dd2a1 TCGA-BH-A0H6 TCGA-BRCA +DO20849 4edd92e1-5da5-49a3-80e4-fa456aa425b8 TCGA-B9-4116 TCGA-KIRP +DO20863 a6e960af-3ca5-42fb-b546-451c2583282d TCGA-B9-4617 TCGA-KIRP +DO20867 75a53bff-bf0f-4ad8-9b8e-14a27914d4f2 TCGA-B3-3925 TCGA-KIRP +DO20905 a7bf785f-0539-4899-ac21-47d1fc347ee8 TCGA-AL-3473 TCGA-KIRP +DO20915 df135722-b98d-4dec-906d-77da13df39b9 TCGA-B9-4114 TCGA-KIRP +DO20923 42d044ae-1416-477d-ac80-779c49bc8d35 TCGA-AL-3472 TCGA-KIRP +DO20928 9fc40b03-0b56-4b1f-93ed-dbc455ae3fd9 TCGA-AL-3466 TCGA-KIRP +DO2096 4e6edfe6-adcb-4c12-8ff4-38a79f5887e8 TCGA-B6-A0X5 TCGA-BRCA +DO2114 96457ebf-ae88-41f4-97e5-cdc3163418aa TCGA-BH-A0H0 TCGA-BRCA +DO219106 9f45a227-1fd5-478c-a194-4a5c61b68cc3 TCGA-QG-A5YW TCGA-COAD +DO219580 25e893ad-a0fd-45ae-9aa9-9bcd7c2ec8e5 TCGA-MR-A520 TCGA-LIHC +DO219585 c7a2f394-3e3f-4c90-9f1e-f2be3e5b0d6b TCGA-DJ-A13L TCGA-THCA +DO219944 ea0e5ae7-1f15-4504-aea3-efd1f6733304 TCGA-DX-A1KU TCGA-SARC +DO219945 cd73a55f-03c1-4da0-9fa7-1ae4a263664e TCGA-DX-A1KW TCGA-SARC +DO219949 854e42bc-95b2-4b77-a4b2-ffbd6c28b281 TCGA-DX-A1L0 TCGA-SARC +DO219951 73d2e8ef-4e10-416a-845b-f48374e48ecb TCGA-DX-A1L2 TCGA-SARC +DO219952 9517413b-b5b8-4130-8ba1-d86d44abe2ec TCGA-DX-A1L3 TCGA-SARC +DO219954 b5f27002-4277-418d-90b0-cc5a2742692e TCGA-DX-A23R TCGA-SARC +DO219960 3608bfbb-2a92-4ec2-a330-bf4790a8b65c TCGA-DX-A240 TCGA-SARC +DO219961 6b189eff-b919-49d6-8775-dbf32c9ccbba TCGA-DX-A2IZ TCGA-SARC +DO219964 f75d2589-fd7b-49a1-a1c8-0b4726b81f3c TCGA-DX-A2J4 TCGA-SARC +DO219965 39f68467-3d37-4b19-aa43-ed7b120c619d TCGA-DX-A3LS TCGA-SARC +DO219966 e7a72f7e-77b1-4d95-8f37-39a6cffc35dc TCGA-DX-A3LT TCGA-SARC +DO219967 e786de34-4c21-460f-89ab-008de4347049 TCGA-DX-A3LU TCGA-SARC +DO219968 2cacd9d0-8414-4f17-8259-c9d838597f39 TCGA-DX-A3LW TCGA-SARC +DO219969 3505f91d-bc66-4732-84b6-18c5b32ca6b8 TCGA-DX-A3LY TCGA-SARC +DO219970 b85e2156-d5ac-486c-93e1-86525ea41a64 TCGA-DX-A3M1 TCGA-SARC +DO219972 8f3f6f7c-4b50-467e-a6e8-d836735913f6 TCGA-DX-A3U5 TCGA-SARC +DO219974 ea3a524b-6ee3-4904-8924-29fcaa20985c TCGA-DX-A3U7 TCGA-SARC +DO220118 7bab6742-4e3d-4fcf-9b90-d3e0218b4d79 TCGA-FX-A3NJ TCGA-SARC +DO220122 7f9031da-124a-4a38-83e6-878a50e58c24 TCGA-FX-A3RE TCGA-SARC +DO220126 12bfd446-2ee5-4664-a492-c023d134f60f TCGA-FX-A48G TCGA-SARC +DO220138 9c9b53e5-cdbe-4510-a839-4a3b5039681b TCGA-HB-A43Z TCGA-SARC +DO220140 c5b0e6be-48b5-40e5-9d90-f9112050aa88 TCGA-HB-A5W3 TCGA-SARC +DO220153 c5e23947-253b-40d2-916a-277b2a580be2 TCGA-IE-A4EI TCGA-SARC +DO220157 590fdb76-20ef-4f95-9783-7d80e7a25e5a TCGA-IE-A4EK TCGA-SARC +DO220163 4e8c570c-3cbe-4878-8b2c-b966868dd60d TCGA-IF-A4AJ TCGA-SARC +DO220168 5a64f7e9-3833-4a3c-962b-13eb3e305c69 TCGA-IS-A3K7 TCGA-SARC +DO220172 257dc623-41e1-4fa3-859c-c787db10bd60 TCGA-IS-A3KA TCGA-SARC +DO220174 0da5ee6c-d012-4a57-9992-9d74937aaa5f TCGA-IW-A3M4 TCGA-SARC +DO220176 c20682a1-f340-430d-99f9-077fe093ed19 TCGA-IW-A3M5 TCGA-SARC +DO220221 8654c2b5-27d7-43c5-9a39-7fdefd7a0324 TCGA-MO-A47R TCGA-SARC +DO220331 83b3060c-2449-4581-88ef-817d126e4525 TCGA-KL-8323 TCGA-KICH +DO220333 6101ffe6-2ef6-4256-9d6b-a7c545836995 TCGA-KL-8325 TCGA-KICH +DO220334 03f3dd32-e1d1-485b-a968-3f55798f4d46 TCGA-KL-8326 TCGA-KICH +DO220335 02979422-5149-4750-ad5f-483e0bec6ac5 TCGA-KL-8327 TCGA-KICH +DO220336 452135f2-6de6-4593-a091-ddf6344ee431 TCGA-KL-8328 TCGA-KICH +DO220338 0f3c4f8f-0766-46b0-81b4-18ca4f6ea20a TCGA-KL-8330 TCGA-KICH +DO220339 8ac7f218-bd7b-4dda-8879-2cec49748e42 TCGA-KL-8331 TCGA-KICH +DO220340 3ec3bcb3-492c-4942-be02-76a67f57e38f TCGA-KL-8332 TCGA-KICH +DO220341 02ec7a3a-2812-4a65-93cd-07bf655cc91b TCGA-KL-8333 TCGA-KICH +DO220342 9aabcbd7-a26a-49b1-abf1-02c0dc3adcf2 TCGA-KL-8334 TCGA-KICH +DO220348 3129e869-bb2f-423a-bd20-d0daae33b11a TCGA-KL-8340 TCGA-KICH +DO220349 70e306ee-9584-49d6-81a1-a49d837ab6aa TCGA-KL-8341 TCGA-KICH +DO220350 c689ae1d-4a6b-45db-b4d1-6b34c5c61522 TCGA-KL-8342 TCGA-KICH +DO220351 b4a0bf74-1381-4c17-b97d-206810d9e59a TCGA-KL-8343 TCGA-KICH +DO220352 a5af1391-89d3-4f41-930e-7e8272afec98 TCGA-KL-8344 TCGA-KICH +DO220354 a122d361-835e-4cf2-850f-af7864aa112c TCGA-KL-8346 TCGA-KICH +DO220355 23656226-2d99-4548-959f-5c2eed803e33 TCGA-KM-8438 TCGA-KICH +DO220356 739deed4-d1dc-458f-bd0f-1e24d2bfd305 TCGA-KM-8439 TCGA-KICH +DO220357 872092b3-d31e-44d7-bd03-e29f52f8ab5a TCGA-KM-8440 TCGA-KICH +DO220358 f31ba8ec-a3dd-4dd8-8d15-f6c5975ebe44 TCGA-KM-8441 TCGA-KICH +DO220359 40f79cfe-a51f-497b-b06e-7a4f9004f092 TCGA-KM-8442 TCGA-KICH +DO220360 2f1c4db6-88a7-4a5d-a5fe-94d85f569440 TCGA-KM-8443 TCGA-KICH +DO220361 4e4faf51-2cbf-4910-a093-186660e4a19b TCGA-KM-8476 TCGA-KICH +DO220362 9680b5ff-1882-434f-9981-e8ac33764b93 TCGA-KM-8477 TCGA-KICH +DO220363 2ea84e23-6a0a-4c5d-9c70-d5355a85df35 TCGA-KM-8639 TCGA-KICH +DO220423 f3ba1cf8-ac67-4da8-96f3-8680dc55f247 TCGA-KN-8418 TCGA-KICH +DO220433 dc04b82c-e1e8-4917-ba19-c2fea4a1b69a TCGA-KN-8419 TCGA-KICH +DO220439 2fc175d1-9d95-44fd-a53e-820f6f404551 TCGA-KN-8421 TCGA-KICH +DO220449 4a88bc85-4c6e-4fac-8dbb-d3fc5f36b35e TCGA-KN-8422 TCGA-KICH +DO220468 8c9e1f60-1838-4aa4-a667-c1cef7dddbdf TCGA-KN-8424 TCGA-KICH +DO220474 cc316faf-dcb1-4dea-b8b0-a17d000ff40f TCGA-KN-8425 TCGA-KICH +DO220478 b08dfba8-6afb-4217-9259-72be6f1f3363 TCGA-KN-8426 TCGA-KICH +DO220479 955d428c-8fe8-456b-b86d-65ea6c2a1b18 TCGA-KN-8427 TCGA-KICH +DO220509 04e9a63a-d742-488c-8564-bf9d79aa2091 TCGA-KN-8431 TCGA-KICH +DO220510 af1111af-9384-4e37-9b16-227d38007ebd TCGA-KN-8432 TCGA-KICH +DO220519 e1e15226-9f8e-4035-8f08-eb1fed57080e TCGA-KN-8434 TCGA-KICH +DO220525 f2b58595-dec0-4194-ab33-7c7dc6c058a6 TCGA-KN-8435 TCGA-KICH +DO220535 f1c1502e-1bd6-425a-b651-454363809668 TCGA-KN-8437 TCGA-KICH +DO220549 d52c3cb6-921e-4f26-b49c-e6e4f11f9c06 TCGA-KO-8405 TCGA-KICH +DO220554 9aa02d46-294b-4e99-b9ee-56edcbfb5275 TCGA-KO-8406 TCGA-KICH +DO220556 151f07f5-1529-4fcd-912c-e55b2ff8abef TCGA-KO-8407 TCGA-KICH +DO220577 ee8c1919-17a9-4df1-8aa5-79546621b23c TCGA-KO-8411 TCGA-KICH +DO22106 799e9106-e62e-4681-8861-137f474ee8b8 TCGA-HT-7602 TCGA-LGG +DO22117 01c4ca12-ac08-41dd-9e4f-bfb0971688a2 TCGA-DU-7009 TCGA-LGG +DO22145 fc3b7596-f515-446f-81db-fed0154ca2c5 TCGA-EZ-7264 TCGA-LGG +DO22195 400b2fa3-baf7-4baa-99b4-88b9f25b5049 TCGA-E1-5318 TCGA-LGG +DO2222 49717f75-0f2d-4e1c-9a12-f1cd7877b80a TCGA-AN-A04D TCGA-BRCA +DO222303 67325322-483f-443f-9ffa-2a20d108a2fb TCGA-FF-8046 TCGA-DLBC +DO222304 f978cb0f-d319-4c01-b4c5-23ae1403a106 TCGA-FF-8047 TCGA-DLBC +DO222305 1f971af1-6772-4fe6-8d35-bbe527a037fe TCGA-FF-8041 TCGA-DLBC +DO222306 f0a326d2-1f3e-4a5d-bca8-32aaccc52338 TCGA-FF-8042 TCGA-DLBC +DO222307 c8cde9ea-89e9-4ee8-8a46-417a48f6d3ab TCGA-FF-8061 TCGA-DLBC +DO222308 a8e2df1e-4042-42af-9231-3a00e83489f0 TCGA-FF-8043 TCGA-DLBC +DO222309 0e9fcccc-0630-408d-a121-2c6413824cb7 TCGA-FF-8062 TCGA-DLBC +DO22285 6a0bcf0c-fa4c-4119-99d2-f722b781d20f TCGA-FG-8182 TCGA-LGG +DO22293 648252da-dd86-48d1-ae75-4257e3142b0b TCGA-FG-5964 TCGA-LGG +DO22295 70939245-41e0-4845-a473-5fef719b9828 TCGA-DU-6407 TCGA-LGG +DO22324 42251e02-b687-4bf1-a1b6-e3dd978542e4 TCGA-HW-7487 TCGA-LGG +DO22356 f86fa219-34e9-4812-8edf-2b66d7219779 TCGA-CS-5395 TCGA-LGG +DO22382 84ae73e4-3b39-4444-9bfb-3fc6057eab32 TCGA-DU-7301 TCGA-LGG +DO22417 5017c115-39b8-4c5c-8be6-2cb298d7c641 TCGA-HT-7695 TCGA-LGG +DO22490 97cb8701-7d63-4142-b404-006fbb3ecb0b TCGA-DU-6401 TCGA-LGG +DO22496 010aac75-3bfe-4bf2-b866-af0f2d92f125 TCGA-DB-5278 TCGA-LGG +DO22524 1cc53ade-73c0-492b-a7dc-328269fd0e66 TCGA-DU-5874 TCGA-LGG +DO22536 84569c1a-6baa-4380-ae3c-707df1be4618 TCGA-E1-5319 TCGA-LGG +DO22548 a5d3f3fb-6541-419d-b47e-720d438f1bff TCGA-CS-6668 TCGA-LGG +DO22720 a5f851c5-6130-4ea1-9f86-be59331a92f7 TCGA-IK-7675 TCGA-LGG +DO22736 55a67639-ed05-4143-af96-a377bc5489ee TCGA-HW-7486 TCGA-LGG +DO22836 68015cc7-d3a1-4c90-b8fb-5d14143d9960 TCGA-DD-A1EB TCGA-LIHC +DO22838 728aedb3-551a-4c16-9be3-7733a9539e9e TCGA-ES-A2HS TCGA-LIHC +DO22839 39eb1db0-2722-4af4-8254-d0e10256c64d TCGA-G3-A3CK TCGA-LIHC +DO22854 b7d6621c-8bce-4a01-8588-9d7761d9372b TCGA-EP-A26S TCGA-LIHC +DO22860 ca8afb2c-be3d-41e8-8dbc-19fe7ab6154b TCGA-CC-5260 TCGA-LIHC +DO22866 f435ec00-0db8-46b4-96bf-890a9d931df3 TCGA-G3-A25V TCGA-LIHC +DO22875 1ad7f17e-b3b8-40e9-aab8-5a17e0aec408 TCGA-CC-5261 TCGA-LIHC +DO22884 f6e5979d-d6a9-4408-bd1b-1b37d1064e83 TCGA-G3-A25T TCGA-LIHC +DO22887 dfa15fbe-bb14-40ce-afc6-70694112e3e2 TCGA-BC-A10Q TCGA-LIHC +DO22896 6dcfc418-ab28-4365-95ea-c1ae254f2341 TCGA-BC-A217 TCGA-LIHC +DO22914 bfd20d71-18a5-4f01-8d56-84378ad92359 TCGA-FV-A3I1 TCGA-LIHC +DO22929 4bbdbe77-f3cb-4b64-9826-72debb734784 TCGA-BC-A216 TCGA-LIHC +DO22932 18c08546-bed3-4ed9-b223-479afe633c8e TCGA-EP-A2KB TCGA-LIHC +DO22935 f15279dd-5d12-460b-9bd1-b4f95681ca65 TCGA-EP-A3RK TCGA-LIHC +DO22953 1b428ada-eb3f-40c2-b4bf-1bbf6b61729d TCGA-CC-5262 TCGA-LIHC +DO22956 7dd3bda9-3a0a-46fa-b052-a6d24b847804 TCGA-ES-A2HT TCGA-LIHC +DO22959 3acf7438-fd65-442c-8a30-68b6714537f3 TCGA-DD-A1EI TCGA-LIHC +DO22970 06ac44e9-1226-4202-828b-6f224766d7ac TCGA-G3-A25W TCGA-LIHC +DO22978 3dd7c93d-8194-44ee-8014-10dd81ef6441 TCGA-FV-A3R3 TCGA-LIHC +DO22986 cab5b264-abab-4bdb-816e-9d2c0ac30629 TCGA-DD-A1EH TCGA-LIHC +DO22994 52292ffc-0902-4d97-b461-20723987a177 TCGA-DD-A1EJ TCGA-LIHC +DO22998 0ff6364a-e6f4-47ca-a7e7-14705bf0a8df TCGA-DD-A1ED TCGA-LIHC +DO23018 b865dec4-f051-4fbe-9405-f832ff2010d7 TCGA-G3-A25S TCGA-LIHC +DO23023 440375d7-cdcd-418a-8992-0a8ca42d290d TCGA-EP-A2KA TCGA-LIHC +DO23028 670c1cbc-4494-45f1-bb8b-18db82d4f7e0 TCGA-CC-A1HT TCGA-LIHC +DO23048 2854ffe8-1f77-497a-bf49-11e0189fee35 TCGA-DD-A1EL TCGA-LIHC +DO23073 239b4298-b090-480a-a19f-9758a827aa30 TCGA-G3-A25Y TCGA-LIHC +DO2323 96312510-c126-485d-8109-ed81844a1dc3 TCGA-AN-A0G0 TCGA-BRCA +DO2341 d9dc3b59-613d-469e-8b4f-6c5a557eb26a TCGA-EW-A3U0 TCGA-BRCA +DO23584 97831d28-ab41-4c18-bfc2-c4c6bc757d13 TCGA-91-6847 TCGA-LUAD +DO23608 d5326429-9805-47f9-97b0-fbda658e3f01 TCGA-91-6840 TCGA-LUAD +DO23610 ebcba7f2-ce13-4bae-97cd-91a6b1dcd465 TCGA-50-5932 TCGA-LUAD +DO23648 df576520-a6b6-4c9b-8d06-3f59cc5342fd TCGA-75-7030 TCGA-LUAD +DO23651 028e99e9-5b9a-4954-bb6e-6d4709a3cea8 TCGA-55-6986 TCGA-LUAD +DO23687 f063ddbb-1668-40df-9ec2-b0a23ca2c389 TCGA-55-6984 TCGA-LUAD +DO23702 9b132e4f-7e35-4cc5-8711-43ad62b906d0 TCGA-50-6591 TCGA-LUAD +DO23735 a1e65587-24c1-4b41-92a7-4e1f15fffd78 TCGA-49-4512 TCGA-LUAD +DO23744 bf15f7ad-9d92-473b-91d1-f24aa373ab97 TCGA-44-2659 TCGA-LUAD +DO23786 61c655ec-52b5-453f-a6cc-b2aba445b027 TCGA-49-4486 TCGA-LUAD +DO23838 9f81c602-8afa-4588-b0b6-6e5a1a128d5a TCGA-05-4396 TCGA-LUAD +DO23942 63bd2175-4b7c-44c9-aef3-9efc8f79837b TCGA-05-5429 TCGA-LUAD +DO24130 397d3f69-1453-4057-b177-8723eec923d1 TCGA-97-8171 TCGA-LUAD +DO24307 a3de401d-91fe-49a2-bb07-81c1a06506e6 TCGA-05-4389 TCGA-LUAD +DO24467 cd0dc947-b708-464e-beb7-954c9d4583e7 TCGA-73-4659 TCGA-LUAD +DO24497 d2824e6d-3784-45c2-9b0f-52b17356b5da TCGA-75-5147 TCGA-LUAD +DO24646 1ab7ad70-0f80-41d8-8efa-3baebc8223cc TCGA-64-1680 TCGA-LUAD +DO24675 3d2aa654-1b5f-4eb4-a1c2-af31f5760069 TCGA-05-4420 TCGA-LUAD +DO24729 501c987e-d1eb-48a9-89eb-72a5062c90b4 TCGA-78-7158 TCGA-LUAD +DO24860 46592b7b-6968-42a6-83af-0917c9f4a9a5 TCGA-78-7535 TCGA-LUAD +DO25010 c4d1e105-28b7-48df-abef-fbe09782fdb2 TCGA-55-6972 TCGA-LUAD +DO2503 74f31744-0aed-4633-beec-4e203315f0b7 TCGA-BH-A0DT TCGA-BRCA +DO25068 dbd5b0de-94c9-45dd-afb3-6820a7ecaca2 TCGA-67-6215 TCGA-LUAD +DO2509 5fd9552a-c742-4388-940d-295d1107ae00 TCGA-AR-A1AY TCGA-BRCA +DO25104 f77cf4d3-9829-4fd3-95a1-b29aed9e94b8 TCGA-64-1678 TCGA-LUAD +DO25189 368e23f0-e573-4547-bf5a-14080baf737b TCGA-50-5930 TCGA-LUAD +DO25320 6dfd47d2-831a-4386-9051-f78199a16bb5 TCGA-05-4397 TCGA-LUAD +DO25444 762dea8a-5b41-4058-979a-b7876ed13d7e TCGA-05-4395 TCGA-LUAD +DO2551 4da999a0-ef41-4a0b-b1d1-446b39cc855a TCGA-EW-A1PH TCGA-BRCA +DO25622 21fb46f9-4bbb-441c-af19-a687e9138344 TCGA-49-6742 TCGA-LUAD +DO25737 8994c8c4-9ea0-46cb-b4a5-0055d7da5bfa TCGA-44-6148 TCGA-LUAD +DO25797 0df573ee-28f0-4244-b434-09e6ca59fbf0 TCGA-67-3771 TCGA-LUAD +DO25847 31c96e35-5e2f-429c-b12a-7bc5a497a300 TCGA-05-4398 TCGA-LUAD +DO25887 a8d6694c-a213-4544-ac0b-63bce16d8f4e TCGA-75-6203 TCGA-LUAD +DO2593 993103b1-e5a1-4c33-8629-be53ebc41d64 TCGA-A8-A094 TCGA-BRCA +DO25985 659668e8-f0d9-4ff2-bbc8-9246f2ef49ab TCGA-55-7281 TCGA-LUAD +DO26062 0d66bf6c-eed0-4726-bd5b-3bf6d610b4e0 TCGA-50-6597 TCGA-LUAD +DO26230 4c31127e-d095-4978-9dac-35153c27f6ed TCGA-73-4666 TCGA-LUAD +DO26280 35cb7841-9b09-465a-90c5-e3b8a9faad49 TCGA-55-6982 TCGA-LUAD +DO2629 0a2a3529-f645-4967-9a58-89ee20b8bb62 TCGA-E2-A15E TCGA-BRCA +DO26354 cc4bd56a-25c5-4c48-b583-ac3aeb778ca6 TCGA-38-4628 TCGA-LUAD +DO26427 9a874b64-d0d6-416e-97bc-e9071ed0b16b TCGA-18-3415 TCGA-LUSC +DO26434 7d905f8f-3967-4ea8-96c8-17b1b03fbec3 TCGA-43-5670 TCGA-LUSC +DO26445 9293e197-e38a-4e19-a7d0-1b45d1ad48bd TCGA-43-3920 TCGA-LUSC +DO26459 395babb3-3f5d-4e71-a675-af4443f23028 TCGA-85-8052 TCGA-LUSC +DO26461 3edf933a-be1a-46e8-bfb0-acdff30e64a0 TCGA-56-1622 TCGA-LUSC +DO26463 0e2ee54a-51c9-4868-842d-a2a1c1cfb016 TCGA-21-1082 TCGA-LUSC +DO26469 fd9ee494-65fe-4de4-adff-7952a059b17f TCGA-60-2698 TCGA-LUSC +DO2647 27b05b15-a44b-45ed-a6e3-e7d1ca488ea9 TCGA-A8-A09X TCGA-BRCA +DO26475 aecf85cc-058c-46f3-9cdc-3573ac3b8438 TCGA-94-7943 TCGA-LUSC +DO26476 e58e8850-154f-4695-bee0-005c76410327 TCGA-60-2713 TCGA-LUSC +DO26492 4ca13c92-84b4-4edf-842a-b20b7e713415 TCGA-96-7545 TCGA-LUSC +DO26512 3666bc65-8e40-409e-9a1f-41583dd6d978 TCGA-68-7755 TCGA-LUSC +DO26540 a3e1ac67-a1f2-44fb-8343-a7e8239fc24a TCGA-60-2724 TCGA-LUSC +DO26553 305f72a8-069b-410b-bbe4-4ecb761c748d TCGA-22-5485 TCGA-LUSC +DO26563 e7d1b042-5b1c-4c29-b501-a5ef4810ecd2 TCGA-34-5240 TCGA-LUSC +DO26572 d4bc755a-2585-4529-ae36-7e1d88bdecfe TCGA-18-3408 TCGA-LUSC +DO26596 ce8612ab-3149-4a6a-b424-29c0c21c9b8b TCGA-43-3394 TCGA-LUSC +DO26599 220be009-8f1e-4a77-ab04-3220b5f87ea6 TCGA-21-1076 TCGA-LUSC +DO26653 0ab8d063-62b4-4d47-82aa-e3351a60029d TCGA-21-1078 TCGA-LUSC +DO26692 9dff11cd-0784-4a60-9c77-43f8957d3dfa TCGA-90-7767 TCGA-LUSC +DO26710 da93a143-5799-4856-a5f4-3ff3b9284311 TCGA-66-2795 TCGA-LUSC +DO26719 8e6020a1-68df-4d57-ad4e-33acc0861fb3 TCGA-85-8277 TCGA-LUSC +DO26769 0398eae1-7216-4595-80a5-6b117d96e070 TCGA-21-1083 TCGA-LUSC +DO26798 6fd72426-f6c8-47ca-a500-d5d3600b9b15 TCGA-60-2722 TCGA-LUSC +DO26806 1f6b2aca-7357-40d1-ba7a-99227d9900a2 TCGA-37-4135 TCGA-LUSC +DO26894 bd15f523-45c4-4d7c-a3cc-4fb56abb0e54 TCGA-22-5477 TCGA-LUSC +DO26934 adc853b1-b8bf-488c-a1e2-e95603459b55 TCGA-98-8022 TCGA-LUSC +DO2694 2a84997d-ccee-4f46-bea2-752534f26416 TCGA-E2-A1LL TCGA-BRCA +DO26957 dc6ceece-efaa-4b9d-9e12-d35863d7a902 TCGA-92-8064 TCGA-LUSC +DO26971 037c57d1-b4a5-45dc-bda4-0550461d321b TCGA-22-1016 TCGA-LUSC +DO26976 b5e2cbda-bbfa-4ef8-a9c4-cb978bef9b23 TCGA-56-7582 TCGA-LUSC +DO26987 1f11de87-a5fe-49fd-80c1-e279b2bc69de TCGA-60-2711 TCGA-LUSC +DO26993 7b982d5e-3a7d-40ac-bd25-6044c62879b6 TCGA-68-8250 TCGA-LUSC +DO27025 1ee543d5-b8c0-4f79-8373-6bb6319f2ee2 TCGA-66-2744 TCGA-LUSC +DO2706 eadf8482-e60e-4307-adb7-d5c3b9fa6cae TCGA-A2-A25B TCGA-BRCA +DO27067 81b7cbc1-c037-426f-95b0-d729a30697da TCGA-52-7812 TCGA-LUSC +DO27081 da61857a-b00e-4a19-ab02-8fbb2ee56bea TCGA-34-2600 TCGA-LUSC +DO27104 19f0cb8c-2e57-4310-967f-a9890f1605db TCGA-18-4721 TCGA-LUSC +DO2712 566792ae-f853-4a47-856d-f02cdcfcb18a TCGA-A8-A07I TCGA-BRCA +DO27120 e21964b8-47c7-4a5f-bfa7-7206222883d2 TCGA-21-5782 TCGA-LUSC +DO27128 43abe847-4ba7-466e-8283-5d7b80b999a7 TCGA-66-2756 TCGA-LUSC +DO2719 1d27253f-b036-44e7-a04d-8da5bbf57419 TCGA-E2-A1LK TCGA-BRCA +DO27208 422a46b2-a67c-4a7e-923f-9b651ced96f8 TCGA-22-5492 TCGA-LUSC +DO27279 87ac0ad4-8c9d-409e-9a86-1d201d01769d TCGA-66-2789 TCGA-LUSC +DO27318 376dfd27-68e8-4a1a-9c4f-5064279b2a9e TCGA-33-4586 TCGA-LUSC +DO27326 e6b72c24-1607-43b9-8b8a-7bf83eea5895 TCGA-66-2759 TCGA-LUSC +DO27334 9af6ed4e-8cdc-4f49-84e9-ba1053b5b3ca TCGA-66-2793 TCGA-LUSC +DO27453 585e6487-b0a3-4828-8a06-46bee01dff74 TCGA-60-2719 TCGA-LUSC +DO27484 3d1f4059-2220-45b4-a4d2-b14f76cec96a TCGA-60-2726 TCGA-LUSC +DO27588 c2598334-f866-4cc3-93ec-2d2e8b85d319 TCGA-77-6843 TCGA-LUSC +DO27636 ee850103-6314-41a3-9734-8b520de95b40 TCGA-77-7139 TCGA-LUSC +DO27747 b913d254-8307-4b8a-8313-d978e32bb38f TCGA-66-2766 TCGA-LUSC +DO2783 05506f4c-e701-4a9d-ae06-97f066aade43 TCGA-AN-A0AT TCGA-BRCA +DO27978 13f5814c-1f99-4ffa-84a4-3bbd8979faae TCGA-13-1477 TCGA-OV +DO28004 c8febeef-8e7e-459e-88cb-8086692dc559 TCGA-13-1491 TCGA-OV +DO28024 15556b28-c6bd-455a-87b6-4b7b3e33d0e4 TCGA-25-1634 TCGA-OV +DO28053 5c127332-5ca0-45f1-a5ac-4876ad94e491 TCGA-25-2400 TCGA-OV +DO28089 f34aa3b6-e966-49c6-bc55-130545772c53 TCGA-24-1103 TCGA-OV +DO28093 f0c353fd-947c-41e2-b643-3ecc0d69796c TCGA-04-1347 TCGA-OV +DO28119 6d10d4ee-6331-4bba-93bc-a7b64cc0b22a TCGA-04-1331 TCGA-OV +DO28233 a5030259-cf9c-4a58-8710-b9da8ee59320 TCGA-10-0937 TCGA-OV +DO28273 ee0a4a13-613e-4c5d-96c3-8083a013702d TCGA-13-0890 TCGA-OV +DO28289 58d34254-4f5b-40a4-9e9f-7160062fb2a4 TCGA-13-0912 TCGA-OV +DO28412 4c18d9cf-4af4-4a86-8b1c-f78795fbbd7e TCGA-24-1562 TCGA-OV +DO2842 359f12f9-5c41-48a4-85bc-fd7e307bf7d8 TCGA-EW-A1P8 TCGA-BRCA +DO28521 c435627c-159d-4a6d-a819-30abac24bf4d TCGA-25-1632 TCGA-OV +DO28763 e9483296-cb91-497a-b955-39a3c3289dac TCGA-04-1514 TCGA-OV +DO28823 90f4b65c-cfd4-4066-a5b9-842885c172e2 TCGA-13-0727 TCGA-OV +DO28952 21f6bf91-dfc6-4a59-8b76-88a0f95c7b47 TCGA-24-1544 TCGA-OV +DO28962 d3164236-c14a-4230-b527-ecd1a3992f02 TCGA-24-1614 TCGA-OV +DO29146 7ca97692-0bf5-4bbd-81ce-10a051d04bd5 TCGA-10-0938 TCGA-OV +DO29308 4b930a10-4b12-4428-84f9-3255b4a3bc4f TCGA-04-1349 TCGA-OV +DO29398 9fb1ba57-2007-4477-b000-2d36f163efd2 TCGA-23-1110 TCGA-OV +DO29760 c9e28934-8379-4511-817c-d787f2c4ca3a TCGA-13-1487 TCGA-OV +DO29850 098acb76-0bf5-44e5-bcae-f919cf5fa5e5 TCGA-24-1466 TCGA-OV +DO29940 29949b65-2913-4c86-bbd7-b85c5df6f15e TCGA-61-2000 TCGA-OV +DO29980 ff530f28-0ec0-4494-bb54-44bb055bae1c TCGA-13-0723 TCGA-OV +DO30060 ba68f2cf-9271-41fd-9655-1fac7681f588 TCGA-24-1558 TCGA-OV +DO30090 2b14123b-8fcd-402c-9399-4e7c47f20252 TCGA-24-2290 TCGA-OV +DO30220 700e91bb-d675-41b2-bbbd-935767c7b447 TCGA-23-1118 TCGA-OV +DO30340 eeb9d147-608d-4692-8adf-2f601d23a8ff TCGA-36-1574 TCGA-OV +DO3037 784de7ac-8424-42eb-83d4-a1bebaa42b97 TCGA-A1-A0SM TCGA-BRCA +DO30550 5e18b17d-4626-4b6d-8ac6-e560cee0376c TCGA-24-1419 TCGA-OV +DO30750 7dcc809b-e33a-4453-b92a-c00786f48cb0 TCGA-24-1557 TCGA-OV +DO30970 30b8f4cd-9245-4496-a8a8-c3e59093bc0a TCGA-04-1367 TCGA-OV +DO3110 8bd68696-5eeb-46d7-9dc6-c39c849029a4 TCGA-A8-A08L TCGA-BRCA +DO31150 e43d3769-e25f-4fb7-9080-ea26defaf094 TCGA-13-0906 TCGA-OV +DO3140 2779fa01-ac93-4e80-a997-3385f72172c3 TCGA-A8-A08S TCGA-BRCA +DO31420 66c92b9e-de3c-4d1a-bb69-46ffbc6caf33 TCGA-24-1548 TCGA-OV +DO31551 62379be5-13f0-474b-94d3-6f944ec4ee96 TCGA-24-1552 TCGA-OV +DO3158 b5b99291-507e-4b68-a039-9a0f571f55df TCGA-AO-A12H TCGA-BRCA +DO31729 0c1a2e7d-e7e4-481e-a012-ef214c444497 TCGA-04-1542 TCGA-OV +DO31769 f5033d52-ec36-4e67-88d8-b6d898b81b2f TCGA-25-2391 TCGA-OV +DO3182 a824b3bd-34d5-4cc1-a92f-f9d6ac0f1814 TCGA-C8-A12Q TCGA-BRCA +DO31869 3c037acf-f453-4513-a6dd-129163ddde2a TCGA-09-2050 TCGA-OV +DO3188 1557183c-aedd-4df6-b98a-3654667f7b69 TCGA-A8-A09I TCGA-BRCA +DO32237 0a48873a-092a-4b25-822b-b0b3c18c08a8 TCGA-24-2024 TCGA-OV +DO32391 4eac2c98-86d2-4ee6-a1d3-157d013c78dc TCGA-09-1666 TCGA-OV +DO32420 3768d34f-1527-40db-b116-cd80cee5ec3a TCGA-09-2045 TCGA-OV +DO32596 8a6d2ce3-cc57-451b-9b07-8263782aa23f TCGA-23-1124 TCGA-OV +DO32646 206ce0ed-36a4-47ab-ac8e-f6fca6ac5c18 TCGA-24-0982 TCGA-OV +DO3412 f51af6db-2655-47fb-9ffc-83a503a728ea TCGA-AN-A0XR TCGA-BRCA +DO3458 1174f6e4-ffbe-4e59-a000-8d861c968369 TCGA-AO-A03L TCGA-BRCA +DO3482 08da7c4c-3067-4bcf-9d7a-78566df72e69 TCGA-B6-A0RE TCGA-BRCA +DO3500 2d29a4ac-98e7-4663-9dd6-5681bc32ac2e TCGA-A2-A3KC TCGA-BRCA +DO3614 d67cd793-2931-429a-9084-2f3c4c8be7ad TCGA-A2-A0EY TCGA-BRCA +DO36226 5dd69c3e-9a3d-450c-9f10-10b4d6daeeba TCGA-HC-7075 TCGA-PRAD +DO36248 dc40cf49-b208-4f59-b5f0-cfbdd02970b2 TCGA-CH-5750 TCGA-PRAD +DO36253 c039014f-821c-43d6-9d3c-914f9c8d98fa TCGA-CH-5788 TCGA-PRAD +DO36254 3acd8328-76f7-4f43-9660-394b95748c3d TCGA-CH-5771 TCGA-PRAD +DO36262 91e249a6-b0b0-487e-9fad-0e9f5798dcae TCGA-HC-7740 TCGA-PRAD +DO36273 87959988-bb25-47ee-afc1-9af92592bbe7 TCGA-G9-7522 TCGA-PRAD +DO36283 f6381367-142c-45d0-92b3-c1727d1813ce TCGA-HC-7744 TCGA-PRAD +DO36285 73f875cb-bcfc-43ba-ab97-4e94caf836d4 TCGA-HC-7737 TCGA-PRAD +DO36345 6845076e-e940-4437-8618-81db4a447544 TCGA-G9-6365 TCGA-PRAD +DO36360 4fe004e5-49f9-4336-a690-6ebad512c813 TCGA-EJ-7791 TCGA-PRAD +DO36372 59207bcd-f127-481f-b576-8716da7d7d97 TCGA-G9-6336 TCGA-PRAD +DO36386 72802851-9340-4c5f-a90b-23b7afd7577a TCGA-AG-4008 TCGA-READ +DO36431 e2d395f8-bf13-4fe4-8031-683b0eedd875 TCGA-AG-3890 TCGA-READ +DO36459 1388bba1-02ba-40df-93e5-c3d58b7bbae0 TCGA-AG-4015 TCGA-READ +DO36556 8feb0e8f-d09b-437d-8651-0cdecfe776bf TCGA-F5-6814 TCGA-READ +DO36572 1586cb0a-081a-4364-9470-46f9d0f0167d TCGA-AG-3901 TCGA-READ +DO3662 5d1d00c6-fcae-479e-ae1e-de76efd41d98 TCGA-BH-A0BW TCGA-BRCA +DO36801 10290e91-9784-4f94-8daa-b5d9a8b0efb9 TCGA-AG-3593 TCGA-READ +DO36822 335d0351-3740-4847-9ff4-be2f078c70f7 TCGA-AF-2689 TCGA-READ +DO36945 0bb786ee-07cf-4853-9b49-ec95cd14e282 TCGA-AG-A032 TCGA-READ +DO36969 830fd359-abd4-4c99-89e4-2d6bb9ce5637 TCGA-AF-2691 TCGA-READ +DO37074 8e00e7e7-ffaf-44f0-91a7-172671f18e08 TCGA-AG-4007 TCGA-READ +DO37259 9af35239-ffb9-4d4f-889e-f743da271eef TCGA-AG-3896 TCGA-READ +DO37409 40e17bf8-6c19-4e5e-b60d-ad54f504a970 TCGA-DA-A1HY TCGA-SKCM +DO37418 b6691e3c-ee3f-49a2-8bd4-2086413e5d8d TCGA-DA-A3F8 TCGA-SKCM +DO37424 7b92fcd4-f9be-4b97-94a2-b2e8348aeaba TCGA-DA-A1I8 TCGA-SKCM +DO37425 b5e37b9b-6264-4cc8-9b44-1d8ac0692d6c TCGA-EE-A2M5 TCGA-SKCM +DO37426 ce28f65d-4ae8-4e63-815f-5b8d4389e092 TCGA-ER-A19J TCGA-SKCM +DO37427 85400bf1-23e0-4e0f-8df7-f0d108641aee TCGA-D3-A3MO TCGA-SKCM +DO37432 3429967f-4f77-4894-bf27-c4a22698ca92 TCGA-GN-A266 TCGA-SKCM +DO37435 d8eac750-9daf-4465-b672-9e31fd369057 TCGA-DA-A3F5 TCGA-SKCM +DO37439 81cedb9e-ee43-476b-9ed9-8d0fdb346065 TCGA-EE-A2A0 TCGA-SKCM +DO37440 d7ee2693-39a8-4e3a-b6f6-e4729800e34e TCGA-EB-A24D TCGA-SKCM +DO37445 8c54fcfd-999f-43c5-b31f-26d006f5fff3 TCGA-D3-A1Q1 TCGA-SKCM +DO37474 1404791b-86bf-4dfd-85b4-c8ff356f109c TCGA-FS-A1ZK TCGA-SKCM +DO37515 e36c9411-773b-4a8e-9b43-895b3cb044c7 TCGA-GN-A262 TCGA-SKCM +DO37527 76cc1d42-acec-46b3-9663-5e4de2550353 TCGA-GN-A264 TCGA-SKCM +DO37531 c061ae6c-7a02-47c3-aecd-a56487c01603 TCGA-EE-A2MI TCGA-SKCM +DO37587 5f6df69c-4ebf-4811-8b28-ad393dfbe1a3 TCGA-ER-A19L TCGA-SKCM +DO37599 a01a350d-64be-4af3-91c3-7d681c3db11d TCGA-EE-A185 TCGA-SKCM +DO37607 3582fb74-f43d-4fcd-8f22-bb1506fcff1c TCGA-DA-A3F3 TCGA-SKCM +DO37627 c2948fe9-5798-43bc-a743-511e94e417f8 TCGA-FS-A1ZD TCGA-SKCM +DO37659 210726bd-db19-48ae-9a3d-e6ac44480e85 TCGA-DA-A1HV TCGA-SKCM +DO37703 4cdb18f5-47bf-475f-b34c-fa4912515268 TCGA-FS-A1ZP TCGA-SKCM +DO37719 68f75ab5-9d77-40c5-a8ce-f56016bbe9e4 TCGA-DA-A1I0 TCGA-SKCM +DO37723 7ce09a09-074e-4f38-bed7-98ab34347fbe TCGA-FS-A1ZU TCGA-SKCM +DO37750 3409f118-1798-48d6-9556-12a5c77ebefb TCGA-ER-A19D TCGA-SKCM +DO37758 01ad975d-c2ed-4e4d-bd3b-c9512fc9073c TCGA-DA-A1I2 TCGA-SKCM +DO37782 fa099993-ef30-4eff-a8fb-aa41b5ea671a TCGA-GN-A26C TCGA-SKCM +DO37827 238111f9-a08b-49b4-9035-cc43026330e7 TCGA-D3-A1Q5 TCGA-SKCM +DO37861 5bc515c8-7727-4d69-98e9-31bbcb748550 TCGA-GN-A26A TCGA-SKCM +DO37896 be73fde2-d843-426b-9306-a02d3f2a754b TCGA-EE-A2GN TCGA-SKCM +DO38129 800d3a17-269f-4d97-94ff-8585f20038d8 TCGA-DA-A1HW TCGA-SKCM +DO38226 647dc54f-a51b-40ea-9d97-c93598c2af71 TCGA-HF-7136 TCGA-STAD +DO38229 af1086cc-b330-4ad6-ae1c-3b5c7259949a TCGA-BR-6564 TCGA-STAD +DO38238 e37cae9e-0656-4b89-b326-739a5cc859e0 TCGA-D7-6519 TCGA-STAD +DO38380 8fccff71-3710-4cb9-b8af-61707aebe5d8 TCGA-CG-4442 TCGA-STAD +DO38388 8d3e3d09-bc61-4d5e-b8d6-5aee6cb488ae TCGA-BR-7722 TCGA-STAD +DO3840 d6f7afc0-1558-43ad-acb1-2b5311ed2264 TCGA-GM-A2DF TCGA-BRCA +DO38404 27988b3e-cde1-4f4b-82a1-6d8ad08db8e9 TCGA-BR-6452 TCGA-STAD +DO38423 e2aaabd9-f4ba-4763-a28a-c14d46a4bc5d TCGA-D7-6822 TCGA-STAD +DO38535 39dd61a8-cd17-4df3-ae17-61e239eb00bd TCGA-CG-4474 TCGA-STAD +DO38547 35496e4e-598f-4ffc-b11e-8edaeb676079 TCGA-D7-6528 TCGA-STAD +DO38571 8b5746f9-dbee-40bd-9141-1081960cf286 TCGA-D7-5579 TCGA-STAD +DO38603 53086e20-a727-40d6-b42d-8030109b5130 TCGA-D7-6815 TCGA-STAD +DO38635 cb622cd6-776d-4681-8a67-4addd61c66b8 TCGA-D7-6527 TCGA-STAD +DO38655 1ca3c5e0-32b0-4467-8ec0-ca212e35d2b3 TCGA-BR-6456 TCGA-STAD +DO38663 e0916597-e489-45ef-a8f7-653a3c6591ef TCGA-BR-4280 TCGA-STAD +DO38671 8a24f1a3-4bab-4591-9676-7bee01d096c1 TCGA-BR-A4J4 TCGA-STAD +DO38731 283776cd-32e8-4a79-bed1-1411c9d3a9e0 TCGA-F1-6875 TCGA-STAD +DO3874 60df7543-6da5-4c75-943b-5800c1e08234 TCGA-E2-A14X TCGA-BRCA +DO38775 c991697f-77f9-412a-9bd1-0e47f7387cb2 TCGA-CD-5799 TCGA-STAD +DO38779 8cb0144b-be6b-40a1-86a2-708f96d9b615 TCGA-CG-4443 TCGA-STAD +DO38791 956017ba-5e4b-42c7-b93d-42316874ba93 TCGA-BR-4255 TCGA-STAD +DO38807 2d291fa3-f0fd-4bc1-91a6-783863714190 TCGA-EQ-5647 TCGA-STAD +DO38831 87c217d4-66f4-46ca-8244-7856ce658fd3 TCGA-IN-7806 TCGA-STAD +DO38847 0b4e8d73-459d-4930-9d1a-abb8d53e1844 TCGA-CG-5724 TCGA-STAD +DO38871 0aecac64-5982-4d76-8f31-958f6a00951d TCGA-CD-5802 TCGA-STAD +DO38899 4fc8e011-4433-4537-b03f-457a3a70240f TCGA-EL-A4K6 TCGA-THCA +DO38901 a09d0be0-fd41-4b30-b488-9a9f2abef8e7 TCGA-EL-A3TB TCGA-THCA +DO38903 fb83f7d7-2182-4fb7-8e3e-8ad7f1feac72 TCGA-EL-A3CV TCGA-THCA +DO38911 6abc861a-376d-446b-acee-fb0f03a82c09 TCGA-EM-A22O TCGA-THCA +DO38946 c31b3adf-3fc6-4a72-83b0-bbc0f295b5e7 TCGA-DE-A0Y3 TCGA-THCA +DO38966 e35b2928-8784-45d0-a71a-a2df161542fa TCGA-DJ-A2Q1 TCGA-THCA +DO38978 77bb90f0-e923-4e28-bb88-4d6b420e0b5c TCGA-DE-A2OL TCGA-THCA +DO39016 8d54a2ed-03f1-4a23-bdd2-f3395f5d3716 TCGA-EL-A3H1 TCGA-THCA +DO39020 b0cb6ed7-3d9f-4d5e-8df5-edcc921c43b4 TCGA-FY-A2QD TCGA-THCA +DO39022 f9cf605f-f287-4eba-8e8f-8ba47e14ecc4 TCGA-DJ-A2Q8 TCGA-THCA +DO39046 2f59f12d-ca0c-4ac6-9718-98a6d455af46 TCGA-EM-A2CP TCGA-THCA +DO39070 bed1fb3c-4459-460e-8190-29211857fa73 TCGA-BJ-A28V TCGA-THCA +DO39072 7629f864-a009-4aa9-bf7e-2d99e00d294e TCGA-ET-A4KN TCGA-THCA +DO39084 e70feafd-1390-4f4c-a635-da65c1a7daeb TCGA-FE-A233 TCGA-THCA +DO39127 785803bb-42a8-4afb-b7b4-58c114894408 TCGA-FK-A3S3 TCGA-THCA +DO39210 63b01b43-a68a-4046-89b0-736ba4d08504 TCGA-FK-A3SD TCGA-THCA +DO39216 6ee95dc8-2bac-4691-afa2-8222e07506ad TCGA-DJ-A3US TCGA-THCA +DO39306 f9ceffc0-d544-418d-b4a9-bd3c84e37026 TCGA-BJ-A0Z2 TCGA-THCA +DO39363 dc4d8468-6f34-40cb-a826-8d582cd86fde TCGA-BJ-A0ZB TCGA-THCA +DO39481 ebbd54ad-7b97-4081-ad11-7f7f3ce74eef TCGA-E8-A418 TCGA-THCA +DO39534 fef9c64f-5959-4da0-aaa2-66b56fc7b4c3 TCGA-EM-A3FQ TCGA-THCA +DO3958 084c6aec-94f7-4090-8f3f-59fa9e89721a TCGA-A2-A0D1 TCGA-BRCA +DO39610 7746be87-7ad0-48a7-84fb-08a81b220bdb TCGA-EL-A3MY TCGA-THCA +DO39635 8a233b01-2959-467f-8950-02be16309156 TCGA-EL-A3T9 TCGA-THCA +DO39684 55e03ffc-f006-4ec8-8e5d-a738ff091f8a TCGA-EM-A2OV TCGA-THCA +DO39798 afb00cd4-e5da-4251-86fb-532c3f65324c TCGA-FK-A3SE TCGA-THCA +DO39841 3e787a0e-b5d2-401e-b1e5-077376090b3a TCGA-DJ-A13R TCGA-THCA +DO39847 78d14f86-896a-4f98-9274-b2e8e387add9 TCGA-FE-A22Z TCGA-THCA +DO39886 4aed5116-d27b-4911-b310-05c2248b861f TCGA-EM-A2CN TCGA-THCA +DO39910 085ffcdc-8600-42e2-852a-506465f716a2 TCGA-DE-A3KN TCGA-THCA +DO40069 acb80b76-2394-4250-8687-60e40e16e83b TCGA-DJ-A4UT TCGA-THCA +DO40147 e1ff7694-64e1-47ed-b0d8-34acfb4ffef1 TCGA-L6-A4ET TCGA-THCA +DO40167 2501ee46-8d38-448b-8765-e9c9706cbbe8 TCGA-BJ-A191 TCGA-THCA +DO4018 ef4cbd38-bc79-4d60-a715-647edd2ebe9e TCGA-E2-A1LG TCGA-BRCA +DO40193 39d90499-8549-4da5-b391-c421f18facf1 TCGA-FE-A3PD TCGA-THCA +DO40274 e4b08e66-0772-4956-8670-e5c6553eecbe TCGA-DJ-A1QL TCGA-THCA +DO40405 a00da907-e527-45f9-9afb-2fad2c7b0940 TCGA-DJ-A2Q2 TCGA-THCA +DO40484 e8f56d0f-eee4-4def-a43a-dec91f4382a1 TCGA-EM-A3AQ TCGA-THCA +DO40496 b46d88e9-8477-4900-8da9-5e7dbded0653 TCGA-ET-A3DV TCGA-THCA +DO40735 934a58af-0ffc-487e-9e95-fa7643b7d5cb TCGA-EL-A3CX TCGA-THCA +DO40831 24118b8f-59fe-482a-aed3-7b1bde1571b7 TCGA-B5-A0K8 TCGA-UCEC +DO40930 95e6d327-9ab7-45de-9527-f5634b3995c3 TCGA-AX-A06B TCGA-UCEC +DO41056 a8614c49-c7f1-4de9-bd2f-4764b094d604 TCGA-A5-A0GG TCGA-UCEC +DO41157 c88e3901-f3ee-436a-ba80-143c8eab7d69 TCGA-AX-A2H5 TCGA-UCEC +DO41217 d39df936-4bc6-404d-9e1d-f6416cc5a34c TCGA-AJ-A23M TCGA-UCEC +DO41343 de4e13c8-5ad0-46d2-ae54-b9ca22b939d8 TCGA-EO-A1Y8 TCGA-UCEC +DO4138 5dd423e8-feaa-4568-a750-500948c41d6c TCGA-B6-A0I1 TCGA-BRCA +DO41398 9205c164-7975-421a-90c3-edfe8def595c TCGA-E6-A1LZ TCGA-UCEC +DO41446 d17668c4-5d66-4e16-8b70-b54616581aae TCGA-AP-A0LL TCGA-UCEC +DO4155 d3d545b3-457f-4389-821f-704cb24aff7f TCGA-AR-A256 TCGA-BRCA +DO41716 8d9e4917-334b-4c76-aee1-1e22be772db0 TCGA-BK-A139 TCGA-UCEC +DO41911 944a102f-3475-45aa-9e30-4b92d1449f00 TCGA-DI-A1NN TCGA-UCEC +DO41951 f7e06a10-74ca-4f52-a944-b3f22c016998 TCGA-AP-A05D TCGA-UCEC +DO42169 9aa8c2b4-db32-4899-bf0c-d578cb175e90 TCGA-AP-A0L8 TCGA-UCEC +DO42209 13f31bcb-befd-4ed2-adc0-dc9e9ce52b75 TCGA-BS-A0TG TCGA-UCEC +DO42304 45d58088-d3f4-4eb0-b75c-bcb98c0981dc TCGA-A5-A0G9 TCGA-UCEC +DO42400 7b7d2d2b-abe6-4f32-add0-3715096995d1 TCGA-BS-A0TC TCGA-UCEC +DO42432 ffaa98a0-2b69-46dc-aee5-c5c3f2abbc38 TCGA-BS-A0TD TCGA-UCEC +DO42464 66e305ca-8a16-44f0-b837-85e529b7ad55 TCGA-BG-A18C TCGA-UCEC +DO42536 f8ac5fd2-11f2-4b3a-a3a6-ff1a6d8e1cae TCGA-EY-A1GW TCGA-UCEC +DO42937 5c22c8d6-5805-455c-8e1c-e6b8006738e4 TCGA-D1-A17K TCGA-UCEC +DO43074 16e69011-c295-479f-b521-86e66fba498d TCGA-D1-A16G TCGA-UCEC +DO43162 c10b2885-70ea-4d32-bc48-8b85530df702 TCGA-EY-A1GS TCGA-UCEC +DO43210 94817e78-78f5-4556-9394-610df65605e2 TCGA-A5-A0GJ TCGA-UCEC +DO43314 79ecc142-8f6e-4260-a7ce-c8e2a6212c35 TCGA-B5-A0K3 TCGA-UCEC +DO43338 bad34c92-1e64-4187-a0b4-061105fd13ba TCGA-AX-A1C8 TCGA-UCEC +DO43514 d2aa7d70-a25f-4a9e-9d52-37e372461824 TCGA-B5-A11G TCGA-UCEC +DO4359 5ed024e8-d05e-4c65-9441-eda9930ccc82 TCGA-A7-A26G TCGA-BRCA +DO43739 5ffa2ec6-2d94-4b09-8fb9-3591cf38fa4f TCGA-BS-A0U9 TCGA-UCEC +DO43811 e56491af-718f-4ca1-b3cd-0e5b3154697b TCGA-D1-A1NU TCGA-UCEC +DO43859 5d2f1a0f-a01b-48c2-b749-1ca710aa2bb4 TCGA-B5-A1MY TCGA-UCEC +DO44086 490eabca-4575-4bd6-9604-6b6e23d069ce TCGA-NH-A50T TCGA-COAD +DO44087 857a5889-19e7-4208-8362-f1377c408037 TCGA-A6-A567 TCGA-COAD +DO44088 c83d38fc-e011-4f75-a100-96513611f3e9 TCGA-QG-A5YV TCGA-COAD +DO44089 e8ec26bd-c88d-42b9-ac8a-ab121b8140f8 TCGA-AD-A5EK TCGA-COAD +DO44091 d8e9ea52-ff28-4626-baa3-71350fd53ddf TCGA-AY-A54L TCGA-COAD +DO44092 790266bc-c252-43f4-85d1-1b06ee529cf8 TCGA-A6-A566 TCGA-COAD +DO44093 300fff5a-0f61-402d-b845-3938ba7b1294 TCGA-QG-A5YX TCGA-COAD +DO44094 d3b7ada6-5395-4212-a921-8e97c92b605d TCGA-NH-A50V TCGA-COAD +DO44095 74a0264d-1d31-430d-9a88-e7334c8aa96c TCGA-A6-A56B TCGA-COAD +DO44097 cb664eef-be6f-44ef-aa90-6f6da0e6440a TCGA-BT-A3PJ TCGA-BLCA +DO44103 53886143-c1c6-40e9-88e6-e4e5e0271fc8 TCGA-A2-A3XX TCGA-BRCA +DO44111 65cac997-4d39-4501-85ec-4fcb328a8eb5 TCGA-GM-A3XL TCGA-BRCA +DO44273 174850b4-5ec2-462b-a890-89bd1716b3c2 TCGA-A2-A3Y0 TCGA-BRCA +DO44285 14c65927-b9dd-4ef6-9917-db7bc431ff2f TCGA-C5-A0TN TCGA-CESC +DO44462 1db41963-a520-47f0-828c-ed5c626507b1 TCGA-IA-A40Y TCGA-KIRP +DO44465 8b3c068d-5090-4030-8e7e-39da005dee79 TCGA-A4-A4ZT TCGA-KIRP +DO44469 925ae8ff-3200-4564-be76-94db0984b027 TCGA-A4-A48D TCGA-KIRP +DO44478 0c151271-2cf9-43f4-9e9b-fe927de43131 TCGA-IA-A40X TCGA-KIRP +DO44496 b809c2f2-185c-496a-a1d4-589294443a1d TCGA-B9-A44B TCGA-KIRP +DO44515 36b46540-472f-4a4d-a801-b8b6ae660dc9 TCGA-HE-A5NF TCGA-KIRP +DO44517 d80f601b-1064-4ff0-9c51-7f6afb22d1f5 TCGA-A4-A57E TCGA-KIRP +DO44520 2c084454-18e9-4837-9bba-7e91a73a48e0 TCGA-GL-A4EM TCGA-KIRP +DO44525 78412f82-cc55-43d5-96ca-1d02cf957725 TCGA-GL-A59R TCGA-KIRP +DO4473 5580b21a-2cdb-4777-ad79-6e06654144f5 TCGA-E2-A15H TCGA-BRCA +DO44872 c95ed407-bcc2-44aa-bc59-bf7549f2d665 TCGA-FV-A23B TCGA-LIHC +DO44875 c4c4820b-6f4c-4cdd-b233-4d22f5933fd9 TCGA-DD-A3A6 TCGA-LIHC +DO44877 65e430e1-0c9a-4045-83a9-653b7cff811d TCGA-FV-A4ZQ TCGA-LIHC +DO44883 69d6881b-9e9a-407b-8b02-bad5fc2455bf TCGA-DD-A3A8 TCGA-LIHC +DO44885 89791f8f-dc06-4a04-b166-afbfd6cf53de TCGA-DD-A4NG TCGA-LIHC +DO44888 457cca7c-a8b4-414e-a3c0-7fccf0ee2034 TCGA-DD-A3A9 TCGA-LIHC +DO44889 bce25281-502e-4599-9679-32dc8462ffb1 TCGA-DD-A4NE TCGA-LIHC +DO44890 bd48fa5e-0152-46f0-a4fa-8cf723ff2dcd TCGA-FV-A495 TCGA-LIHC +DO44919 d680df09-368e-42b5-b540-45c41ed31042 TCGA-FV-A2QQ TCGA-LIHC +DO44930 bee9d6c8-948f-4b76-97ca-b985064249ef TCGA-DD-A4NA TCGA-LIHC +DO44966 8afa3140-dcbd-44c0-ab13-4ada6a0444d4 TCGA-FV-A496 TCGA-LIHC +DO44974 56403463-1727-4c23-a709-de5f168c4073 TCGA-ED-A459 TCGA-LIHC +DO44978 9a56d8a8-dadd-442e-bc09-732cd619aef7 TCGA-G3-A5SL TCGA-LIHC +DO44982 614c0ecf-6a84-4c86-8706-79be5bb00991 TCGA-DD-A3A7 TCGA-LIHC +DO44986 c0bf9278-9cbb-4361-b4d8-ab172b67e276 TCGA-DD-A4ND TCGA-LIHC +DO44994 17833039-0d7e-47f0-90e6-8cabae94c124 TCGA-FV-A3R2 TCGA-LIHC +DO45009 a8c2aa08-9156-4845-ad77-5e060858c757 TCGA-FV-A3I0 TCGA-LIHC +DO4635 790ff5db-b7f3-4946-aaed-305f66b1dd6a TCGA-BH-A0AV TCGA-BRCA +DO46813 e2b33861-2247-413f-b2e5-9be3b604666c TCGA-HC-7233 TCGA-PRAD +DO4695 88db1340-e4bf-451a-87c0-6e9168296f5e TCGA-EW-A1PB TCGA-BRCA +DO4719 3fc3755d-a3f8-4e2c-813f-ff124f2a75c1 TCGA-A2-A259 TCGA-BRCA +DO472 9c274536-3ca1-4f0e-93a8-1688074d862f TCGA-DK-A1A6 TCGA-BLCA +DO47255 905ea29f-588a-497b-9f06-2eef3bdc9442 TCGA-EE-A29B TCGA-SKCM +DO47268 296f44f9-3ecb-4b40-ae51-c3a06e3fbe00 TCGA-EE-A3JI TCGA-SKCM +DO47331 d6283ab0-9019-4ad2-93aa-2e6b39cd9641 TCGA-EE-A3J5 TCGA-SKCM +DO47451 d32a4219-4cca-470f-b6b3-11447d9f5e0b TCGA-ER-A2NF TCGA-SKCM +DO47542 3fea3f02-2db3-44e0-984e-c61130d92c9b TCGA-ER-A19E TCGA-SKCM +DO4766 f0787165-6f58-4d67-b510-928eea2c4882 TCGA-E2-A109 TCGA-BRCA +DO477 904a8757-e0c5-41ef-b583-c8f170caaac1 TCGA-DK-A1A5 TCGA-BLCA +DO47737 d5efcde2-21e6-4071-9aa4-3c5688a45b3f TCGA-BR-8682 TCGA-STAD +DO47747 30a82e9f-2a2d-4a66-bdb1-26b881c21d01 TCGA-HU-A4G6 TCGA-STAD +DO47753 5b7e557a-6aee-4077-b3cd-c96179e643ae TCGA-CD-8529 TCGA-STAD +DO47785 fe45eaee-ff0c-4d52-b18c-04dc2d64d3b4 TCGA-HU-A4H0 TCGA-STAD +DO47817 aef8d98f-7a1c-47f3-b5a7-a3cb299c70de TCGA-BR-8381 TCGA-STAD +DO47841 c6316fb5-99a7-4adf-832f-dce44ddcd3aa TCGA-HU-8608 TCGA-STAD +DO47859 60879669-96c2-4710-82e0-0ae9d0e5b242 TCGA-BR-8373 TCGA-STAD +DO47863 b8a1c8c7-3945-4d84-8f58-3a1a1eb10730 TCGA-BR-8486 TCGA-STAD +DO479 1e308b12-0590-4dae-94d0-a539fcf25df7 TCGA-CF-A3MF TCGA-BLCA +DO47911 337819d2-5281-4b41-9583-a476924bc837 TCGA-BR-8690 TCGA-STAD +DO47929 4e9d13da-438d-4b5e-8e86-67432b5e9471 TCGA-HU-8245 TCGA-STAD +DO4796 38a8b734-9acc-42f9-b5b7-e51b0dfc6504 TCGA-A8-A092 TCGA-BRCA +DO47997 2b4f014f-74e9-4dda-a1ad-abcbd124fc44 TCGA-D7-A4YX TCGA-STAD +DO48040 1e1cdf72-820f-48a4-8fb6-89575d07cde9 TCGA-D7-8570 TCGA-STAD +DO48065 cfdd2155-57a9-499c-92e1-fb82b577b050 TCGA-EM-A3FL TCGA-THCA +DO48084 8cfb88f6-7b37-4f1e-882b-3a368b604167 TCGA-BJ-A45K TCGA-THCA +DO48103 2b84cfbc-7cf6-4cb1-a1e3-ba85dafb89f4 TCGA-EM-A3FR TCGA-THCA +DO48159 0c50a2c2-1d4c-45ca-9127-f96d4db18daf TCGA-EM-A3AL TCGA-THCA +DO483 3ed614e7-f356-4d87-985b-d3bbbae3bb40 TCGA-DK-A1A7 TCGA-BLCA +DO48474 a40beabe-496e-4324-a1d1-5a7d8b002aab TCGA-MH-A560 TCGA-KIRP +DO48482 f0f7e57f-1e9d-4f24-a685-a0811aacbbed TCGA-MH-A55Z TCGA-KIRP +DO48495 45bdcfd6-1e3f-4be8-b843-ae949e8e43eb TCGA-MH-A562 TCGA-KIRP +DO48505 c52abac9-07da-4b3d-bf8a-2161cacf9682 TCGA-ED-A4XI TCGA-LIHC +DO48509 e6e6a773-b155-45a9-989f-75c12ce1702a TCGA-MH-A561 TCGA-KIRP +DO48516 2030e10b-a4c2-460f-9e29-90c8a76ea05e TCGA-MH-A55W TCGA-KIRP +DO48518 6c79f64b-63fe-46f5-bf7b-77208231b0f2 TCGA-DD-A1EG TCGA-LIHC +DO48552 1983fade-44f1-41e0-8983-f2ce8d649655 TCGA-DD-A4NB TCGA-LIHC +DO48977 0809ba8b-4ab6-4f43-934c-c1ccbc014a7e TCGA-EK-A2PK TCGA-CESC +DO48987 00bca18c-b3d4-45a3-8f19-034cc40449a4 TCGA-C5-A2LV TCGA-CESC +DO48991 262311a0-7a10-4934-8fba-11ee581ad738 TCGA-C5-A1BF TCGA-CESC +DO48993 110cd100-5730-4873-9fab-2ae6c8732062 TCGA-C5-A1M8 TCGA-CESC +DO49009 7941d1c7-5319-4e8d-a15f-60b61cb68a9b TCGA-C5-A1BJ TCGA-CESC +DO49014 a3be445a-a592-4073-815b-47ae41f9c4a0 TCGA-C5-A1M9 TCGA-CESC +DO496 5d54c742-5a8e-4c40-8d62-95e75e210ab8 TCGA-GD-A2C5 TCGA-BLCA +DO49659 c7f1b389-aec7-4993-8afb-b7a56064f8fc TCGA-BW-A5NQ TCGA-LIHC +DO49695 2d00f09f-0074-46c8-b357-b5f33567762a TCGA-HP-A5MZ TCGA-LIHC +DO49728 b06cdf61-61c9-4035-a55d-1ef21dc21069 TCGA-AL-A5DJ TCGA-KIRP +DO49736 072507cf-1249-4b28-86d4-3ab97822a74c TCGA-B1-A47M TCGA-KIRP +DO49738 f1ec5a7b-0973-4c38-9908-862196df9541 TCGA-B1-A47N TCGA-KIRP +DO49768 8d28f586-3de5-4033-a9fd-4f67be406fb4 TCGA-HE-A5NJ TCGA-KIRP +DO498 493a4ff2-37a5-4b79-928d-83dbfe534556 TCGA-DK-A1AE TCGA-BLCA +DO49815 da5b0169-e4f4-4674-b348-220af03ff4b5 TCGA-HC-8258 TCGA-PRAD +DO49880 fb66b7f5-22ca-4d4d-97c2-30e6edfff781 TCGA-C5-A2LT TCGA-CESC +DO49954 61eb08d0-ea01-45e7-884b-ac3d6070172a TCGA-EK-A2RL TCGA-CESC +DO49956 c13215d9-26a8-4c8b-b4fe-5da6aa3222af TCGA-EK-A2RM TCGA-CESC +DO5046 f05d314c-5ec5-4e2a-b785-9a702716f111 TCGA-E2-A156 TCGA-BRCA +DO51251 b85b14d7-3b5a-4800-af12-622ec03b9fe5 TCGA-ER-A19T TCGA-SKCM +DO51551 6a6c1569-1045-4ed3-acdb-5c45e47614ca TCGA-B1-A47O TCGA-KIRP +DO51592 374f614a-bcd1-4283-ac6f-fc27ddc79987 TCGA-C5-A1ML TCGA-CESC +DO51594 7c93642d-7e05-40f8-b1ef-a014edcfba42 TCGA-DG-A2KJ TCGA-CESC +DO5200 233b02f3-c4f0-4a67-9db5-e68d5cdaccb6 TCGA-BH-A18U TCGA-BRCA +DO52029 bd75d8ee-916b-4abc-bf33-bce6a4217076 TCGA-BA-A4IH TCGA-HNSC +DO52171 ee63c44b-91f4-4004-8dea-192eab527036 TCGA-BW-A5NO TCGA-LIHC +DO52172 56a82a56-0241-4d3d-9de2-696b0c36df91 TCGA-BW-A5NP TCGA-LIHC +DO522 cede5274-279d-4021-8aca-5ecc4bf94d66 TCGA-H4-A2HQ TCGA-BLCA +DO52307 1987b453-f97d-45c7-9c89-b9a33313d645 TCGA-PD-A5DF TCGA-LIHC +DO5249 733a2a0f-b37a-4b81-b49e-3c0f30d1eb37 TCGA-A2-A04X TCGA-BRCA +DO5347 75113445-d2d6-44a0-866c-c9175e6d214b TCGA-A2-A0D4 TCGA-BRCA +DO536 556fcbc8-172a-4af1-8822-ae036e8d68e8 TCGA-BL-A13J TCGA-BLCA +DO548 56d8a1b8-1268-4362-a7f1-77b76fbf1b59 TCGA-DK-A1AA TCGA-BLCA +DO555 178b28cd-99c3-48dc-8d09-1ef71b4cee80 TCGA-DK-A3IL TCGA-BLCA +DO561 797a71ee-372d-41e5-aeee-5ab3c4661110 TCGA-BT-A2LA TCGA-BLCA +DO5626 44bec761-b603-49c0-8634-f6bfe0319bb1 TCGA-AO-A0JM TCGA-BRCA +DO5661 b97bf89a-7a85-4eef-ae7e-f787aead1f0a TCGA-BH-A1FC TCGA-BRCA +DO5745 9f6be944-83de-42ab-8738-f0022f475e61 TCGA-GI-A2C9 TCGA-BRCA +DO575 8c1dd7f7-b74a-4fa2-b6a7-86f0348d2567 TCGA-FD-A3N6 TCGA-BLCA +DO5808 9435447e-d65f-408b-863b-6576b1d652dd TCGA-AO-A03N TCGA-BRCA +DO6144 5d8ed961-a012-4b6b-bffb-9468d3953646 TCGA-EW-A1J5 TCGA-BRCA +DO6231 f45210d3-9e66-4f5e-bef1-5ee5547cc893 TCGA-EW-A1PC TCGA-BRCA +DO6249 941f75a1-fea4-4539-ba69-60bb11608f6d TCGA-C5-A1MI TCGA-CESC +DO6252 8a4da4f0-30f2-497e-8a7c-988fd6b813cb TCGA-C5-A1BN TCGA-CESC +DO6255 6b74c07a-9932-4a80-8296-9db08d0c6964 TCGA-EX-A1H5 TCGA-CESC +DO6314 e733f289-a0ec-478c-aad9-1378ad26ddfa TCGA-C5-A2LY TCGA-CESC +DO6332 dfd30f94-19e3-4838-95df-14a3763419af TCGA-DS-A0VL TCGA-CESC +DO639 a658becf-3934-49f2-9e74-291dca61d4bb TCGA-FD-A3N5 TCGA-BLCA +DO669 ca554128-da9d-4f37-9560-ca083509e01d TCGA-C4-A0F7 TCGA-BLCA +DO675 f48937ed-e294-41f0-8872-7010391167a9 TCGA-BT-A20P TCGA-BLCA +DO689 24f21425-b001-4986-aedf-5b4dd851c6ad TCGA-BT-A20V TCGA-BLCA +DO695 419deaac-ea45-4bdd-9fa0-b5cd8429b44f TCGA-BT-A20Q TCGA-BLCA +DO720 cce62116-3dcd-400d-ada0-9839ca02466e TCGA-FT-A3EE TCGA-BLCA +DO7801 df5ab6cc-6f68-4b6b-95e2-954c6b57ba9c TCGA-CA-6718 TCGA-COAD +DO7831 a06c09a8-0951-49f4-a765-5147bbcae109 TCGA-AA-3664 TCGA-COAD +DO7990 fafd4576-0709-49e1-926f-ad1ee09907ec TCGA-A6-2683 TCGA-COAD +DO804 096b4f32-10c1-4737-a0dd-cae04c54ee33 TCGA-BT-A20T TCGA-BLCA +DO8047 dca004bf-9c14-45a1-b186-ab3759f6c7fe TCGA-AA-3685 TCGA-COAD +DO822 f8c0a69c-df53-4a0a-a30e-4424c0ce3dd5 TCGA-DK-A1AG TCGA-BLCA +DO8264 268c01b3-d2ce-44c0-a2fe-ea846a1253cc TCGA-CA-6717 TCGA-COAD +DO828 7fee9d6f-c0ea-43c2-89c2-66d965a077a8 TCGA-CF-A27C TCGA-BLCA +DO8360 5e7074a6-5011-4dc5-950c-41365dbc6d09 TCGA-AD-6964 TCGA-COAD +DO856 932bcdd4-5145-49e8-9a95-8a546396f572 TCGA-BT-A3PH TCGA-BLCA +DO8658 750770db-53cf-4021-a8c2-5e0ddb97fef5 TCGA-AA-3994 TCGA-COAD +DO8730 33e989b9-fac4-47d8-b429-269044ac772a TCGA-AA-3977 TCGA-COAD +DO8858 ca22edd4-eb7c-4677-84aa-6f0354459a19 TCGA-AA-A00R TCGA-COAD +DO8898 36eaf209-c332-4321-b6aa-51a3368ecd79 TCGA-AA-3555 TCGA-COAD +DO9034 5116e3b4-2bac-40f5-8046-b9c1783faaa5 TCGA-A6-3807 TCGA-COAD +DO9042 a512701f-88bb-4b05-8c94-fd285b9dd13e TCGA-AA-3666 TCGA-COAD +DO9074 ff1407c6-9174-4bae-a19b-d34ca71b898c TCGA-A6-2680 TCGA-COAD +DO9106 84e34cf2-b3bb-4a1a-8fb7-6c2c604c4008 TCGA-AA-A01S TCGA-COAD +DO9732 bfb07784-693b-4c25-874e-4ad6e04a5d46 TCGA-AA-3529 TCGA-COAD +DO9788 7d8eab0a-e6c8-4449-9ebf-50c41db94a06 TCGA-A6-2681 TCGA-COAD +DO9876 e457344d-76fb-46bf-b362-61a6e811d131 TCGA-AA-A00N TCGA-COAD +DO9940 df228c56-89ac-4122-93b3-6c0740c4ec09 TCGA-AA-3534 TCGA-COAD diff --git a/hla_loh/escape_genes.parquet b/hla_loh/escape_genes.parquet new file mode 100644 index 0000000000000000000000000000000000000000..82ea4ef43615c073082c19cf2f2815a5dcb99600 --- /dev/null +++ b/hla_loh/escape_genes.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:116948dcec63d66c344a72a63183a09919cc62a3b60fe33e2c49e98ea948fc89 +size 3626 diff --git a/hla_loh/gie_allele.parquet b/hla_loh/gie_allele.parquet new file mode 100644 index 0000000000000000000000000000000000000000..8e661580a10fb67cf924eb5e63777f8792e4e030 --- /dev/null +++ b/hla_loh/gie_allele.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b43ee1c870e053bec6bf29d223e5bb38df653729501edbb7fbbd5860b2e8de1 +size 145724 diff --git a/hla_loh/gie_sample.parquet b/hla_loh/gie_sample.parquet new file mode 100644 index 0000000000000000000000000000000000000000..8d910713757e09717ceb334f72e54e71d22d6c4c --- /dev/null +++ b/hla_loh/gie_sample.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10579459c2a274671a4f61d539492ce284737d4fc9ce71ceefcd77c506da4077 +size 71045 diff --git a/hla_loh/harmonize.py b/hla_loh/harmonize.py new file mode 100644 index 0000000000000000000000000000000000000000..4d0bbc9aca8fbbc1da485be691751e7b0050016f --- /dev/null +++ b/hla_loh/harmonize.py @@ -0,0 +1,177 @@ +#!/usr/bin/env python3 +# 2026-09-16 -- harmonise the Martinez-Jimenez 2023 HLA-I immune-escape supplements into parquet. +# +# The three publisher workbooks are copyrighted Springer Nature supplementary material and are not +# ours to redistribute, so they live in the gitignored `raw/hla_loh/`. The *facts* they carry are +# not copyrightable, so this script writes neutral, long-format tables that do ship. Same call, and +# the same shape, as `icb_cohorts/`: harmonised output published, publisher workbook kept back. +# +# Re-run after refreshing raw/hla_loh/*.xlsx: python hla_loh/harmonize.py +# +# Source: Martinez-Jimenez F, Priestley P, Shale C, Baber J, Rozemuller E, Cuppen E. "Genetic immune +# escape landscape in primary and metastatic cancer." Nat Genet 55(5):820-831, 2023. +# doi:10.1038/s41588-023-01367-1 (verified against crossref, 2026-09-16). +# MOESM4 = Supplementary Tables 1-2, MOESM5 = Supplementary Data 1, MOESM6 = Supplementary Data 2. + +from __future__ import annotations + +import os + +import openpyxl +import polars as pl + +HERE = os.path.dirname(os.path.abspath(__file__)) +RAW = os.path.join(os.path.dirname(HERE), "raw", "hla_loh") +SLOTS = ("A1", "A2", "B1", "B2", "C1", "C2") +CALLERS = ("LILAC", "xHLA", "polysolver") +TISSUE = {"n": "normal", "t": "tumour"} + + +def sheet(fn: str, name: str) -> tuple[list[str], list[tuple]]: + """Header and non-empty data rows of one worksheet, values only.""" + wb = openpyxl.load_workbook(os.path.join(RAW, fn), read_only=True, data_only=True) + ws = wb[name] + it = ws.iter_rows(values_only=True) + hdr = [str(h) for h in next(it) if h is not None] + rows = [r for r in it if r and r[0] is not None] + wb.close() + return hdr, rows + + +def _b(x): + """openpyxl hands these back as the strings 'True'/'False'; 'None' and None are both missing.""" + if x is None or x in ("None", ""): + return None + return str(x) == "True" + + +def _f(x): + if x is None or x in ("None", ""): + return None + try: + return float(x) + except (TypeError, ValueError): + return None + + +def _s(x): + return None if x is None or x == "None" else str(x) + + +def _allele(x): + """The workbooks write `A*02:01`; every other table in this compendium -- and `mhcmatch` -- + writes `HLA-A*02:01`. Two vocabularies for one molecule is the recurring hazard here, so the + published column carries the repo's spelling and SOURCES.md records the workbook's.""" + a = _s(x) + return None if a is None else (a if a.startswith("HLA-") else f"HLA-{a}") + + +def _df(rows): + """Scan every row before fixing a dtype. polars infers from the first 100 by default, and the + sparsest column here (`checkpoint_pathway_detail`) is null until row 250 -- inferred Null, then + a string arrives and the build fails. Cheap at this size, and it cannot silently mistype.""" + return pl.DataFrame(rows, infer_schema_length=None) + + +def gie_tables(): + """Supplementary Data 2, sheet `GIE per sample`: 6,319 samples, 4,439 Hartwig + 1,880 PCAWG. + + Split into one row per (sample, allele slot) and one row per sample, because the sheet is two + tables side by side: a six-slot allele record and a per-sample escape record. `*_MUT` is a + mutation COUNT (0.0/0.5/1.0/2.0), not a flag -- casting it to bool loses the 2.0 and the 0.5. + """ + hdr, rows = sheet("41588_2023_1367_MOESM6_ESM.xlsx", "GIE per sample") + i = {c: n for n, c in enumerate(hdr)} + g = lambda r, c: r[i[c]] if c in i else None # noqa: E731 + + allele = [ + {"sample_id": _s(g(r, "sample_id_2")), "cohort": _s(g(r, "cohort")), + "cancer_type": _s(g(r, "cancer_type")), "cancer_type_code": _s(g(r, "cancer_type_code")), + "locus": f"HLA-{s[0]}", "slot": s, "allele": _allele(g(r, s)), + "copy_number": _f(g(r, f"{s}_CN")), "n_mutations": _f(g(r, f"{s}_MUT"))} + for r in rows for s in SLOTS] + + flags = ("loh_lilac", "mut_hla_lilac", "systemic_app_pathway", "cd58_pathway", + "ifn_gamma_pathway", "checkpoint_pathway", "epigenetic_regulators_pathway", + "targeted_escape", "non_targeted_escape", "genetic_immune_escape") + # `copy_number_neutral_loh_hla` LOOKS like a flag beside the others and is not: it carries a + # comma-separated locus list (`HLA-A,HLA-C,HLA-B`). Read as bool it would have nulled every + # real value. The four LOH columns are all locus lists, kept verbatim as the source wrote them. + texts = ("copy_number_neutral_loh_hla", "loh_hfocal", "loh_focal", "loh_nonfocal", + "systemic_app_pathway_detail", "cd58_pathway_detail", "ifn_gamma_pathway_detail", + "checkpoint_pathway_detail", "epigenetic_regulators_pathway_detail") + sample = [] + for r in rows: + row = {"sample_id": _s(g(r, "sample_id_2")), "cohort": _s(g(r, "cohort")), + "cancer_type": _s(g(r, "cancer_type")), + "cancer_type_code": _s(g(r, "cancer_type_code"))} + for loc in "ABC": + row[f"ploidy_minor_HLA-{loc}"] = _f(g(r, f"ploidy_minor_HLA-{loc}")) + row[f"ploidy_major_HLA-{loc}"] = _f(g(r, f"ploidy_major_HLA-{loc}")) + row[f"del_HLA-{loc}"] = _f(g(r, f"del_HLA-{loc}")) + row |= {c: _b(g(r, c)) for c in flags} + row |= {c: _s(g(r, c)) for c in texts} + sample.append(row) + return _df(allele), _df(sample) + + +def typing_callers(): + """Supplementary Data 1, sheets `Hartwig` and `PCAWG`: three callers x normal/tumour x six + slots, one column each. Long, so a caller comparison is a group_by rather than 36 column names.""" + out = [] + for sh in ("Hartwig", "PCAWG"): + hdr, rows = sheet("41588_2023_1367_MOESM5_ESM.xlsx", sh) + i = {c: n for n, c in enumerate(hdr)} + for r in rows: + sid, mut = _s(r[i["sampleId"]]), _b(r[i["is_mutated"]]) + for caller in CALLERS: + for t, tname in TISSUE.items(): + for s in SLOTS: + col = f"{caller}_{t}_{s}" + if col not in i: + continue + out.append({"sample_id": sid, "cohort": sh, "is_mutated": mut, + "caller": caller, "tissue": tname, "slot": s, + "locus": f"HLA-{s[0]}", "allele": _allele(r[i[col]])}) + return _df(out) + + +def sample_metadata(): + """Supplementary Table 2: per-sample cohort, cancer type, treatment history and mutation load.""" + hdr, rows = sheet("41588_2023_1367_MOESM4_ESM.xlsx", "Supp. Table 2 - sample metadata ") + num = {"ploidy", "sv_load", "smnv_load", "sbs_load", "dbs_load", "indel_load", + "sbs_load.clonal", "dbs_load.clonal", "indel_load.clonal", + "sbs_load.subclonal", "dbs_load.subclonal", "indel_load.subclonal"} + bul = {"blacklisted_gie", "is_selected", "has_lilac", "has_group_cohort", + "has_group_mut_signatures", "has_rna", "whole_genome_duplication", "had_radiotherapy", + "had_other_treatment", "had_chemotherapy", "had_hormone_therapy", "had_targeted_therapy", + "had_immunotherapy", "treatment_info_available"} + return _df([ + {c: (_f(r[n]) if c in num else _b(r[n]) if c in bul else _s(r[n])) + for n, c in enumerate(hdr)} for r in rows]) + + +def escape_genes(): + """Supplementary Table 1: the immune-escape gene panel the paper's calls are made over.""" + hdr, rows = sheet("41588_2023_1367_MOESM4_ESM.xlsx", "Supp. Table 1 - genes") + keys = ["gene", "pathway_specific", "pathway_general", "references", "mutation_types", + "selected"] + return _df([ + {k: (_f(r[n]) if k == "selected" else _s(r[n])) for n, k in enumerate(keys)} + for r in rows]) + + +def main() -> int: + out = {"gie_allele.parquet": None, "gie_sample.parquet": None} + out["gie_allele.parquet"], out["gie_sample.parquet"] = gie_tables() + out["hla_typing_callers.parquet"] = typing_callers() + out["sample_metadata.parquet"] = sample_metadata() + out["escape_genes.parquet"] = escape_genes() + for name, df in out.items(): + df.write_parquet(os.path.join(HERE, name), compression="zstd") + print(f"{name:28s} {df.height:>7,} rows x {df.width:>2} cols") + return 0 + + +if __name__ == "__main__": + raise SystemExit(main()) diff --git a/hla_loh/hla_typing_callers.parquet b/hla_loh/hla_typing_callers.parquet new file mode 100644 index 0000000000000000000000000000000000000000..818574a4e36f2e435ea664fce8b9aadb5a8da3bf --- /dev/null +++ b/hla_loh/hla_typing_callers.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d4881daeaa39eb998db01ec15579581e65f68066746932f2a7f99328f053bfb5 +size 148451 diff --git a/hla_loh/perallele_tcga_loh.parquet b/hla_loh/perallele_tcga_loh.parquet new file mode 100644 index 0000000000000000000000000000000000000000..160e517b845f8a370bc1548924dc36469a7edd84 --- /dev/null +++ b/hla_loh/perallele_tcga_loh.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4d507a72aabb96e5bdf30f9f481666554a2c4e7b93bf5b2162bb2dd440f6e95c +size 7649 diff --git a/hla_loh/sample_metadata.parquet b/hla_loh/sample_metadata.parquet new file mode 100644 index 0000000000000000000000000000000000000000..16e19e7c39988e3cb5b44d94ef2ae8011d1c5804 --- /dev/null +++ b/hla_loh/sample_metadata.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c579034e9bc484225c8b68d0d97dfb72ac80b9a49526c6dd0248b59548d6ed6a +size 340067 diff --git a/homology/HOM_MouseHumanSequence.rpt.gz b/homology/HOM_MouseHumanSequence.rpt.gz new file mode 100644 index 0000000000000000000000000000000000000000..3cb354f3bd2a80ae483a64987af4ad4c1dc18b4d --- /dev/null +++ b/homology/HOM_MouseHumanSequence.rpt.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9985b691a867025f24bda5af0f28a91b1aa2d28d3e06d4ee7d0db54fa217c12c +size 3243831 diff --git a/homology/SOURCES.md b/homology/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..1f58bde61d4f3bcf701727d60793189f5ff38735 --- /dev/null +++ b/homology/SOURCES.md @@ -0,0 +1,50 @@ +# `homology/` — sources + +One file, and it exists to answer exactly one question: **given a human protein, which mouse gene +is its orthologue, spelled the way MGI spells it.** + +## `HOM_MouseHumanSequence.rpt.gz` + +| field | value | +|---|---| +| origin | `https://www.informatics.jax.org/downloads/reports/HOM_MouseHumanSequence.rpt` | +| format | tab-separated, 13 columns, gzipped here (3,243,831 bytes; 15,110,808 raw) | +| retrieved | 2026-09-04; the served file was stamped `last-modified: 2026-08-31` | +| provenance | **derived / computed — not experimental.** Ortholog assignments are *inferred*, not measured | +| rows | 46,521 data rows over **21,928 homology classes** | +| organisms | 24,592 `human` rows, 21,929 `mouse, laboratory` rows | +| distinct symbols | 19,366 human, 21,926 mouse | +| human rows carrying a SWISS_PROT ID | 23,569 of 24,592 (95.8 %) | + +```zsh +curl -sL -o HOM_MouseHumanSequence.rpt \ + https://www.informatics.jax.org/downloads/reports/HOM_MouseHumanSequence.rpt +gzip -9 HOM_MouseHumanSequence.rpt +``` + +**MGI is a redistributor here, not the source of the inference.** `informatics.jax.org/homology.shtml` +states that MGI "uses homology class data from the Alliance of Genome Resources", and does **not** +state the method by which the Alliance computes an ortholog assignment. So the right claim to make +downstream is "MGI/Alliance homology class", never "orthologue" as though it were measured — and a +one-to-many class is a real answer, not a failure to resolve. + +**Why this file and not `HOM_ProteinCoding.rpt` or a UniProt round-trip.** Two properties, and both +are load-bearing: + +1. **It carries `SWISS_PROT IDs`**, so a human UniProt accession reaches a mouse gene in *one* hop. + `HOM_ProteinCoding.rpt` has no accession column and would need the proteome FASTA first. +2. **Its `Symbol` column is the MGI official symbol**, which is the vocabulary + `expression/toil_matrix_mmu.npz` is keyed on (26,737 genes). + +That second one is the whole reason a curated table was fetched at all. The obvious cheap route -- +read `GN=` out of `proteome/{human,mouse}.fasta.gz` and match UniProt's shared entry-name mnemonic +(`P53_HUMAN` / `P53_MOUSE`) -- resolves 14 of 21 accessions and returns UniProt's `GN=Tp53` for +human TP53. **`Tp53` is absent from `toil_matrix_mmu.npz`; the MGI official symbol `Trp53` is +present.** A lookup that returns nothing and a lookup that returns the wrong thing fail the same way +when the layer below drops silently, which is this package's recurring hazard. Through this file the +same 21 accessions resolve 16 directly, and 21 of 21 once a human-symbol fallback is added. + +**Two accessions on `source_species = human` rows of `neoantigens/neoag_tested_mmu.tsv.gz` are +*mouse* accessions** -- `O54916` (Reps1) and `Q8VDL4` (Adpgk). They are not xenoantigens; the +deposit's species label is wrong on those two rows. A resolver that reports which route answered +sees this; one that silently falls through does not. diff --git a/icb_cohorts/COHORTS.md b/icb_cohorts/COHORTS.md new file mode 100644 index 0000000000000000000000000000000000000000..ffe012bb5e0701a828adb2df723819dd9adf4e04 --- /dev/null +++ b/icb_cohorts/COHORTS.md @@ -0,0 +1,77 @@ +# The cohort table — what each one can carry, and what it is missing + +Reconciled 2026-09-10 from the screen recorded in `2026-mhcmatch-benchmark/bench/results/cohort_availability.md` +and the provenance in `SOURCES.md` beside this file. The analysis plan that consumes this table is +`2026-mhcmatch-benchmark/REPLICATION_PLAN.md`. + +**Three things are needed per patient**: somatic mutation calls at protein level (a tumour mutational +burden count is not enough), four-digit class-I HLA-A/B/C, and a clinical outcome. Tumour RNA is a +fourth that improves two of the nine EPIC terms and is not required — `expr_lvl` and `expr_norm` fall +back to the GTEx/TCGA tumour-type median with `expr_imputed` recording that they did. + +**Typing is the scarce leg.** Of 200 distinct cohorts screened, 74 are blocked on typing alone. +`harmonized/typing.parquet` is the result of searching for it: 52,492 patients, 11 cohorts, +360 distinct allotypes. + +## Tier 1 — mutations, true typing and outcome, all public + +| cohort | patients | tumour | therapy | mutation calls | clone or CCF layer | typing source | tumour RNA | outcomes | +|---|--:|---|---|---|---|---|---|---| +| MSK-IMPACT checkpoint union | 2,655 | pan-cancer | PD-1, PD-L1, CTLA-4, combination | cBioPortal panel MAF | FACETS cancer-cell fraction | `msk_impact_50k_2026` germline, six alleles | none | overall survival 2,655; progression 1,699 | +| SU2C-MARK (Ravi 2023) | 309 | non-small-cell lung | anti-PD-(L)1 | exome MAF | PhyloGic cluster per mutation | Polysolver, Table S9 | subset | overall survival, progression-free survival, RECIST | +| Van Allen 2015 | 110 | melanoma | anti-CTLA-4 | exome MAF | none | cBioPortal sample attribute | none | overall survival, disease-free survival, best response | +| Anagnostou 2020 | 89 | non-small-cell lung | anti-PD-(L)1 | supplementary S3 | cancer-cell fraction per mutation | supplementary S11, with per-locus loss of heterozygosity | none | response, progression-free survival, overall survival | +| Hellmann 2018 | 75 | non-small-cell lung | nivolumab + ipilimumab | exome MAF | none | cBioPortal sample attribute | none | progression-free survival, best response, durable benefit | +| Riaz 2017 | 68 | melanoma | nivolumab | open exome `SRP095809` | none | Chowell Table S1 | yes | response, progression-free survival, overall survival | +| Snyder 2014 | 64 | melanoma | anti-CTLA-4 | exome MAF | none | cBioPortal sample attribute, class II too | none | overall survival, response, durable benefit | +| Hugo 2016 | 38 | melanoma | anti-PD-1 | exome MAF | none | cBioPortal sample attribute | yes | overall survival, response | +| Rizvi 2015 | 34 | non-small-cell lung | pembrolizumab | *Science* supplement | none | Chowell Table S1 | none | response, progression-free survival | +| Forde 2018 | 13 | non-small-cell lung, neoadjuvant | nivolumab | supplementary S6 | none | recoverable from supplementary S8 | none | pathological response | + +The eight exome cohorts below SU2C-MARK pool to **491 patients** with deep calls and true typing. +That pool, not MSK, is the like-for-like replication of the SU2C-MARK arm. + +## Tier 2 — mutations and outcome public, typing absent, imputation available + +| cohort | patients with calls | patients with outcome | tumour | randomised comparison | mutation calls | clone or CCF layer | tumour RNA | outcomes | +|---|--:|--:|---|---|---|---|---|---| +| Jammihal 2025 (CheckMate-214 + JAVELIN Renal 101) | 953 | 950 | renal cell | 465 checkpoint against 488 tyrosine-kinase inhibitor | 129,726 mutations with `Protein_Change` | ABSOLUTE `ccf_hat` and a clonality call | 733 | objective response, RECIST, progression-free survival, overall survival | +| Braun 2020 (CheckMate 009/010/025) | 454 | 1,006 | renal cell | 609 nivolumab against 397 everolimus | 68,030 mutations, 26,434 missense | none | 311 | objective response, benefit, progression-free survival, overall survival, immune-related progression-free survival | + +Both are blocked on typing alone and both carry a **randomised control arm**, which no tier-1 cohort +does. Under imputed genotypes the patient-level ordering survives at Spearman rho +0.9433 at k = 20 +and the top-k peptide sets do not (mean Jaccard 0.143) — `bench/results/hla_impute_check.md`. So +these two carry an outcome analysis and not a design claim. + +## Tier 3 — one email away + +| ask | patients returned | who | what to request | +|---|--:|---|---| +| Jammihal and Braun typing | ~1,400 | David A. Braun (Yale), Sachet A. Shukla — senior authors of both | the POLYSOLVER class-I calls, keyed to `SUBJID` and `UUID`. Promotes both tier-2 rows to tier 1 | +| IMvigor210 | 348 | Genentech | class-I genotypes; exome MAF, RECIST and overall survival are public | +| IMmotion150 | 263 | David F. McDermott (BIDMC) | Polysolver calls; the apparently public typing is an MHCflurry argmax complete for 45 of 157 | +| Liu 2019 | 144 | Dana-Farber / Essen | the POLYSOLVER calls run for the paper's own loss-of-heterozygosity analysis, never published | + +## Tier 4 — typing computable from open reads, no email needed + +| cohort | patients | reads | already public | +|---|--:|---|---| +| MORRISON-1 (CheckMate 038/064/067) | 427 | `PRJNA923698`, 160 open exome runs | variant allele frequency, logCPM, RECIST at `ParkerICI/MORRISON-1-public` | +| Zhao 2019 glioblastoma | 66 | `PRJNA482620`, 91 of 91 runs open | mutations, response, survival | +| Kim 2018 gastric | 61 | `PRJEB25780`, 95 exomes and 71 RNA | mutations, objective response, progression-free survival | +| Hong 2022 hepatocellular | 60 | `PRJEB34724`, 89 exomes and 40 RNA | mutations, objective response, progression-free survival, overall survival | + +## Tier 5 — typed and public, but not checkpoint blockade + +Survival cohorts beside TCGA rather than beside SU2C-MARK: PCAWG / ICGC 1,028 donors closing all +three legs; TRACERx 421 with 89 typed and relapse-free survival; ICGC OV-AU ovarian 93 donors. +Hartwig 4,439 and PCAWG 1,839 carry LILAC typing in the open; Hartwig's mutations and outcome need +an application. + +## Ruled out, so nobody looks twice + +Gide 2019 and Kang/Lee 2024 sequenced no exome. IMmotion151 publishes no per-mutation calls. +Chowell Cohort 2 has typing and overall survival but its identifiers cannot reach per-mutation +calls. Litchfield CPI1000+ is anonymised. AACR GENIE BPC releases no germline. Cloughesy 2019, +Khan 2020, Rose 2021, MOSCATO-01 and TOPACIO release too little. The 525-patient NeoPrecis +compendium publishes no typing — it stays a directional-rule arm only, `bench/results/cassette_ici.md`. diff --git a/icb_cohorts/SOURCES.md b/icb_cohorts/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..5a7cdb4323af6f7ec6db250e2baefec782ede9e9 --- /dev/null +++ b/icb_cohorts/SOURCES.md @@ -0,0 +1,45 @@ +# ICB cohort assembly — sources + +Staged 2026-09-10. **`raw/` is gitignored and not published** -- publisher supplementary workbooks, +copyrighted and not ours to redistribute. **`harmonized/` is published** (2026-09-11): tables we +derive from them -- our own six-allele genotypes over public typings -- carrying no publisher content. + +The analysis being assembled for needs, per patient: somatic mutation calls at protein level, +four-digit class-I HLA-A/B/C, and a clinical outcome. **Typing is the scarce leg**, so it is +harmonised first, by `harmonize_typing.py`. + +## `raw/` — what each file is, and where it came from + +| file | bytes | source | what it contributes | +|---|--:|---|---| +| `chowell2018_TableS1.xlsx` | 196,759 | Chowell D et al., *Science* 2018;359:582-587, doi:10.1126/science.aao4572, PMID 29301960, PMC6057471 — Supplementary Table S1 (`NIHMS980063-supplement-Supplemental_Table_S1.xlsx`) | **experimental** typing. `Cohort 1` 371 patients (Van Allen 100, Riaz 68, Rizvi_CUMC 67, Snyder 64, Hugo 38, Rizvi 2015 34) with `HLA Class I Alleles`, `OS_Months`, `OS_Event`, `Drug Class`, `MutCnt`; `Cohort 2` 1,166 MSK-IMPACT with the same; `HLA-II` 304; a `TCR CDR3s` sheet besides. 370 of 371 and 1,166 of 1,166 carry six four-digit alleles | +| `forde2018_nejm_S4-S10.xlsx`, `forde2018_nejm_supp1.pdf` | 3,077,810 / 1,662,342 | Forde PM et al., *N Engl J Med* 2018;378:1976-1986 — neoadjuvant nivolumab in resectable NSCLC (`NIHMS980037`) | **experimental**. `S6` 1,609 somatic alterations over 13 samples with hg19 coordinates, `Amino Acid (protein)`, CCDS transcript and consequence; `S8` 19,585 neoantigen rows carrying `HLA Allele` — six alleles per sample, so the genotype is recoverable for all 13 | +| `anagnostou2020_natcancer_S.xlsx` | 9,649,675 | Anagnostou V et al., *Nat Cancer* 2020;1:99-111, doi:10.1038/s43018-019-0008-8, PMC7514475 — Supplementary Tables | **experimental**. `S11` *HLA class I genomic variation*: 89 patients x six four-digit alleles plus per-locus LOH; `S3` per-mutation calls with cancer cell fraction | +| `braun2020_natmed_MOESM2.xlsx` | 125,569,016 | Braun DA et al., *Nat Med* 2020;26:909-918, doi:10.1038/s41591-020-0839-y — CheckMate 009/010/025 ccRCC | **experimental**. `S1_Clinical_and_Immune_Data` 1,006 patients x 121 columns (`Arm` nivolumab 609 / everolimus 397; `Cohort` CM-025 803, CM-010 168, CM-009 35; `ORR`, `Benefit`, `PFS`+`PFS_CNSR`, `OS`+`OS_CNSR`, `irPFS`); `S2_WES_Data` 68,030 somatic mutations (26,434 missense) over 454 tumours with `Protein_Change`; `S4A_RNA_Expression` on 311. **No HLA sheet — read every sheet, 2026-09-10** | +| `jammihal2025_rcc_S1-S6.xlsx` | 195,455,480 | Jammihal T et al., *Nat Cancer* 2025;6:372-384, doi:10.1038/s43018-024-00896-w, PMID 39789182 — CheckMate-214 + JAVELIN Renal 101 | **experimental**. `Supp. Table 1` 1,958 patients (arms, ORR, RECIST, PFS/OS with censoring); `Supp. Table 4` 129,726 mutations over 953 patients with `Protein_Change` and ABSOLUTE `ccf_hat`/clonality; `Supp. Table 3` expression on 733. **No HLA sheet** | +| `mjimenez2023_hla_suppl1.xlsx` | 882,368 | Martinez-Jimenez F et al., *Nat Genet* 2023;55:820-831, doi:10.1038/s41588-023-01367-1, PMID 37165135 — Supplementary Data 1 | **derived from experimental reads**: LILAC / xHLA / Polysolver class-I typing on normal *and* tumour separately. Hartwig 4,439 samples, PCAWG 1,839, plus TRACERx typing (491 rows) and LOH sheets | +| `neoprecis_nci_alleles.csv` | 12,622 | `cartercompbio/NeoPrecis`, `data/nci_alleles.csv` | **experimental** typing for the 75 NCI Surgery Branch GI patients, class I and class II | + +Fetched with `curl -A ` from `static-content.springer.com`; PMC serves a bot challenge +for the same files and must not be used. Chowell and Forde are NIHMS deposits and arrived through +the author's own browser download. + +## `harmonized/` + +`typing.parquet` — one row per patient, columns `cohort`, `study`, `patient`, `A1 A2 B1 B2 C1 C2`, +every allele written `HLA-A*02:01`. A homozygous locus repeats its allele. Built 2026-09-10: + +| source | patients | breakdown | +|---|--:|---| +| Chowell 2018 | 1,536 | `chowell_c1` 370, `chowell_c2` 1,166 | +| MSK-IMPACT 50K (cBioPortal API) | 44,227 | germline, full field, cut to two | +| Martinez-Jimenez 2023 | 6,278 | Hartwig 4,439, PCAWG 1,839 (LILAC **normal** calls, not tumour) | +| cBioPortal checkpoint-blockade studies | 287 | Van Allen 110, Hellmann 75, Snyder 64, Hugo 38 | +| Anagnostou 2020 | 89 | | +| NCI Surgery Branch | 75 | | +| **total** | **52,492** | 11 cohorts, 360 distinct allotypes | + +Mutation and outcome tables are deliberately **not** built yet: the selection is open, and +harmonising 100k-row MAFs for cohorts that may be dropped is work spent before the decision. + +The reconciled cohort table is `COHORTS.md` beside this file; the analysis plan is `2026-mhcmatch-benchmark/REPLICATION_PLAN.md`. diff --git a/icb_cohorts/harmonize_typing.py b/icb_cohorts/harmonize_typing.py new file mode 100644 index 0000000000000000000000000000000000000000..94dd076e0a8ee2ce60f54656a1ac409f461fb303 --- /dev/null +++ b/icb_cohorts/harmonize_typing.py @@ -0,0 +1,199 @@ +#!/usr/bin/env python3 +# 2026-09-10 -- one class-I typing table across every cohort whose genotypes we can reach. +"""Class-I typing is the scarce leg of every checkpoint-blockade cohort, so it is harmonised first. + +Writes `harmonized/typing.parquet` with one row per (cohort, patient, locus, allele_1, allele_2) +flattened to six columns, every allele in `HLA-A*02:01` form, plus `harmonized/manifest.tsv` +naming what each source contributes. Mutation and outcome tables are deliberately NOT built yet: +the manuscript's cohort selection is open, and harmonising 100k-row MAFs for cohorts that may be +dropped is work spent before the decision. + + uv run --with openpyxl --with polars python3 harmonize_typing.py +""" +from __future__ import annotations + +import collections +import csv +import json +import os +import re +import urllib.request + +HERE = os.path.dirname(os.path.abspath(__file__)) +RAW = os.path.join(HERE, "raw") +OUT = os.path.join(HERE, "harmonized") +LOCI = ("A", "B", "C") + + +def norm(a: str) -> str | None: + """Every spelling seen across these sources -> ``HLA-A*02:01``, or None if it is not one.""" + if a is None: + return None + s = str(a).strip() + if not s or s.upper() in ("NA", "NAN", "-", "NONE"): + return None + s = s.replace("HLA-", "").replace("hla_", "").replace("HLA", "") + m = re.match(r"^([ABC])\*?(\d{2}):?(\d{2})", s) # A*02:01 / A0201 / A02:01 + if m: + return f"HLA-{m.group(1)}*{m.group(2)}:{m.group(3)}" + m = re.match(r"^_?([abc])_(\d{2})_(\d{2})", s) # hla_a_02_01_01_01 + if m: + return f"HLA-{m.group(1).upper()}*{m.group(2)}:{m.group(3)}" + return None + + +def six(alleles: list[str]) -> dict | None: + """Bucket a patient's calls by locus and require two per locus (a homozygote repeats).""" + by = collections.defaultdict(list) + for a in alleles: + n = norm(a) + if n: + by[n[4]].append(n) + row = {} + for L in LOCI: + v = sorted(set(by[L])) or [] + if len(v) == 1: + v = v * 2 + if len(v) != 2: + return None + row[f"{L}1"], row[f"{L}2"] = v + return row + + +def chowell() -> list[dict]: + import openpyxl + wb = openpyxl.load_workbook(f"{RAW}/chowell2018_TableS1.xlsx", read_only=True, data_only=True) + out = [] + for sheet, tag in (("Cohort 1", "chowell_c1"), ("Cohort 2", "chowell_c2")): + ws = wb[sheet] + it = ws.iter_rows(values_only=True) + hdr = [h for h in next(it) if h] + ia, ir = hdr.index("HLA Class I Alleles"), (hdr.index("Reference") if "Reference" in hdr else None) + for r in it: + if not r or not r[0]: + continue + g = six(str(r[ia]).split(",")) + if g: + out.append({"cohort": tag, "study": (r[ir] if ir is not None else tag), + "patient": str(r[0]), **g}) + return out + + +def anagnostou() -> list[dict]: + import openpyxl + wb = openpyxl.load_workbook(f"{RAW}/anagnostou2020_natcancer_S.xlsx", read_only=True, + data_only=True) + ws = wb["S11"] + rows = [r for r in ws.iter_rows(values_only=True) if r and any(x is not None for x in r)] + hdr = [str(x) for x in rows[1]] + ip = hdr.index("Patient ID") if "Patient ID" in hdr else 0 + cols = [i for i, h in enumerate(hdr) if h.startswith("HLA-") and "Allele" in h] + out = [] + for r in rows[2:]: + g = six([r[i] for i in cols]) + if g: + out.append({"cohort": "anagnostou2020", "study": "Anagnostou 2020 Nat Cancer", + "patient": str(r[ip]), **g}) + return out + + +def hartwig_pcawg() -> list[dict]: + import openpyxl + wb = openpyxl.load_workbook(f"{RAW}/mjimenez2023_hla_suppl1.xlsx", read_only=True, + data_only=True) + out = [] + for sheet, tag in (("Hartwig", "hartwig"), ("PCAWG", "pcawg")): + ws = wb[sheet] + it = ws.iter_rows(values_only=True) + hdr = [str(h) for h in next(it)] + cols = [i for i, h in enumerate(hdr) if h.startswith("LILAC_n_")] # normal, not tumour + for r in it: + if not r or not r[0]: + continue + g = six([r[i] for i in cols]) + if g: + out.append({"cohort": tag, "study": "Martinez-Jimenez 2023 Nat Genet", + "patient": str(r[0]), **g}) + return out + + +def nci() -> list[dict]: + out = [] + with open(f"{RAW}/neoprecis_nci_alleles.csv") as f: + for row in csv.DictReader(f): + cols = [v for k, v in row.items() if k and re.match(r"^[ABC]_\d$", k)] + g = six(cols) + if g: + out.append({"cohort": "nci_gi", "study": "NCI Surgery Branch (via NeoPrecis)", + "patient": row.get("Patient") or row.get("sample") or "", **g}) + return out + + +def cbioportal() -> list[dict]: + """The four checkpoint-blockade studies whose typing is a SAMPLE-level attribute.""" + api = "https://www.cbioportal.org/api/studies/{}/clinical-data?clinicalDataType=SAMPLE&projection=SUMMARY&pageSize=100000" + studies = {"skcm_dfci_2015": "Van Allen 2015", "skcm_mskcc_2014": "Snyder 2014", + "mel_ucla_2016": "Hugo 2016", "nsclc_mskcc_2018": "Hellmann 2018"} + out = [] + for sid, name in studies.items(): + with urllib.request.urlopen(api.format(sid), timeout=120) as fh: + data = json.load(fh) + S = collections.defaultdict(dict) + for r in data: + S[r["sampleId"]][r["clinicalAttributeId"]] = r["value"] + for sample, d in S.items(): + vals = [] + for k, v in d.items(): + if k.startswith("HLA_") and "LOH" not in k: + vals += str(v).split("/") + g = six(vals) + if g: + out.append({"cohort": sid, "study": name, "patient": sample, **g}) + return out + + +def msk50k() -> list[dict]: + api = ("https://www.cbioportal.org/api/studies/msk_impact_50k_2026/clinical-data" + "?clinicalDataType=PATIENT&attributeId={}&projection=SUMMARY&pageSize=200000") + H = collections.defaultdict(list) + for a in ("HLA-A1", "HLA-A2", "HLA-B1", "HLA-B2", "HLA-C1", "HLA-C2"): + with urllib.request.urlopen(api.format(a), timeout=300) as fh: + for r in json.load(fh): + H[r["patientId"]].append(r["value"]) + out = [] + for p, vals in H.items(): + g = six(vals) + if g: + out.append({"cohort": "msk_impact_50k", "study": "MSK-IMPACT 50K", "patient": p, **g}) + return out + + +SOURCES = {"chowell2018": chowell, "anagnostou2020": anagnostou, + "martinez_jimenez2023": hartwig_pcawg, "nci_neoprecis": nci, + "cbioportal_icb": cbioportal, "msk_impact_50k": msk50k} + + +def main() -> None: + import polars as pl + os.makedirs(OUT, exist_ok=True) + rows, man = [], [] + for name, fn in SOURCES.items(): + try: + got = fn() + except Exception as e: # noqa: BLE001 + print(f"# {name}: FAILED {e}") + man.append({"source": name, "patients": 0, "note": f"failed: {e}"}) + continue + rows += got + by = collections.Counter(g["cohort"] for g in got) + print(f"# {name}: {len(got):,} patients {dict(by)}") + man.append({"source": name, "patients": len(got), "note": json.dumps(dict(by))}) + df = pl.DataFrame(rows) + df.write_parquet(f"{OUT}/typing.parquet") + pl.DataFrame(man).write_csv(f"{OUT}/manifest.tsv", separator="\t") + print(f"\n# typing.parquet: {df.height:,} patients over {df['cohort'].n_unique()} cohorts, " + f"{len(set(df['A1']) | set(df['A2']) | set(df['B1']) | set(df['B2']) | set(df['C1']) | set(df['C2']))} distinct allotypes") + + +if __name__ == "__main__": + main() diff --git a/icb_cohorts/harmonized/manifest.tsv b/icb_cohorts/harmonized/manifest.tsv new file mode 100644 index 0000000000000000000000000000000000000000..11157b72173349604f08e91f69684fb747b06fae --- /dev/null +++ b/icb_cohorts/harmonized/manifest.tsv @@ -0,0 +1,7 @@ +source patients note +chowell2018 1536 "{""chowell_c1"": 370, ""chowell_c2"": 1166}" +anagnostou2020 89 "{""anagnostou2020"": 89}" +martinez_jimenez2023 6278 "{""hartwig"": 4439, ""pcawg"": 1839}" +nci_neoprecis 75 "{""nci_gi"": 75}" +cbioportal_icb 287 "{""skcm_dfci_2015"": 110, ""skcm_mskcc_2014"": 64, ""mel_ucla_2016"": 38, ""nsclc_mskcc_2018"": 75}" +msk_impact_50k 44227 "{""msk_impact_50k"": 44227}" diff --git a/icb_cohorts/harmonized/typing.parquet b/icb_cohorts/harmonized/typing.parquet new file mode 100644 index 0000000000000000000000000000000000000000..9b4bc604e236dad0b299ecb04487f3f99d32b0cf --- /dev/null +++ b/icb_cohorts/harmonized/typing.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b8a7494b2d3c73142a7f5cc707bb093732b9c6e23473b0d425842d0daa5669c +size 289060 diff --git a/immunogenicity/SOURCES.md b/immunogenicity/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..37ef8692d17c78aa3213775bd85082478a474dbc --- /dev/null +++ b/immunogenicity/SOURCES.md @@ -0,0 +1,726 @@ +# `immunogenicity/` — sources, provenance, filtering and a complete audit + +General immunogenicity corpora: **is this presented peptide seen by a T cell?** Neoantigen datasets +moved to [`../neoantigens/`](../neoantigens/SOURCES.md) — this directory no longer holds NCI, +`neoag_*` or any tumour-specific screen. + +Every citation resolved against **PubMed**; PMIDs and DOIs given. Audit regenerated by +`bench/neoag/audit_pmhc_data.py` in `2026-mhcmatch-benchmark`; corpora rebuilt by +`bench/neoag/corpus_chowell.py` + `bench/neoag/deposit_corpora.py`. + +Written 2026-08-17. + +--- + +## Complete audit + +Measured from the files, not copied from any prior description. + +| file | rows | distinct peptides | pos | neg | prevalence | HLA alleles | length | +|---|--:|--:|--:|--:|--:|--:|:--:| +| `chowell_rebuilt.tsv.gz` | 511,301 | 315,191 | 19,866 | 491,435 | 0.0389 | 210 | 8–11 | +| `chowell_rebuilt_hla_matched.tsv.gz` | 115,592 | 99,763 | 19,866 | 95,726 | 0.1719 | 183 | 8–11 | +| `kesmir_rebuilt.tsv.gz` | 65,737 | 36,155 | 22,613 | 43,124 | 0.3440 | 196 | 8–11 | +| `kesmir_rebuilt_hla_matched.tsv.gz` | 59,941 | 35,202 | 22,613 | 37,328 | 0.3773 | 188 | 8–11 | +| `iedb_labeled.tsv.gz` | 790,868 | 516,092 | 34,362 | 756,506 | 0.0435 | 250 | 2–50 | +| `hla_pop_freqs.tsv.gz` | 7,171 | — | — | — | — | 4,234 | — | + +No donor/patient identifier exists in any file here — these are peptide-level corpora aggregated +over studies, not per-donor cohorts. (Donor counts are meaningful in `../neoantigens/` and are +reported there.) + +--- + +## `chowell_rebuilt.tsv.gz` — immunogenic vs presented **self** + +Chowell's construction, rebuilt from the current IEDB dump rather than read off the 2015 +spreadsheet, which is copyrighted supplementary material and frozen at 2015. + +> Chowell D, Krishna S, Becker PD, Cocita C, Shu J, Tan X, Greenberg PD, Klavinskis LS, +> Blattman JN, Anderson KS. **TCR contact residue hydrophobicity is a hallmark of immunogenic CD8+ +> T cell epitopes.** *Proc Natl Acad Sci USA* 2015;112(14):E1754–62. +> PMID [25831525](https://pubmed.ncbi.nlm.nih.gov/25831525/) · +> doi:[10.1073/pnas.1500973112](https://doi.org/10.1073/pnas.1500973112) + +### Filtering procedure, in order + +1. **Source** — `iedb_labeled.tsv.gz` (see below), itself an IEDB export. +2. **Class** — `mhc_class == "MHCI"`. +3. **Host** — `host_species ∈ {human, mouse}`; kept as a column and **never pooled**. The two hosts + have different MHC and different thymic repertoires, so a fit across them without the split is + fitting a mixture. +4. **Peptide** — uppercased, whitespace-stripped; length **8–11**; canonical 20 amino acids only. +5. **Exclusions** — `SLLMWITQV`, `KLGGALQAK` (NY-ESO-1 / MAGE-A3 epitopes that appear on both sides + of the literature with contradictory annotation). +6. **Positives** — any peptide with a positive T-cell assay (`dataset_origin == "iedb_tcell"`). +7. **Negatives** — eluted ligands whose `source_species` equals the **host** (self), plus the HLA + Ligand Atlas thymus immunopeptidome for the human arm. +8. **Label is a property of the (peptide, host), not the row.** 1,075 peptides are both eluted and + positively assayed. Elution is evidence of *presentation*; it is never evidence against a T-cell + response. Assigning per row makes the same sequence positive under one allele and negative under + another — it sits on both sides of the fit. Computing the positive set once per (peptide, host) + and applying it everywhere raised self-sourced positives by **57%**. +9. **Aggregation** — one row per `(peptide, allele, host_species)`, label = max. + +### Not controlled, deliberately + +**Allele and length.** Immunogenic and eluted sets genuinely differ in both: some HLAs are +protective and some are not, and immunogenic peptides have a real length distribution that differs +from bulk ligands. Matching those away removes effects, not confounds. `allele` and `length` ship +as columns so a stratified analysis is a filter, not a rebuild — and the HLA-matched variant below +exists for when allele composition *is* the thing to remove. + +--- + +## `kesmir_rebuilt.tsv.gz` — immunogenic vs presented **non-self** + +The Calis/Kesmir contrast. Same shape, same rules, one difference: negatives are eluted ligands +from a source organism that is **not** the host. + +> Calis JJA, Maybeno M, Greenbaum JA, Weiskopf D, De Silva AD, Sette A, Keşmir C, Peters B. +> **Properties of MHC class I presented peptides that enhance immunogenicity.** +> *PLoS Comput Biol* 2013;9(10):e1003266. +> PMID [24204222](https://pubmed.ncbi.nlm.nih.gov/24204222/) · +> doi:[10.1371/journal.pcbi.1003266](https://doi.org/10.1371/journal.pcbi.1003266) + +**This arm requires `../ligandome/viral_foreign_iedb.tsv.gz`, and is wrong without it.** Every +eluted-ligand negative in the IEDB table itself is self-sourced — 449,449 human-self and 41,986 +mouse-self against 149 + 63 cross-species, and zero from any other organism. Filtering +`source_species != "human"` on a **mouse** host therefore selects mouse *self* ligands and rebuilds +Chowell under another name. The foreign ligandome supplies 41,294 genuine human non-self negatives. + +Kesmir removes the self/non-self axis that Chowell confounds with immunogenicity, and is the harder +problem: prevalence 0.344 against Chowell's 0.039, and every negative is a foreign peptide that was +presented and still not attacked. + +--- + +## The HLA-matched variants + +`*_rebuilt_hla_matched.tsv.gz`, built by `deposit_corpora.py --ratio 5 --seed 20260817`. + +There are 5–25× more negatives than positives and the two sides do not share an allele +distribution, so a classifier can separate them on **allele composition alone**. For each allele, +negatives are drawn without replacement to **5 per positive of that same allele**: + +* alleles with no positives contribute **no** negatives — keeping them would reintroduce exactly the + imbalance being removed; +* alleles with too few negatives contribute all they have, and the shortfall stays visible in the + per-allele counts rather than being topped up from another allele; +* the draw is seeded (`20260817`), so the file is reproducible rather than one sample. + +Effect: Chowell 491,435 → 95,726 negatives, 210 → 183 alleles, prevalence 0.039 → 0.172. Kesmir +43,124 → 37,328, 196 → 189 alleles. + +Use the matched file when the question is about peptide sequence; use the full file when allele +composition is part of the signal you want. Neither is "the" corpus. + +--- + +## Two residue-level artefacts, measured + +Both are properties of how the data was generated. A model will fit either as if it were biology. + +| corpus | Cys % positives | Cys % negatives | ratio | C-term R/K % pos | C-term R/K % neg | +|---|--:|--:|--:|--:|--:| +| `chowell_rebuilt` | 10.69 | **1.55** | **6.9×** | 9.21 | 9.50 | +| `kesmir_rebuilt` | 10.94 | 14.50 | 0.8× | 9.91 | **20.33** | +| `iedb_labeled` | 14.56 | **2.52** | **5.8×** | 10.90 | 13.98 | + +### Cysteine — an MS detection artefact, and it hits Chowell hardest + +Cysteine is systematically **under-detected in mass-spectrometry immunopeptidomics** unless samples +are alkylated. Where positives are T-cell-assayed (synthetic peptides, no MS step) and negatives are +MS-eluted, the corpus carries a Cys gradient with no biological content. Chowell shows **6.9×**; +finer-grained, the thymus MS subset runs at 0.17% against 11.59% in the assayed positives. + +Fitted freely on the Chowell corpus, Cys took the **single largest coefficient** in a position-role +naive Bayes (+1.84 anchor / +2.05 TCR-facing). Masking it costs grouped-CV AUROC 0.712 → 0.690 and +*improves* transfer to neoantigen cohorts. The shipped `mhcmatch.posbayes` tables zero it. + +**`kesmir_rebuilt` does not have this problem** (0.8×) — its negatives are IEDB-curated foreign +ligands rather than a single MS pipeline. That makes it the safer corpus for fitting residue-level +terms, and is a reason to run both. + +### C-terminal Arg/Lys — genuine anchor preference, or tryptic carry-over + +A basic C-terminus is the real anchor preference of **HLA-A\*03, A\*11, A\*31, A\*68 and B\*27** — +and Arg/Lys are exactly the residues **trypsin cleaves after**. An elevated rate can be either, and +the two are **not separable from sequence alone**. + +`kesmir_rebuilt` negatives run at **20.33%** against 9.91% in positives — a 2× gradient in the +corpus that is otherwise the cleaner of the two. Before attributing anything to a C-terminal basic +residue on this corpus, stratify by allele: if the effect survives within A\*03/A\*11/B\*27 and +within non-basic-preferring alleles separately, it is anchor preference; if it only appears pooled, +it is composition or contamination. + +--- + +## `iedb_labeled.tsv.gz` + +The IEDB export both corpora are built from. **This is also the `ipred` training set**, so any +`ipred` figure quoted on it is in-sample. + +790,868 rows · 516,092 distinct peptides · 34,362 immunogenic · 250 alleles · lengths 2–50 +(unfiltered — the 8–11 restriction is applied downstream). + +> Vita R, Mahajan S, Overton JA, Dhanda SK, Martini S, Cantrell JR, Wheeler DK, Sette A, Peters B. +> **The Immune Epitope Database (IEDB): 2018 update.** *Nucleic Acids Res* 2019;47(D1):D339–D343. +> PMID [30357391](https://pubmed.ncbi.nlm.nih.gov/30357391/) · +> doi:[10.1093/nar/gky1006](https://doi.org/10.1093/nar/gky1006) + +Provenance: **experimental** — curated T-cell assay and MHC-ligand records. Raw export kept at +`raw/immunogenicity/iedb_tcell_assays_29032026.tsv.gz`. + +Thymus negatives come from the HLA Ligand Atlas (`thymus/thymus_immunopeptidome.tsv.gz`); its +`mhc_a` is null by design — the deposit is not allele-deconvolved. + +--- + +## `hla_pop_freqs.tsv.gz` + +Class-I allele frequencies (A/B/C) for EUR/ASN/AFR populations, 7,171 rows over 4,234 distinct +alleles. Source: Allele Frequency Net Database (AFND), . +Provenance: **experimental** (population genotyping), aggregated. + +--- + +## Regenerate + + # in ~/vcs/projects/2026-mhcmatch-benchmark + python bench/neoag/corpus_chowell.py # plain build (Chowell arm) + python bench/neoag/corpus_chowell.py --with-viral # + foreign ligandome (Kesmir arm) + python bench/neoag/deposit_corpora.py --ratio 5 --seed 20260817 + python bench/neoag/audit_pmhc_data.py + +`corpus_chowell.parquet` must keep its exact row counts (human 464,310 / 14,712 pos; mouse 47,203 / +5,154) — the shipped `mhcmatch.posbayes` tables are fitted on it, and folding extra sources into it +in place breaks that provenance silently. That is why `--with-viral` writes a separate file. + +## `chowell_iedb_full.tsv.gz` / `chowell_iedb_full_matched.tsv.gz` (2026-08-18) + +Rebuilt from the **full** IEDB export (build 2026-08-11) rather than `iedb_labeled.tsv.gz`. That +harmonised table kept only *positive* T-cell assays, so its `assay_type` is perfectly collinear with +its label and it cannot supply a measured negative at all; the full export carries 361,962 negative +T-cell assay rows, for mouse as well as human. + +`chowell_rebuilt.tsv.gz` and `kesmir_rebuilt.tsv.gz` are **unchanged and still current** — the +shipped `mhcmatch.complement` and `posbayes` tables are fitted on them, they are named in the +library's `REFERENCE_FILES`, and every AUROC in `bench/results/complementarity.md` belongs to them. +These two files are additional, not replacements. + +### The rules, in the order applied + +1. MHC class I. +2. Host ∈ {human, mouse}; a column, never pooled. +3. Length 8–11, canonical 20 amino acids. +4. Exclude `SLLMWITQV`, `KLGGALQAK`. +5. **Immunogenic** — the peptide has ≥1 positive T-cell assay in that host (`Positive`, + `Positive-Low`, `-High`, `-Intermediate`). +6. **Non-immunogenic** — an eluted **self** ligand appearing in no positive T-cell assay in any + host. 2,086 rows are eluted but positive elsewhere and are excluded from the negatives rather + than counted as such. +7. The label is a property of `(peptide, host)`, never of an assay row. +8. The restriction is resolved, and **no row is dropped for lacking one**: 1,677,439 of 3,278,355 + surviving class-I rows name only `HLA class I`. netMHCpan-4.2 picks the best `%Rank_EL` among + whatever the recorded string still permits — an allele group or serotype restricts to that + group, an H-2 haplotype to the molecules it carries, a bare `HLA class I` to a 30-allele + population panel built per locus. A **known** non-human/non-mouse MHC on a human or mouse host + (2,812 rows) is a curation error, not a missing value, and is excluded rather than imputed. +9. One row per `(peptide, allele group, host)`, label = max. The key is the allele **group** + (`HLA-A*02`), which is what lets serotype-only records join. + +### Counts + +| file | rows | positive | negative | peptides | allele groups | +|---|--:|--:|--:|--:|--:| +| `chowell_iedb_full` (human) | 789,047 | 24,293 | 764,754 | 579,322 | 75 | +| `chowell_iedb_full` (mouse) | 65,472 | 7,187 | 58,285 | 49,253 | 43 | +| `chowell_iedb_full_matched` (human) | 20,360 | 10,180 | 10,180 | 17,181 | 18 | +| `chowell_iedb_full_matched` (mouse) | 9,132 | 4,566 | 4,566 | 7,969 | 25 | + +Columns add `allele_2f`, `allele_group`, `allele_source` ∈ {`reported`, `imputed_group`, +`imputed_panel`} and `allele_rank_el`, so an analysis can exclude imputed restrictions and say how +much sample it lost. Human: 375,462 reported against 413,585 imputed; mouse 51,003 against 14,469. + +The matched arm resamples both classes to one profile at 1:1 so the allele group carries no signal +about the label — human to population frequency from `hla_pop_freqs.tsv.gz`, mouse to the positives' +own profile, there being no H-2 frequency table. It is **small**: 52 of 75 human allele groups are +too scarce to fill their population weight and are dropped, which is the cost of the matching, not a +bug. + +Regenerate, in `~/vcs/projects/2026-mhcmatch-benchmark`: + + python immunogenicity/fetch.py # stage the export from ~/hf/pmhc_data/dump + bench/neoag/corpus_iedb.py --workers 12 # ~55 min, 14.7M netMHCpan predictions + bench/neoag/deposit_arms.py + +Arm counts, the filter cascade and the selection tree are `bench/results/corpus_arms.md`. + +**Caveat carried forward.** These negatives are still *inferred* — an eluted ligand nobody has +tested is assumed non-immunogenic. 38,322 rows in the rebuild have a measured negative T-cell assay +instead, held under evidence `tcell_negative` and deliberately **not** merged into the arm: swapping +the negative definition changes what every AUROC recorded on these corpora means. + +### `immunogenicity_legacy_arms.tsv.gz` (2026-08-19) + +The published deposits and the legacy Kešmir arms in one table, keyed `corpus` ∈ {`chowell_vanilla`, +`kesmir_S1`, `kesmir_S2`, `kesmir_S3`}, so a held-out evaluation can be reproduced without the +copyrighted spreadsheets. 15,991 rows. + +**Species is keyed on the restricting allele, not on the deposit's own `Species` column.** The two +disagree on 777 of the 2,508 Calis rows (31%), which carry `Species = Mus` with an HLA restriction — +HLA-transgenic mice, murine host and human restricting molecule. Both readings describe the +experiment; they build different corpora. Keyed on the host, the Calis human arm is 1,113 +immunogenic against **one** non-immunogenic; keyed on the allele it is **1,619 against 272** and the +mouse arm is H-2 throughout. The reported host is retained as `host_reported`. + +Regenerate with `bench/neoag/corpus_legacy.py` then `deposit_arms.py`. + +### What the recognition model is fitted on + +`chowell_iedb_full` is the training arm for `mhcmatch.recognition` (0.15.0). Two choices were made +by measurement and are recorded in `bench/results/recognition_model.md` and +`feature_importance.md`: + +- **The unmatched arm ships.** Resampling negatives to match population HLA usage costs 0.019 + (human) and 0.058 (mouse) AUROC on held-out published deposits. +- **ESM2 and the physicochemical features are both kept.** ESM2 adds +0.022/+0.042 over physchem + alone; physchem adds +0.009/+0.019 over ESM2 alone. Permutation importance on held-out data ranks + the two ESM face-pools first (0.19–0.27), then `kf_tcr`; the whole-peptide Kidera aggregate is + worth 0.0165/0.0012 and is not shipped, being exactly the sum of the two role columns. + +## `staged/` — the IEDB full export, as staged (2026-08-19) + +The **root of the provenance chain**. Every rebuilt corpus in this directory +(`chowell_iedb_full*`, `immunogenicity_legacy_arms`) is derived from these two tables, and until +now they lived only in a gitignored working directory — one bad checkout from being unreproducible. +Deposited so the chain closes without re-downloading 450 MB from IEDB. + +| file | provenance | rows | +|---|---|--:| +| `staged/iedb_tcell.parquet` | **experimental** — IEDB `tcell_full_v3_tsv.zip`, build 2026-08-11, downloaded 2026-08-18 | 570,434 linear-peptide rows, 233,636 distinct peptides | +| `staged/iedb_ligand.parquet` | **experimental** — IEDB `mhc_ligand_full_tsv.zip`, same build | 5,773,495 linear-peptide rows, 1,704,996 distinct peptides | +| `staged/fetch.py` | the script that produced both | — | +| `staged/FETCH.md` | its record: zip mtimes, export row counts, alignment failures, class and qualitative-measurement breakdowns | — | + +Ten columns are kept from each export; nothing is filtered beyond `Linear peptide` and a non-empty +name, because length, alphabet, class and host filters are corpus-construction rules and belong +where they can be stated as rules. **0 rows failed to align** in either table — IEDB ships two +header lines and embedded newlines in comment fields, so `fetch.py` reads with `csv.reader` rather +than splitting on tabs, and reports the misalignment count rather than assuming it is zero. + +The T-cell export is what makes a *measured* negative possible: 361,962 `Negative` rows across all +hosts, where the older harmonised `iedb_labeled.tsv.gz` keeps only positive T-cell assays and its +`assay_type` is therefore perfectly collinear with its label. + +Regenerate (needs `~/hf/pmhc_data/dump`, which is gitignored — the raw zips are not deposited): + +```zsh +python immunogenicity/staged/fetch.py +``` + +## `corpus_iedb_mhc2.parquet` — the MHC class-II arm (2026-08-19) + +**Derived/computed**, from `staged/iedb_tcell.parquet` and `staged/iedb_ligand.parquet` — the same +IEDB build 2026-08-11 export the class-I corpus is built from, by the same rules, so the two arms are +comparable by construction rather than by assertion. This is what the shipped class-II +complementarity model (`mhcmatch.complement`, `cls="mhc2"`, v0.16.0) is fitted on. + +| field | value | +|---|--:| +| rows (peptide, allele group, host) | 1,096,034 | +| distinct peptides | 643,744 | +| immunogenic peptides | 77,943 | +| human rows | 1,036,041 | +| human positives | 65,486 | +| human distinct peptides | 603,781 | +| human allele groups | 121 | +| mouse rows | 59,993 | +| mouse positives | 12,457 | +| mouse distinct peptides | 50,258 | +| mouse allele groups | 40 | +| peptide length | 11–25, median 15 | + +Three rules differ from the class-I arm and nothing else does: + +- **Length** is 11–25 rather than 8–11. +- **Restriction is parsed, never imputed.** Class-II restriction is written a dozen ways across IEDB + (`HLA-DRB1*15:01`, `HLA-DPA1*01:03/DPB1*04:01`, `HLA-DR15`, serotypes, `H2-IAb`), so the corpus + carries a parser and keeps whatever it cannot resolve as an explicit `unresolved` group rather than + guessing an allele. It parses **1,003,554 / 1,974,413 rows (50.8 %)**; the remainder are retained, + not dropped, because the complementarity fit reads only peptide and label. There is therefore **one + class-II arm where class I has two** — no HLA-matched variant is built, since matching needs a + resolved allele on both sides. +- **Rule 8** (pathogen epitopes) is applied by parsing the source organism rather than by set-match. + +`evidence` records why each row is what it is: `eluted_but_positive_elsewhere=5,305`, +`eluted_foreign=103,351`, `eluted_self=1,156,574`, `eluted_unknown=440,458`, `tcell_negative=110,312`, +`tcell_positive=158,413`. + +`corpus_iedb_mhc2.log` is the generator's own run record, deposited beside it; the counts in the table +above are read from it. The build is deterministic — regenerated 2026-08-19 from a clean checkout, every +count reproduced exactly. + +Regenerate (needs this mirror at `$MHCMATCH_PMHC_DIR`, and the benchmark repo): + +```zsh +python bench/neoag/corpus_iedb.py --cls mhc2 +``` + +--- + +## `neoprecis_anchor_triplets.parquet` — added 2026-08-24 + +14,140 rows × 12. Anchor / positive / negative peptide triplets by allele, with the mutated position +and the reference and alternate residue for each of the positive and negative members. A **motif +resource** rather than a screen: the triplets were constructed to train an allele-conditioned +recognition model, so `positive_peptide` and `negative_peptide` are **derived** designations, not +assay outcomes. `mhc` and `allele` are as-reported. + +Deposited by `bench/neoag/ingest_vaccines.py` from Supplementary Data 3 of Lee KH, Sears TJ, +Zanetti M, Carter H, "NeoPrecis: enhancing immunotherapy response prediction through integration of +qualified immunogenicity and clonality-aware neoantigen landscapes", *Nat Commun* 2026;**17**(1), +doi:[10.1038/s41467-026-68651-6](https://doi.org/10.1038/s41467-026-68651-6), +PMID [41577704](https://pubmed.ncbi.nlm.nih.gov/41577704/), PMC12932759. + +--- + +## `pathogen_tcell_{mhc1,mhc2}_{human,mouse}.tsv.gz` — the four pathogen fits' corpora (2026-09-05) + +The training corpora for the `{cls}.{species}.pathogen` cells of `mhcmatch`'s aggregate scorer. +One row per **peptide**, not per assay: a peptide's assays are folded to a single label, and the +counts that produced it travel with it. + +### Source — the full IEDB T-cell export, not a filtered download + +`dump/tcell_full_v3_tsv.zip` (`tcell_full_v3.tsv`, 1,346,206,535 bytes, **577,219 assay rows**, +IEDB build dated 2026-08-11). This is the complete export. The earlier +`raw/immunogenicity/iedb_tcell_assays_29032026.tsv.gz` (397,648 rows) and +`raw/immunogenicity/mouse_datasets/iedb_mouse.tsv.zip` (25,160 rows, **all `Positive*`, all class +I**) are query-filtered downloads; the mouse one carries no negatives at all, which is why the +mouse cells looked unfittable until the full export was read. + +The export carries a **two-row header** — row 1 is the block (`MHC Restriction`), row 2 the field +(`Name`) — so a column is identified by the *pair*, never by index. Three blocks repeat the same +field names (`Epitope`, `Related Object`, `Assay Antigen` each carry `Species IRI`). + +> Vita R, Blazeska N, Marrama D, et al. **The Immune Epitope Database (IEDB): 2024 update.** +> *Nucleic Acids Res* 2025;53(D1):D436–D443. +> PMID [39558162](https://pubmed.ncbi.nlm.nih.gov/39558162/) · +> doi:[10.1093/nar/gkae1092](https://doi.org/10.1093/nar/gkae1092) + +Organism classification is **NCBI Taxonomy**, fetched by taxid from the export's own +`Epitope / Species IRI` (`NCBITaxon_`) via E-utilities `efetch?db=taxonomy`; 897 of 897 +distinct taxids resolved. Cached at `bench/pathogen/_work/lineage.parquet` in the benchmark repo. + +### Filtering procedure, in order + +1. **Source organism is a pathogen** — the NCBI lineage is one of `virus`, `bacteria`, `archaea`, + `fungi`, `protozoa` (Apicomplexa, Euglenozoa, Amoebozoa, Fornicata, Parabasalia, + Heterolobosea), `helminth` (Platyhelminthes, Nematoda, Acanthocephala). **Read off the tree, + never off the organism's name.** 577,219 → 363,242 rows. + Deliberately excluded, and each is a real category rather than noise: `mammal` (97,019 rows — + self and xeno), `arthropod` (26,679 — house-dust mite, cockroach), `plant` (23,854 — timothy + grass and the other pollens), `other_metazoa` (2,738 — **`Gallus gallus` ovalbumin, 1,650 rows, + and `Tetronarce californica` AChR, 915**, the two canonical model antigens), and `unassigned` + (63,556, of which 61,664 record no source organism at all). +2. **Linear peptide, canonical residues** — `Epitope / Object Type == "Linear peptide"` and the + sequence matches `^[ACDEFGHIKLMNPQRSTVWY]+$`, which drops modified and non-peptidic epitopes. + 363,242 → 361,795. +3. **The assay is labelled** — `Positive`, `Positive-Low/Intermediate/High` → 1; `Negative` → 0; + nothing else exists in this column. 361,795 rows, **126,593 positive**. The grade survives as + the `grade` column (`Positive-High` 4 … `Positive-Low` 1, 0 for a negative) rather than being + collapsed away. +4. **The restriction resolves** — `MHC Restriction / Name` must resolve **exactly** to one groove + under `mhcmatch.pseudoseq.resolve_allele` *and* be present in the shipped panel, or name an H-2 + haplotype (`H2-b class I`). 247 distinct names resolve onto the panel, 19 name a haplotype, 73 + resolve off-panel (all `BoLA-*`, cattle), 286 do not resolve. + **`exact=True` is load-bearing.** `HLA-A2` is a serotype; `resolve_allele("HLA-A2", "mhc1")` + answers `("HLA-A2301", False)` — a *different molecule*, on 3,963 rows. An inexact hit here does + not raise and does not return nothing; it scores the wrong groove. +5. **Length window** — class I 8–11, class II 12–20. 17,306 rows fall outside. +6. **Fold to one row per peptide per cell** — `label = max` over the peptide's assays. + `n_assay`, `n_pos`, `n_neg` are kept so a stricter rule (majority, or drop-on-conflict) stays + available; 6,205 peptides across the four cells carry both a positive and a negative assay. + +### The cells + +| file | peptides | positive | prevalence | groove-specific | haplotype-only | allotypes | +|---|--:|--:|--:|--:|--:|--:| +| `pathogen_tcell_mhc1_human.tsv.gz` | 16,841 | 7,018 | 0.417 | 16,841 | 0 | 116 | +| `pathogen_tcell_mhc1_mouse.tsv.gz` | 10,445 | 2,203 | 0.211 | 8,306 | 2,139 | 8 | +| `pathogen_tcell_mhc2_human.tsv.gz` | 7,948 | 5,150 | 0.648 | 7,948 | 0 | 86 | +| `pathogen_tcell_mhc2_mouse.tsv.gz` | 11,732 | 3,326 | 0.283 | 3,948 | 7,784 | 7 | + +Restricted to peptides whose restriction is not **exclusively** prediction-derived +(`MHC binding prediction`, `Statistically inferred by motif or alleles present`): + +| cell | peptides | positive | +|---|--:|--:| +| `mhc1.human.pathogen` | 14,108 | 5,894 | +| `mhc1.mouse.pathogen` | 6,763 | 1,456 | +| `mhc2.human.pathogen` | 4,695 | 3,347 | +| `mhc2.mouse.pathogen` | 10,833 | 2,928 | + +`mhc_evidence` ships as a column and the filter is the **caller's**, because a restriction that is +itself an MHC binding prediction is circular for a model whose leading term is `binder` — and it is +the single largest evidence code in `mhc1.mouse` (3,872 of 10,445 peptides). + +### Schema + +`peptide`, `label`, `n_assay`, `n_pos`, `n_neg`, `grade`, `alleles`, `haplotypes`, +`allele_source`, `mhc_evidence`, `assay_methods`, `source_species`, `kingdom`, +`source_molecules`, `pmids`, `length`. Multi-valued fields are `;`-joined and sorted. + +**Not** the compendium's shared 19-column schema: these are peptide-level with assay counts, and +they carry no `mhc_a`/`mhc_b`/`affinity`. + +### `allele_source` is a key, not a convention + +A restriction is either an allotype (`H-2Kb`) or a haplotype (`H2-b class I`), and the two are +told apart by the `allele_source` column rather than by a rule the reader has to carry. Haplotype +rows keep the **letter** in `haplotypes` and leave `alleles` empty; expanding a haplotype to its +allotypes is a **model** decision and is not baked in. Two reasons it is not: + +- The panel's own spelling differs between classes — class I is `H-2Kb`, class II is `H-2-IAb`, + and `resolve_allele` returns a third form (`H-2-Kb`). Every join here folds through + `normalize_allele` on both sides. +- A mechanical superscript expansion is **wrong for class II**: the panel carries `H-2-IEb`, but + the b haplotype does not express I-E. Class-I K/D/L expansion has no such problem. The + haplotype-restricted best-of arm measured in `bench/results/pathogen_allele_policy_arms.md` + is where that decision belongs. + +### Two skews to check before pooling anything over these + +- `mhc2.mouse`'s groove-specific rows are **3,474 of 3,957 allotype mentions `H-2-IAb` (87.8%)**. + A pooled null over this cell is the I-Ab motif. The haplotype rows are what carry d, k and s. +- `mhc1.human` is 5,237 `HLA-A*02:01` of 16,841 peptides (31.1%). + +### Regenerate + +```zsh +cd ~/vcs/projects/2026-mhcmatch-benchmark && source .venv/bin/activate +python bench/pathogen/extract_tcell.py # 577,219 x 15, ~9 s +python bench/pathogen/fetch_lineage.py # 897 taxids from NCBI, 5 efetch calls +python bench/pathogen/build_pathogen_corpus.py # writes the four files here +python bench/pathogen/proofread_pathogen_corpus.py # 10 invariants per cell, exits 1 on any +``` + +### What was fitted on them (added 2026-09-05, after the fits) + +Three of the four cells became **shipped** `mhcmatch` artifacts in release **1.15.0**, completing +`rank.AGGREGATE_ARTIFACTS` at eight of eight `(cls, species, mode)` cells: + +| corpus file | artifact | peptides fitted | positives | prevalence | allotypes | ROC-AUC | PR-AUC | +|---|---|--:|--:|--:|--:|--:|--:| +| `pathogen_tcell_mhc1_mouse.tsv.gz` | `aggregate_mhc1_mouse_pathogen.json` | 10,404 | 2,196 | 0.2111 | 8 | 0.5564 | 0.2665 | +| `pathogen_tcell_mhc2_human.tsv.gz` | `aggregate_mhc2_human_pathogen.json` | 7,946 | 5,148 | 0.6479 | 80 | 0.6064 | 0.7361 | +| `pathogen_tcell_mhc2_mouse.tsv.gz` | `aggregate_mhc2_mouse_pathogen.json` | 11,725 | 3,324 | 0.2835 | 7 | 0.6224 | 0.4098 | +| `pathogen_tcell_mhc1_human.tsv.gz` | `aggregate_mhc1_pathogen.json` (from 2026-09-06) | 16,790 | 7,002 | 0.4170 | 104 | 0.5988 | 0.5147 | + +Both figures are in-sample, pooled off the logit, each against its own prevalence. Fitted peptide +counts are below the file's row counts because `drop_corpus_members` removes peptides that are exact +members of a corpus table the fit's own `C_corpus_*` channels read (36 on mouse class I, 51 on human +class I) and because a peptide whose candidate allotypes all fail to score is dropped. + +**`mhc1.human.pathogen` was on the Kesmir corpus, and moved to this one on 2026-09-06.** Until +then the shipped fit for that cell stayed on `kesmir_rebuilt_hla_matched.tsv.gz`, whose negatives +are *eluted ligands with no positive T-cell record* -- presentation held fixed on both sides. Here a +negative was *assayed and did not respond*, of unknown presentation, so `binder` carries +presentation again: **+0.1454 against +0.3418** for the same cell on the two corpora. The move was +the author's call, taken with the refit that put every pathogen cell on one corpus and one balanced +design; the paragraph above this one described the earlier state and is corrected rather than +deleted, because a recorded number cites the corpus it was produced on. + +Two things not to carry over from the Kesmir arm. Its PR-AUC is **not** comparable to this one -- +0.0905 at prevalence 0.0691 against 0.5147 at 0.4146, which is 1.31x and 1.24x lift respectively, +i.e. the same. And its five terms are not these two: `C_phys_buried` and `C_corpus_thymus` were +each highly resolved there and are absent here, because `C_corpus_thymus` and `C_corpus_self` +correlate at +0.683 to +0.792 and the five-term fits were fitting their difference. + +Fitted by `bench/pathogen/fit_tcell.py` in `2026-mhcmatch-benchmark`; per-cell reports at +`bench/results/pathogen_tcell_fit_{cls}_{species}.md`, the head-to-head at +`bench/results/pathogen_mhc1_human_v1_vs_v2.md`, and the shipped-artifact provenance in +`src/mhcmatch/data/PROVENANCE.md`. + +```zsh +MHCMATCH_MODEL_RELEASE=1.15.0 python bench/pathogen/fit_tcell.py --cls mhc1 --species mouse +MHCMATCH_MODEL_RELEASE=1.15.0 python bench/pathogen/fit_tcell.py --cls mhc2 --species human +MHCMATCH_MODEL_RELEASE=1.15.0 python bench/pathogen/fit_tcell.py --cls mhc2 --species mouse +``` + +**One consequence for anyone scoring with these artifacts.** Each was fitted on the +best-**presenting** allotype among those the peptide's own record supports, so each declares +`allele_policy = {"kind": "panel", "select": "presentation", "source": "record"}` and `mhcmatch` +refuses a bare `rank pairs` run rather than letting the coefficients meet a shifted column. +`rank fasta --alleles A,B` and `rank pairs --allele-panel A,B` are the supported forms. Measured on +the fitted designs, the best-of-record `binder` sits a mean **−0.2146** (sd 0.5262) below the same +peptide's single-recorded-allele `binder` on `mhc1_human`, and −0.1403 / −0.1248 / −0.0533 on +`mhc1_mouse` / `mhc2_human` / `mhc2_mouse`; 62.3 % / 73.7 % / 63.5 % / 52.7 % of peptides carry +exactly one candidate and are unaffected. + +## `pathogen_tcell_balanced_{mhc1,mhc2}_{human,mouse}.tsv.gz` — allele-balanced fitting sets (2026-09-05) + +**Derived**, not experimental. One row per peptide, carrying the allotype it is attributed to, its +T-cell label, and a prior weight that removes the allotype-composition difference between the two +labels. Built from `pathogen_tcell_{mhc1,mhc2}_{human,mouse}.tsv.gz` (entry above) by +`bench/pathogen/netmhcpan_best.py --cls {mhc1,mhc2} --species {human,mouse} --min-records 5` in +`repseq/2026-mhcmatch-code`. Class I goes through NetMHCpan-4.2, class II through +NetMHCIIpan-4.3. + +| column | meaning | +|---|---| +| `peptide` | the epitope, as deposited | +| `mhc_a` | NetMHCpan's own best allotype for that peptide, panel spelling | +| `label` | `pos` / `neg`, the max over that peptide's IEDB T-cell assays | +| `weight` | prior (case) weight for a weighted GLM; see below | + +**Why the allele comes from NetMHCpan.** The balancing variable must not be the model's own choice. +`fit_tcell.score` picks the best-*presenting* allotype with mhcmatch, so weighting on it would +condition the design on the term being fitted (`binder`). NetMHCpan-4.2 is external to every term in +these fits, so its `%Rank_EL` argmin over each record's **own recorded candidate allotypes** is the +allele used. It is not a free argmin over the whole panel — that would invent a restriction the +record does not carry. + +**The two classes spell an allotype differently, and the resolver tries both.** The class-II +deposits already store NetMHCIIpan's own names (`DRB1_0101`, `HLA-DPA10103-DPB10401`, `H-2-IAb`), +so putting them through the canonicaliser converts them *away* from what the tool accepts; class I +is the opposite, the deposit writes `HLA-A*02:01` and the tool wants `HLA-A02:01`. The builder +resolves from an ordered list — as deposited, then canonical — and prints which spelling answered. +All 116 human and 8 mouse class-I candidate allotypes and all 86 human and 7 mouse class-II ones +resolve; none is dropped for want of a name. + +**Order: predict, then filter, then weight.** Every (peptide, candidate allotype) the record +supports goes through NetMHCpan — one call per allotype, never one per peptide. Each peptide takes +its `%Rank_EL` argmin. An allotype with fewer than 5 peptides **in either class** is then dropped +together with its peptides: a frequency ratio resting on one or two peptides on one side is noise, +and an allotype absent from a class has no ratio at all. 0 of 116 human and 0 of 8 mouse candidate +allotypes lack a NetMHCpan name; the thin-tail cut removes 64 of 107 human and 2 of 8 mouse +class-I best-allele allotypes, 53 of 82 human class-II ones, and 0 of 7 mouse class-II ones. + +**The weight.** The target is the **pooled** allotype distribution `p(a) = n(a)/N` over both labels, +and each label is normalised against it: `w = p(a)/p_pos(a)` on positives and `p(a)/p_neg(a)` on +negatives. Each label's weights sum to its own count identically — `sum_a n_c(a)·p(a)/p_c(a) = +N_c·sum_a p(a) = N_c` — so the prevalence is unchanged and only the allotype mix within each label +moves. **Nothing is subsampled:** every peptide surviving the two filters is present, which is why +this ships as a weight column rather than as a matched subset. + +This is the same operation as cross-batch V/J-gene usage correction in `vdjtools` +(`preprocess.correct_vj_usage`, batch → label, gene → allotype). Two deliberate differences: that +implementation targets the *geometric* grand mean of `log p` and keeps rare genes with a Laplace +pseudocount, where this targets the arithmetic pooled marginal and drops the thin tail; and its +`apply_vj_correction` roulette-wheel *resamples* the table, where this reweights it. + +| file | rows | pos | neg | prevalence | allotypes | balanceable | rows at w=1 | min w | max w | max \|Δp\| all, raw | all, weighted | subset, weighted | +|---|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:| +| `pathogen_tcell_balanced_mhc1_human.tsv.gz` | 16,790 | 7,002 | 9,788 | 0.4170 | 107 | 43 | 1,236 | 0.4596 | 7.0060 | 0.1255 | 0.0608 | 1.1e-16 | +| `pathogen_tcell_balanced_mhc1_mouse.tsv.gz` | 10,404 | 2,196 | 8,208 | 0.2111 | 8 | 6 | 6 | 0.2825 | 3.1045 | 0.0955 | 0.0017 | 1.6e-15 | +| `pathogen_tcell_balanced_mhc2_human.tsv.gz` | 7,946 | 5,148 | 2,798 | 0.6479 | 82 | 29 | 659 | 0.4208 | 7.0918 | 0.0616 | 0.0185 | 3.3e-16 | +| `pathogen_tcell_balanced_mhc2_mouse.tsv.gz` | 11,725 | 3,324 | 8,401 | 0.2835 | 7 | 7 | 0 | 0.4191 | 2.2146 | 0.2736 | 0.0000 | 1.1e-15 | + +**Rebuilt 2026-09-06 under one weighting implementation, and the thin tail is now kept rather than +dropped.** These four files were built by an inline copy of `pooled_weights` that *dropped* an +unbalanceable allotype together with its peptides, while the two `neoag_tcell_balanced_*` files +below were built by the shared `balance.py`, which *keeps* them at weight 1. One entry in this file +described both. Every caller goes through `balance.py` now, at `drop_thin=False`, which readmits +1,236 / 6 / 659 / 0 peptides and 554 / 5 / 608 / 0 positives. + +**So the residual over the FULL allotype set is not zero, and the honest claim is the narrow one:** +within the balanceable subset the two weighted distributions agree to 1e-15 or better in every +cell, and over all allotypes the imbalance falls but does not vanish -- 0.1255 to 0.0608 on human +class I, where 64 of 107 allotypes are thin. `mhc2.mouse` reaches exactly zero because it is the +one cell with no thin allotype at all. The largest raw imbalance is **mouse class II at 0.2736**, +the `H-2-IAb` dominance this compendium already documents on the ligand side. + +Kish effective sample size, the price of the reweighting: + +| file | positives, n → ESS | negatives, n → ESS | +|---|---|---| +| `..._mhc1_human` | 6,448 → 4,606 (71.4%) | 9,106 → 7,942 (87.2%) | +| `..._mhc1_mouse` | 2,191 → 2,025 (92.4%) | 8,207 → 8,078 (98.4%) | +| `..._mhc2_human` | 4,540 → 4,019 (88.5%) | 2,747 → 1,980 (72.1%) | +| `..._mhc2_mouse` | 3,324 → 2,839 (85.4%) | 8,401 → 8,135 (96.8%) | + +The weights are consumed by `bench/pathogen/fit_tcell.py --balance`, which passes them as prior +weights to `bench/epic/optimize.py::irls(pw=)`. `pw=None` is bit-identical to the unweighted fit. +Per-allotype composition and weights are tabulated in +`bench/results/pathogen_tcell_allele_balance_{mhc1,mhc2}_{human,mouse}.md`. + +**These four sets are what the 1.15.0 pathogen models are fitted on**, all four with `--balance`: +class I on `binder + C_corpus_self`, class II on `binder + C_phys_buried + C_phys_charge` (class II +has no corpus block — the mimicry tables are class-I TCR-face geometry). + +## `neoag_tcell_balanced_mhc2_{human,mouse}.tsv.gz` — allele-balanced class-II neoantigen fitting sets (2026-09-06) + +**Derived**, not experimental. The same weighting as the four `pathogen_tcell_balanced_*` files +above, on a different corpus and a different balancing variable. Built by +`bench/mhc2_neoag/allele_balance.py --species {human,mouse} --min-records 5` in +`repseq/2026-mhcmatch-code`, which calls the two class-II neoantigen fits' own loaders +(`fit_human_mhc2.load`, `fit_mouse.deposit` ∪ `fit_mouse.cedar`) rather than re-reading the +deposits, so each balanced corpus spans exactly the population its fit sees. + +| column | meaning | +|---|---| +| `peptide` | the epitope, as deposited | +| `mhc_a` | the allotype the row is attributed to; **the deposit's own spelling**, see below | +| `label` | `pos` / `neg`, the deposited `immunogenicity` | +| `weight` | prior (case) weight for a weighted GLM; the pooled-marginal weight defined above | + +**The unit is the `(peptide, allotype)` pair, and there is no NetMHCpan argmin here.** That is the +one substantive difference from the pathogen corpora, and it is a property of the deposits. The +pathogen deposits carry a peptide-level label and an unresolved *set* of candidate restrictions, so +an argmin resolves them. These deposits carry a measured restriction per assay: 589 of 804 human +peptides and 795 of 808 mouse ones appear with exactly one allotype, only 10 human and 25 mouse +peptides disagree on the label across their allotypes, and both fits already score each row against +its own allele. An argmin would collapse 1,112 human pairs to 804 and replace a *measured* +restriction with a predicted one. + +| species | balancing variable | why it is external to every fitted term | +|---|---|---| +| human | CEDAR's deposited `allele`, DRA folded | an assay annotation; nothing mhcmatch computed | +| mouse | `mhc_a_pred` | already NetMHCIIpan-4.3's argmin over the strain's candidate alleles, with any reported allele added to the set — see "Predicted restriction" in the benchmark repo's `SOURCES.md` | + +**The DRA fold is not tidiness.** `HLA-DRA*01:01/DRB1*04:04` and `HLA-DRB1*04:04` are one molecule — +DRA is effectively monomorphic — and as deposited they are two allotypes splitting one allotype's +counts, which is exactly what a thin-tail cut punishes. 72 spellings, 68 molecules. + +**Three allele vocabularies now appear across the six balanced files, one per deposit, and each is +the spelling its own fit consumes.** The pathogen class-II files carry NetMHCIIpan's names +(`DRB1_0101`, `HLA-DPA10103-DPB10401`, `H-2-IAb`); this human file carries the panel spelling +(`HLA-DRB1*04:01`, `HLA-DQA1*03:01/DQB1*03:02`); this mouse file carries the deposit's +(`H2-IAb`, without the second hyphen). All 11 spellings across these two files resolve to an exact +match through `mhcmatch.pseudoseq.resolve_allele`; none is dropped for want of a name. + +**The thin tail is kept at weight 1, not cut -- and this entry described a cut until 2026-09-06.** +An allotype with zero rows in one class cannot be balanced by any weight: the target `p_pool(a)` is +strictly positive while the achievable `p(a | that class)` is identically zero, and a Laplace +pseudocount makes the *ratio* finite without supplying the missing rows. There are then two ways +not to lose those rows, and the author's call is the second: drop them, or keep them at weight 1 and +compute the target marginal within the balanceable subset. **The files always held the second and +this entry always described the first**, so its counts (693 / 271 / 7 allotypes human, 538 / 214 / 4 +mouse) were the cut that was never shipped, and its claim that both weighted distributions "are +equal to floating-point noise in every cell of both files" is true only *within* the balanceable +subset. The cascade a cut would have cost is reported at four thresholds in +`bench/results/neoag_tcell_allele_balance_mhc2_{human,mouse}.md`. + +**The human corpus is where keeping rather than cutting matters most.** Whole allotypes in CEDAR +are all-positive — `HLA-DRB1*04:05` 31/31, `HLA-DRB1*07:01` 24/24, `HLA-DRB1*15:02` 22/22, +`HLA-DRB5*01:01` 17/17 — because a study that reports a responder's restriction reports no +negatives for it, while the two `DQB1*03:02` composites are ~90 % negative from a type-1-diabetes +screen. Allotype and label are confounded in the deposit, and those are exactly the allotypes no +weight can touch. Cutting them would have cost **59 % of the positives**; keeping them at weight 1 +costs a residual imbalance of **0.1273** over all 72 allotypes instead, against 5.6e-17 within the +7 that can be balanced. Neither number alone describes the file, so both are in the table below. + +| file | rows | pos | neg | prevalence | allotypes | balanceable | rows at w=1 | min w | max w | max \|Δp\| all, raw | all, weighted | subset, weighted | Kish n_eff | +|---|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:| +| `neoag_tcell_balanced_mhc2_human.tsv.gz` | 1,112 | 656 | 456 | 0.5899 | 72 | 7 | 419 | 0.5281 | 4.2690 | 0.2427 | 0.1273 | 5.6e-17 | 830.7 (74.7 %) | +| `neoag_tcell_balanced_mhc2_mouse.tsv.gz` | 627 | 261 | 366 | 0.4163 | 7 | 4 | 46 | 0.5010 | 3.0430 | 0.2800 | 0.0994 | 1.1e-16 | 532.1 (84.9 %) | + +**Nothing is retained-against: every row each fit sees is present.** 1,112 of 1,112 human rows and +627 of 627 mouse ones, all 656 and 261 positives, over all 72 and 7 allotypes. A cut at +`min_records = 5` would have kept 693 / 271 / 7 and 538 / 214 / 4. +positives (41.3 %) over 7 of 68 allotypes; mouse 538 of 572 rows (94.1 %) and 214 of 247 positives +(86.6 %) over 4 of 7 allotypes. The two weighted allotype distributions are equal to floating-point +noise in every cell of both files. diff --git a/immunogenicity/chowell_iedb_full.tsv.gz b/immunogenicity/chowell_iedb_full.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..caa7718758877a37493dbfdaf28b6ed0aa96ae3d --- /dev/null +++ b/immunogenicity/chowell_iedb_full.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7b1553218b974a848b1cb77f66ecece7af4b3149936d23a5ed93634392c91307 +size 7231939 diff --git a/immunogenicity/chowell_iedb_full_matched.tsv.gz b/immunogenicity/chowell_iedb_full_matched.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..b87aaa974f09b23b8ed70e77418d3ebba1bb9994 --- /dev/null +++ b/immunogenicity/chowell_iedb_full_matched.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:180f6480e2f518816f78613ca7959577cfc21c5e864ce8a8798032d694c16162 +size 266797 diff --git a/immunogenicity/chowell_rebuilt.tsv.gz b/immunogenicity/chowell_rebuilt.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..c727a69f985c5de5271c70b7e639c3a53b8d4151 --- /dev/null +++ b/immunogenicity/chowell_rebuilt.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:585983d4c9f9e3400a2b1bb3cdc6fc08ca8145b4e7627d668bd4fbd216bde42a +size 4432513 diff --git a/immunogenicity/chowell_rebuilt_hla_matched.tsv.gz b/immunogenicity/chowell_rebuilt_hla_matched.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..d2a32d3cfa860eff49fcc831aeae9fe68becf775 --- /dev/null +++ b/immunogenicity/chowell_rebuilt_hla_matched.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2411a25837bb531bcd7d76aaf81c2c3d87a5a14285947d0f04db805a670886ed +size 928735 diff --git a/immunogenicity/corpus_iedb_mhc2.log b/immunogenicity/corpus_iedb_mhc2.log new file mode 100644 index 0000000000000000000000000000000000000000..79e5d0edcf3f09b4452717bc71daa93f709d06c9 --- /dev/null +++ b/immunogenicity/corpus_iedb_mhc2.log @@ -0,0 +1,24 @@ +# tcell: 570,434 staged rows +# class II: 247,456 +# host human or mouse: 239,196 +# 11-25mer, canonical: 221,787 +# after artefact exclusions: 221,787 (115,263 peptides) +# ligand: 5,773,495 staged rows +# class II: 1,966,913 +# host human or mouse: 1,905,785 +# 11-25mer, canonical: 1,752,626 +# after artefact exclusions: 1,752,626 (538,321 peptides) +# T-cell assayed (peptide, host) pairs: 119,615 +# positive: 39,818 | measured negative: 79,797 +# peptides positive in any host: 38,320 +# elution pool by origin: foreign=145,186, self=1,165,489, unknown=441,951 +# rows entering allele resolution: 1,974,413 (1,324,898 distinct peptide/host/restriction) +# class-II restriction parsed on 1,003,554/1,974,413 rows (50.8%); the rest are kept as `unresolved` +# top groups: unresolved=970,859, HLA-DRB1*15=174,134, HLA-DRB1*13=163,275, HLA-DRB1*01=105,737, HLA-DRB1*03=83,487, HLA-DRB1*04=52,007, HLA-DRB1*07=39,065, HLA-DPA1*01/DPB1*04=37,066 +# evidence: eluted_but_positive_elsewhere=5,305, eluted_foreign=103,351, eluted_self=1,156,574, eluted_unknown=440,458, tcell_negative=110,312, tcell_positive=158,413 +# out: 1,096,034 (peptide, allele group, host) rows over 643,744 peptides, 77,943 immunogenic +# human: 1,036,041 rows, 65,486 positive, 121 allele groups, 603,781 peptides +# length 11-25, median 15 +# mouse: 59,993 rows, 12,457 positive, 40 allele groups, 50,258 peptides +# length 11-25, median 15 +# -> /Users/mikesh/vcs/projects/2026-mhcmatch-benchmark/bench/neoag/corpus_iedb_mhc2.parquet (3s) diff --git a/immunogenicity/corpus_iedb_mhc2.parquet b/immunogenicity/corpus_iedb_mhc2.parquet new file mode 100644 index 0000000000000000000000000000000000000000..42e9c4a78f2fa5dc89c6e6d84560d7be6e78b878 --- /dev/null +++ b/immunogenicity/corpus_iedb_mhc2.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63fe69c10d635f5c17bebae135bd390dc475ee7cffbde59e04c2dcdfca639708 +size 8509426 diff --git a/immunogenicity/hla_pop_freqs.tsv.gz b/immunogenicity/hla_pop_freqs.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..0ed0a68e24a3f4cff04cd1d33c076e5ccf0eccc4 --- /dev/null +++ b/immunogenicity/hla_pop_freqs.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0beccfd30d62e5934f6335c971416ce165c84301e8e262a0c418d7cc9f8ba5de +size 56665 diff --git a/immunogenicity/iedb_labeled.tsv.gz b/immunogenicity/iedb_labeled.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..7db1904a8f551d7791a8f95d03058c7cdadf08a5 --- /dev/null +++ b/immunogenicity/iedb_labeled.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eccc87b7108a0131c5b39f65629dc16950820599b5208503f9d40f356b8183b9 +size 9833065 diff --git a/immunogenicity/immunogenicity_legacy_arms.tsv.gz b/immunogenicity/immunogenicity_legacy_arms.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ddb40f05178c3989567f8978b790b115c9a5c333 --- /dev/null +++ b/immunogenicity/immunogenicity_legacy_arms.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:76f0fc21892fdec615a3437e3f2e70575fc301bae4518ac5fd139c61f7cd93fb +size 119696 diff --git a/immunogenicity/kesmir_rebuilt.tsv.gz b/immunogenicity/kesmir_rebuilt.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..38bc4c68937da3f11cb4e75048b4d974e9d4ee28 --- /dev/null +++ b/immunogenicity/kesmir_rebuilt.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bf4dbaf44f0409109e21aa37502224c895ea29d668c31d711a7dd43b05a70c38 +size 575885 diff --git a/immunogenicity/kesmir_rebuilt_hla_matched.tsv.gz b/immunogenicity/kesmir_rebuilt_hla_matched.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..db0eb82838cc5d28fc5e94fb74d983b45d3c2147 --- /dev/null +++ b/immunogenicity/kesmir_rebuilt_hla_matched.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c413096c1752caf14712700f896f49716c8d1cc8f941c0d832183c1338a47794 +size 501103 diff --git a/immunogenicity/neoag_tcell_balanced_mhc2_human.tsv.gz b/immunogenicity/neoag_tcell_balanced_mhc2_human.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ee70c6d46974f859d020ca80e20cc6e4511b3b45 --- /dev/null +++ b/immunogenicity/neoag_tcell_balanced_mhc2_human.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:33ad0bb4e15d95207426134fb5646826e9f5b95a97005b095fd8574991a8484c +size 10222 diff --git a/immunogenicity/neoag_tcell_balanced_mhc2_mouse.tsv.gz b/immunogenicity/neoag_tcell_balanced_mhc2_mouse.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..8e93d35afbb0198957bad14d03d31c55198a8a5a --- /dev/null +++ b/immunogenicity/neoag_tcell_balanced_mhc2_mouse.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ee40aa234f0944716a5e97b9b2d6f744921e2eff925088241a7a8acc71432c7 +size 6725 diff --git a/immunogenicity/neoprecis_anchor_triplets.parquet b/immunogenicity/neoprecis_anchor_triplets.parquet new file mode 100644 index 0000000000000000000000000000000000000000..3c0ba9a1c5bc6527adc5ede6ed9a3a98123d377a --- /dev/null +++ b/immunogenicity/neoprecis_anchor_triplets.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:99d8975f702d76d0fd174071d90cce593fb42ee2004f4af3c2371ffabf9f5f4a +size 78504 diff --git a/immunogenicity/pathogen_tcell_balanced_mhc1_human.tsv.gz b/immunogenicity/pathogen_tcell_balanced_mhc1_human.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..b1b0dcc9c05bb842a6eda186c3550a09bf5f01ea --- /dev/null +++ b/immunogenicity/pathogen_tcell_balanced_mhc1_human.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6ef3daf2d93b155d08a14ea126be5e8078e02d51ec59aea46bfab55182559e93 +size 100626 diff --git a/immunogenicity/pathogen_tcell_balanced_mhc1_mouse.tsv.gz b/immunogenicity/pathogen_tcell_balanced_mhc1_mouse.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..3e1ecfade485fecd4b828268bce513a88d5730ab --- /dev/null +++ b/immunogenicity/pathogen_tcell_balanced_mhc1_mouse.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7462e341e2c4d07aaee62a7c7f9f367226de740b61efc88d2c68314826ac31b8 +size 57422 diff --git a/immunogenicity/pathogen_tcell_balanced_mhc2_human.tsv.gz b/immunogenicity/pathogen_tcell_balanced_mhc2_human.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..24621da9566ac8ea6451822bb80d5b31bc547d3d --- /dev/null +++ b/immunogenicity/pathogen_tcell_balanced_mhc2_human.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:24a319e076109d8c70a353f3214b7621fdbe5905f52b09f8cca77ca13bf42bec +size 74067 diff --git a/immunogenicity/pathogen_tcell_balanced_mhc2_mouse.tsv.gz b/immunogenicity/pathogen_tcell_balanced_mhc2_mouse.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..1d66449834f3b8be135de3835aba01517631c5a2 --- /dev/null +++ b/immunogenicity/pathogen_tcell_balanced_mhc2_mouse.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd97fd6e3ae9898d42bd6999a2e1b1df8a29b67665aceb8cdb527b2a5cfd2b6f +size 108061 diff --git a/immunogenicity/pathogen_tcell_mhc1_human.tsv.gz b/immunogenicity/pathogen_tcell_mhc1_human.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..95d8c264e8e74fdb72285a1577a9785fd8f13e10 --- /dev/null +++ b/immunogenicity/pathogen_tcell_mhc1_human.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:64ef62a61d8cc103b68b44684e5410bf34542d377630f8da972475a04d2324a4 +size 416999 diff --git a/immunogenicity/pathogen_tcell_mhc1_mouse.tsv.gz b/immunogenicity/pathogen_tcell_mhc1_mouse.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..904c1edbe79df4df57cfef01836d2b172a9cb29d --- /dev/null +++ b/immunogenicity/pathogen_tcell_mhc1_mouse.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d70fc911aed3a4f92c1e2beba5a6c76815a6d7b5749d1362d06f161b01dbf106 +size 223885 diff --git a/immunogenicity/pathogen_tcell_mhc2_human.tsv.gz b/immunogenicity/pathogen_tcell_mhc2_human.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..3d9541fde779b0ee0547e709f4262aa006cabeb1 --- /dev/null +++ b/immunogenicity/pathogen_tcell_mhc2_human.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d2fe6f4c47aaeac61c8026658ce711ce0de261686d975ae8e7bff3018cc6372a +size 202438 diff --git a/immunogenicity/pathogen_tcell_mhc2_mouse.tsv.gz b/immunogenicity/pathogen_tcell_mhc2_mouse.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..51918d963d7fcfbec5e55d3150dbb1ce62ad601e --- /dev/null +++ b/immunogenicity/pathogen_tcell_mhc2_mouse.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a33d115bca010f6c5c6993ab6ecf4dc1ede456f45c461ed3e5e81ed767e1ca8f +size 296469 diff --git a/immunogenicity/staged/FETCH.md b/immunogenicity/staged/FETCH.md new file mode 100644 index 0000000000000000000000000000000000000000..94dfae014617f526e650e5824d4bba294856e06b --- /dev/null +++ b/immunogenicity/staged/FETCH.md @@ -0,0 +1,56 @@ +# IEDB export staged from ~/hf/pmhc_data/dump +Generated by immunogenicity/fetch.py + +## `iedb_tcell.parquet` + +| field | value | +|---|---| +| source zip | `tcell_full_v3_tsv.zip` | +| inner file | `tcell_full_v3.tsv` | +| downloaded | 2026-08-18 20:54 | +| data rows in export | 577,219 | +| rows the reader could not align | 0 | +| linear-peptide rows kept | 570,434 | +| distinct peptides | 233,636 | + +| Assay - Qualitative Measurement | rows | +|---|--:| +| Negative | 361,962 | +| Positive | 180,370 | +| Positive-Low | 17,478 | +| Positive-High | 5,653 | +| Positive-Intermediate | 4,971 | + +| MHC Restriction - Class | rows | +|---|--:| +| II | 247,456 | +| I | 185,547 | +| (blank) | 136,441 | +| non classical | 990 | + +## `iedb_ligand.parquet` + +| field | value | +|---|---| +| source zip | `mhc_ligand_full_tsv.zip` | +| inner file | `mhc_ligand_full.tsv` | +| downloaded | 2026-08-18 20:57 | +| data rows in export | 5,774,266 | +| rows the reader could not align | 0 | +| linear-peptide rows kept | 5,773,495 | +| distinct peptides | 1,704,996 | + +| Assay - Qualitative Measurement | rows | +|---|--:| +| Positive | 5,012,284 | +| Negative | 603,798 | +| Positive-Low | 55,595 | +| Positive-High | 55,113 | +| Positive-Intermediate | 46,705 | + +| MHC Restriction - Class | rows | +|---|--:| +| I | 3,797,753 | +| II | 1,966,913 | +| non classical | 8,829 | + diff --git a/immunogenicity/staged/fetch.py b/immunogenicity/staged/fetch.py new file mode 100755 index 0000000000000000000000000000000000000000..969473be8dacc580231b8b2cbe4bae5962f0e4e0 --- /dev/null +++ b/immunogenicity/staged/fetch.py @@ -0,0 +1,132 @@ +#!/usr/bin/env -S uv run --script +# /// script +# requires-python = ">=3.11" +# dependencies = ["polars>=1.0"] +# /// +# 2026-08-18 -- stage the IEDB full export from ~/hf/pmhc_data/dump into this repo. +# +# The harmonised ~/hf/pmhc_data/immunogenicity/iedb_labeled.tsv.gz that every current corpus is +# built from keeps only POSITIVE T-cell assays, so its assay_type is perfectly collinear with its +# label and it can never supply a measured negative. The full export does: tcell_full_v3 carries +# Negative rows for every host species, mhc_ligand_full carries the elution pool at ~10x the +# harmonised row count. +# +# Selects the ten columns the corpus builder reads and writes them as parquet. No filtering beyond +# "Linear peptide" and a non-empty name -- length, alphabet, class and host filters are the corpus +# builder's job and belong where they can be stated as rules (bench/neoag/corpus_chowell.py). + +import csv, io, os, sys, time, zipfile +import polars as pl + +DUMP = os.path.expanduser("~/hf/pmhc_data/dump") +HERE = os.path.dirname(os.path.abspath(__file__)) + +# zip -> (output stem, {output column: source column}) +TABLES = { + "tcell_full_v3_tsv.zip": ("iedb_tcell", { + "peptide": "Epitope - Name", + "object_type": "Epitope - Object Type", + "source_organism":"Epitope - Source Organism", + "source_species": "Epitope - Species", + "host": "Host - Name", + "assay_method": "Assay - Method", + "qualitative": "Assay - Qualitative Measurement", + "mhc_allele": "MHC Restriction - Name", + "mhc_class": "MHC Restriction - Class", + "pmid": "Reference - PMID", + }), + "mhc_ligand_full_tsv.zip": ("iedb_ligand", { + "peptide": "Epitope - Name", + "object_type": "Epitope - Object Type", + "source_organism":"Epitope - Source Organism", + "source_species": "Epitope - Species", + "host": "Host - Name", + "assay_method": "Assay - Method", + "qualitative": "Assay - Qualitative Measurement", + "mhc_allele": "MHC Restriction - Name", + "mhc_class": "MHC Restriction - Class", + "pmid": "Reference - PMID", + }), +} +CHUNK = 200_000 + + +def read_table(zip_name, colmap): + """Stream one IEDB zip, keeping colmap's columns. + + IEDB ships two header lines -- group, then field -- which flatten to ' - '. + Comment fields contain embedded newlines, so this goes through csv.reader rather than + splitting on tabs; the returned dropped count is how many rows the reader could not align. + """ + path = os.path.join(DUMP, zip_name) + zf = zipfile.ZipFile(path) + inner = zf.namelist()[0] + rows, chunks, dropped, seen = [], [], 0, 0 + t0 = time.time() + with zf.open(inner) as fh: + text = io.TextIOWrapper(fh, encoding="utf-8", errors="replace", newline="") + reader = csv.reader(text, delimiter="\t") + group, field = next(reader), next(reader) + header = [f"{a} - {b}".strip(" -") for a, b in zip(group, field)] + idx = {} + for out, src in colmap.items(): + if src not in header: + sys.exit(f"{zip_name}: column {src!r} not in the export") + idx[out] = header.index(src) + width = len(header) + for rec in reader: + seen += 1 + if len(rec) != width: + dropped += 1 + continue + rows.append([rec[i] for i in idx.values()]) + if len(rows) >= CHUNK: + chunks.append(pl.DataFrame(rows, schema=list(idx), orient="row")) + rows = [] + print(f"# {inner}: {seen:,} rows {time.time()-t0:.0f}s", flush=True) + if rows: + chunks.append(pl.DataFrame(rows, schema=list(idx), orient="row")) + df = pl.concat(chunks, how="vertical") if chunks else pl.DataFrame(schema=list(idx)) + return df, inner, seen, dropped + + +def main(): + lines = ["# IEDB export staged from ~/hf/pmhc_data/dump", + f"Generated by {os.path.relpath(sys.argv[0], os.path.dirname(HERE))}", ""] + for zip_name, (stem, colmap) in TABLES.items(): + src = os.path.join(DUMP, zip_name) + if not os.path.exists(src): + sys.exit(f"missing {src}") + mtime = time.strftime("%Y-%m-%d %H:%M", time.localtime(os.path.getmtime(src))) + print(f"# {zip_name} ({os.path.getsize(src)/1e6:.0f} MB, downloaded {mtime})", flush=True) + df, inner, seen, dropped = read_table(zip_name, colmap) + kept = df.filter( + (pl.col("object_type") == "Linear peptide") & (pl.col("peptide").str.len_chars() > 0) + ).drop("object_type") + out = os.path.join(HERE, f"{stem}.parquet") + kept.write_parquet(out, compression="zstd") + print(f"# -> {out} {kept.height:,} linear-peptide rows of {seen:,}", flush=True) + + qual = (kept.group_by("qualitative").len().sort("len", descending=True)) + cls = (kept.group_by("mhc_class").len().sort("len", descending=True)) + lines += [f"## `{stem}.parquet`", "", + f"| field | value |", "|---|---|", + f"| source zip | `{zip_name}` |", + f"| inner file | `{inner}` |", + f"| downloaded | {mtime} |", + f"| data rows in export | {seen:,} |", + f"| rows the reader could not align | {dropped:,} |", + f"| linear-peptide rows kept | {kept.height:,} |", + f"| distinct peptides | {kept['peptide'].n_unique():,} |", "", + "| Assay - Qualitative Measurement | rows |", "|---|--:|"] + lines += [f"| {r[0] or '(blank)'} | {r[1]:,} |" for r in qual.iter_rows()] + lines += ["", "| MHC Restriction - Class | rows |", "|---|--:|"] + lines += [f"| {r[0] or '(blank)'} | {r[1]:,} |" for r in cls.iter_rows()] + [""] + + with open(os.path.join(HERE, "FETCH.md"), "w") as fh: + fh.write("\n".join(lines) + "\n") + print(f"# -> {os.path.join(HERE, 'FETCH.md')}") + + +if __name__ == "__main__": + main() diff --git a/immunogenicity/staged/iedb_ligand.parquet b/immunogenicity/staged/iedb_ligand.parquet new file mode 100644 index 0000000000000000000000000000000000000000..d3de3dbe5cdea0283bc4e7a653157c5b81af5a52 --- /dev/null +++ b/immunogenicity/staged/iedb_ligand.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:033cb363890a6623c0a6a43e41b975efa73788b8cd2316c64b092954e79b279f +size 29097041 diff --git a/immunogenicity/staged/iedb_tcell.parquet b/immunogenicity/staged/iedb_tcell.parquet new file mode 100644 index 0000000000000000000000000000000000000000..38de504ab589f8fde35007b49ac30136c2478c17 --- /dev/null +++ b/immunogenicity/staged/iedb_tcell.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e870b28b99d1710adc65b7ff3c74e0093358d8a8b69b3b335292e66581067a06 +size 3662436 diff --git a/ligandome/SOURCES.md b/ligandome/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..ce686c0799758a77140865e844e657ab5c707ff7 --- /dev/null +++ b/ligandome/SOURCES.md @@ -0,0 +1,417 @@ +# `ligandome/` — sources, provenance and citations + +Four reference sets used as *background* rather than as labelled training data: what a foreign +peptide looks like, what a novel viral ORF looks like, which self genes are tumour-restricted +enough to be targeted safely, and what a mouse presents outside the thymus. + +Every citation below was resolved against **PubMed** and carries its PMID and DOI. Where a source +could not be resolved, that is stated as unresolved rather than filled in with a plausible guess. + +Audited and rewritten 2026-08-17. The previous root-level entries were one-liners with no +citations, and the row `#` was wrong for all three files. + +--- + +## `viral_foreign_iedb.tsv.gz` + +Foreign (pathogen-derived) peptides observed presented on MHC — the non-self reference. + +| | | +|---|---| +| rows | **82,368** (44,993 distinct peptides) | +| columns | `peptide`, `mhc_a`, `mhc_class`, `mhc_species`, `source_organism`, `dataset_origin`, `tcell_positive_any`, `tcell_positive_host` | +| class | MHC-I 55,084 · MHC-II 19,762 · unlabelled 7,522 | +| host | HomoSapiens 71,207 · MusMusculus 3,639 · unlabelled 7,522 | +| source organisms | 909 distinct | +| provenance | **experimental** — MS-eluted / assayed ligands as curated by IEDB | + +### The two evidence classes, separated 2026-09-05 + +**`dataset_origin` is the single literal `iedb_viral` on every row, so nothing in this file said +which of its peptides are also known epitopes.** **5,533 of the 44,993 distinct peptides (12.3 %) +carry a positive IEDB T-cell assay.** The columns below record that; they are **not** a filter, and +the file's meaning has not changed. + +**This is composition, not contamination.** This deposit is the non-self **presented** reference, +and it is defined by the *nature* of its compartment — foreign in origin, observed presented on +MHC. A peptide that is also T-cell positive is still foreign and still presented, so it belongs +here; it is a labelled subset of the compartment, not something that leaked into it. + +**And the complement is not a class.** IEDB's T-cell export is **positives-only** — 27,497 +peptides, none recorded as tested-negative — so `tcell_positive_any = 0` means *no positive record +exists*, which for most of these peptides means **nobody measured it**. It does not mean presented +and not attacked. Do not read the 0 arm as a negative set, and do not filter the `viral` corpus +channel on this column: that would silently redefine the channel from a compartment to an assay +outcome, against a complement that has no defined membership. + +Two columns rather than one, because IEDB's T-cell export is keyed on **(peptide, host)** and is +positives-only (27,497 peptides, none recorded as tested-negative), so "positive" always means +"positive somewhere" and *where* is the whole question. Of the 676 Kesmir human negatives that are +T-cell positive at all, **671 are positive only in mouse and 5 in human** — a single flag could not +report which of those it meant. + +| column | meaning | 1 | 0 | null | +|---|---|--:|--:|--:| +| `tcell_positive_any` | positive T-cell assay under **any** host | 13,720 | 68,648 | — | +| `tcell_positive_host` | positive under **this row's** `mhc_species` | 10,653 | 64,193 | 7,522 | + +`tcell_positive_host` is null exactly where `mhc_species` is unlabelled: the question has no answer +there, and filling 0 would assert "not positive in its host" for 7,522 rows the data cannot speak +for. + +Per class and host, distinct peptides: + +| class | host | presentation only | T-cell positive | +|---|---|--:|--:| +| MHCI | HomoSapiens | 19,043 | 3,346 | +| MHCI | MusMusculus | 1,592 | 703 | +| MHCII | HomoSapiens | 13,143 | 1,451 | +| MHCII | MusMusculus | 372 | 107 | +| unlabelled | unlabelled | 6,756 | 488 | + +**Where the columns are load-bearing: a fit whose negatives come from here.** The Chowell/Kesmir +immunogenicity corpora draw their non-self negatives from this deposit, so any model fitted on them +that also carries `C_corpus_viral` scores those negatives against a k-mer table **that contains +them**. That is circularity between a fit's negative class and its reference table — a property of +that fit, not a defect in this file or in the channel — and it is why `mhcmatch`'s `pathogen` mode +drops the corpus block outright rather than picking a channel. + +For scale when reading such a fit: on the class-I 8–11mers feeding the `viral` table the +T-cell-positive subset is **15.8 % of the 3-mer window mass** (8,733 of 55,116 windows, 3,876 of +23,984 distinct peptides). + +**Nothing downstream changed.** No corpus table was rebuilt and no shipped artifact moved: the +`viral` channel in `aggregate_mhc1.json` (human, +0.2091) and `aggregate_mhc1_mouse.json` (mouse, +−0.3120) is the same table it was. These columns are a record of what the deposit holds. + +Regenerate: `python bench/neoag/annotate_viral_tcell.py` in `2026-mhcmatch-benchmark` +(`--dry-run` audits without writing). + +**Origin.** Export of the Immune Epitope Database restricted to non-human source organisms. + +> Vita R, Mahajan S, Overton JA, Dhanda SK, Martini S, Cantrell JR, Wheeler DK, Sette A, Peters B. +> **The Immune Epitope Database (IEDB): 2018 update.** *Nucleic Acids Res* 2019;47(D1):D339–D343. +> PMID [30357391](https://pubmed.ncbi.nlm.nih.gov/30357391/) · +> doi:[10.1093/nar/gky1006](https://doi.org/10.1093/nar/gky1006) + +**Why it matters here.** This is the only source of *non-self eluted ligands* we hold. Every +eluted-ligand negative in `immunogenicity/iedb_labeled.tsv.gz` is self-sourced (449,449 human-self +and 41,986 mouse-self against 149 + 63 cross-species), so a Calis/Kesmir-style corpus — immunogenic +vs presented-**foreign**-but-not-immunogenic — cannot be built without this file. + +**Caveat.** 7,522 rows (9.1%) carry no `mhc_class` or `mhc_species`. Filter on those columns +explicitly; do not assume MHC-I. + +--- + +## `viral_orfs_gse272406.tsv.gz` + +Translated **unannotated** viral ORFs — proteins the canonical viral proteome does not contain. + +| | | +|---|---| +| rows | **5,673** ORFs | +| columns | `orf_id`, `protein`, `virus`, `virus_family`, `gene`, `orf_type`, `dataset_origin` | +| coverage | 365 viruses, 31 viral families | +| ORF type | uORF 1,954 · mORF 1,304 · N_ext 779 · iORF 751 · ouORF 603 · N_trunc 282 | +| provenance | **experimental** — ribosome-profiling-derived ORF calls, translated to protein | + +**Origin.** GEO series [GSE272406](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE272406). + +> Weingarten-Gabbay S, Bauer MR, Stanton AC, Yu Y, Freije CA, Welch NL, Boehm CK, Klaeger S, +> Verzani EK, López D, Hensley LE, Clauser KR, Carr SA, Abelin JG, Rice CM, Sabeti PC. +> **Pan-viral ORFs discovery using massively parallel ribosome profiling.** +> *Science* 2025;388(6752):1218–1224. +> PMID [40504907](https://pubmed.ncbi.nlm.nih.gov/40504907/) · +> doi:[10.1126/science.ado6670](https://doi.org/10.1126/science.ado6670) + +**Correcting a prior mislabel.** An earlier spec filed this as thymus *self*. It is neither thymic +nor self: these are viral ORFs, and the source paper reports peptides from noncanonical viral ORFs +presented on class-I HLA in infected cells. Using it as a self reference would invert its meaning. + +--- + +## `cancer_targets_tsarina.tsv.gz` + +Tumour-antigen genes with **tissue-restriction and safety annotation** — not merely a gene list. +The previous description ("curated shared tumor-antigen genes", 3 columns) understated it by an +order of magnitude. + +| | | +|---|---| +| rows | **393** genes, **51** columns | +| restriction call | TESTIS 268 · SOMATIC 79 · REPRODUCTIVE 26 · PLACENTAL 11 · NO_DATA 9 | +| confidence | MODERATE 167 · HIGH 152 · LOW 65 · NO_DATA 9 | +| with immunopeptidomics evidence | 222 genes carry `ms_pmids` (178 distinct PMIDs) | +| provenance | **mixed — see below** | + +**Column provenance, which is the point of this file and must not be flattened:** + +| group | example columns | provenance | +|---|---|---| +| identity | `Symbol`, `Aliases`, `Ensembl_Gene_ID`, `Canonical_Transcript_ID`, `biotype` | reference annotation | +| membership | `source_databases` | **curated** — which CTA database(s) list this gene | +| expression | `rna_testis_ntpm`, `rna_max_somatic_ntpm`, `protein_testis`, … | **experimental** (HPA) | +| filters / calls | `passes_filters`, `rna_*_pct_filter`, `restriction`, `restriction_confidence`, `safety_flags` | **derived/computed** from the above | +| MS evidence | `ms_pmids`, `ms_healthy_somatic_tissues`, `ms_restriction` | **experimental** — per-gene PMIDs of immunopeptidomics studies observing the gene as an HLA ligand | + +**Pipeline origin.** Built by [`pirl-unc/tsarina`](https://github.com/pirl-unc/tsarina) (Apache-2.0), +which delegates CTA membership and reproductive-tissue sets to +[`pirl-unc/oncoref`](https://github.com/pirl-unc/oncoref). **Neither repository declares a DOI or +PMID for itself** — cite the tool by URL, and the underlying data by the references below. + +### Underlying sources, as tagged in `source_databases` + +`CTpedia` · `CTexploreR_CT` / `CTexploreR_CTP` · `daSilva2017` / `daSilva2017_protein` · +`placental_antigen` · `paralog:` + +> **daSilva2017** — da Silva VL, Fonseca AF, Fonseca M, da Silva TE, Coelho AC, Kroll JE, +> de Souza JES, Stransky B, de Souza GA, de Souza SJ. +> **Genome-wide identification of cancer/testis genes and their association with prognosis in a +> pan-cancer analysis.** *Oncotarget* 2017;8(54):92966–92977. +> PMID [29190970](https://pubmed.ncbi.nlm.nih.gov/29190970/) · +> doi:[10.18632/oncotarget.21715](https://doi.org/10.18632/oncotarget.21715) + +> **CTexploreR** — Loriot A, Devis J, Gatto L, De Smet C. +> **A survey of human cancer-germline genes: Linking X chromosome localization, DNA methylation and +> sex-biased expression in early embryos.** *PLoS Genet* 2025;21(10):e1011734. +> PMID [41091706](https://pubmed.ncbi.nlm.nih.gov/41091706/) · +> doi:[10.1371/journal.pgen.1011734](https://doi.org/10.1371/journal.pgen.1011734) +> — the paper describing the R/Bioconductor package and its reference list of 146 cancer-germline +> genes. + +> **Human Protein Atlas** — source of every `protein_*` / `rna_*` column. +> Uhlén M, Fagerberg L, Hallström BM, Lindskog C, Oksvold P, Mardinoglu A, et al. +> **Tissue-based map of the human proteome.** *Science* 2015;347(6220):1260419. +> PMID [25613900](https://pubmed.ncbi.nlm.nih.gov/25613900/) · +> doi:[10.1126/science.1260419](https://doi.org/10.1126/science.1260419) + +> **`ms_pmids`** — per-gene immunopeptidomics evidence, 178 distinct PMIDs. Spot-verified against +> PubMed; the four most frequent are +> [35051231](https://pubmed.ncbi.nlm.nih.gov/35051231/) (influenza-infected human tissue +> immunopeptidome, 98 genes), +> [32938616](https://pubmed.ncbi.nlm.nih.gov/32938616/) (melanoma spliced-peptide immunopeptidome, +> 96), +> [31844290](https://pubmed.ncbi.nlm.nih.gov/31844290/) (Sarkizova et al., mono-allelic HLA +> peptidome / HLAthena, 76) and +> [29557506](https://pubmed.ncbi.nlm.nih.gov/29557506/) (glioblastoma stem-like cell HLA ligandome, +> 76). Resolve any individual PMID directly; they are not reproduced here. + +### Unresolved + +**`CTpedia` has no citation I could verify.** It is not resolvable from the tsarina or oncoref +repositories, and a PubMed search did not return a record that unambiguously matches this tag. +Recorded as unresolved rather than guessed — do not cite it until someone confirms which resource +is meant (candidates include the CTdatabase / CT antigen database lineage, but that is a hypothesis, +not a verification). + +--- + +## `tissue_self_mmu.tsv.gz` + +Mouse **self** peptides eluted from 18 normal tissues — the non-thymic counterpart to +`thymus/thymus_immunopeptidome_mmu.tsv.gz`, and the first mouse self-ligandome in this compendium. + +| | | +|---|---| +| rows | **53,992** (24,177 distinct peptides) | +| columns | `peptide`, `mhc_a`, `mhc_class`, `mhc_species`, `source_protein`, `species`, `tissue`, `dataset_origin` | +| class | MHC-I 13,139 distinct · MHC-II 4,348 distinct | +| host | MusMusculus, 100% — **two strains**: C57BL/6 (PXD008733) and NOD (PXD031966 pancreas) | +| alleles | `H-2Kb`, `H-2Db` (MHC-I) · `I-Ab` (MHC-II) | +| tissues | 18 | +| provenance | **experimental** — MS-eluted ligands; identifications and the FDR cutoff **derived** | + +**Origin.** PRIDE [PXD008733](https://www.ebi.ac.uk/pride/archive/projects/PXD008733), the deposit's +own Trans-Proteomic Pipeline PSM tables (`raw_first_batch.csv`, `raw_second_batch.csv`, +`raw_third_batch.csv`, `raw_orbi.csv`, `raw_lumos.csv` — 850,396 PSMs, 151 MB), not its raw spectra. + +> Schuster H, Shao W, Weiss T, Pedrioli PGA, Roth P, Weller M, Campbell DS, Deutsch EW, Moritz RL, +> Planz O, Rammensee HG, Aebersold R, Caron E. +> **A tissue-based draft map of the murine MHC class I immunopeptidome.** +> *Sci Data* 2018;5:180157. +> PMID [30084848](https://pubmed.ncbi.nlm.nih.gov/30084848/) · +> doi:[10.1038/sdata.2018.157](https://doi.org/10.1038/sdata.2018.157) + +**The FDR is recomputed here, not inherited.** The PSM tables carry `DECOY_`-prefixed entries, so +distinct peptides were ranked by best `probability_ip` (iProphet) and the largest set with +decoy/target ≤ 0.01 kept: **`probability_ip` ≥ 0.974142 → 35,470 target peptides against 354 decoy +peptides, 0.0100 estimated peptide-level FDR**. Requiring an `sp|` protein then drops the `iRT` +spike-ins, `CONT_*` and `LMP_*` entries, leaving 32,827 distinct peptides across all 19 tissues — +against the 28,448 the paper reports, so the cutoff is neither loose nor lossy. + +**`mhc_a` is the immunoprecipitating antibody, as reported — not a prediction.** The `spectrum` +field names the run and the run name names the antibody: **Y3 → H-2Kb**, **B22 → H-2Db**, +**M5 / M15 / M15classII → I-Ab** (C57BL/6 is I-Ab only; *Ea* is deleted in the b haplotype). The +mapping is checked rather than assumed — on 9-mers it yields H-2Kb P5 = F 48%, P9 = L 49% and +H-2Db P5 = N 77%, P9 = L/I/V 48/22/10%, the canonical Kb and Db anchor motifs. + +**Why it matters here.** `mimicry`'s `self` channel is a *proteome* — being encoded is not being +presented. This is the first mouse set that says which self peptides are actually presented, and by +which allele, in tissue a T cell can reach. Paired with the thymic deposit it separates two +different tolerance arguments: met during negative selection (thymus) versus reachable in the +periphery (here). + +### The NOD pancreas arm — `pride_pxd031966_pancreas` + +7,658 peptides, added 2026-08-21 from the PXD031966 re-search. `NOD4` and `NOD5` carry 160 and 145 +acinar zymogen peptides and **no thymus-exclusive protein at all**, while the third run of that +deposit carries `Prss16` and `Psmb11` and went to `thymus/`; the rule and its evidence are in +`thymus/SOURCES.md`. This takes pancreas from 355 distinct peptides to 8,013 — a 22× increase in the +tissue that matters most for type 1 diabetes. + +**Read `dataset_origin` before pooling.** These are **NOD** mice (H-2^g7: K^d, D^b), not the +C57BL/6 (K^b, D^b) of the rest of the file, and `mhc_a` is empty because the immunoprecipitation used +the pan-H-2 antibody M1/42.3.9.8. Their FDR is mhcquant's 1% peptide-level, not the recomputed +iProphet cutoff used for the atlas. + +**Caveats.** +- **Two strains and two FDR regimes live in this file**, keyed on `dataset_origin`. Pancreas in + particular is a mixture: 355 C57BL/6 peptides and 7,658 NOD ones. +- Cell lines in the source atlas (EL4, B16F10, GL261, LLC/LLC1/LLC2) are **excluded** — tumour + lines, not normal tissue. +- Coverage is very uneven, and it is sampling depth, not biology: small_intestine carries 5,635 + MHC-I peptides and spinal_cord 239. Do not read a tissue's peptide count as presentation breadth. +- MHC-II coverage is incidental — only a few runs used an anti-I-A/I-E antibody, so 15 of 18 tissues + have under 200 MHC-II peptides and five have fewer than 10. +- `source_protein` is **re-derived** by exact lookup against `proteome/mouse.fasta.gz`, not taken + from the deposit. I and L are not collapsed: MS cannot distinguish them, so an I/L-collapsed match + would assert an identification the data does not carry. 3 rows map to no mouse protein. +- 28,833 rows were dropped by the length filter (MHC-I 8–11, MHC-II 11–25, matching + `mhcmatch.mimics._LEN`) because every consumer drops them silently anyway. + +--- + +## Re-fetch / regenerate + +None of these three are regenerated by this repository — they are deposited as fetched. + + # viral_foreign_iedb: IEDB export, non-human source organisms + # https://www.iedb.org/ (manual export) + # viral_orfs_gse272406: + # https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE272406 + # cancer_targets_tsarina: + # pip install tsarina # https://github.com/pirl-unc/tsarina (CTA table export) + # tissue_self_mmu: regenerated, unlike the three above. + # cd ~/vcs/projects/public_expression_peptidomes + # aria2c -i src/urls_results.txt -d ~/hf/pmhc_data/raw/thymus_mouse -x4 -j3 --continue + # ./src/ingest_pxd008733.py && ./src/build_tissue_mmu.py + +## Audit note + +The root-level table that used to index this repo carried an undocumented `#` column which matched +neither column count, row count nor dataset count — for these three files it read 2, 1 and 3 +against actual column counts of 6, 7 and 51. That table is gone; the rows below carry real counts. + +## `self_ligandome_mhc2_{hsa,mmu}.tsv.gz` and `viral_ligandome_mhc2_pan.tsv.gz` — class-II corpus references (2026-09-06) + +**Derived** from deposits already in this dataset; nothing is downloaded to build them. Written by +`bench/mhc2_neoag/build_mhc2_corpus.py` in `repseq/2026-mhcmatch-code`. They replace two class-II +corpus references that were the wrong object for class II — **class I is untouched**, and +`mimics.CLS_REFS` is what keeps the two apart. + +| file | from | rule | distinct peptides | +|---|---|---|--:| +| `self_ligandome_mhc2_hsa.tsv.gz` | `immunogenicity/staged/iedb_ligand.parquet` (IEDB `mhc_ligand_full`) | class II, host species == source species, `Positive*`, L 11–25 | 421,174 | +| `self_ligandome_mhc2_mmu.tsv.gz` | same | same | 25,314 | +| `viral_ligandome_mhc2_pan.tsv.gz` | `ligandome/viral_foreign_iedb.tsv.gz` | class II, L 11–25, **every** `mhc_species` | 14,645 | + +**`self` was the host PROTEOME, and a proteome cannot be a tolerance channel at class II.** Its +class-II table occupied **8,000 of 8,000** 3-mer cells at 110,932,623 (human) and 101,989,053 +(mouse) window mass — a saturated table is near-uniform, so the channel could barely discriminate +by construction. A proteome says what the host *encodes*; tolerance is about what it *presents*. +Divergence from uniform, proteome → ligandome: **0.4859 → 0.6606 bits** (human, CV 0.9291 → 1.0764) +and **0.4759 → 0.9226 bits** (mouse, CV 0.9134 → 1.4058). + +Self is `host species == source species`, which IEDB records per assay — not a proteome-membership +test, which would re-derive a worse version of a column already on the row. 99.3 % of the human and +99.6 % of the mouse distinct peptides are already `Positive*`, and 419,260 / 24,951 of them carry an +MS-elution assay, so the `Positive*` filter buys an unambiguous object for free. + +**`viral` is pan-species at class II, and the deposit carries no `mhc_species` column** — that is the +mechanism, not an omission: `mimics.load_peptides` keeps a row whose species field is absent, so a +deposit that names no species is pan-species with no flag, no argument and no branch in the loader. +Split per species the mouse cell stood on **464** peptides against human's 14,468 and occupied 6,058 +of 8,000 cells; pooled it is 14,645 peptides over 179 source organisms at **7,852** cells and +1,134,720 window mass, 7.0× the mass. A viral k-mer density is not species-specific the way a +proteome is. The mouse split table's *higher* apparent divergence (1.0872 bits) was small-sample +lumpiness estimated from 464 peptides, against 0.5314 from 14,645. + +Consumed through `mhcmatch.mimics.CLS_REFS` (keyed `(category, cls, species)`, consulted before +`SPECIES_REFS`) and vendored into the wheel's `corpus_tables.npz` by `mhcmatch build corpus`, which +reported exactly four moved tables — the class-II `self` and `viral` cells — and six unchanged +class-I ones. + +## `hla_ligand_atlas_*.tsv.gz` — the atlas as 29 per-tissue corpora, both classes (2026-09-18/19) + +Seventy files, 16 MB. They re-partition one deposit that was already here — they add no new +measurement — and they exist because the corpus channel in `mhcmatch`'s EPIC model turned out not +to be reading anything *thymic*. Splitting the atlas by tissue is what let that be measured +instead of assumed. + +**Source, and it is the same one the shipped thymus deposit came from.** +`raw/thymus/hla_ligand_atlas_bundle/HLA_aggregated.tsv` (63 MB, 223,247 peptides) with +`HLA_protein_map.tsv` for the source-protein column. HLA Ligand Atlas, query of 2021-04-23, +CC-BY 4.0. **Experimental** in origin — every peptide is an MS observation off an MHC molecule +from benign tissue — and **derived** in this form: the partition, the pooling and the filtering +are computed here and nothing is re-measured. + +| family | files | what it is | +|---|--:|---| +| `hla_ligand_atlas_.tsv.gz` | 29 | one benign tissue, class I | +| `hla_ligand_atlas_presented{,_nothymus,_xthymus}.tsv.gz` | 3 | class-I pooled, see below | +| `hla_ligand_atlas_match_{presented,thymus}_s.tsv.gz` | 6 | size- and length-matched random proteome draws, 3 seeds each | +| `hla_ligand_atlas_mhc2_.tsv.gz` | 29 | one benign tissue, class II | +| `hla_ligand_atlas_mhc2_presented{,_nothymus,_xthymus}.tsv.gz` | 3 | class-II pooled | + +The three poolings are **not** three names for one thing. `presented` is the union over all 29 +tissues, one row per distinct peptide. `presented_nothymus` is the union over the 28 non-thymic +tissues — a peptide those tissues happen to share with thymus is legitimately in it, because this +is "what the rest of the body presents" and not "thymic peptides removed". `presented_xthymus` +*is* that removal: every peptide the atlas never observed in thymus at all. + +| pooled corpus | class I | class II | +|---|--:|--:| +| `presented` | 80,621 | 132,818 | +| `presented_nothymus` | 67,022 | 116,312 | +| `presented_xthymus` | 54,730 | 104,831 | + +**Class membership is a strict equality, and the difference matters.** The atlas marks a peptide +seen in both experiment types `HLA-I+II`, and the shipped `thymus/thymus_immunopeptidome.tsv.gz` +excludes those; `hla_class == "HLA-I"` reproduces it and `"HLA-I" in hla_class` does not. For +thymus the loose test rebuilds class I at 18,117 against the correct 17,546. + +**Provenance gate, and nothing is written if it fails.** Rebuilding `Thymus` must reproduce the +shipped deposit exactly: **17,546** class-I 9-mers, and **27,987** class-II peptides at every +length. Both hold, and the class-II peptide sets are identical member-for-member. + +Schema is the thymus deposit's, unchanged, so no reader needs to know these exist: +`peptide, mhc_a, mhc_class, mhc_species, source_protein, species, dataset_origin`. `mhc_a` is +empty throughout — the atlas types its donors, not its peptides. + +**The `match_*` draws are computed, not experimental.** Each is a random sample of distinct human +proteome windows drawn to the same peptide count and the same length composition as the atlas +corpus it is named for, written in the same schema and read by the same builder at the same kappa. +They exist to separate "not presented" from "smaller and rougher table": the released `self` +channel enumerates 121,968,158 proteome windows against the atlas's ~450,000, so depth had to be +ruled out as the explanation. Three seeds per size, because one draw is one sample of a +distribution. + +### Re-fetch / regenerate + +Both builders live in `~/vcs/projects/2026-mhcmatch-benchmark` and read only files already in this +compendium: + +``` +./.venv/bin/python bench/immuno/build_atlas_corpora.py --cls mhc1 +./.venv/bin/python bench/immuno/build_atlas_corpora.py --cls mhc2 +./.venv/bin/python bench/immuno/build_proteome_match.py +``` + +### Who reads them + +`mimics.CLS_REFS` points human class-II `thymus` at `hla_ligand_atlas_mhc2_presented.tsv.gz` from +mhcmatch 2026-09-19, so that file is a **released dependency**: `corpus_tables.npz`'s +`mhc2|thymus|human|3` cell is built from it and `store.fetch_file` resolves it from this +repository. The rest are read by the benchmark's own analyses +(`bench/results/epic_corpus_presented*.md`) and are not on any library code path. diff --git a/ligandome/cancer_targets_tsarina.tsv.gz b/ligandome/cancer_targets_tsarina.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..dd04274af7c8c97fe4c5892d04d3568d6edfd1e9 --- /dev/null +++ b/ligandome/cancer_targets_tsarina.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dbddaa46e98d45d38af7b8b39df5e1cb7914f670745865c4e898ba8688e2aaf2 +size 26883 diff --git a/ligandome/hla_ligand_atlas_adrenal_gland.tsv.gz b/ligandome/hla_ligand_atlas_adrenal_gland.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..b7b5237a9319bc48bea75de731a863bd5589189b --- /dev/null +++ b/ligandome/hla_ligand_atlas_adrenal_gland.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fc7cda8f28bcbde656065d7019628a4c374ab565f8d135ab1e79f0dd8e7d9af1 +size 122458 diff --git a/ligandome/hla_ligand_atlas_aorta.tsv.gz b/ligandome/hla_ligand_atlas_aorta.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..877710202525a4147bb806144679854d92f3da8c --- /dev/null +++ b/ligandome/hla_ligand_atlas_aorta.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:062aa5abfe0a6baa6746cf44557962aec6d7bec782b781c634af302ba128fc37 +size 51872 diff --git a/ligandome/hla_ligand_atlas_bladder.tsv.gz b/ligandome/hla_ligand_atlas_bladder.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..65dd2c8d11f9db7f07a32b5ceae2ba39a64ddd17 --- /dev/null +++ b/ligandome/hla_ligand_atlas_bladder.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a8b8c4e8d19afbc80c9cac5b0471f2c3e35560897ec2d49a9f12190029367a3 +size 67055 diff --git a/ligandome/hla_ligand_atlas_bone_marrow.tsv.gz b/ligandome/hla_ligand_atlas_bone_marrow.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..2922cbd29e5aa6e5de879846784af41fa5e25683 --- /dev/null +++ b/ligandome/hla_ligand_atlas_bone_marrow.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1fd37e35c5b653fdfe94cea22808cc67beea6080a640eb0840715a5ecbaf7c95 +size 175076 diff --git a/ligandome/hla_ligand_atlas_brain.tsv.gz b/ligandome/hla_ligand_atlas_brain.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..11c8d30123230a663dcdee3645463b8ed0bff4dc --- /dev/null +++ b/ligandome/hla_ligand_atlas_brain.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2fcb314bb06a762e588fb3e6713bb852d877818580cb6852f3e91c7214f55c2 +size 46710 diff --git a/ligandome/hla_ligand_atlas_cerebellum.tsv.gz b/ligandome/hla_ligand_atlas_cerebellum.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..4094252166fb10db8c895809f73e4bf8e94aff10 --- /dev/null +++ b/ligandome/hla_ligand_atlas_cerebellum.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f1c18724b3fefea1187b8df3efadac351d10d3b253d039a52146ea532fda26e8 +size 79428 diff --git a/ligandome/hla_ligand_atlas_colon.tsv.gz b/ligandome/hla_ligand_atlas_colon.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..8767bb3360845e2e77bcc77fc0e8b1dba5fe492b --- /dev/null +++ b/ligandome/hla_ligand_atlas_colon.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:749592d0def8506fa1f3573285ec6be975abc4c415889f62ffd556af6d72133d +size 116577 diff --git a/ligandome/hla_ligand_atlas_esophagus.tsv.gz b/ligandome/hla_ligand_atlas_esophagus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..8acd088050469e0ca6e32ccc5ea8b3a260c72532 --- /dev/null +++ b/ligandome/hla_ligand_atlas_esophagus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c4645cda561ed426a81baffc7f38fa2ebf65dcc1fda7f3903d860ddbaea2b61a +size 100605 diff --git a/ligandome/hla_ligand_atlas_gallbladder.tsv.gz b/ligandome/hla_ligand_atlas_gallbladder.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..1b2351c44f0cf0b37e6cd09fb6e098b0358c527c --- /dev/null +++ b/ligandome/hla_ligand_atlas_gallbladder.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d05a0f773debedbd6523819bd85f1ffe6a90716c3af67618b0429590aeda25b0 +size 22020 diff --git a/ligandome/hla_ligand_atlas_heart.tsv.gz b/ligandome/hla_ligand_atlas_heart.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..d30348316551ae2d285aadf8ea9ae60380c5d5b1 --- /dev/null +++ b/ligandome/hla_ligand_atlas_heart.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:91f5e06e114b4897fc640e1ba5d1f04bf255bd2c93882ccb0364f0d3ee8a2530 +size 58504 diff --git a/ligandome/hla_ligand_atlas_kidney.tsv.gz b/ligandome/hla_ligand_atlas_kidney.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6d7cd8f96869aa4362b300366cb1e91f58ce75c0 --- /dev/null +++ b/ligandome/hla_ligand_atlas_kidney.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:336a2e64353fbfbf975feac18c323555163f8815c8dc3b214f3c2efb597d6c74 +size 157822 diff --git a/ligandome/hla_ligand_atlas_liver.tsv.gz b/ligandome/hla_ligand_atlas_liver.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..650b7fef5456d476c16af661dd70b16d526763f1 --- /dev/null +++ b/ligandome/hla_ligand_atlas_liver.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:981a8ca2c655ebf14e32c751a89b1577f818882f6e612945759178938cad88ab +size 162265 diff --git a/ligandome/hla_ligand_atlas_lung.tsv.gz b/ligandome/hla_ligand_atlas_lung.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..33b2b3b65fd2a480443dcc2cdd0e3362ff80f1fd --- /dev/null +++ b/ligandome/hla_ligand_atlas_lung.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b6ac41e6d9afc33c1dbb3f92873fa2500a95a66c9af857427b6ee9e083806a3 +size 260908 diff --git a/ligandome/hla_ligand_atlas_lymph_node.tsv.gz b/ligandome/hla_ligand_atlas_lymph_node.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..b756e426f3a04a3c43cfeea5a9b1809810d95b87 --- /dev/null +++ b/ligandome/hla_ligand_atlas_lymph_node.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d94f2e6d598cf94d0a3556724ef3faae47a52af64dbe6748284f4badbee3717e +size 98647 diff --git a/ligandome/hla_ligand_atlas_mamma.tsv.gz b/ligandome/hla_ligand_atlas_mamma.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..8502ef90c05d47aca1394c8b5a6d0ef41c834df6 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mamma.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2c7c94c1417eeac7d7490b3726c9329409664e76daddbbf524fa2b1837200bc +size 11996 diff --git a/ligandome/hla_ligand_atlas_match_presented_s20260918.tsv.gz b/ligandome/hla_ligand_atlas_match_presented_s20260918.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ef44744d6190f2b9d96ca7fc6a41b5e8e7f84acd --- /dev/null +++ b/ligandome/hla_ligand_atlas_match_presented_s20260918.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b021d86667a07a6778f167bbad4c85c7920cb741812053e9c51e3bd158fedcf9 +size 559358 diff --git a/ligandome/hla_ligand_atlas_match_presented_s20260919.tsv.gz b/ligandome/hla_ligand_atlas_match_presented_s20260919.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..d522c78250f9d360f13cfd5188dfb9d699d2615b --- /dev/null +++ b/ligandome/hla_ligand_atlas_match_presented_s20260919.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:781bb4f5b68c5c402daa20f24953a05edbd8eb2f232559a11e6fd62747b6a6cf +size 559624 diff --git a/ligandome/hla_ligand_atlas_match_presented_s20260920.tsv.gz b/ligandome/hla_ligand_atlas_match_presented_s20260920.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..741621e9d63f152c5aaf4d69029f0c906dacc6ea --- /dev/null +++ b/ligandome/hla_ligand_atlas_match_presented_s20260920.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:31e01c34287aabb143f835633dea79650f0092f5c529c966a393e1922113a491 +size 559792 diff --git a/ligandome/hla_ligand_atlas_match_thymus_s20260918.tsv.gz b/ligandome/hla_ligand_atlas_match_thymus_s20260918.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..5709a575342e2545f5393d4cbc58947a0bd6d561 --- /dev/null +++ b/ligandome/hla_ligand_atlas_match_thymus_s20260918.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c2c9eb1f4063ebc692af593a34df53bfa84559ba1ec1a9f7714570c73e45383 +size 181030 diff --git a/ligandome/hla_ligand_atlas_match_thymus_s20260919.tsv.gz b/ligandome/hla_ligand_atlas_match_thymus_s20260919.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..fb3912fa1b504c3dcdeae6e98cd6fcf24c33fb44 --- /dev/null +++ b/ligandome/hla_ligand_atlas_match_thymus_s20260919.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:970a7b867fb0f0765d3a59a272aad8b74c019adfb74c9fceb0065001a57a9ddc +size 180911 diff --git a/ligandome/hla_ligand_atlas_match_thymus_s20260920.tsv.gz b/ligandome/hla_ligand_atlas_match_thymus_s20260920.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..728c51440afb259d5bbdb42eb2b19e0909a21b81 --- /dev/null +++ b/ligandome/hla_ligand_atlas_match_thymus_s20260920.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ee4d39d64ab3b0c34949da9171117276806a023344e23f5b8c3b40c828870ca +size 181046 diff --git a/ligandome/hla_ligand_atlas_mhc2_adrenal_gland.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_adrenal_gland.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..e4e7391e681dcaf1183206eadceebbd82aa6ae4e --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_adrenal_gland.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6faece2d5494d38033df42fd34cbd082368bdfae3af9a9f8a219a3cf34c3963b +size 169950 diff --git a/ligandome/hla_ligand_atlas_mhc2_aorta.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_aorta.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..5bb1f92a2b708cc5c75bc816b2197e03e10c334f --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_aorta.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29ae1f6c105815f579d9c6c337b2f22a688c3a70f5cd750801a603e6fc75c28f +size 113370 diff --git a/ligandome/hla_ligand_atlas_mhc2_bladder.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_bladder.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6763453439589dc35aadf77b1a797d4bd1336c8f --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_bladder.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6816119741e5c2792ef77ea1b9b5a4a5a178db1e143a63fb38ecec0ae667dec0 +size 123281 diff --git a/ligandome/hla_ligand_atlas_mhc2_bone_marrow.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_bone_marrow.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..3e66c1d08bbd666d5f3e3a5adb381bdc0d5909a1 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_bone_marrow.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe445e878757d8f721c44529fd01c566dcd3230c1d6d69779f4a127eb5883f97 +size 178234 diff --git a/ligandome/hla_ligand_atlas_mhc2_brain.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_brain.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..d8b87634053f18aa59399f230497c4df5f75f058 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_brain.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:730a585fc19a92af01a227c2566b962e3aac327589b2aca819b4d1e8cc66968c +size 121206 diff --git a/ligandome/hla_ligand_atlas_mhc2_cerebellum.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_cerebellum.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..2f51d219138457a87e1d1f6aef77fdb64db9ed05 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_cerebellum.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b89aa4afce68f3a71f3ce2b88d7033b7238a243cef4a42a67c7ab34065a3d34c +size 150217 diff --git a/ligandome/hla_ligand_atlas_mhc2_colon.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_colon.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..1029ae7de63b5601ee5f5df7f3f8c2f53219cb9a --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_colon.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d15a4bba9b6ebbf851cd81d2ed5396d55b3586135cf34f5b41c47fa1fdcc413 +size 152956 diff --git a/ligandome/hla_ligand_atlas_mhc2_esophagus.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_esophagus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..32d73f5d67387223c05f32f6ce62b15432511986 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_esophagus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:92a562a13216f75f3734e24beac24e7030e5ae644fa84203d60a16ef23de87b3 +size 169485 diff --git a/ligandome/hla_ligand_atlas_mhc2_gallbladder.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_gallbladder.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..5ee9d9a865d7a9c3366660e2983b8df49f2bcab2 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_gallbladder.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3431172de1b9f36692078998323d717f6bc8c62cb0727d9a96b55d9ccc03e4c +size 4425 diff --git a/ligandome/hla_ligand_atlas_mhc2_heart.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_heart.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..4fde84fa3e247962476ba2580e235e0faf5ab4c6 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_heart.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2f41b1240a9af5722724d28f5b4efa4f1255544dc218d018f2f29ba180494457 +size 129801 diff --git a/ligandome/hla_ligand_atlas_mhc2_kidney.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_kidney.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..1e8e3220e16edb38e27f4838d8a7b50c1a0106af --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_kidney.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2b73a976b39fd94604b6a3c5bc12fc06122fe924823226a325af6a12bee35bff +size 243515 diff --git a/ligandome/hla_ligand_atlas_mhc2_liver.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_liver.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..40125bfeec92023a96269ae7573676dda4efd633 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_liver.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e89de145db1b0ffaef6c4c3780af61cedb8ddb31cf6aa513a84b58b5c6547ab +size 189398 diff --git a/ligandome/hla_ligand_atlas_mhc2_lung.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_lung.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6261b324c2717c572a24cba4cd2a6c8858668653 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_lung.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:093fa3fd4b6642ce5d3fd180cd16f372341bd90ee40de3a26b0a423cbc3bdc92 +size 400789 diff --git a/ligandome/hla_ligand_atlas_mhc2_lymph_node.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_lymph_node.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..58f987924411faafa3243a95f2b8cac6e5f0b1d3 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_lymph_node.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f695624c010c966cfab69923b7b1d3ab80406dc693afdae6ba6893a861b3b310 +size 140857 diff --git a/ligandome/hla_ligand_atlas_mhc2_mamma.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_mamma.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..f1807f3333591320069fe51c43b8c193dc0ac584 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_mamma.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c091ae9270122f5dcbc0d945bd7359d90bdb31f3edac6b3bff85e1c713cbaae +size 16527 diff --git a/ligandome/hla_ligand_atlas_mhc2_muscle.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_muscle.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6b7d9ca32afbd427715a265701cb067bee9f1e79 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_muscle.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:89d488e62f2359cf52285c66db5cbd1dd03cb52e2c4f2505ca97b07bf1aa3738 +size 78651 diff --git a/ligandome/hla_ligand_atlas_mhc2_myelon.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_myelon.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ae4f0a3ecef5941b9bda1927ecf13e61d40d1f7c --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_myelon.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:193584a08de02b781753ff740cbf159884ea813a1f964b5ecd3cd43aeacc1687 +size 2574 diff --git a/ligandome/hla_ligand_atlas_mhc2_ovary.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_ovary.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..56ce5da3c838a8776bc1f9dcb088c517130300a0 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_ovary.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:121d0181f4628a4497f08b3d6d2e588832629486bb7140b455ac18084d9f4f5f +size 45572 diff --git a/ligandome/hla_ligand_atlas_mhc2_pancreas.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_pancreas.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..fccf631805add18821bffed20107ffb947688c87 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_pancreas.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:275dc3033f7a271c690016ed8f0353cbebd757fb894f95360c5f6218696c54db +size 161257 diff --git a/ligandome/hla_ligand_atlas_mhc2_presented.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_presented.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..5fd3a7d79d5489d8a7697ab3798889561a406d07 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_presented.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f308e0bef6d6d47d51a5cffd5101c9c6470df9e3fe87a46b05a771c51f3ba61d +size 1687091 diff --git a/ligandome/hla_ligand_atlas_mhc2_presented_nothymus.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_presented_nothymus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ec2deaaf5fbe5aa0aa8c9075e062935123ea8c6c --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_presented_nothymus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:07e450415736f179852b6f50c662606917d0e117c870fe6e1599f480826c2995 +size 1508531 diff --git a/ligandome/hla_ligand_atlas_mhc2_presented_xthymus.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_presented_xthymus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..3ac9df471b9c5376aa1a29b7fddfede300febe64 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_presented_xthymus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:79327c4018eb941a8b9be51b37dd865ddfb8c53a1ae011d4d2da9a6e88a0cb14 +size 1363825 diff --git a/ligandome/hla_ligand_atlas_mhc2_prostate.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_prostate.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..cc32c8f5b1c30686409a1e1357810b22cb340261 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_prostate.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:666299a3f1fd56f8d6cf63fd3250a722ed30468a22ba68e8b730f3c1e3c47940 +size 24997 diff --git a/ligandome/hla_ligand_atlas_mhc2_skin.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_skin.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..67de9b00c74be32a0734497701e450538f348a86 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_skin.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a70c5cc50ac760028f1bdfdf1676522d0d4519ceae1dc0e7f5a5dbadc4f581f3 +size 131295 diff --git a/ligandome/hla_ligand_atlas_mhc2_small_intestine.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_small_intestine.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..7204dca7a7aa9d7e775131b614c1d4c63ae39c50 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_small_intestine.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7d5d38af0d1561a9de90ac37d4ae867f3994e56b435327aaf1839357ad44dd94 +size 278825 diff --git a/ligandome/hla_ligand_atlas_mhc2_spleen.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_spleen.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..c871a6d8c8af2719665a1b92434556c53a97e23d --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_spleen.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e83d3b3a87346a089ae4e35c560d81f456aa63fbbdf862d1ff6d107e4dbe15b3 +size 331060 diff --git a/ligandome/hla_ligand_atlas_mhc2_testis.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_testis.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..02b41b3ed08e2e154b9772174386b9fbb373602d --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_testis.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bcce1e5db48b33a4706a066624e5e3c2e6728f768e69c39d6517dd614699ccde +size 112453 diff --git a/ligandome/hla_ligand_atlas_mhc2_thymus.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_thymus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ce513fff934462fa3ea8e38a5481954594fd381b --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_thymus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1d898e98cfb72021593183faac937ef60a40215e86b99386f8f776acd44625a4 +size 347629 diff --git a/ligandome/hla_ligand_atlas_mhc2_thyroid.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_thyroid.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..3b0c90ae376586b82a873e87b4c692409065b351 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_thyroid.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4df904e8ceb4ae2028c3574e7456a511853d02052dd45a4d56e423c77a233c32 +size 126568 diff --git a/ligandome/hla_ligand_atlas_mhc2_tongue.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_tongue.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..31b992a7b6c974391369c5ea74152c4cff206acc --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_tongue.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11c6990335bf25b826623bd16a2303b8d79d9de4014a8adb4b9c1a6520f6f3ea +size 118123 diff --git a/ligandome/hla_ligand_atlas_mhc2_trachea.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_trachea.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..4a75fc232ab6be0de9600cdacd4979ac4e30a1a6 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_trachea.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b561c46945a56305962691a0b47fbbcd125393df146a019cc912a2fc718bedae +size 144906 diff --git a/ligandome/hla_ligand_atlas_mhc2_uterus.tsv.gz b/ligandome/hla_ligand_atlas_mhc2_uterus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..817ef38f5260ab5a659d5f1007fbdf480c282812 --- /dev/null +++ b/ligandome/hla_ligand_atlas_mhc2_uterus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bd4ca9b737949c102035b77f9ba0bf003f8fe5cc50790767c8259fdb0db29d12 +size 46553 diff --git a/ligandome/hla_ligand_atlas_muscle.tsv.gz b/ligandome/hla_ligand_atlas_muscle.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..1f7cdfe6d60cdc96dfa5248137db283c322d3caa --- /dev/null +++ b/ligandome/hla_ligand_atlas_muscle.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b6c633ebd3d3c274a47dc8e0cd845ab7e011f201d24dc18bcc3f463d685752e +size 37212 diff --git a/ligandome/hla_ligand_atlas_myelon.tsv.gz b/ligandome/hla_ligand_atlas_myelon.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..caee22dea56cfd15c2249893d91ce7aec686509c --- /dev/null +++ b/ligandome/hla_ligand_atlas_myelon.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83a2b74f15a05d603dc417c95c628fb5e8b5b1195d6de0e35486949224d604c4 +size 3237 diff --git a/ligandome/hla_ligand_atlas_ovary.tsv.gz b/ligandome/hla_ligand_atlas_ovary.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..32fc6d0318a0640632fe71a1b3671e2c7860a2b3 --- /dev/null +++ b/ligandome/hla_ligand_atlas_ovary.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7a13a7d3540c6382ee5c968e8e086ff63e5987012481ab37e4277fe25f99e72b +size 145603 diff --git a/ligandome/hla_ligand_atlas_pancreas.tsv.gz b/ligandome/hla_ligand_atlas_pancreas.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..1a321e4a32f362f10c6d8a45dc01dc597d3a738d --- /dev/null +++ b/ligandome/hla_ligand_atlas_pancreas.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3efa9bf7cb679285f2e1c3ea070fc8df860f3a3a0c7534ac2fc1243829d4f3c4 +size 46096 diff --git a/ligandome/hla_ligand_atlas_presented.tsv.gz b/ligandome/hla_ligand_atlas_presented.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..e13ef9b91b92ea1e61acb51d0bb0a69d13f73584 --- /dev/null +++ b/ligandome/hla_ligand_atlas_presented.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:02093b7011a686c5599719a7009dae701560ce20ed3f367b412ba10aea29b277 +size 912735 diff --git a/ligandome/hla_ligand_atlas_presented_nothymus.tsv.gz b/ligandome/hla_ligand_atlas_presented_nothymus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..b7bf9eee5f355c19b53c9df6c8dd1be16218a4e0 --- /dev/null +++ b/ligandome/hla_ligand_atlas_presented_nothymus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3affc01fa6020562ea491e374770aa4e89646e79964a539703c65da9aa0bf039 +size 777453 diff --git a/ligandome/hla_ligand_atlas_presented_xthymus.tsv.gz b/ligandome/hla_ligand_atlas_presented_xthymus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..7bcdb2500f78fcc1691fec6af6400ca2c5f1fbbf --- /dev/null +++ b/ligandome/hla_ligand_atlas_presented_xthymus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:71d74864b2560ae1b45f1f2c522f4bf5419af20a2535a55a26b1cc381bb3b688 +size 636676 diff --git a/ligandome/hla_ligand_atlas_prostate.tsv.gz b/ligandome/hla_ligand_atlas_prostate.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..b8ab8d360d172ff92397a7d3f0cdc47b31ea402b --- /dev/null +++ b/ligandome/hla_ligand_atlas_prostate.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8547cb38007edfecc6850a01771e8d530168f2f4a99006fe3d0491a5d9083c51 +size 40366 diff --git a/ligandome/hla_ligand_atlas_skin.tsv.gz b/ligandome/hla_ligand_atlas_skin.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..08c91e1927ead8115497eab98bdc123e46476297 --- /dev/null +++ b/ligandome/hla_ligand_atlas_skin.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6f8faa23f4192503e282ccc155d69282c8b1755e4f95c24912960a514008be7 +size 46583 diff --git a/ligandome/hla_ligand_atlas_small_intestine.tsv.gz b/ligandome/hla_ligand_atlas_small_intestine.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6c6eac3f6b1f1914f91416b79a8f621b9f3760b9 --- /dev/null +++ b/ligandome/hla_ligand_atlas_small_intestine.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2cb57f04b3a3514451544d8218992acea6d1e37b592106d9394059e2a003288e +size 150099 diff --git a/ligandome/hla_ligand_atlas_spleen.tsv.gz b/ligandome/hla_ligand_atlas_spleen.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..d7852acb6fd0548236952ccd7e4d818e6c135712 --- /dev/null +++ b/ligandome/hla_ligand_atlas_spleen.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9acd51f7916a468ae9090026598c8d01ed49fcf069ea8c5c01662dcbfcdadcdb +size 298225 diff --git a/ligandome/hla_ligand_atlas_testis.tsv.gz b/ligandome/hla_ligand_atlas_testis.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..611b5dc26036624e305a8d5b7d83b631d190eeff --- /dev/null +++ b/ligandome/hla_ligand_atlas_testis.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a0e5cf0f8054bc87913acd679b57c81c160fb897b20f2fca9f27d0f1c7783f6d +size 65930 diff --git a/ligandome/hla_ligand_atlas_thymus.tsv.gz b/ligandome/hla_ligand_atlas_thymus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..833fd206247e5a3acebb90465b8c6932b75bf34e --- /dev/null +++ b/ligandome/hla_ligand_atlas_thymus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dceb4177e86215138bc82d39ecc3c6c3e5a1a21ddfded2f10d022a923b31efa6 +size 285647 diff --git a/ligandome/hla_ligand_atlas_thyroid.tsv.gz b/ligandome/hla_ligand_atlas_thyroid.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..27bd72445681543d8c85c559592ec7640a6a78d4 --- /dev/null +++ b/ligandome/hla_ligand_atlas_thyroid.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:460469068ef4f8a4d94a7b48e6590c9e9d5f59db394281e93ec25be63ae9be82 +size 115221 diff --git a/ligandome/hla_ligand_atlas_tongue.tsv.gz b/ligandome/hla_ligand_atlas_tongue.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..e63e216376afc95c1ac6b8b1cc2956ffe3b77dd1 --- /dev/null +++ b/ligandome/hla_ligand_atlas_tongue.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0dd75d6a6d06dd3c1e69daa007410e999384ad341628af220937290acb8eb28c +size 104483 diff --git a/ligandome/hla_ligand_atlas_trachea.tsv.gz b/ligandome/hla_ligand_atlas_trachea.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6cdaf1431b5bc57fd733c04ecd718bd349dbedff --- /dev/null +++ b/ligandome/hla_ligand_atlas_trachea.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b9e273b43f4b31744b271c3a18dbfa33e0cbc91ce0fa9b712aea20c17b410702 +size 77386 diff --git a/ligandome/hla_ligand_atlas_uterus.tsv.gz b/ligandome/hla_ligand_atlas_uterus.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ddb11ec62d6dd7b763f2109f532d30dc66be6d2d --- /dev/null +++ b/ligandome/hla_ligand_atlas_uterus.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:08c0e916bdd98f56218d5e20eda539631b9370e904275f46b20f5a82f17d8a99 +size 54939 diff --git a/ligandome/self_ligandome_mhc2_hsa.tsv.gz b/ligandome/self_ligandome_mhc2_hsa.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..834898c70d10476cf97d53e8ca759167c6dcdf29 --- /dev/null +++ b/ligandome/self_ligandome_mhc2_hsa.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1e785d33903073f3eed5f6272d059cba5e43509a6d06b7660c480b1d310b90 +size 3434565 diff --git a/ligandome/self_ligandome_mhc2_mmu.tsv.gz b/ligandome/self_ligandome_mhc2_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..f0e9ef4711ea42ab27379a89a09a4600551c76a9 --- /dev/null +++ b/ligandome/self_ligandome_mhc2_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:697fd171eb7dc776ab0e0d775115f3abc1734bfda15688b41b4ba8e936a5bf0d +size 212599 diff --git a/ligandome/tissue_self_mmu.tsv.gz b/ligandome/tissue_self_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..e76ce07886e42f265f1f8b2afd2c2e1b0ef3c67b --- /dev/null +++ b/ligandome/tissue_self_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:34462bddd798399f51caa30a8f899a2b24cdaddf9d2aaef009942cfd803f28cb +size 788062 diff --git a/ligandome/viral_foreign_iedb.tsv.gz b/ligandome/viral_foreign_iedb.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..7fb0412c8c2763f855289b717e3e398156e5520d --- /dev/null +++ b/ligandome/viral_foreign_iedb.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6c4311e48d4472d20b688b61b4f4a4562c743e082c8296c9637a611acdd6d5e6 +size 1029879 diff --git a/ligandome/viral_ligandome_mhc2_pan.tsv.gz b/ligandome/viral_ligandome_mhc2_pan.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..2c0522eda86f9abe65948586d1423686ced9368d --- /dev/null +++ b/ligandome/viral_ligandome_mhc2_pan.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5859d0878f035d703d4f72d288b34afb2fc89580f37543a21f3860e631a9d184 +size 173212 diff --git a/ligandome/viral_orfs_gse272406.tsv.gz b/ligandome/viral_orfs_gse272406.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..d225d1457bd8d5ee67d70b42241f9c747e20690f --- /dev/null +++ b/ligandome/viral_orfs_gse272406.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b6c6d204103cd29add8a2d3bc1717611a2595e5bd89638ab8a46a4331cc0cb0 +size 789117 diff --git a/neoantigens/SOURCES.md b/neoantigens/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..56b2418445feac5fbc258ca5c05e7a043db5f1b8 --- /dev/null +++ b/neoantigens/SOURCES.md @@ -0,0 +1,345 @@ +# `neoantigens/` — sources, provenance, citations and a complete audit + +Every tumour-neoantigen dataset. Moved here from `immunogenicity/` on 2026-08-17 so that directory +holds only general immunogenicity corpora; `nci_complete.tsv.gz` was renamed +`nci_complete_minigenes.tsv.gz` because its unit is a minigene, not an epitope, and the old name hid +that. + +Citations resolved against **PubMed**, PMIDs and DOIs given. Anything unverified is marked as such +rather than filled in. Audit regenerated by `bench/neoag/audit_pmhc_data.py`. + +--- + +## Complete audit + +| file | rows | distinct peptides | pos | neg | prevalence | HLA | donors | length | +|---|--:|--:|--:|--:|--:|--:|--:|:--:| +| `neoantigens_tested_peptides.tsv.gz` | 423,085 | 420,196 | 178 | 422,907 | 0.00042 | 82 | 98 | 8–12 | +| `neoag_tested.tsv.gz` | 525,989 | 491,495 | 29,951 | 496,038 | 0.05694 | 601 | 127 | 3–50 | +| `neoag_candidates.tsv.gz` | 1,355,926 | 1,344,070 | 0 | 0 | — | 82 | 99 | 8–12 | +| `neoag_tested_hsa.tsv.gz` | 717 | 535 | 373 | 344 | 0.52022 | 34 | — | 8–35 | +| `neoag_tested_mmu.tsv.gz` | 1,977 | 1,757 | 740 | 1,237 | 0.37430 | 17 | — | 8–35 | +| `nci_gartner_nmers.tsv.gz` | 9,541 | 9,492 | 144 | 9,397 | 0.01509 | — | 70 | 11–168 | +| `nci_gartner_mmp.tsv.gz` | 147 | 144 | 147 | 0 | 1.00000 | 43 | 77 | 8–12 | +| `nci_gartner_samples.tsv.gz` | 98 | — | — | — | — | 85 | 97 | — | +| `nci_complete_minigenes.tsv.gz` | 1,326 | 1,303 | 28 | 1,298 | 0.02112 | **0** | 46 | 13–316 | + +`neoag_candidates` is the **untested** candidate pool — 0 positives and 0 negatives is correct, its +`immunogenicity` is null throughout. `nci_gartner_mmp` is positives-only by construction (it is the +confirmed-epitope table); its negatives live in `nci_gartner_nmers`. `nci_complete_minigenes` has +**zero** alleles because the deposit is not allele-deconvolved. + +### Residue composition, by label + +| file | Cys % pos | Cys % neg | C-term R/K % pos | C-term R/K % neg | +|---|--:|--:|--:|--:| +| `neoantigens_tested_peptides.tsv.gz` | 15.17 | 22.21 | 14.04 | 11.10 | +| `neoag_tested.tsv.gz` | 19.50 | 23.03 | 11.40 | 11.38 | +| `neoag_tested_hsa.tsv.gz` | 20.38 | 18.60 | 10.19 | 7.27 | +| `neoag_tested_mmu.tsv.gz` | 20.81 | 26.27 | 5.27 | 6.71 | +| `nci_gartner_nmers.tsv.gz` | 40.97 | 39.68 | 11.81 | 11.13 | +| `nci_complete_minigenes.tsv.gz` | 25.00 | 39.21 | 3.57 | 11.86 | + +**No cysteine artefact here, and that is the point of contrast.** The MS-elution depletion that +gives `immunogenicity/chowell_rebuilt.tsv.gz` a 6.9× Cys gradient (10.69% positives vs 1.55% +negatives) is absent: these are screened synthetic peptides, so both classes go through the same +assay. Cys rates are *high* (15–41%) simply because longer peptides are more likely to contain one — +the 25-mer minigene tables sit near 40%. + +A model fitted on Chowell and applied here therefore meets an inverted cysteine distribution. That +is a corpus-transfer artefact, not biology, and it is one measured reason a Chowell-trained residue +term does not transfer to a neoantigen screen. See `../immunogenicity/SOURCES.md` for the full +treatment of both the Cys and the C-terminal Arg/Lys issue. + +--- + +## `neoantigens_tested_peptides.tsv.gz` — the screens at epitope resolution + +423,085 screened **8–12mers** with wild-type counterpart and restricting alleles; 178 CD8-positive; +98 patients. Three datasets in one file with the authors' own split: `NCI` 420,786 · `HiTIDE` 1,563 · +`TESLA` 736. + +Harmonized from `raw/immunogenicity/NCI_dataset_only_tested.txt` by `bench/neoag/ingest_tested.py`. +Canonical columns as-reported; every predicted quantity carries a `pred_` prefix. + +Provenance: **experimental** (CD8 T-cell screening) for `peptide`, `peptide_wt`, `mhc_a`, +`immunogenicity`; **derived/computed** for every `pred_*` column. + +> **TESLA** — Wells DK, van Buuren MM, Dang KK, Hubbard-Lucey VM, Sheehan KCF, Campbell KM, et al. +> **Key Parameters of Tumor Epitope Immunogenicity Revealed Through a Consortium Approach Improve +> Neoantigen Prediction.** *Cell* 2020;183(3):818–834.e13. +> PMID [33038342](https://pubmed.ncbi.nlm.nih.gov/33038342/) · +> doi:[10.1016/j.cell.2020.09.015](https://doi.org/10.1016/j.cell.2020.09.015) + +**Prevalence 4.2 × 10⁻⁴ — one immunogenic peptide in 2,376.** This is the operational base rate a +ranking model faces, and it is ~75× lower than the training-corpus prevalence of +`chowell_rebuilt` (0.039). Reading a corpus-prevalence probability as an operational one overstates +it by that factor. + +--- + +## `nci_gartner_*.tsv.gz` — the NCI screen, deconvolved + +Three tables, deliberately not merged; they are different units of observation. + +* **`nci_gartner_mmp`** — 144 distinct confirmed CD8+ **minimal mutant peptides** with restricting + HLA. 120 carry the wild-type mmp, so agretopicity is **measured, not predicted**. `split` + preserves the paper's train/test. +* **`nci_gartner_nmers`** — 9,541 screened mutations as WT/mutant 25-mers (144 CD8+): the negative + universe at the unit the assay used. +* **`nci_gartner_samples`** — 98 patients with full class-I genotype and histology. This is what + makes a per-patient allele-aware scan possible at all. + +> Gartner JJ, Parkhurst MR, Gros A, Tran E, Jafferji MS, Copeland A, Hanada KI, Zacharakis N, +> Lalani A, Krishna S, Sachs A, Prickett TD, Li YF, Florentin M, Kivitz S, Chatmon SC, +> Rosenberg SA, Robbins PF. **A machine learning model for ranking candidate HLA class I +> neoantigens based on known neoepitopes from multiple human tumor types.** +> *Nat Cancer* 2021;2(5):563–574. +> PMID [34927080](https://pubmed.ncbi.nlm.nih.gov/34927080/) · +> doi:[10.1038/s43018-021-00197-6](https://doi.org/10.1038/s43018-021-00197-6) + +Supplementary Tables 1, 3, 11, 12, 18, 20 of `43018_2021_197_MOESM2_ESM.xlsx`, ingested by +`bench/neoag/ingest_gartner.py`. + +**One join caveat, recorded.** `(ID, Variant key)` is not unique in Suppl. Tables 11 or 12; an outer +join inflates 9,541 rows to 10,521. The ingest deduplicates the annotation side and left-joins with +`validate="m:1"` so the row count stays the screened-nmer count. + +--- + +## `nci_complete_minigenes.tsv.gz` — the same screen, minigene-level + +1,326 mutant 25-mer minigenes across 46 GI-cancer patients, 28 CD8-immunogenic. **`mhc_a` is blank +by design** — not epitope-deconvolved. Formerly `immunogenicity/nci_complete.tsv.gz`. + +> Koşaloğlu-Yalçın Z, Lee J, Greenbaum J, Schoenberger SP, Miller A, Kim YJ, Sette A, Nielsen M, +> Peters B. **Combined assessment of MHC binding and antigen abundance improves T cell epitope +> predictions.** *iScience* 2022;25(2):103850. +> PMID [35128348](https://pubmed.ncbi.nlm.nih.gov/35128348/) · +> doi:[10.1016/j.isci.2022.103850](https://doi.org/10.1016/j.isci.2022.103850) — Table S6. + +**Do not promote its k-mers to positives.** The immunogenicity readout is per minigene: a positive +minigene contains one epitope and dozens of innocent k-mers. Use `nci_gartner_mmp` when the epitope +identity is needed, or score this table max-over-k-mers at the minigene level. + +--- + +## `neoag_tested.tsv.gz` / `neoag_candidates.tsv.gz` — the aggregated workhorse + +`neoag_tested`: 525,989 rows over 9 cohorts — CEDAR 91,651 · Neopep 422,292 · Gfeller 7,758 · +Gfeller_GBM 2,948 · TESLA 605 · Bjerregaard 291 · ITSNdb 199 · GBM 151 · VACCIMEL 94. +`neoag_candidates`: the untested Neopep candidate pool. + +Deduplication rule: primary datasets take precedence over the aggregated compendia (CEDAR, Neopep) — +a duplicate peptide–MHC already present in a primary dataset is dropped from CEDAR/Neopep. + +**`neoag_tested_mmu` / `_hsa` split on `mhc_species`, not host.** H-2 restrictions go to `_mmu`, +HLA restrictions (HLA-transgenic mice *and* human hosts) to `_hsa`; `host_species` / +`source_species` preserve the distinction within each file. `_hsa` has no populated `patient_id`. + +### Per-cohort citations — **partially unresolved** + +Verified: + +* **TESLA** — Wells et al. 2020, PMID [33038342](https://pubmed.ncbi.nlm.nih.gov/33038342/) + (above). + +* **GBM** — Additional file 5 of Buckley PR, Lee CH, Ma R, Woodhouse I, Woo J, Tsvetkov VO, + Shcherbinin DS, Antanaviciute A, Shughay M, Rei M, Simmons A, Koohy H. "Evaluating performance of + existing computational models in predicting CD8+ T cell pathogenic epitopes and cancer + neoantigens." *Brief Bioinform* 2022;**23**(3):bbac141. + doi:[10.1093/bib/bbac141](https://doi.org/10.1093/bib/bbac141) · PMID + [35471658](https://pubmed.ncbi.nlm.nih.gov/35471658/) · PMC9116217. + Resolved 2026-09-03: `raw/immunogenicity/GBM_Peptides.tsv.gz` is **byte-identical** + (md5 `ec00febf6fa68b2378390530e785404f`) to `GBM_Benchmark_A201_Train/GBM_Peptides.tsv` and + `GBM_Benchmark_PanHLA_Train/GBM_Peptides.tsv` in `hkoohy/Immunogenicity_Classifier` at commit + `a8e9807`, whose README reads *"GBM Dataset (Additional file 5) (produced by Margarida Rei and + Rui Ma)"*. **Data**: 153 rows, 4 donors (HGG2, HGG7, HGG11, HGG13), **HLA-A\*02:01 on every + row**, 8–11mers, two response columns — `tcell_response_old` (49 positive) and + `tcell_response_updatedNov19` (33 positive, the one the harmonisation uses). + **It is not the Keskin et al. 2019 NeoVax glioblastoma trial** (PMID 30568305), which had been + inferred from the cohort name downstream: that trial vaccinated 8 patients with 15–30mer long + peptides, median 12 per vaccine, read out per pool of 3–5, and its text names none of HGG, + BDP1, GRIN3A, AMER3 or `MHCrank`. + +**Not yet verified, and deliberately not written out:** CEDAR, Neopep, Gfeller, Gfeller_GBM, +ITSNdb, VACCIMEL, Bjerregaard. Raw inputs for each are under `raw/immunogenicity/` +(`cedar_tcell_assays_25-03-2026.tsv`, `Gfeller_peptides.tsv.gz`, `Gfeller_GBM.tsv.gz`, +`ITSNdb.tsv`, `GBM_Peptides.tsv.gz`, `VACCIMEL.tsv`, `Bjerregaard_Peptides.csv.gz`), which is the +provenance that is actually established. Resolve each against PubMed before citing it anywhere +downstream — do not infer a reference from a cohort name. + +### Cohort overlap — these are not independent holdouts + +Measured: TESLA and NCI/Neopep share **71.8%** of the smaller set (30 of TESLA's 37 positives); +Gfeller and Gfeller_GBM share **96.5%**. Four "independent" cohorts are two families. Treat them as +two, not four, in any leave-one-cohort-out design. + +--- + +## Regenerate + + # in ~/vcs/projects/2026-mhcmatch-benchmark + python bench/neoag/ingest_tested.py # -> neoantigens_tested_peptides.tsv.gz + python bench/neoag/ingest_gartner.py # -> nci_gartner_{mmp,nmers,samples}.tsv.gz + python bench/neoag/audit_pmhc_data.py # -> the audit tables above + +`neoag_tested*` / `neoag_candidates` / `nci_complete_minigenes` are deposited as fetched and are not +regenerated by this repository. + +## Sahin TNBC 2025 — the full supplementary set (2026-08-19) + +Sahin et al., *Nature* 2025, "Individualized mRNA vaccines evoke durable T cell immunity in adjuvant +TNBC", [doi:10.1038/s41586-025-10004-2](https://doi.org/10.1038/s41586-025-10004-2). Published under +**CC BY 4.0**, verified against Crossref, so the *processed* tables are redistributable here with +attribution. The source `.xlsx` and the article PDF stay in `~/hf/tnbc` and are **not** deposited. + +| file | from | rows | provenance | +|---|---|--:|---| +| `sahin_tnbc_2025.tsv.gz` | MOESM3, sheet `n=251 targets` | 251 | **experimental** — one row per vaccinated target; `ex_vivo` and `post_ivs` are two separate ELISpot read-outs, `mean_spot_count` the magnitude | +| `sahin_tnbc_2025_drivers.tsv.gz` | MOESM3, sheet `Breast cancer driver genes` | 6 | **experimental** — driver-gene targets and their per-patient responses | +| `sahin_tnbc_2025_deg.tsv.gz` | MOESM4 | 24,517 | **computed** — DESeq2 tumour vs normal on the same cohort: `baseMean, log2FoldChange, lfcSE, stat, pvalue, padj` | +| `sahin_tnbc_2025_gsea.tsv.gz` | MOESM5 | 43 | **computed** — gene-set enrichment over those genes | + +`sahin_tnbc_2025_deg.tsv.gz` is a **log2 fold change, not TPM**, and is deposited under its own name +so that distinction stays visible: it is not a drop-in for the `expr` feature, which is an absolute +expression level. It is the only tumour-vs-normal contrast in this directory measured on the same +patients as a prospective holdout. + +This cohort is **held out by construction** — it was published after every model in the benchmark +repository was fitted, and enters no fit. + +Regenerate: + +```zsh +python bench/neoag/ingest_sahin.py # -> sahin_tnbc_2025.tsv.gz +python bench/neoag/ingest_sahin_extra.py # -> the drivers, deg and gsea tables +``` + +--- + +## Added 2026-08-24 — the NeoPrecis pass, and the leakage flag it forced + +Four parquet tables deposited by `bench/neoag/ingest_vaccines.py`, from the supplements of +Lee KH, Sears TJ, Zanetti M, Carter H, "NeoPrecis: enhancing immunotherapy response prediction +through integration of qualified immunogenicity and clonality-aware neoantigen landscapes", +*Nat Commun* 2026;**17**(1), +doi:[10.1038/s41467-026-68651-6](https://doi.org/10.1038/s41467-026-68651-6), +PMID [41577704](https://pubmed.ncbi.nlm.nih.gov/41577704/), PMC12932759. + +| file | rows × cols | what | provenance | +|---|--:|---|---| +| `nci_parkhurst_gi.parquet` | 7,384 × 48 | per-mutation NCI GI TIL screen, 75 patients | **mixed** — `patient`, `chromosome`, `start`, `ref`, `alt`, `refgene_name`, `consequence`, `wt_seq`, `mt_seq`, `immune_activity`, `cd8`, `cd4`, `dna_af`, `rna_af`, `rna_exp_qrt` are **experimental**; every `phbr_*`, `agretopicity_*`, `foreignness_*`, `*dist_*`, `crd_*`, `np_immuno_*`, `prime_*`, `icerfire_*`, `deepneo_*` column is **derived** by NeoPrecis; `in_epic_fit` is **derived** here | +| `neoprecis_ici_cohort.parquet` | 525 × 50 | checkpoint-blockade cohort | **mixed** — clinical and outcome columns experimental, landscape scores derived | +| `neoprecis_ranker_panel.parquet` | 4,263 × 22 | four published rankers on labelled peptides | **mixed** — `label`/`measurement` experimental, scores derived | +| `neoprecis_allele_benefit.parquet` | 547 × 4 | per-allele benefit scores | **derived** | + +### `nci_parkhurst_gi` overlaps `neoantigens_tested_peptides` — use `in_epic_fit` + +Both come from the same NCI screening programme, and `neoantigens_tested_peptides.tsv.gz` (420,786 +of whose 423,085 rows are `NCI`) is in the corpus the shipped EPIC scorer was fitted on. Measured +overlap: **29 of the 75 patients are in both**, and 7,865 of 8,000 sampled corpus peptides on those +patients occur inside an `mt_seq` 25-mer here — the same mutations, tiled into 8–12mers. + +| half | patients | mutations | CD8⁺ | CD4⁺ | +|---|--:|--:|--:|--:| +| `in_epic_fit = false` | 46 | 4,019 | 15 | 45 | +| `in_epic_fit = true` | 29 | 3,365 | 42 | 21 | + +Any figure pooled over all 75 patients is part training-set performance, and the CD8 positives are +concentrated in the fitted half. The flag ships as a column so this cannot be missed by a reader who +has not read a script. + +`nci_complete_minigenes.tsv.gz` above is the **same underlying screen at minigene resolution** +(46 patients, 1,326 minigenes, via Koşaloğlu-Yalçın et al., *iScience* 2022); this table is the +mutation-level view of it via NeoPrecis, and it carries the abundance and derived-feature columns the +minigene table does not. + +Original screen: Parkhurst MR, Robbins PF, Tran E, Prickett TD, Gartner JJ, Jia L, Ivey G, Li YF, +El-Gamil M, Lalani A, …, Rosenberg SA. "Unique Neoantigens Arise from Somatic Mutations in Patients +with Gastrointestinal Cancers." *Cancer Discov* 2019;**9**(8):1022-1035. +doi:[10.1158/2159-8290.CD-18-1494](https://doi.org/10.1158/2159-8290.CD-18-1494), +PMID [31164343](https://pubmed.ncbi.nlm.nih.gov/31164343/), PMC7138461. 75 patients, 124 +neoantigen-reactive TIL populations, 1.6 % of somatic mutations immunogenic. + +Shipped vaccine cassettes from the same staging directory are deposited in **`vaccines/`**, which has +its own `SOURCES.md`. + +--- + +## Cleanup 2026-09-04 — this directory now holds neoantigens only + +Run by `bench/pmhc_data/clean_neoantigens.py` in `2026-mhcmatch-benchmark-ext`; `--audit` reruns +every count below without writing. Three rules, applied to every table here: + +1. **No pathogen epitopes.** A tumour neoantigen and a pathogen epitope are answered by different + mechanisms, which is why `mhcmatch.rank.AGGREGATE_MODES` carries the distinction as a *key* and + not a covariate. Pooling them was measured before this rule existed: mouse class-I held-out + within-reference AUROC **0.5930 → 0.5241** on peptide-grouped folds, 1,703 rows against the 923 + that had done better. +2. **Cross-species only across human and mouse.** An HLA-transgenic mouse, or a human antigen in a + syngeneic model, is a real design and `bench/mouse_neoag/ortholog.py` resolves the gene across + it. Any other pair is a mislabel. +3. **A source organism on every row**, resolved against `proteome/human.fasta.gz` (147,506 + accessions) and `proteome/mouse.fasta.gz` (54,857) rather than guessed. A row whose accessions + disagree, or match neither, keeps its null and is removed rather than filled by majority vote. + +| table | rows before | rows after | removed | what was removed | +|---|--:|--:|--:|---| +| `neoag_tested_hsa.tsv.gz` | 717 | 709 | 8 | HCMV pp65 (`Q6SW59`) — `NLVPEVATV` and three single-residue analogs, filed `source_species = "other"` | +| `neoag_tested_mmu.tsv.gz` | 1,977 | 1,974 | 3 | no accession, no molecule, no organism — unattributable | + +21 rows in `_hsa` and 19 in `_mmu` carried a null `source_species` and were **resolved, not +dropped**, from their deposited `uniprot_id`. Cross-species rows kept: 3 mouse-source in `_hsa`, +98 human-source in `_mmu`. + +### New — the class-II deposits, one per species + +`neoantigens/` carried **34** human and **172** mouse MHC-II rows, which is too thin to fit six +terms against. `neoag_tested.tsv.gz` does hold 23,262 human class-II rows and they are **not +usable for a fit**, for four reasons worth recording so they are not rediscovered: + +* `reference_id` is **null on every one** — no intercept group, no held-out reference; +* **14,143 of 23,262** are filed as `HLA class II` with no allele at all; +* there is no `gene_name` column in that table, so the expression block has nothing to key on; +* **1,328 rows spelled `H2-d class II` / `H2-b class II` / `H2 class II` carry + `mhc_species = HomoSapiens`** — mouse restrictions in a human-labelled table. + +So both class-II deposits are rebuilt from the CEDAR export named above, which has all four fields. + +| file | rows | positives | prevalence | alleles | references | source | +|---|--:|--:|--:|--:|--:|---| +| `cedar_neoag_mhc2_hsa.tsv.gz` | 1,112 | 656 | 0.590 | 72 | 157 | `bench/mhc2_human/build_cedar_human_mhc2.py` | +| `cedar_neoag_mhc2_mmu.tsv.gz` | 415 | 135 | 0.325 | 7 | 49 | `bench/mouse_neoag/build_cedar_mouse.py` | + +Both are **derived**, not experimental: the assay results are CEDAR's, the label rule +(positive if any assay on that `(peptide, allele)` read `Positive*`, negative only when every one +read `Negative`) and the gene resolution are ours. + +> **The human table was rebuilt after its first commit and this row records the rebuild.** The +> initial build collapsed duplicate `(peptide, allele)` pairs first-wins, which let row order decide +> the label of the 75 pairs whose assays disagree, and it carried no `disease` column. The rule +> above is the any-positive one, and applying it moved **581 -> 656 positives** (prevalence 0.522 +> -> 0.590) with the row count, allele count and reference count unchanged. The shipped +> `aggregate_mhc2_human.json` (`fit.positives` 656, and a `population` block naming 46 diseases) +> was fitted on the rebuild, so the earlier numbers describe bytes nothing downstream ever used. + +The mouse table lost **969 of 1,528** CEDAR rows to rule 1 — *Bordetella pertussis* 147, HPV16 103, +*Coxiella burnetii* 97, LCMV 88, Merkel cell polyomavirus 48, vaccinia 45, HBV 78 across two +subtypes, *Plasmodium* 30, and 149 with no attribution at all. **Ovalbumin goes too**: `Gallus +gallus`, 7 rows. OVA/SIINFEKL is the canonical mouse model antigen and it is still not a +neoantigen; it is named here because it is the one removal an immunologist would ask about. +257 of the 415 kept rows are human-source xenoantigen, which rule 2 allows. + +## Caveat — mouse rows labelled as human in `neoag_tested.tsv.gz` + +6,186 CEDAR rows carry mouse MHC alleles (`H2-Db` 1,020, `H2-Kb` 932, `H2-b class I` 719, +`H2-Kd` 525, `H2-d class II` 505, …) while their `mhc_species` reads `HomoSapiens` and +`host_species` reads `human`. `SIINFEKL`, the canonical H-2Kb ovalbumin epitope, is among them. + +**Filtering this table on `mhc_species` therefore silently admits mouse pMHC.** The defect is +upstream and is deliberately left unfixed, so the table stays byte-identical to what was audited; +`neoag_tested_mmu.tsv.gz` is the correctly-labelled murine set. + +`neoag_tested_{mmu,hsa}` also overlap this table by design — all 2,273 peptides appear in both, +since both are IEDB-derived — so dedupe on peptide × MHC when joining. diff --git a/neoantigens/cedar_neoag_mhc2_hsa.tsv.gz b/neoantigens/cedar_neoag_mhc2_hsa.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..5bcc87fe97ba37109b6cb3f9806c817a34aa0404 --- /dev/null +++ b/neoantigens/cedar_neoag_mhc2_hsa.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7b8bb0f1973fb0b93aec914554a9d2f55d0a756ecaebb174d8e2391f94cbaaec +size 14774 diff --git a/neoantigens/cedar_neoag_mhc2_mmu.tsv.gz b/neoantigens/cedar_neoag_mhc2_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..3269d1ecf5cbcc64e697bf76aaef9201a8654bce --- /dev/null +++ b/neoantigens/cedar_neoag_mhc2_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:98edaba95b9d9b244100a6dfd5db4489cdee229dccdee926d9653a9b56e070f4 +size 3899 diff --git a/neoantigens/nci_complete_minigenes.tsv.gz b/neoantigens/nci_complete_minigenes.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..25bf031ad23e35f6b93628720045d258641d94c5 --- /dev/null +++ b/neoantigens/nci_complete_minigenes.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8af04aafec01dde959400945ecba939b3a64fb9ad8c94449272f77d4e178efc +size 38870 diff --git a/neoantigens/nci_gartner_mmp.tsv.gz b/neoantigens/nci_gartner_mmp.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..6dc0a1d81624e307ff496beddbdadaa9accb5e50 --- /dev/null +++ b/neoantigens/nci_gartner_mmp.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:52b331decad1704df4c93086714bda18c4aa6b93c73d583d71aa7c344d3f6aa4 +size 19142 diff --git a/neoantigens/nci_gartner_nmers.tsv.gz b/neoantigens/nci_gartner_nmers.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..5aa5730b23ed0413397a9cc876e33579c2f1244b --- /dev/null +++ b/neoantigens/nci_gartner_nmers.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c385fa2af4427d4e98d62ba2bba49a26ec1f669658c903007ea37f1ad7d4f7d +size 1134674 diff --git a/neoantigens/nci_gartner_samples.tsv.gz b/neoantigens/nci_gartner_samples.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..bf3293c5a9d9d0a99b44e5ebfcc7fe0db32cf584 --- /dev/null +++ b/neoantigens/nci_gartner_samples.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd58143296b2885f33271a3fc3f387f25b803a59fc8cbc71e5a9966b66233073 +size 3481 diff --git a/neoantigens/nci_parkhurst_gi.parquet b/neoantigens/nci_parkhurst_gi.parquet new file mode 100644 index 0000000000000000000000000000000000000000..19d5af66af8d41327466ae42b30d526d64457d56 --- /dev/null +++ b/neoantigens/nci_parkhurst_gi.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:67c4c11d26a7b221eb1bcbdbefbb4d83b43c1b7c4c2c7685eef2c32d46b6c207 +size 865607 diff --git a/neoantigens/neoag_candidates.tsv.gz b/neoantigens/neoag_candidates.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..8a67a85b2f5351cb14b34da549d6740309977e9c --- /dev/null +++ b/neoantigens/neoag_candidates.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54851cf9d574e0fe085a2f80c3f3c06cd84c60805612bb1e08da48e29d671d3c +size 17531721 diff --git a/neoantigens/neoag_tested.tsv.gz b/neoantigens/neoag_tested.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..809fd9acbd7e28b0557b4400cc6bc986c5fd204c --- /dev/null +++ b/neoantigens/neoag_tested.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:422ba89167bf327e8251f9bf8bb2c928435ad77b9706a724c5bab83e32e6540d +size 7053481 diff --git a/neoantigens/neoag_tested_hsa.tsv.gz b/neoantigens/neoag_tested_hsa.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..8367aca2eb21647f9cc0f9ff66a3acb9900c01d8 --- /dev/null +++ b/neoantigens/neoag_tested_hsa.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:839430e926a12009119d4b811aa12e807acaee202b3e0547eb68bd78cc8f510d +size 30057 diff --git a/neoantigens/neoag_tested_mmu.tsv.gz b/neoantigens/neoag_tested_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..527d797010f9a3beb79627e84361b91fc8ad390e --- /dev/null +++ b/neoantigens/neoag_tested_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f40701057ec5510914b0a10d9e8c6f90f6c53fdea47b0d56f815329d997438cc +size 161122 diff --git a/neoantigens/neoantigens_tested_peptides.tsv.gz b/neoantigens/neoantigens_tested_peptides.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..d618b511760df6c525103669d5ecac5776bb9d39 --- /dev/null +++ b/neoantigens/neoantigens_tested_peptides.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f732baee6f8b0dcb909eaea3720ec51be9b990152bebac6bde6a0a5bcd3afd69 +size 42751480 diff --git a/neoantigens/neoprecis_allele_benefit.parquet b/neoantigens/neoprecis_allele_benefit.parquet new file mode 100644 index 0000000000000000000000000000000000000000..a5a2e1bb417ddd5290cab756197ebadf43896fd8 --- /dev/null +++ b/neoantigens/neoprecis_allele_benefit.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1539646a69b8e365c6d74909c1a9b6b553c894649487ee137a6db928014876ac +size 4882 diff --git a/neoantigens/neoprecis_ici_cohort.parquet b/neoantigens/neoprecis_ici_cohort.parquet new file mode 100644 index 0000000000000000000000000000000000000000..3a14095335996e1a2b7b0e671038033dc74b5e29 --- /dev/null +++ b/neoantigens/neoprecis_ici_cohort.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7798b634738fe2667123b457c23b2e073c9e4460d6abf17ddb064fa43e749bab +size 130131 diff --git a/neoantigens/neoprecis_ranker_panel.parquet b/neoantigens/neoprecis_ranker_panel.parquet new file mode 100644 index 0000000000000000000000000000000000000000..8358ec2d3936792d75c2d8146db647e7ab1268da --- /dev/null +++ b/neoantigens/neoprecis_ranker_panel.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6cebcb0c485ef7730e13525b4cfc8258aee8375061dd7f0c28f21d422af2d4c9 +size 213863 diff --git a/neoantigens/sahin_tnbc_2025.tsv.gz b/neoantigens/sahin_tnbc_2025.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..333caae41202b43d06f636519514b449ccc43b60 --- /dev/null +++ b/neoantigens/sahin_tnbc_2025.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bce12324657f0a326bd8189a10194972b83f3bf9e995744c05f10c47ad1c942d +size 4697 diff --git a/neoantigens/sahin_tnbc_2025_deg.tsv.gz b/neoantigens/sahin_tnbc_2025_deg.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..00ea6c35344783d05b789caaba6f10be8613fa60 --- /dev/null +++ b/neoantigens/sahin_tnbc_2025_deg.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6bec570eef94b1f1e3bf1a06b15dba9643ea33182865b35ec1d6efbf8b0771de +size 1392118 diff --git a/neoantigens/sahin_tnbc_2025_drivers.tsv.gz b/neoantigens/sahin_tnbc_2025_drivers.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..51b3ea6fdbd6b4dca6f84cdc4a1c9986f0c13c80 --- /dev/null +++ b/neoantigens/sahin_tnbc_2025_drivers.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ea2ff9cf00dcc8f6d21c7a7662155c18d9bc3b4724f6018376c7ccbff47e261c +size 301 diff --git a/neoantigens/sahin_tnbc_2025_gsea.tsv.gz b/neoantigens/sahin_tnbc_2025_gsea.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..5206d52ddbce5b50d952a9c29a78bc1fc3dafccc --- /dev/null +++ b/neoantigens/sahin_tnbc_2025_gsea.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e93e8d6bbf33736aa11cf24cfe90e9a7a86dae0775ea1c2f3c4e5114a5339ec9 +size 6975 diff --git a/pmhc/SOURCES.md b/pmhc/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..44e100fb0eb383348b05c3ce650bcdd6cc8422f3 --- /dev/null +++ b/pmhc/SOURCES.md @@ -0,0 +1,30 @@ +# `pmhc/` — sources + +The reference ligand panel every `mhcmatch.Store` is built from. **Experimental**: positive IEDB +MHC-ligand records, harmonised; nothing here is predicted. + +| file | rows | what it is | +|---|--:|---| +| `pmhc_full.tsv.gz` | 1,482,188 | every positive IEDB epitope–MHC record | +| `pmhc_shortlist.tsv.gz` | 645,102 | the subset supported by **≥2 references** | + +Schema: `epitope`, `gene`, `species`, `mhc_a`, `mhc_b`, `mhc_class`, `mhc_species`, `reference_id`, +`n_references`. + +`pmhc_full.tsv.gz` composition: + +| class | host | rows | +|---|---|--:| +| MHCI | HomoSapiens | 1,040,631 | +| MHCII | HomoSapiens | 327,487 | +| MHCI | MusMusculus | 97,601 | +| MHCII | MusMusculus | 16,469 | + +**The two tiers are a precision/recall choice, not a quality ranking.** `shortlist` drops +single-reference pairs, which removes both the noise and the rare alleles that only one study ever +reported — so it is the better background and the worse coverage. `full` is the default; +`Store.from_pmhc(tier="shortlist")` selects the other. + +```zsh +mhcmatch bootstrap --tier all # fetches both, ~16 MB +``` diff --git a/pmhc/pmhc_full.tsv.gz b/pmhc/pmhc_full.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..4ae0b29efd68db50d02f14673dc444f9d48a0cf1 --- /dev/null +++ b/pmhc/pmhc_full.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0752bb5af6fbe8eb14521a1e9a6bcff3e6b68ddc4602a21b59bf332dc2b97f7 +size 11743092 diff --git a/pmhc/pmhc_shortlist.tsv.gz b/pmhc/pmhc_shortlist.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..61dc27be55b3dccdd3c69d01e5dba0638b528733 --- /dev/null +++ b/pmhc/pmhc_shortlist.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cf0d2a09d68b6c30c56a742a06a2b8329b23cf8e82df5c94fc79b8b94b5ea3b5 +size 4329305 diff --git a/proteome/README.md b/proteome/README.md new file mode 100644 index 0000000000000000000000000000000000000000..c5c2b91fda19e2362c5dc90febf825a58e058e8d --- /dev/null +++ b/proteome/README.md @@ -0,0 +1,36 @@ +# Reference proteomes + +UniProt proteomes used for source-protein lookup and peptide-flank extraction. All gzipped, git-LFS tracked. + +`human.fasta.gz` (UP000005640, 147,506 seqs) and `mouse.fasta.gz` (UP000000589, 54,857 seqs) are the full +proteomes **including isoforms and TrEMBL** entries. The files below are canonical **reference** proteomes, +fetched from UniProt (REST `?query=(proteome:)&format=fasta`, or the reference_proteomes FTP). + +| file | organism | proteome | taxid | sequences | bytes (gz) | +|---|---|---|---|---|---| +| `ecoli_K12_UP000000625.fasta.gz` | *Escherichia coli (strain K12)* | UP000000625 | 83333 | 4,403 | 988,444 | +| `egallinarum_UP000254807.fasta.gz` | *Enterococcus gallinarum* | UP000254807 | 1353 | 3,635 | 775,627 | +| `mgnavus_UP000018690.fasta.gz` | *Mediterraneibacter gnavus (strain CC55_001C)* | UP000018690 | 1073375 | 2,942 | 666,125 | +| `saureus_UP000008816.fasta.gz` | *Staphylococcus aureus (strain NCTC 8325 / PS 47)* | UP000008816 | 93061 | 2,889 | 577,723 | +| `lreuteri_UP000001991.fasta.gz` | *Limosilactobacillus reuteri (strain DSM 20016)* | UP000001991 | 557436 | 1,865 | 401,750 | +| `cmv_HHV5_UP000000938.fasta.gz` | *Human cytomegalovirus (strain Merlin)* | UP000000938 | 295027 | 168 | 42,838 | +| `hhv6a_UP000009295.fasta.gz` | *Human herpesvirus 6A (strain Uganda-1102)* | UP000009295 | 10370 | 99 | 29,596 | +| `ebv_UP000153037.fasta.gz` | *Epstein-Barr virus (strain B95-8)* | UP000153037 | 10377 | 92 | 27,660 | +| `hhv7_UP000098510.fasta.gz` | *Human herpesvirus 7 (strain RK)* | UP000098510 | 262398 | 84 | 26,566 | +| `ebv_AG876_UP000007639.fasta.gz` | *Epstein-Barr virus (strain AG876)* | UP000007639 | 82830 | 79 | 26,243 | +| `hsv1_HHV1_UP000009294.fasta.gz` | *Human herpesvirus 1 (strain 17)* | UP000009294 | 10299 | 73 | 25,505 | +| `vzv_HHV3_UP000002602.fasta.gz` | *Varicella-zoster virus (strain Dumas)* | UP000002602 | 10338 | 69 | 24,236 | +| `adenovirusC_HAdV2_UP000008167.fasta.gz` | *Human adenovirus C serotype 2* | UP000008167 | 10515 | 45 | 9,028 | +| `fluA_PR8_UP000009255.fasta.gz` | *Influenza A virus (strain A/Puerto Rico/8/1934 H1N1)* | UP000009255 | 211044 | 13 | 3,378 | +| `bkpyv_UP000008475.fasta.gz` | *BK polyomavirus* | UP000008475 | 1891762 | 6 | 1,238 | +| `parvoB19_UP000006624.fasta.gz` | *Human parvovirus B19 (strain HV)* | UP000006624 | 648237 | 5 | 1,342 | + +Totals: 16,467 sequences across 16 proteomes (3,627,299 bytes gzipped). + +## Note on Epstein-Barr virus and HHV-7 + +Two **EBV** strains are included: B95-8 (UP000153037, taxid 10377) and AG876 (UP000007639, taxid 82830). +UP000007639 is EBV strain AG876 -- **not** HHV-7, despite what its proteome ID may suggest by adjacency. +The genuine HHV-7 reference proteome is **UP000098510** (strain RK, taxid 262398, 84 proteins); +UP000009246 (strain JI) exists but is flagged *Non Reference* by UniProt. + diff --git a/proteome/SOURCES.md b/proteome/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..d0f7e45c91d324fbcccd8cf557790433ba0a80ee --- /dev/null +++ b/proteome/SOURCES.md @@ -0,0 +1,26 @@ +# `proteome/` — sources + +**Provenance lives in [`README.md`](README.md)** beside this file, and is not repeated here: it +carries every proteome's organism, UniProt proteome ID, taxid, sequence count and gzipped size, plus +the exact fetch form (UniProt REST `?query=(proteome:)&format=fasta`, or the +reference_proteomes FTP). All 19 files are **experimental** sequence, unmodified. + +Summary: `human.fasta.gz` (UP000005640, 147,506 seqs) and `mouse.fasta.gz` (UP000000589, 54,857) are +the **full** proteomes including isoforms and TrEMBL; the other 16 are canonical **reference** +proteomes — 5 bacterial and 11 viral, 16,467 sequences in total — for molecular-mimicry scans and +peptide-flank extraction. + +```zsh +mhcmatch bootstrap --proteome human,mouse,ecoli_K12_UP000000625 +``` + +Two traps the README records and that are easy to reintroduce: + +- **UP000007639 is EBV strain AG876, not HHV-7**, despite the adjacency of the IDs. The genuine + HHV-7 reference is UP000098510 (strain RK, 84 proteins). +- Two EBV strains are present on purpose (B95-8 and AG876); they are not duplicates. + +**Size is the operational fact.** The human proteome has 68,389,335 nine-mer windows — every +position of every protein, because a source lookup is a question about *where* — and one seqtree +index over them costs ~12.6 GB. Ask for the lengths you need +(`mhcmatch.proteome.Proteome.find_sources` builds one index per distinct query length). diff --git a/proteome/adenovirusC_HAdV2_UP000008167.fasta.gz b/proteome/adenovirusC_HAdV2_UP000008167.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..9f311be96393dd84eefae38137070291d01c1bf9 --- /dev/null +++ b/proteome/adenovirusC_HAdV2_UP000008167.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b1eabff2a4b7e2cefb5dbc51a61914b904e005711901ab277f157d309d021e41 +size 9028 diff --git a/proteome/bkpyv_UP000008475.fasta.gz b/proteome/bkpyv_UP000008475.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..381fcff5d46d113e3e6084d08b6746680ff50397 --- /dev/null +++ b/proteome/bkpyv_UP000008475.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:56ee6f1f221c2b079df8908bf62510da8c34ef1fbfb2f8ef2d8506c0827cbf4e +size 1238 diff --git a/proteome/cmv_HHV5_UP000000938.fasta.gz b/proteome/cmv_HHV5_UP000000938.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..d8dc47f1428abf007968636f0f1b914c8fb5688b --- /dev/null +++ b/proteome/cmv_HHV5_UP000000938.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ea38683920247574a05632931142d57c61106212aafa4a738c41b6cfb04eb5d4 +size 42838 diff --git a/proteome/ebv_AG876_UP000007639.fasta.gz b/proteome/ebv_AG876_UP000007639.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..ba05485b47c595f8108b557b6774426cf2615478 --- /dev/null +++ b/proteome/ebv_AG876_UP000007639.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a37dc317bad5aff21b7698d2daa1b3e71b3de6d69519fd9eb258bfdedff57fd8 +size 26243 diff --git a/proteome/ebv_UP000153037.fasta.gz b/proteome/ebv_UP000153037.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..67eed1d6910ae1fdcb9f018def752e63f3820913 --- /dev/null +++ b/proteome/ebv_UP000153037.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:db6757bcdab194763f04d13e5deda8ae1df9af9d0f933a3c7ed42f83831f871e +size 27660 diff --git a/proteome/ecoli_K12_UP000000625.fasta.gz b/proteome/ecoli_K12_UP000000625.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..a8c23dd0718401e8d46d4a0acd7f9805b3434433 --- /dev/null +++ b/proteome/ecoli_K12_UP000000625.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61404bd12b6e35a69da2727efbd3a79ee7fad227c856a1866628c423c01c6c27 +size 988444 diff --git a/proteome/egallinarum_UP000254807.fasta.gz b/proteome/egallinarum_UP000254807.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..ef55e0f6e0de2907eaa4e176bc6d2b292be473bd --- /dev/null +++ b/proteome/egallinarum_UP000254807.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b563a293116491dfbbb1724b576925c2db045d3571b2089147ad7b24bf7ad3bd +size 775627 diff --git a/proteome/fluA_PR8_UP000009255.fasta.gz b/proteome/fluA_PR8_UP000009255.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..294e5e5c06b8d0b9c51c1cbd0ac355bcce425f42 --- /dev/null +++ b/proteome/fluA_PR8_UP000009255.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87236a912871ab45735becb74a4e8f3d298e9b2bfd2122016270ca5ba8ac92d9 +size 3378 diff --git a/proteome/hhv6a_UP000009295.fasta.gz b/proteome/hhv6a_UP000009295.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..90db7ce90f1848f8df5ee95fe4191e48f9777dc8 --- /dev/null +++ b/proteome/hhv6a_UP000009295.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:755d2d660bcd0e776ec6dd97dd8a363e7d30d06551f9f630ae6bb9185d3c2dcf +size 29596 diff --git a/proteome/hhv7_UP000098510.fasta.gz b/proteome/hhv7_UP000098510.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..1efddfecd142357d44e07b50e74004be85a56aa6 --- /dev/null +++ b/proteome/hhv7_UP000098510.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:de24382b53fb242cb7de8c86716f555324f476b54e629ee9e4ea724f56d93bb6 +size 26566 diff --git a/proteome/hsv1_HHV1_UP000009294.fasta.gz b/proteome/hsv1_HHV1_UP000009294.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..1c4eb1fce325cf389892872fd904c6d4c63099f0 --- /dev/null +++ b/proteome/hsv1_HHV1_UP000009294.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fb91d8970094a4ba70d3f557691bbec3219fd2b61bd2f799e3893f9cf1f3e414 +size 25505 diff --git a/proteome/human.fasta.gz b/proteome/human.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..e84199998a3d4ceda5454115a0fa23197bb239a8 --- /dev/null +++ b/proteome/human.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6068b7ad09ebc57f1b7335abd936d74e7fdd58a67e9019efe8a30968d10e04b6 +size 38512271 diff --git a/proteome/lreuteri_UP000001991.fasta.gz b/proteome/lreuteri_UP000001991.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..3c9ab1d697f5e61db1fa18d333bd9ae15e16935d --- /dev/null +++ b/proteome/lreuteri_UP000001991.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:153d70bdd29eecbe5f9f0acfde9d53ba7208afd5058cef9774a223aa329608ab +size 401750 diff --git a/proteome/mgnavus_UP000018690.fasta.gz b/proteome/mgnavus_UP000018690.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..3f94e81f03e4970556475b7992ae0f8bce64f040 --- /dev/null +++ b/proteome/mgnavus_UP000018690.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59066c0e5d284518ed1ee27c270c3b8a9908fdb7e3caf99857cca8898a9357ea +size 666125 diff --git a/proteome/mouse.fasta.gz b/proteome/mouse.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..c7a7b8fe006341d3c7a10452c79e0660206eb562 --- /dev/null +++ b/proteome/mouse.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0e3ca4d8f01504c968ae9e2550484ad3adec32b5e0a50b7a4c74f89efabd1b75 +size 14956114 diff --git a/proteome/parvoB19_UP000006624.fasta.gz b/proteome/parvoB19_UP000006624.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..c48cfc51bf0b8a05975ec9de07c49426eadb1bc0 --- /dev/null +++ b/proteome/parvoB19_UP000006624.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c7cbd0d64c860b211a6de242e21c2d7204710167f70b6271ef9df32a662d1ac +size 1342 diff --git a/proteome/saureus_UP000008816.fasta.gz b/proteome/saureus_UP000008816.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..e7469b0bf405b963224df9a85ce6c7690a790500 --- /dev/null +++ b/proteome/saureus_UP000008816.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a83758f5b58896369e5fcfc242ce2d25c5c67524cf80e1574ee0bb1f69f4cd15 +size 577723 diff --git a/proteome/vzv_HHV3_UP000002602.fasta.gz b/proteome/vzv_HHV3_UP000002602.fasta.gz new file mode 100644 index 0000000000000000000000000000000000000000..3f11940493f22d2766dec1728568b3af7067c258 --- /dev/null +++ b/proteome/vzv_HHV3_UP000002602.fasta.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8f5df515a159e5ee8d2f5102a028781c581045116ecf61f3e7a8df896eeb562a +size 24236 diff --git a/tcga/SOURCES.md b/tcga/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..19973de9f4b42200a11d8e45269285e297eb19eb --- /dev/null +++ b/tcga/SOURCES.md @@ -0,0 +1,26 @@ +# `tcga/` — sources + +Pan-cancer neoantigen predictions over open-access TCGA. **Derived, not measured** — every row is a +prediction, and the only experimental inputs behind it are the TCGA somatic calls, expression and +per-barcode HLA genotypes. + +| file | rows | what it is | +|---|--:|---| +| `tcga_neoantigens.tsv.gz` | 2,244,779 | expressed mutant peptide × HLA binders, 8,505 donors | +| `hla_coverage.tsv.gz` | 353 | per-allele donor counts, donor fraction, allele frequency, epitope counts | + +`tcga_neoantigens` schema: `sample`, `peptide`, `mhc_a`, `affinity`, `expression`, `ensp`, +`mhc_class`, `mhc_species`, `source_species`, `host_species`, `immunogenicity`, `dataset_origin`, +`cancer_type`. + +Derived from open-access TCGA somatic + expression data, with per-barcode HLA genotypes and +neoantigen calls as published by TCIA / Thorsson et al. 2018. **No controlled-access data is used or +reproduced here**, which is why this directory is public while the NCI intermediates (dbGaP +phs001003) never leave local disk. + +**`immunogenicity` in this file is a prediction, not an assay result.** Nothing in TCGA was +T-cell-screened. Use it as a candidate universe and a per-allele prevalence reference; a model +validated against it has been validated against another model. + +`hla_coverage.tsv.gz` is what makes an allele-aware scan honest: it says how many donors carry each +allele, so a per-allele result can be weighted by how much of the cohort it actually covers. diff --git a/tcga/hla_coverage.tsv.gz b/tcga/hla_coverage.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..994fe1c18dda3dd1bf18d18af6a2ec06fce7f3b7 --- /dev/null +++ b/tcga/hla_coverage.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:684714f4c987fb7fd9d72b7f9c81ed96bbbd3a5ff70bc90ca1a14d01f7b0bf32 +size 2951 diff --git a/tcga/tcga_neoantigens.tsv.gz b/tcga/tcga_neoantigens.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..8dfc7885c6629704b8540304c104bea7af42f1a7 --- /dev/null +++ b/tcga/tcga_neoantigens.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b60780d1ef1769a46ba7f99820bbb4c1655cfae415bd2e221eefc20377106df9 +size 30010965 diff --git a/thymus/SOURCES.md b/thymus/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..71c8fbdd793b64478dcd871ee4a178fcba8484b4 --- /dev/null +++ b/thymus/SOURCES.md @@ -0,0 +1,259 @@ +# `thymus/` — sources + +The central-tolerance reference: what a developing T-cell repertoire was negatively selected +against. All three files are **experimental** (mass spectrometry / antibody-based expression); the +harmonisation into this schema, and every FDR cutoff, is computed. + +Human and mouse are separate files rather than one two-species table, matching the `_mmu` / `_hsa` +convention in `neoantigens/`. Every consumer already filters on `mhc_species`, and the human deposit +is an LFS object all of them read. + +## `thymus_immunopeptidome.tsv.gz` — 53,878 rows + +Thymus-eluted self peptides, **HLA Ligand Atlas** (`dataset_origin = hla_ligand_atlas_thymus`). + +| `mhc_class` | rows | +|---|--:| +| MHCII | 27,987 | +| MHCI | 25,891 | + +Schema: `peptide`, `mhc_a`, `mhc_class`, `mhc_species`, `source_protein`, `species`, +`dataset_origin`. Human only. `mhc_a` is empty throughout — the atlas gives donor-level candidate +alleles, not a per-peptide deconvolution. + +**A hit here is a tolerance argument and an autoimmunity flag, and those are the same fact read two +ways**: a neoantigen resembling a thymically presented peptide had its reactive clones plausibly +deleted (lower expected immunogenicity), and a vaccine built on it risks cross-reacting with self. +`mhcmatch.mimics` keeps it as its own category for exactly that reason and never sums it with the +foreign ones — see `mhcmatch.mimics.KINDS`. + +## `thymus_immunopeptidome_mmu.tsv.gz` — the mouse counterpart + +Same schema, column for column. `mhc_species = species = MusMusculus` throughout. Added 2026-08-21; +this file is what the `## Not here` block of this document used to say was impossible. + +**8,151 rows / 6,791 distinct peptides** as deposited here. + +| `dataset_origin` | `mhc_class` | `mhc_a` | peptides | +|---|---|---|--:| +| `massive_msv000087031_thymus` | MHCII | `I-Ab` | 1,488 | +| `pride_pxd008733_thymus` | MHCI | `H-2Db` | 1,574 | +| `pride_pxd008733_thymus` | MHCI | `H-2Kb` | 1,089 | +| `pride_pxd008733_thymus` | MHCII | `I-Ab` | 2 | +| `pride_pxd031966_thymus` | MHCI | *(empty)* | 1,694 | +| `pride_pxd042241_tec` | MHCI | *(empty)* | 2,304 | + +1,490 class-II and 5,301 class-I distinct peptides. 148 rows carry an empty `source_protein`. + +MHC-I lengths run 8–11 with a median of 9. The two MHC-II rows are all that survive the 11–25 window +from PXD008733's 16 thymic anti-I-A/I-E PSMs — that immunoprecipitation returned mostly 9–10mers, +which is the class-I-length contamination the Nanaware paper is about. 121 rows carry an empty +`source_protein`; they are PXD042241 peptides from non-canonical ORFs that are not in UP000000589. + +Unlike the human file, `mhc_a` is populated where the deposit records an immunoprecipitating +antibody, in the `pmhc/`+`ligandome/` spelling `H-2Kb` / `H-2Db` / `I-Ab` — **not** the `H2-Kb` form +found in the contaminated CEDAR rows of `neoantigens/neoag_tested.tsv.gz`. It is as-reported +attribution; nothing is predicted, and it is empty where the immunoprecipitation was pan-H-2. + +### `pride_pxd008733_thymus` — whole thymus, H-2Kb and H-2Db + +The thymus slice of the murine MHC-I tissue atlas; the other 18 tissues are in +`ligandome/tissue_self_mmu.tsv.gz`, which carries the full provenance and the FDR derivation. + +> Schuster H, Shao W, Weiss T, Pedrioli PGA, Roth P, Weller M, Campbell DS, Deutsch EW, Moritz RL, +> Planz O, Rammensee HG, Aebersold R, Caron E. +> **A tissue-based draft map of the murine MHC class I immunopeptidome.** +> *Sci Data* 2018;5:180157. +> PMID [30084848](https://pubmed.ncbi.nlm.nih.gov/30084848/) · +> doi:[10.1038/sdata.2018.157](https://doi.org/10.1038/sdata.2018.157) + +Taken from the deposit's own Trans-Proteomic Pipeline PSM tables, not its raw spectra. Peptide-level +FDR recomputed from the deposit's `DECOY_` entries at 1%: `probability_ip` ≥ 0.974142, 35,470 target +against 354 decoy peptides. + +### `pride_pxd042241_tec` — sorted cortical and medullary thymic epithelial cells + +The cells that actually induce central tolerance, which whole thymus dilutes. + +> Larouche JD, Laumont CM, Trofimov A, Vincent K, Hesnard L, Brochu S, Côté C, Humeau JF, +> Bonneil É, Lanoix J, Durette C, Gendron P, Laverdure JP, Richie ER, Lemieux S, Thibault P, +> Perreault C. +> **Transposable elements regulate thymus development and function.** +> *eLife* 2024;12:e91037. +> PMID [38635416](https://pubmed.ncbi.nlm.nih.gov/38635416/) · +> doi:[10.7554/eLife.91037](https://doi.org/10.7554/eLife.91037) + +`mhc_a` is empty: the deposit does not name the immunoprecipitating antibody. + +**Its FDR is not recomputable, and the cutoff is a stated convention.** The mzid declares +`no threshold` at protocol level and carries no decoy entries, and its `passThreshold` attribute is +not a quality filter — it passes PSMs down to −10lgP = 5.00 while rejecting others up to 35.62. +The retained set is rank-1 PSMs at **PEAKS −10lgP ≥ 20**. That cutoff is doing real work: 8–11mers +rise from 67.2% to 82.0% of distinct peptides and 9-mers from 28.0% to 38.1%. It is **not** a +measured 1% FDR and must not be described as one. + +Source proteins in this deposit are Ensembl and transposable-element identifiers from a personalized +mTEC/cTEC database, so peptides derived from non-canonical ORFs carry an empty `source_protein` +after re-mapping. + +### `pride_pxd031966_thymus` — NOD mouse, whole thymus + +The NOD-mouse MHC-I deposit covers thymus *and* pancreas, and **its run labels name neither**: +`NOD2_3`, `NOD4`, `NOD5`. PMID 36210013 is closed-access with no PMC copy, and the deposited tables +give only 30–50 peptides per run, far too few to tell the two apart. This was recorded here as an +open labelling problem, and its peptides were deliberately withheld rather than guessed. + +The re-search settled it. Searching the nine raw files against the whole mouse proteome (rather than +the deposit's own predicted-binder FASTA) gives 1,700–5,800 peptides per run — enough for the source +proteins to speak: + +| run | thymus-exclusive proteins | acinar zymogen peptides | call | +|---|---|--:|---| +| `NOD2_3` | **`Prss16` ×3, `Psmb11` ×1** | 4 | **thymus** | +| `NOD4` | none | 160 | pancreas | +| `NOD5` | none | 145 | pancreas | + +**The call is made on presence, not on a ratio.** `Prss16` is the thymus-specific serine protease and +`Psmb11` the thymoproteasome subunit β5t, confined to cortical thymic epithelium; neither has a +tissue of expression outside thymus in the mouse, so a peptide from either cannot have been eluted +from pancreas. The converse holds for the exocrine zymogen program. A ratio is the wrong instrument +and would have got this wrong — on a 3× rule `NOD2_3` reads "unresolved" at 11 thymic against 4 +pancreatic markers, while carrying three `Prss16` peptides and a `Psmb11` one. + +An earlier, looser marker set inflated the thymic signal in *both* pancreas runs, because it counted +`Krt5`/`Krt8`/`Krt14` — thymic-epithelium markers in a thymus context but generic epithelial keratins +elsewhere — and `Lck`/`Zap70`/`Tcf7`, which mark the T-cell infiltrate that is the defining lesion of +the NOD pancreas. Those genes are now explicitly excluded; the rule is in +`src/assign_pxd031966_tissue.py`. + +Only `NOD2_3` is deposited here. `NOD4` and `NOD5` are in `ligandome/tissue_self_mmu.tsv.gz` as +pancreas. + +> Wang L, Li X, Yang S, Chen X, Li J, Wang S, Zhang M, Zheng Z, Zhou J, Wang L, Wu Y. +> **Proteomic identification of MHC class I-associated peptidome derived from non-obese diabetic +> mouse thymus and pancreas.** +> *J Proteomics* 2022;270:104746. +> PMID [36210013](https://pubmed.ncbi.nlm.nih.gov/36210013/) · +> doi:[10.1016/j.jprot.2022.104746](https://doi.org/10.1016/j.jprot.2022.104746) + +`mhc_a` is empty: the immunoprecipitation used the pan-H-2 antibody M1/42.3.9.8, which resolves no +allele. NOD is H-2^g7 (K^d, D^b), so these peptides are **not** C57BL/6 K^b/D^b ligands — read +`dataset_origin` before pooling them with the PXD008733 rows. + +**Re-searched, not taken as deposited.** The authors searched a predicted-MHC-binder FASTA at 5% PSM +FDR; a database of peptides someone already predicted to bind can only return peptides someone +already predicted to bind. This is nf-core/mhcquant 3.2.0 against UP000000589 at 1% peptide-level +FDR. + +### `massive_msv000087031_thymus` — I-Ab, the MHC class II arm + +**1,488 distinct peptides — the only mouse class-II thymic ligands in the compendium.** No +identifications were deposited, so this is a re-search of the raw spectra with nf-core/mhcquant +3.2.0: unspecific cleavage, Comet, MS²Rescore (DeepLC + MS²PIP `Immuno-HCD`), Percolator, 1% +peptide-level FDR, against UP000000589 plus contaminants. + +**Checked as class-II ligands rather than assumed.** Median length **16 aa** (mean 16.0, range +11–25), matching the ~16 aa thymic peak the source paper reports. **49% of the peptides (733 of +1,488) share a nested core with another peptide in the set** — ragged N- and C-termini around a +common binding core is the defining signature of MHC class-II processing and does not arise for +class I, where the termini are trimmed to the closed groove. + +**The deposit's 11–25 window removes a peak the paper itself flags.** That paper is titled for this +problem: the thymic I-Ab length distribution is bimodal, with a second mode near 10 aa that is +class-I-length material co-purifying with a limiting-abundance class-II preparation. The window +applied here (`mhcmatch.mimics._LEN["mhc2"]`) drops it by construction — 1,303 rows fell outside the +per-class windows across all arms. That is a filter, not a measurement, and it is why the +distribution here is unimodal where the paper's is not. + +> Nanaware PP, Jurewicz MM, Clement CC, Lu L, Santambrogio L, Stern LJ. +> **Distinguishing Signal From Noise in Immunopeptidome Studies of Limiting-Abundance Biological +> Samples: Peptides Presented by I-A^b in C57BL/6 Mouse Thymus.** +> *Front Immunol* 2021;12:658601. +> PMID [33995376](https://pubmed.ncbi.nlm.nih.gov/33995376/) · +> doi:[10.3389/fimmu.2021.658601](https://doi.org/10.3389/fimmu.2021.658601) + +Twelve of the deposit's 37 runs — the four thymus samples (`120116WTII`, `020617WT_MHCII`, +`061417MHCII_WT1`, `061417MHCII_WT2`) in three technical replicates each — of which **eleven were +usable**. `061417MHCII_WT1_1.raw` is unreadable: it carries the size MassIVE declares +(757,890,890 bytes) and a valid Thermo `01a1` "Finnigan" header, but ThermoRawFileParser fails on it +with `RAW file cannot be processed`, so the damage is mid-file. The other eleven convert cleanly. +It costs one of three *technical* replicates of Thymus 3 — not a sample and not a condition. The splenic B-cell and +dendritic-cell arms and the human LCL runs are deliberately **not** included: this is a thymic +reference, and pooling peripheral APCs into it would destroy the distinction the file exists to make. + +## `thymus_expression_mmu.tsv.gz` — 15,306 rows + +The mouse side of the table below, added 2026-08-21. Schema `gene`, `species`, `evidence`, `value`; +`species = MusMusculus` throughout, and unlike the human file `value` is populated — it carries the +measurement, so a caller can threshold rather than take mere presence as evidence. + +| `evidence` | rows | what it is | +|---|--:|---| +| `fantom5_mouse_thymus_tpm` | 11,862 | adult-thymus CAGE median TPM > 0 (E-MTAB-3579) | +| `geiger_silac_thymus_ppb` | 3,444 | thymus SILAC protein abundance > 0 (E-PROT-11) | + +12,646 distinct genes. The two lines of evidence are kept in a column rather than pooled, exactly as +in the human file: one says the gene is transcribed in thymus, the other that its protein was +detected there. Full provenance, citations and caveats — including why `Aire` reads zero — are in +`expression/SOURCES.md`. + +## `thymus_expression.tsv.gz` — 26,772 rows + +Thymically expressed self, by gene. Schema: `gene`, `species`, `evidence`, `value`. + +| `evidence` | rows | what it is | +|---|--:|---| +| `hpa_thymus_ntpm` | 16,279 | **HPA thymus nTPM > 0 — the measured value**, added 2026-08-21 | +| `hla_ligand_atlas_thymus_source` | 10,367 | source proteins of the peptides above | +| `hpa_thymus_cluster` | 126 | Human Protein Atlas thymus-cluster genes | + +26,772 rows. The evidence lines are kept in a column rather than pooled, because they are three +different statements: a gene is *transcribed* in thymus, a peptide from its protein was *eluted* +from thymus, and it is thymus-*enriched* by cluster assignment. They support each other; they are +not interchangeable. + +**Why the `hpa_thymus_ntpm` line was added.** Before it, this table was not quantitative: 10,367 of +its 10,493 rows carried an **empty `value`** — they record which proteins peptides came from, which +is not an expression measurement — and the only rows with content held a text cluster label. The +mouse table had numbers and the human one did not. Note also that **GTEx has no thymus at all**, so +the compendium's main safety read cannot answer thymic questions; `expression/SOURCES.md` records +what was searched and why HPA is the source that works. + +## Caveats on the mouse file + +- **It is not the human file's equal in size, and cannot be.** The human deposit is one large + multi-donor atlas; the mouse side is assembled from three smaller studies with different + antibodies, instruments and search pipelines. Read the `dataset_origin` column before pooling. +- **Three different FDR regimes are present**, one per `dataset_origin` — a recomputed 1% + peptide-level FDR (PXD008733), a stated PEAKS score convention (PXD042241), and mhcquant's 1% + peptide-level FDR (MSV000087031). They are not interchangeable; the column is the handle. +- `source_protein` is **re-derived** by exact lookup against `proteome/mouse.fasta.gz`, not taken + from the deposits, because the three name their proteins in three incompatible ways. I and L are + not collapsed — MS cannot distinguish them, so an I/L-collapsed match would assert an + identification the data does not carry. +- Lengths are clipped to `mhcmatch.mimics._LEN` (MHC-I 8–11, MHC-II 11–25) because every consumer + drops the rest silently. + +## Not here + +**The mouse thymus expression gap is closed** — see `thymus_expression_mmu.tsv.gz` above. What +is still absent is a *promiscuous-expression* (AIRE-driven TRA) table: the Sansom mTEC-TRA gene list +(GSE53111) is journal-supplement-only and the mouse thymus proteome (PXD007288) is a 12 GB MaxQuant +archive. Bulk data cannot substitute — `Aire` itself reads 0 TPM in whole thymus because it is +confined to a rare mTEC subset, so the promiscuous-expression question needs sorted mTEC, not this +file. **Absent, not overlooked.** + +**PXD031966 is now here, and the tissue labels were settled by measurement.** See the +`pride_pxd031966_thymus` section above. + +## Re-fetch / regenerate + + cd ~/vcs/projects/public_expression_peptidomes + # results-only deposits (671 MB); verify every file against PRIDE's SHA-1 `checksum` field + aria2c -i src/urls_results.txt -d ~/hf/pmhc_data/raw/thymus_mouse -x4 -j3 --continue + ./src/ingest_pxd008733.py + ./src/ingest_pxd042241.py + # MSV000087031 raw spectra + re-search, on aldan3 + aldan3 slurm submit src/fetch_raw.sbatch -- + aldan3 slurm submit src/mhcquant.sbatch -- mhc2 /projects/hla_epitope/mouse_thymus + ./src/build_thymus_mmu.py diff --git a/thymus/thymus_expression.tsv.gz b/thymus/thymus_expression.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..c46e6c75266c90146ae568c0df51ee6ca00becca --- /dev/null +++ b/thymus/thymus_expression.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4de6b82d1f57859e4059249eae201246e1b227bab241b85e205b21f919cef10f +size 114898 diff --git a/thymus/thymus_expression_mmu.tsv.gz b/thymus/thymus_expression_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..2436c38f1c49239f5fe2ee00e7f1c9941669327b --- /dev/null +++ b/thymus/thymus_expression_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6da4508ec4d61851a29a02b3eaa043429a5c63c3be739b2e7662337d36bbdb75 +size 95367 diff --git a/thymus/thymus_immunopeptidome.tsv.gz b/thymus/thymus_immunopeptidome.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..1021d859cf74988046d3006b732de781598c20e8 --- /dev/null +++ b/thymus/thymus_immunopeptidome.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e12f5dd5ef5605eccfa8d6477739a063be822ebae0a57ce649c2218ab06868a1 +size 763343 diff --git a/thymus/thymus_immunopeptidome_mmu.tsv.gz b/thymus/thymus_immunopeptidome_mmu.tsv.gz new file mode 100644 index 0000000000000000000000000000000000000000..ec8c7ba39750aa122f2fc22b45035f75466a68c3 --- /dev/null +++ b/thymus/thymus_immunopeptidome_mmu.tsv.gz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:37c4c072e60267d9e05b97402d7a2dd935920ed18c417d96400881a75499acb7 +size 123386 diff --git a/vaccines/SOURCES.md b/vaccines/SOURCES.md new file mode 100644 index 0000000000000000000000000000000000000000..ea9f923ffd800241ce121cb2b263e5c358ab1642 --- /dev/null +++ b/vaccines/SOURCES.md @@ -0,0 +1,175 @@ +# `vaccines/` — sources, provenance and audit + +**Shipped neoantigen-vaccine cassettes**: which units a trial actually manufactured and injected into +which patient, which peptides it then monitored, and — where a trial published it per unit — which +units drew a response. A *cassette* is a selection, so its unit of observation is neither a peptide +(`immunogenicity/`) nor a screened mutation (`neoantigens/`), and it gets its own directory. + +Added 2026-08-24 by `bench/neoag/ingest_vaccines.py` in `2026-mhcmatch-benchmark`. Sources are the +published supplements staged at `$VACCINE_DIR` (`~/hf/mrna_vaccines_dump/SOURCES.md`); the source +PDFs and workbooks are **not** redistributed, only these processed tables. + +Citations resolved against **PubMed**; PMIDs and DOIs given. + +**Parquet rather than gzip-TSV**, unlike the rest of this deposit. `gzip` writes an mtime into its +header, so an unchanged re-ingest produces byte-different content and a spurious LFS diff every run +(measured on `neoantigens/sahin_tnbc_2025.tsv.gz`: identical 4,697 bytes, different OID). Parquet has +no such field — verified byte-identical across two runs. + +--- + +## Complete audit + +| file | rows | distinct peptides | patients | trials | length | +|---|--:|--:|--:|--:|:--:| +| `cassettes.parquet` | 1,506 | 1,496 | 81 | 2 | 14–30 | +| `assayed_units.parquet` | 5,427 | 5,410 | 47 | 2 | 8–17 | +| `unit_responses.parquet` | 19 | 19 | 4 | 1 | 14–17 | +| `ivac_mutanome_units.parquet` | 125 | 125 | 13 | 1 | 14–27 | + +| trial | cassettes | units | cassette size | assayed pools | `EPT` | `ASP` | +|---|--:|--:|:--:|--:|--:|--:| +| `NEO-PV-01-NT001` | 60 | 1,005 | 9–20 | 34 | 2,195 | 1,961 | +| `NEO-PV-01-NT002` | 21 | 501 | 11–30 | 13 | 400 | 871 | +| `IVAC-MUTANOME` | 13 | 125 | 5–10 | 13 | — | — | + +Cassettes outnumber assayed pools because both trials published every patient's vaccine composition +but monitored a subset. NT-002's 21 published cassettes are of 38 patients treated; NT-001's 60 are +of 82 analysed. + +--- + +## Schema + +### `cassettes.parquet` — what was manufactured + +| column | provenance | notes | +|---|---|---| +| `trial` | **derived** | `NEO-PV-01-NT001` or `NEO-PV-01-NT002` | +| `patient` | **experimental** | the trial's own alias: `M1`/`L1`/`B1` in NT-001, `2L1`/`2UV1` in NT-002 | +| `cancer` | **experimental** | Melanoma / Lung / Bladder; null for NT-002, which is non-squamous NSCLC throughout | +| `unit_id` | **experimental** | `IM`, the immunogen. Null for NT-001, whose `Vaccine peptides` sheet carries no identifier | +| `peptide` | **experimental** | the synthetic long peptide as manufactured | +| `peptide_len` | **derived** | residues | +| `construct` | **derived** | `SLP` — both trials are synthetic long peptide | + +### `assayed_units.parquet` — what was monitored + +| column | provenance | notes | +|---|---|---| +| `peptide_id` | **experimental** | `IM.
`; a few name two immunogens (`IM01;IM17.EPT01299`), kept verbatim | +| `unit_id` | **derived** | the part before `.` — joins a monitored peptide to its cassette unit | +| `panel` | **derived** | `EPT` or `ASP`, read off `peptide_id` | +| `peptide`, `peptide_len` | **experimental** / **derived** | | + +**`panel` is the class boundary, and it is a fact of the data rather than an assumption**: `EPT` +peptides are 8–12 residues, `ASP` are 14–17, with no overlap. The ingest asserts it. `EPT` is the +class-I substrate — 2,595 across both trials — and the only material available for inferring HLA, +because **neither trial published per-patient genotypes**. + +### `unit_responses.parquet` — what responded + +NT-002 Table S3 of `mmc1.pdf`, parsed with `pdftotext -layout`. **19 vaccine units** whose induced +T-cell response was tested against the wild-type epitope: `mutant_specific` = 1 for the 18 that +responded to the mutant and not the wild type, `cross_reactive_wt` = 1 for the single unit that +responded to both. `peptide_wt` is the matched wild-type sequence. Every `peptide_id` here joins to +`assayed_units.parquet`; the ingest asserts that too. + +**This is the whole per-unit labelling that exists.** NT-002 reports an average 55 % of vaccine +epitopes generating any T-cell response across its 13 assessed patients (39 % CD4⁺, 31 % CD8⁺), but +as Figure 4B values rather than a table, so 501 units cannot be labelled and 19 can. + +### `ivac_mutanome_units.parquet` — what was manufactured **and** what responded + +The only deposited table carrying a **measured response label on every manufactured unit**. IVAC +MUTANOME assayed all ten units of each patient's cassette after vaccination, so its 50 non-responding +units are measured negatives rather than units nobody looked at. + +| column | provenance | notes | +|---|---|---| +| `trial` | **derived** | `IVAC-MUTANOME` | +| `patient` | **experimental** | `P01`…`P19`, the trial's own alias | +| `unit_id` | **experimental** | position on the pentatope, `A1`–`B5` | +| `mutation_id` | **experimental** | `P04-C7-E258K` — patient, gene, substitution | +| `gene`, `aa_change` | **experimental** | | +| `peptide`, `peptide_len` | **experimental** / **derived** | the 27-mer as manufactured; 11 shorter, at a protein terminus | +| `construct` | **derived** | `RNA-pentatope` — five units per RNA, two RNAs per patient | +| `exon_expression` | **experimental** | the trial's own exon expression level | +| `hla_i_score`, `hla_ii_score` | **experimental** | the trial's own binding scores, computed with each patient's real genotype; smaller is better | +| `pre_existing` | **experimental** | a response detectable before vaccination | +| `responded` | **experimental** | any post-vaccination T-cell response — **75 of 125** | +| `resp_cd4`, `resp_cd8` | **experimental** | 60 and 31; a unit can raise both (19 did). Null on the one responding unit whose class the trial reports as `n.k.`, never a silent zero | + +**The mutation sits at position 14 of a full-length 27-mer**, verified against `aa_change` on every +row by the ingest. Anything scoring this table by window should span that residue: most 8–11mers of a +27-mer contain no mutation at all and describe the germline protein. + +**No IVAC patient is in `neoantigens/neoantigens_tested_peptides.tsv.gz`**, and only 2 of the 125 +units share even an 8–11mer with it, so this table is held out of the EPIC fit by patient and very +nearly by sequence. That is the opposite of `nci_parkhurst_gi.parquet`, which needs `in_epic_fit`. + +--- + +## Citations + +**NT-001.** Ott PA, Hu-Lieskovan S, Chmielowski B, Govindan R, Naing A, Bhardwaj N, Margolin K, +Awad MM, Hellmann MD, Lin JJ, …, Srinivasan L. "A Phase Ib Trial of Personalized Neoantigen Therapy +Plus Anti-PD-1 in Patients with Advanced Melanoma, Non-small Cell Lung Cancer, or Bladder Cancer." +*Cell* 2020;**183**(2):347-362.e24. +doi:[10.1016/j.cell.2020.08.053](https://doi.org/10.1016/j.cell.2020.08.053) · +PMID [33064988](https://pubmed.ncbi.nlm.nih.gov/33064988/) · NCT02897765. + +**NT-002.** Awad MM, Govindan R, Balogh KN, Spigel DR, Garon EB, Bushway ME, Poran A, Sheen JH, +Kohler V, Esaulova E, …, Srinivasan L. "Personalized neoantigen vaccine NEO-PV-01 with chemotherapy +and anti-PD-1 as first-line treatment for non-squamous non-small cell lung cancer." +*Cancer Cell* 2022;**40**(9):1010-1026.e11. +doi:[10.1016/j.ccell.2022.08.003](https://doi.org/10.1016/j.ccell.2022.08.003) · +PMID [36027916](https://pubmed.ncbi.nlm.nih.gov/36027916/). + +**IVAC MUTANOME.** Sahin U, Derhovanessian E, Miller M, Kloke BP, Simon P, Löwer M, Bukur V, +Tadmor AD, Luxemburger U, Schrörs B, …, Türeci Ö. "Personalized RNA mutanome vaccines mobilize +poly-specific therapeutic immunity against cancer." *Nature* 2017;**547**(7662):222-226. +doi:[10.1038/nature23003](https://doi.org/10.1038/nature23003) · +PMID [28678784](https://pubmed.ncbi.nlm.nih.gov/28678784/). Supplementary Table 2; record retrieved +from PubMed. + +--- + +## Companion tables deposited in the same pass + +| file | rows × cols | what | +|---|--:|---| +| `../neoantigens/nci_parkhurst_gi.parquet` | 7,384 × 48 | the NCI gastrointestinal TIL screen, **with `in_epic_fit`** | +| `../neoantigens/neoprecis_ici_cohort.parquet` | 525 × 50 | per-patient checkpoint-blockade table | +| `../neoantigens/neoprecis_ranker_panel.parquet` | 4,263 × 22 | peptides scored by four published rankers, with labels | +| `../neoantigens/neoprecis_allele_benefit.parquet` | 547 × 4 | per-allele benefit scores | +| `../immunogenicity/neoprecis_anchor_triplets.parquet` | 14,140 × 12 | anchor / positive / negative triplets by allele | + +### `in_epic_fit` — read this before scoring `nci_parkhurst_gi` + +`neoantigens/neoantigens_tested_peptides.tsv.gz` is part of the corpus the shipped EPIC scorer was +fitted on, and it keys on the same numeric NCI patient ids. **29 of the 75 patients appear in both**, +and 7,865 of 8,000 sampled corpus peptides on those patients occur inside a `mt_seq` 25-mer from this +table — the same screened mutations, tiled into 8–12mers. Any figure pooled over all 75 patients is +part training-set performance. + +| half | patients | mutations | CD8⁺ | CD4⁺ | +|---|--:|--:|--:|--:| +| `in_epic_fit = false` — held out | 46 | 4,019 | 15 | 45 | +| `in_epic_fit = true` | 29 | 3,365 | 42 | 21 | + +The CD8 positives sit mostly in the fitted half and the CD4 positives mostly in the held-out half, so +which class a result is about decides how much of it is honest. The flag is a column rather than a +note in a script for that reason. + +Parkhurst MR, Robbins PF, Tran E, Prickett TD, Gartner JJ, Jia L, Ivey G, Li YF, El-Gamil M, +Lalani A, …, Rosenberg SA. "Unique Neoantigens Arise from Somatic Mutations in Patients with +Gastrointestinal Cancers." *Cancer Discov* 2019;**9**(8):1022-1035. +doi:[10.1158/2159-8290.CD-18-1494](https://doi.org/10.1158/2159-8290.CD-18-1494) · +PMID [31164343](https://pubmed.ncbi.nlm.nih.gov/31164343/) · PMC7138461. + +Lee KH, Sears TJ, Zanetti M, Carter H. "NeoPrecis: enhancing immunotherapy response prediction +through integration of qualified immunogenicity and clonality-aware neoantigen landscapes." +*Nat Commun* 2026;**17**(1). +doi:[10.1038/s41467-026-68651-6](https://doi.org/10.1038/s41467-026-68651-6) · +PMID [41577704](https://pubmed.ncbi.nlm.nih.gov/41577704/) · PMC12932759. diff --git a/vaccines/assayed_units.parquet b/vaccines/assayed_units.parquet new file mode 100644 index 0000000000000000000000000000000000000000..80f2f0f6acbf2cdd073fac1039bb69e1e81f4deb --- /dev/null +++ b/vaccines/assayed_units.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:739b9a9b9ed9c16e98d417b3549eb976ca928dfbb284fb136c0048908825faff +size 56148 diff --git a/vaccines/cassettes.parquet b/vaccines/cassettes.parquet new file mode 100644 index 0000000000000000000000000000000000000000..6c95477a516ce28eb8e23023321c96b2cea27d5c --- /dev/null +++ b/vaccines/cassettes.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1332d1e60c454b73d64ac0f671a8204d25dbb9537c474db980f4777d8c7ad8ba +size 22746 diff --git a/vaccines/ivac_mutanome_units.parquet b/vaccines/ivac_mutanome_units.parquet new file mode 100644 index 0000000000000000000000000000000000000000..185f1c3a89a864807916e8f2eefecc78a6672b23 --- /dev/null +++ b/vaccines/ivac_mutanome_units.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:df7923bcf06f10a3b2aff13c81d4cb450a011cf97f59763339e0fa62836b54e0 +size 12458 diff --git a/vaccines/unit_responses.parquet b/vaccines/unit_responses.parquet new file mode 100644 index 0000000000000000000000000000000000000000..2a8d8cd7b0d0ff54d844c39059523387444aafcc --- /dev/null +++ b/vaccines/unit_responses.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1c8a85916438b287f6da2fd9168178d46aa98e667a7fd4ee09b3d513e360c5a4 +size 3912