Buckets:
| Name | Size | Uploaded | Xet hash |
|---|---|---|---|
| Arabidopsis_lyrata | 1 items | ||
| Arabidopsis_thaliana | 16 items | ||
| Sorghum_bicolor | 15 items | ||
| README.md | 12.6 kB xet | f5da0be9 |
Curated sample sheets
Sample sheets for the genesis-preprocessing
workflows, as <Species>/<modality>-<source>.tsv, each with its workflow's exact header. Read URLs
point to ENA over HTTPS, or to GSA (China) for CRA studies.
*-curated.tsvsheets gather every public run of a modality that passes the rules below. Each has a*.runs-qc.tsvbeside it: every candidate run, its classes and measurements, and why it was left out.- Single-study sheets (
<modality>-<study>.tsv) are the repository's original sheets, kept unchanged.
| Sheet | Workflow | Rows | Source |
|---|---|---|---|
Arabidopsis_thaliana/rna-seq-curated.tsv |
rna-seq | 14,584 runs (8,267 PE), 1,956 studies | Curated (below) |
Arabidopsis_thaliana/atac-seq-curated.tsv |
atac-seq | 207 runs (PE), 47 studies | Curated (below) |
Arabidopsis_thaliana/scatac-reads-curated.tsv |
scatac-reads | 18 libraries, 9 studies | Curated (below) |
Arabidopsis_thaliana/tss-seq-curated.tsv |
tss-seq | 61 runs, 17 studies | Curated (below) |
Arabidopsis_thaliana/dap-seq-curated.tsv |
dap-seq | 101 treatments and 19 controls, 11 studies | Curated (below): studies beyond GSE60141 and multiDAP |
Arabidopsis_thaliana/atac-seq-fable.tsv |
atac-seq | 439 runs (PE) | Earlier sheet from the "Bulk ATAC Fable Curated" tab, before read checks; superseded by atac-seq-curated.tsv |
Arabidopsis_thaliana/dap-seq-GSE60141.tsv |
dap-seq | 936 (SE) | O'Malley 2016 |
Arabidopsis_thaliana/dap-seq-PRJNA1177481.tsv |
dap-seq | 800 (PE) | multiDAP |
Arabidopsis_thaliana/rna-seq-GSE206323.tsv |
rna-seq | 16 (PE) | Col-0 inflorescence stem, 8 wild-type samples |
Arabidopsis_thaliana/scatac-reads-GSE155304.tsv |
scatac-reads | 2 | Farmer 2021, root nuclei |
Arabidopsis_thaliana/tss-seq-GSE136356.tsv |
tss-seq | 9 (SE CAGE) | Thieffry 2020: wild type, hen2-4 and rrp4-2 |
Arabidopsis_lyrata/dap-seq-PRJNA1177479.tsv |
dap-seq | 405 (PE) | multiDAP |
Sorghum_bicolor/rna-seq-curated.tsv |
rna-seq | 701 runs (542 PE), 235 studies | Curated (below) |
Sorghum_bicolor/atac-seq-curated.tsv |
atac-seq | 9 runs (PE), 3 studies | Curated (below) |
Sorghum_bicolor/scatac-reads-curated.tsv |
scatac-reads | 6 libraries | Curated (below): SRP474955 at 0 h |
Sorghum_bicolor/dap-seq-curated.tsv |
dap-seq | 9 treatments and 6 controls, 3 studies | Curated (below): studies beyond multiDAP |
Sorghum_bicolor/atac-seq-fable.tsv |
atac-seq | 24 runs (PE) | Earlier sheet from the Fable tab; superseded |
Sorghum_bicolor/dap-seq-PRJNA1177471.tsv |
dap-seq | 142 (PE) | multiDAP, RTx430 |
Sorghum_bicolor/dap-seq-PRJNA1177471-test.tsv |
dap-seq | 6 (PE) | Subset for a quick test |
Sorghum_bicolor/rna-seq-GSE184725.tsv |
rna-seq | 6 (SE) | BTx623 shoot, control and sulfur-deficient |
Sorghum_bicolor/scatac-reads-SRP474955.tsv |
scatac-reads | 11 (multiome) | Swift 2024, BTx623 de-etiolation, 0 h and 12 h |
Sorghum_bicolor/tss-seq-ERP126061.tsv |
tss-seq | 4 (RAMPAGE) | RTx436 and Tx2783 |
Sorghum_bicolor/tss-seq-SRP360924.tsv |
tss-seq | 4 (PE Smar2C2) | Murray 2022, shoot and root |
References are Phytozome 447 (TAIR10, Araport11) and 730 (Sorghum v5.1, with RefSeq Mt/Pt), as
described in the repository's samplesheets/README.md.
Rules for the curated sheets
All curated sheets keep runs with:
- the reference genotype (Col-0, or BTx623 for Sorghum), from each sample's attributes and title;
- no treatment, or none stated.
Sample titles. Submitters often give the background (Col-0) as every sample's genotype and name the mutant or the treatment only in the sample title. Within a study, when some runs' titles mark the wild type (or the control) and others do not, the others are left out, as are runs whose own title names a mutant, transgene or treatment. In a check of 50 such A. thaliana RNA-seq runs, about 45 were clearly mutants, transgenics or other accessions; the rest were unclear.
Curated RNA-seq
Bulk short-read RNA-seq runs from ENA, from any tissue: leaf, root, shoot/apex, seedling, flower,
seed, stem, pollen, whole plant, cell culture, or not-stated. Rows are named
<study>_<tissue>_<sample>_<run>; sum a sample's runs for sample-level counts. Single-end runs
that ENA serves as _1/_2 files are kept as paired.
Strandedness is measured for each run, because the metadata rarely states it:
- Measurement: the first 20,000 read-1 reads are aligned with STAR, and the aligned reads are counted on each gene strand with featureCounts.
- Call, from the forward fraction: above 0.8 is
forward, below 0.2reverse, and 0.35–0.65unstranded. - Checks: the pipeline's
forward_fractionagrees on 20 test runs, and calls from 20,000 and 100,000 reads agree for all 1,830 runs measured both ways.
Runs are left out when:
- strandedness is unclear, or too few reads land on genes;
- the assay is not standard RNA-seq (DeLTa-Seq, polysome or translatome, degradome, CAGE, SHAPE/DMS, small RNA, RIP, m6A or other IP, ChIP, Smar2C2 TSS libraries, single-cell, nuclear or chromatin fractions, TRAP, PAT-seq, 3′-end (QuantSeq) and TIF-seq, from the run's own titles or library selection);
- the run duplicates another submission (same reads);
- the run was submitted as paired but ENA serves one file holding both mates;
- a FASTQ is under 1 MB;
- the sample title suggests a mutant or a treatment (above).
Runs with a low gene fraction (rRNA- or organelle-rich) are kept, so filter on it if needed.
Curated bulk ATAC
Candidates are ENA runs with library strategy ATAC-seq plus the runs in the Genesis metadata
sheet's "Bulk ATAC Fable Curated" tab, including GSA (China) runs. Every A. thaliana and Sorghum
study in the sheet's "ATAC" tab is among them. Any tissue is kept. Rows are named
<study>_<tissue>_<run>.
Runs are left out when the metadata show:
- a single-end, single-cell (10x, sci-ATAC) or long-read library;
- a control or another assay in the title (input or genomic DNA, ChIP, MNase, DNase, capture ATAC);
- reads that are not public as a pair.
Reads. Every candidate's first 200,000 read pairs are mapped with chromap (--preset atac),
giving:
- TSS enrichment: as in the pipeline: Tn5 cut sites within ±2 kb of each gene's TSS, the central 100 bp over the outer 100 bp at each end;
- the share of fragments on the organelles, the share of nucleosome-free fragments, and the median fragment length.
Genomic DNA and input controls score 0.97–1.08, so runs below 1.5 are left out as showing no
accessibility signal. Plant ATAC scores are modest by this definition: the median kept run scores
2.4 (A. thaliana) and 2.5 (Sorghum). On 9 test runs the pipeline's TSS enrichment agrees with this
check (differences of 0.03–0.18, and 0.7 for one Sorghum run). Runs with many organelle reads
(often over half, without nuclei purification) are kept unless above 95%, so filter on
organelle_fraction if needed.
Curated scATAC (from reads)
Libraries in the sheet's "scATAC Fable Curated" tab, one row per 10x library with its runs grouped (runs of one library share cell barcodes; separate libraries share under 1%).
- Chemistry comes from the read structure: a 16-base barcode read is
10x-atac-v1(v1 and v2 are handled the same) with the737K-cratac-v1whitelist, and a 24-base one is10x-multiome-atac. Barcodes match the whitelist for 72–97% of reads in the libraries tested. - Left out: sci-ATAC and other libraries without a 10x barcode read, BAM-only duplicates of a FASTQ submission, runs with no reads, and treated libraries (pathogen, ethylene, 12 h of light).
- Sorghum SRP474955: the genotype is not in SRA; the paper's Methods give BTx623.
- Mixed instruments: two SRP699388 libraries (root and shoot rep1) interleave MiSeq and NovaSeq reads whose barcodes read in opposite orientations. The workflow sets one orientation per library from its first reads, so it loses the NovaSeq reads; these two are left out. The other two SRP699388 libraries start with NovaSeq reads and are kept.
Curated TSS mapping
A. thaliana runs from the sheet's "TSS Fable Curated" and "TSS" tabs plus an ENA search, with
methods the workflow supports: CAGE (23 runs), nAnT-iCAGE (9), TSS-seq (14) and csRNA-seq (15).
Rows are named <study>_<tissue>_<run>; for paired libraries the sheet uses read 1. Sorghum has
nothing beyond the two single-study sheets.
- Trim and adapter are set per study from the reads: 20,000 reads mapped at several 5′ trims,
keeping the trim that puts 5′ ends on annotated starts (for example, trim 9 for reads starting
CTTCAGCAG, trim 2 for oligo-capping's GG).-means no trim. - Read check: the first 50,000 reads must map uniquely at 0.2 or more, with at least 0.3 of 5′ ends within 100 bp of an annotated TSS (kept runs: 0.39–0.66; RNA-seq from the same studies: under 0.03).
- Left out: nanoPARE, NAD-RNA capture, TIF-seq, PEAT and 5′GRO-seq (other assays); nanoCAGE runs that fail the read check; RNA-seq runs inside TSS studies; cap-less controls, polysome fractions and duplicate submissions.
- Kept, worth knowing: T87 cell-culture runs (ecotype Col-0), and dark-grown wild type as the untreated sample of light-response studies.
Curated DAP-seq
Public DAP-seq studies (ENA search) beyond the single-study sheets above: 11 A. thaliana studies
and 3 Sorghum studies (BES1 and MSD1 on BTx623 DNA; awn1). Each treatment points to its own
study's control (input, empty vector, GFP/GUS or HaloTag only). Rows are named
<study>_<label>_<genotype>_<run>.
- Left out: methyl-CpG readers (MBD5/MBD6) and a nematode effector, which are not plant TFs; studies without a control (LFY ampDAP, SEP3-AG seq-DAP, EIN6); DNA from 37 °C-treated plants; libraries under 1M reads; single-partner controls of the bZIP heterodimer study; and the "DAP" matches that mean days after pollination.
- Trimmed uploads: PRJNA1095882, PRJNA551415 and PRJNA1354843 uploaded trimmed reads of varying length, which the workflow rejects, so they are left out.
- Single-end where mates can't be paired: PRJNA1090874's input was uploaded as two runs with
different read names, so the study runs single-end on read 1. A paired control used with
single-end treatments also gets a read-1-only row (
_read1), since MACS3 needs the treatment and control in the same layout. - Checked: a subset covering every study ran end to end (Sorghum 10 rows, A. thaliana 32 rows). Peaks: Sorghum 18–127k, A. thaliana 39–31k. In the weaker Sorghum libraries a TF motif is still strongly enriched in the top peaks (BES1: CACGTG; MSD1: the TCP site). Three libraries have under 100 peaks (SbBES1_2, bHLH115, bZIP25); they stay in the sheet.
Corrections
2026-11-18. Checked against ENA after comparing the sheets with genesis sheets build.
Each changed run's reason_left_out in the runs-qc table gives the reason.
- A. thaliana RNA-seq, 269 runs dropped (14,584 → 14,315 rows):
- 193 duplicate submissions whose original run is also in the sheet. ENA lists the same read and base counts, and the first reads match (mostly PRJNA377558, a resubmission of PRJNA376779).
- 51 PAT-seq runs (poly(A) tail sequencing; PRJNA559744, PRJNA866703, PRJNA552176, PRJNA1064557).
- 2 NanoPARE runs (RNA 5' ends; PRJNA704712).
- 23 runs from lines that express a foreign gene (PRJNA1072118, PRJNA765119).
- Sorghum bulk ATAC, CRR145290 dropped: FRiP 0.066 in a full pipeline run, although the 200k-pair check passed it.
- A. thaliana bulk ATAC, CRR1760445 added: it was left out only because its download had failed. On 200k pairs: 51% mapped, TSS enrichment 3.26 (Col-0 seedling, mock).
- A. thaliana TSS mapping, DRR095710 added: it is not a duplicate of DRR072208. It comes from a different sample (SAMD00083164), instrument and depth (231M vs 125M reads).
2026-11-18, second batch. A. thaliana RNA-seq, 84 more runs dropped (14,315 → 14,231 rows). The BioSample records show they are not wild-type bulk RNA-seq:
- 64 runs whose BioSample
isolatenames a mutant or transgenic line ("T-DNA mutant", "CRISPR mutant", "transgenic plants", "OE line"), while the genotype field gives only the background (Col-0). - 12 mutant runs from studies that also have wild-type samples (pgdh3, PRJNA1099156; serk1, PRJNA430092). The studies' wild-type runs stay in the sheet.
- 8 runs of FANS-sorted stem-cell nuclei (PRJNA929034).
- Total size
- 2 GB
- Files
- 49
- Last updated
- Nov 18
- Pre-warmed CDN
- US EU US EU