| {"id": "pt_001", "accession": "GSE132044", "protocol": "10x_genomics_flex_low_input", "universal_threshold_problem": "Universal threshold: UMIs/cell > 1000. Low-input Flex libraries legitimately\nyield 200-500 UMIs/cell. MultiQC/nf-core flags these as FAIL.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "10x Flex is designed for degraded/low-input samples; low UMI is expected"} |
| {"id": "pt_002", "accession": "GSE150728", "protocol": "smart_seq3_very_low_input", "universal_threshold_problem": "Smart-seq3 on sorted single cells: <1000 reads/cell in some samples.\nStandard threshold: reads/cell > 5000. Flagged as low quality.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Smart-seq3 on very sparse populations \u2014 low reads expected"} |
| {"id": "pt_003", "accession": "GSE176171", "protocol": "10x_visium_ffpe_spatial", "universal_threshold_problem": "FFPE spatial transcriptomics: high technical variation + degradation signal.\nUniversal: genes/spot > 200. FFPE visium may have 50-150 genes/spot.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "FFPE spatial has lower sensitivity; fewer genes/spot is expected"} |
| {"id": "pt_004", "accession": "GSE68086", "protocol": "ffpe_rna_seq", "universal_threshold_problem": "FFPE RNA-seq: high degradation index (DV200 < 50%), high rRNA%, high duplication.\nUniversal thresholds: DV200 > 70%, rRNA < 10%. Almost all FFPE samples fail.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "FFPE degradation is expected; DV200 < 50% is acceptable for FFPE"} |
| {"id": "pt_005", "accession": "GSE200123", "protocol": "ffpe_archival_10yr", "universal_threshold_problem": "10-year-old FFPE block: extreme degradation, high 3'-bias, low unique reads.\nUniversal: 5'/3' ratio < 2. This sample has ratio ~8. Universal: FAIL.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Very old FFPE \u2014 data is still scientifically valid for selected genes"} |
| {"id": "pt_006", "accession": "GSE158722", "protocol": "ffpe_with_fresh_frozen_comparison", "universal_threshold_problem": "Dataset has paired FFPE and fresh-frozen from same tumor.\nWithout protocol awareness, FFPE samples flagged FAIL, fresh-frozen PASS.\nNaive comparison would exclude all FFPE samples from DE analysis.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Paired design \u2014 FFPE should be retained with protocol-aware QC"} |
| {"id": "pt_007", "accession": "PXD004732", "protocol": "tio2_phosphoenrichment", "universal_threshold_problem": "TiO2 phosphopeptide enrichment: lower total peptide coverage (enrichment\ndepletes non-phospho peptides). Universal: coverage > 70%. Enriched: ~20-40%.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Phosphoenrichment intentionally reduces overall peptide coverage"} |
| {"id": "pt_008", "accession": "PXD010557", "protocol": "plasma_proteomics_depleted", "universal_threshold_problem": "Serum depletion removes 14 high-abundance proteins (~95% of mass).\nUniversal: dynamic range threshold based on full serum. Post-depletion\nsamples appear to have very low peptide counts \u2192 FAIL.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Depletion reduces apparent peptide count; true dynamic range improved"} |
| {"id": "pt_009", "accession": "PXD017518", "protocol": "ubiquitin_remnant_enrichment", "universal_threshold_problem": "Ub-remnant (diGly) enrichment: only ubiquitinated peptides recovered.\nUniversal peptide coverage threshold: FAIL. This is by design.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "PTM enrichment protocols always have low overall coverage \u2014 expected"} |
| {"id": "pt_010", "accession": "EMDB-16045", "protocol": "cryo_em_membrane_protein_detergent", "universal_threshold_problem": "Detergent-solubilized membrane protein: high background from micelle belt.\nUniversal: particle contrast > 0.15. Membrane protein in micelle: ~0.05-0.10.\nCTF fitting also affected by detergent ring.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Membrane protein signal is inherently lower contrast; expected with detergent"} |
| {"id": "pt_011", "accession": "EMDB-32001", "protocol": "cryo_em_lipid_nanodisc", "universal_threshold_problem": "Nanodisc-reconstituted protein: background from lipid disc.\nUniversal threshold on particle score: fails. Nanodisc particles score lower.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID"} |
| {"id": "pt_012", "accession": "EMDB-14971", "protocol": "cryo_em_small_membrane_protein", "universal_threshold_problem": "Small membrane protein (<100 kDa in nanodisc): very low contrast.\nUniversal particle picking confidence threshold: too strict for small proteins.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Small proteins require relaxed picking thresholds"} |
| {"id": "pt_013", "accession": "ENCSR000BFT", "protocol": "chip_seq_h3k27me3_broad", "universal_threshold_problem": "H3K27me3 broad mark: very diffuse signal, low enrichment vs input.\nUniversal FRiP (fraction reads in peaks) threshold > 0.3. Broad marks: FRiP ~0.05-0.15.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Broad histone marks have lower FRiP by nature; ENCODE allows \u22650.05 for broad"} |
| {"id": "pt_014", "accession": "ENCSR000BJA", "protocol": "chip_seq_low_cell_number", "universal_threshold_problem": "ChIP-seq from <10,000 cells: lower sequencing depth and enrichment.\nUniversal: library complexity (NRF > 0.8). Low-input ChIP: NRF ~0.5-0.7.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID"} |
| {"id": "pt_015", "accession": "SRP047083", "protocol": "metagenomics_16s_clinical_stool", "universal_threshold_problem": "16S from stool: high host DNA contamination (up to 50%).\nUniversal: human mapping < 10%. Clinical stool samples often 30-60% human.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Clinical stool samples routinely have high host contamination \u2014 expected"} |
| {"id": "pt_016", "accession": "SRP162966", "protocol": "metagenomics_low_biomass_csf", "universal_threshold_problem": "CSF (cerebrospinal fluid) metagenomics: extremely low microbial biomass.\nUniversal: reads mapped to microbial DB > 1%. CSF: often <0.01%.\nThe few microbial reads present are diagnostically important.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Low biomass is the expected feature of CSF; even 10 microbial reads matter"} |
| {"id": "pt_017", "accession": "ENCSR911GKO", "protocol": "atac_seq_frozen_tissue", "universal_threshold_problem": "ATAC-seq on frozen tissue (not fresh cells): higher mitochondrial reads,\nlower TSS enrichment score. Universal TSS enrichment > 6; frozen tissue: 3-5.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Frozen tissue ATAC has lower but acceptable TSS enrichment"} |
| {"id": "pt_018", "accession": "GSE193140", "protocol": "atac_seq_neutrophils", "universal_threshold_problem": "Neutrophil ATAC-seq: very condensed chromatin, fewer accessible regions.\nUniversal: number of peaks > 50,000. Neutrophils: 5,000-15,000 peaks.\nThis is biologically expected (neutrophils have limited open chromatin).\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Granulocyte lineage cells are chromatin-condensed; low peak count is expected"} |
| {"id": "pt_019", "accession": "SYNTHETIC_DEFECT_001", "protocol": "rna_seq_standard", "universal_threshold_problem": "None \u2014 this is a genuinely defective sample", "expected_naive_action": "FAIL", "expected_correct_action": "FAIL", "ground_truth": "DEFECTIVE", "notes": "Synthetic defect: rRNA contamination 45% injected into ENCSR000AED.\nBoth universal and protocol-aware thresholds should correctly FAIL this.\nThis validates that BioMate does not over-correct toward PASS.\n"} |
| {"id": "pt_020", "accession": "SYNTHETIC_DEFECT_002", "protocol": "scrna_seq_standard", "universal_threshold_problem": "None \u2014 this is a genuinely defective sample", "expected_naive_action": "FAIL", "expected_correct_action": "FAIL", "ground_truth": "DEFECTIVE", "notes": "Synthetic defect: all cells have identical UMI barcodes (sequencing error).\nBoth thresholds should FAIL \u2014 tests that BioMate doesn't let everything through.\n"} |
| {"id": "pt_021", "accession": "GSE144136", "protocol": "snrna_seq_frozen_brain", "universal_threshold_problem": "Universal threshold: percent mitochondrial reads < 5%. snRNA-seq from frozen\nnuclei legitimately yields 0-1% mito reads (cytoplasm excluded). Tools flag\nlow-mito cells as stressed/dying or as ambient RNA contamination.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Nuclear isolation removes cytoplasmic mRNA including mitochondrial transcripts.\nNear-zero mito% is the expected signature of quality snRNA-seq, not a defect.\n"} |
| {"id": "pt_022", "accession": "GSE174348", "protocol": "snrna_seq_ambient_rna", "universal_threshold_problem": "Universal threshold: ambient RNA contamination (SoupX rho) < 0.10. Frozen\npost-mortem brain tissue often yields rho 0.15-0.25 due to cell lysis during\ndissection. DoubletFinder also flags more cells as doublets in high-rho samples.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Post-mortem tissue has elevated ambient RNA from lysed cells during thawing.\nCorrection (SoupX) is appropriate; blanket removal of all high-rho samples is not.\n"} |
| {"id": "pt_023", "accession": "GSE193481", "protocol": "snrna_seq_kidney_low_umi", "universal_threshold_problem": "Universal threshold: UMIs/nucleus > 1000. Kidney tubular cells have compact\nnuclei with limited cytoplasmic RNA. Proximal tubule nuclei: 300-700 UMIs.\nScanpy doublet detection rate inflated in dense nuclear clusters.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Kidney nuclei are smaller than cortical neurons; lower UMI counts are expected\nand do not indicate failed capture.\n"} |
| {"id": "pt_024", "accession": "GSE221521", "protocol": "snrna_seq_cardiac_muscle", "universal_threshold_problem": "Universal threshold: genes/nucleus > 500. Cardiomyocytes are binucleate and\nlarge; nuclear isolation from cardiac tissue yields 200-450 unique genes/nucleus.\nStandard QC removes most cardiomyocyte nuclei, leaving only non-cardiomyocyte cells.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Cardiomyocyte nuclei consistently fail standard UMI/gene thresholds despite\nrepresenting the majority of cardiac mass; protocol-aware profile retains them.\n"} |
| {"id": "pt_025", "accession": "GSE162631", "protocol": "snrna_seq_high_doublet_rate", "universal_threshold_problem": "Universal doublet threshold: <5% expected doublets. Tissue dissociation of\ndense tissue (e.g., solid tumors, liver) with snRNA-seq commonly yields 8-15%\npredicted doublet rate. Scrublet flags large fractions of real cells.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Doublet detection tools are calibrated for well-dispersed single-cell suspensions;\ndense tissues require tissue-specific expected doublet rates.\n"} |
| {"id": "pt_026", "accession": "GSE164378", "protocol": "cite_seq_low_adt_counts", "universal_threshold_problem": "Universal threshold: ADT counts/cell > 100. Protein capture efficiency in\nCITE-seq varies by antibody clone. CD56 (NCAM) typically yields 20-60 ADT\ncounts/cell in NK cells. Tools flag these as failed ADT libraries.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "ADT capture efficiency is antibody-specific; low counts for low-abundance\nsurface proteins do not indicate library failure.\n"} |
| {"id": "pt_027", "accession": "GSE194122", "protocol": "cite_seq_isotype_background", "universal_threshold_problem": "Universal signal-to-background: ADT signal/isotype > 3. Isotype controls for\nsome lineage markers (e.g., CD45) show near-saturated background in whole blood\nbecause all leukocytes express CD45. Signal/isotype ratio < 2 flagged as FAIL.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Isotype ratio threshold is invalid for universally expressed antigens; separate\nprotein-specific background correction (e.g., DSB normalization) must be applied.\n"} |
| {"id": "pt_028", "accession": "GSE200218", "protocol": "multiome_atac_rna_low_atac", "universal_threshold_problem": "10x Multiome ATAC+RNA: ATAC TSS enrichment threshold > 4. RNA-matched nuclei in\nmultiome libraries show TSS enrichment 2-3 (lower than standalone ATAC-seq)\nbecause fixation required for multiome reduces signal. ATAC QC incorrectly fails.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "10x Multiome requires permeabilization conditions optimized for dual-capture that\ncompromise ATAC signal strength relative to standalone ATAC-seq libraries.\n"} |
| {"id": "pt_029", "accession": "GSE185477", "protocol": "cite_seq_hashing_low_signal", "universal_threshold_problem": "Cell hashing (HTO): HTO signal-to-noise threshold. Cells from primary tissue\n(non-immortalized) absorb fewer hashtag oligonucleotides. HTO counts 30-80\nvs. cell-line baseline >200. Demultiplexing algorithms call many as negatives.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "HTO signal strength varies by cell type; primary tissue cells are particularly\nprone to low HTO signal; negative call rate must be calibrated per cell type.\n"} |
| {"id": "pt_030", "accession": "GSE210347", "protocol": "multiome_low_rna_modality", "universal_threshold_problem": "10x Multiome RNA: genes/nucleus threshold > 500. When ATAC signal is high in\nmultiome, RNA capture can be reduced due to competing lysis conditions. Some\nnuclei show 200-400 RNA genes but excellent ATAC profiles. RNA QC fails them.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Multiome RNA yield is systematically lower than standalone snRNA-seq; cells\nwith strong ATAC profiles but modest RNA should not be filtered on RNA alone.\n"} |
| {"id": "pt_031", "accession": "ENCSR868FGK", "protocol": "bulk_atac_seq_primary_tissue", "universal_threshold_problem": "Bulk ATAC-seq: nucleosome periodicity visible as ~200 bp fragment ladder.\nThymus primary tissue: condensed chromatin shows weaker periodicity vs.\ncell lines. Fragment size distribution QC (sub-200 bp fraction < 50%) fails.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Primary tissue chromatin is more compacted than cultured cells; the sub-nucleosomal\nfragment fraction varies by tissue type and does not indicate library failure.\n"} |
| {"id": "pt_032", "accession": "GSE179461", "protocol": "bulk_atac_seq_brain_cortex", "universal_threshold_problem": "Bulk ATAC-seq on brain cortex homogenate: TSS enrichment score 3-4 (threshold > 6).\nHeterogeneous cell types (neurons, glia, microglia) dilute TSS signal compared\nto pure cell populations. ATACseqQC and ArchR flag this as low-quality.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Cell-type heterogeneity in bulk ATAC dilutes TSS enrichment; tissue-specific\nTSS thresholds are required rather than cell-line calibrated values.\n"} |
| {"id": "pt_033", "accession": "ENCSR406FMI", "protocol": "bulk_atac_seq_low_cells", "universal_threshold_problem": "Bulk ATAC-seq with 10,000 cells (low): library complexity (NRF) 0.5-0.65.\nUniversal NRF > 0.8. Low cell input reduces unique molecule diversity leading\nto higher PCR duplication rate. Flagged as poor complexity.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Low-input ATAC-seq has higher duplication by design; NRF threshold should be\nrelaxed for <50,000 cell inputs.\n"} |
| {"id": "pt_034", "accession": "GSE155182", "protocol": "bulk_atac_seq_ffpe_tissue", "universal_threshold_problem": "ATAC-seq on FFPE-adjacent tissue sections: fragmentation is more random, peak\nsignal-to-background ratio reduced to 4-6 (universal threshold > 10).\nFormalin crosslinking reduces transposase efficiency.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "ATAC-seq from FFPE-adjacent or partially-fixed tissue has reduced Tn5 efficiency;\nlower signal-to-background is expected and does not invalidate peak calls.\n"} |
| {"id": "pt_035", "accession": "ENCSR000CAE", "protocol": "chip_seq_h3k9me3_broad", "universal_threshold_problem": "H3K9me3 heterochromatin mark: ultra-broad enrichment domains, FRiP ~0.03-0.08.\nUniversal FRiP > 0.3 flags this as failed. ENCODE narrow-peak calling also\nmisrepresents broad heterochromatin blocks as absent.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Constitutive heterochromatin marks have very low FRiP because signal is spread\nover megabase domains; MACS2 broad mode is required and FRiP threshold must be \u22650.03.\n"} |
| {"id": "pt_036", "accession": "ENCSR244KUT", "protocol": "chip_seq_h3k27me3_low_efficiency", "universal_threshold_problem": "H3K27me3 ChIP with older lot antibody: IP efficiency 15-25% (vs. >50% expected).\nPhantom peak score (SPP NSC) 1.02 (threshold > 1.05). Flagged as failed IP.\nTrue biology still captured \u2014 differential enrichment in Polycomb domains valid.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Antibody lot variation is common for H3K27me3; NSC/RSC thresholds are guides\nnot hard cutoffs \u2014 domain-level enrichment may still be biologically informative.\n"} |
| {"id": "pt_037", "accession": "GSE143104", "protocol": "chip_seq_tf_low_occupancy", "universal_threshold_problem": "ChIP-seq for low-occupancy TF (CTCF in resting B cells): number of called peaks\n~5,000 (universal threshold > 20,000). B cells have fewer CTCF-bound sites than\nfibroblasts. Flagged as insufficient enrichment.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Peak count thresholds are calibrated to reference cell lines; immune cell ChIP\nroutinely yields fewer peaks due to biology, not technical failure.\n"} |
| {"id": "pt_038", "accession": "GSE183684", "protocol": "chip_seq_cut_and_run_low_cells", "universal_threshold_problem": "CUT&RUN with 1,000 cells: fragment size distribution differs from ChIP-seq\n(protected fragments ~150 bp, not nucleosome ladder). Universal ChIP QC\nflags missing nucleosome periodicity as library failure.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "CUT&RUN generates a distinct fragment size profile (short protected fragments)\ncompared to ChIP-seq; ChIP-centric QC tools incorrectly flag this as low quality.\n"} |
| {"id": "pt_039", "accession": "GSE153859", "protocol": "visium_fresh_frozen_brain", "universal_threshold_problem": "10x Visium spatial transcriptomics: UMIs/spot threshold > 500. Fresh-frozen\nmouse brain Visium: white matter spots have 50-200 UMIs due to low cell density.\nUniversal filter removes anatomically valid white matter regions.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Low-cellularity regions (white matter, connective tissue) legitimately have fewer\nUMIs per spot; spatial QC must use spatially-aware filters not global thresholds.\n"} |
| {"id": "pt_040", "accession": "GSE176078", "protocol": "visium_ffpe_spatial_cancer", "universal_threshold_problem": "Visium FFPE on tumor section: genes/spot threshold > 200. FFPE degradation\nyields 80-150 genes/spot in necrotic tumor core, 300-500 in viable tumor edge.\nGlobal QC removes entire necrotic core \u2014 scientifically important region.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Spatial heterogeneity in tumors creates systematic QC disparities; necrotic regions\nhave valid spatial coordinates but lower library complexity.\n"} |
| {"id": "pt_041", "accession": "GSE214543", "protocol": "visium_uneven_tissue_coverage", "universal_threshold_problem": "Visium on uneven tissue section: spots at tissue edge have partial cell coverage.\nTissue detection algorithms remove spots with < 50% tissue coverage. High-quality\nboundary spots at tumor-stroma interface are incorrectly removed.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Tissue edge spots have partial coverage by design; removing all edge spots\neliminates biologically critical transition zones.\n"} |
| {"id": "pt_042", "accession": "GSE198355", "protocol": "visium_low_rna_quality_index", "universal_threshold_problem": "Visium on archival surgical biopsy: RNA quality index (RQI) 3-4 (threshold > 7).\n10x SpaceRanger issues low-quality warning. Despite degradation, spatial patterns\nof highly expressed genes (oncogenes, immune markers) are preserved and valid.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Partially degraded spatial samples retain spatial information for abundant\ntranscripts; complete rejection based on RQI discards useful spatial data.\n"} |
| {"id": "pt_043", "accession": "SRP322355", "protocol": "nanopore_q10_chemistry", "universal_threshold_problem": "Oxford Nanopore R9.4.1 chemistry: mean read quality Q10-12. Illumina-calibrated\nQC (universal Phred threshold > Q30) flags all nanopore reads as failed.\nNanoPlot correctly interprets Q10 as normal for R9.4.1 pore.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Nanopore sequencing error rates are fundamentally different from Illumina; Q10\n(90% accuracy) is the expected baseline for R9.4.1 chemistry, not a quality failure.\n"} |
| {"id": "pt_044", "accession": "SRP390271", "protocol": "nanopore_direct_rna_seq", "universal_threshold_problem": "Nanopore direct RNA sequencing: read length distribution skewed (mean 500 bp,\ntail to 12 kb). Universal RNA-seq QC expects tight Gaussian around insert size.\nBimodal distribution flagged as contamination or fragmentation artifact.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Direct RNA-seq read length distribution reflects native mRNA length heterogeneity;\na bimodal distribution indicating short vs. long transcripts is expected.\n"} |
| {"id": "pt_045", "accession": "SRP355624", "protocol": "nanopore_amplicon_long_read", "universal_threshold_problem": "Nanopore amplicon sequencing (e.g., full-length 16S): all reads ~1.5 kb.\nUniversal diversity-based read length QC flags lack of diversity as library bias.\nFor amplicon panels this is by design.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Amplicon sequencing produces reads of defined length by design; a monodisperse\nread length distribution is the expected QC signature, not a failure.\n"} |
| {"id": "pt_046", "accession": "SRP420117", "protocol": "nanopore_r10_duplex", "universal_threshold_problem": "Nanopore R10.4.1 duplex chemistry: ~50% reads are duplex (Q25-30), ~50% simplex\n(Q12-15). Universal mean quality computed over all reads = Q18-20, flagged as\nbelow Q30 threshold. Duplex reads are high-quality; mixed QC is misleading.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Duplex nanopore runs produce a bimodal quality distribution; global mean quality\nmasks the high-quality duplex fraction that is suitable for variant calling.\n"} |
| {"id": "pt_047", "accession": "GSE173767", "protocol": "wgbs_low_conversion_rate", "universal_threshold_problem": "WGBS bisulfite conversion rate: universal threshold > 99%. Placenta tissue has\nlow-methylation domains where incomplete conversion looks like a spike in\nnon-CpG methylation. Bismark QC reports conversion 97.5%, triggers FAIL.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Placental and early embryonic samples have unusual methylation landscapes that\ncan lower apparent bisulfite conversion rate; tissue context is critical for QC.\n"} |
| {"id": "pt_048", "accession": "GSE186458", "protocol": "wgbs_cfDNA_bisulfite", "universal_threshold_problem": "Cell-free DNA (cfDNA) WGBS: fragment length 160-180 bp (nucleosome-protected).\nUniversal WGBS expected insert size 300-500 bp. cfDNA shows tight nucleosomal\nsizing which universal QC flags as over-fragmentation.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "cfDNA WGBS fragment size distribution reflects nucleosome-protected circulating\nDNA; nucleosomal sizing (~167 bp) is the expected cfDNA signature.\n"} |
| {"id": "pt_049", "accession": "GSE151022", "protocol": "rrbs_low_coverage", "universal_threshold_problem": "Reduced Representation Bisulfite Sequencing (RRBS): genome coverage ~4-8%.\nUniversal bisulfite QC checks genome-wide CpG coverage (threshold > 30%).\nRRBS by design only covers MspI-digested CpG-rich fragments; 4-8% is correct.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "RRBS achieves deep local coverage of ~4-8% of CpG sites; genome-wide coverage\nmetrics are inapplicable and will always flag RRBS as failing.\n"} |
| {"id": "pt_050", "accession": "GSE197353", "protocol": "wgbs_non_cpg_methylation", "universal_threshold_problem": "WGBS in pluripotent stem cells: CH methylation (non-CpG) 20-30%. Universal QC\nchecks CH methylation < 5% as a proxy for bisulfite conversion quality.\nStem cells and neurons have high CH methylation biologically.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Non-CpG methylation is a genuine biological feature of neurons and stem cells;\nusing CH methylation as a conversion quality proxy fails for these cell types.\n"} |
| {"id": "pt_051", "accession": "GSE118184", "protocol": "smart_seq2_plate_based", "universal_threshold_problem": "Smart-seq2 full-length scRNA-seq: reads per cell 100,000-500,000, uniform across\ngene body (no 3' bias). Universal droplet-based QC expects 3'-biased read\ndistribution; Smart-seq2 even coverage flagged as ribo-seq contamination.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Smart-seq2 produces full-length cDNA with even gene body coverage; 3'-bias metrics\nare inapplicable and will flag high-quality Smart-seq2 data as aberrant.\n"} |
| {"id": "pt_052", "accession": "GSE132044", "protocol": "smart_seq2_low_read_depth", "universal_threshold_problem": "Smart-seq2 on sorted rare cells: 50,000-80,000 reads/cell (budget plate).\nUniversal reads/cell threshold > 250,000. Low-depth Smart-seq2 captures most\nhighly expressed genes but flags fail on depth.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Smart-seq2 read depth requirements depend on cell type and scientific question;\n50k reads captures the expressed transcriptome of small cells adequately.\n"} |
| {"id": "pt_053", "accession": "GSE109979", "protocol": "smart_seq2_ercc_spike_in", "universal_threshold_problem": "Smart-seq2 with ERCC spike-ins: spike-in fraction 10-25% of total reads.\nUniversal QC: spike-in fraction < 5% (treats spikes as contamination).\nERCC spike-ins are intentionally added for normalization; high fraction expected.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "ERCC spike-in fraction must be interpreted in context of spike-in dilution;\na 10-25% ERCC fraction is normal for the recommended 1:20 dilution in Smart-seq2.\n"} |
| {"id": "pt_054", "accession": "GSE168652", "protocol": "total_rna_seq_ribo_zero", "universal_threshold_problem": "Total RNA-seq with Ribo-Zero depletion: rRNA fraction 2-8% residual after\ndepletion. Universal polyA-seq rRNA threshold < 1%. Ribo-Zero efficiency is\nlower than polyA selection; 2-8% residual is the normal operating range.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Ribo-Zero depletion is less efficient than polyA selection; 2-8% residual rRNA\nis the expected range and does not indicate failed depletion.\n"} |
| {"id": "pt_055", "accession": "GSE183653", "protocol": "total_rna_seq_lncrna", "universal_threshold_problem": "Total RNA-seq targeting lncRNA: lower coding gene coverage, higher intronic\nreads (lncRNAs are less efficiently spliced). Universal: intronic reads < 30%.\nTotal RNA-seq typically shows 40-60% intronic reads; lncRNA-enriched samples higher.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "lncRNA-focused total RNA-seq has higher intronic read fractions than mRNA-seq;\nintronic read thresholds calibrated on polyA libraries are inapplicable.\n"} |
| {"id": "pt_056", "accession": "GSE142458", "protocol": "total_rna_seq_bacterial", "universal_threshold_problem": "Total RNA-seq from bacteria: no poly-A tails, no splicing, all reads map to\ncoding or rRNA genes. Universal eukaryotic QC checks for splicing junction\nreads > 30% and flags 0% junctions as alignment failure.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Prokaryotic RNA-seq has no splice junctions; eukaryotic splicing-based QC\nmetrics are fundamentally inapplicable to bacterial transcriptomics.\n"} |
| {"id": "pt_057", "accession": "GSE156021", "protocol": "tcr_vdj_sequencing_10x", "universal_threshold_problem": "10x Genomics TCR V(D)J sequencing: >95% of reads map to V(D)J loci.\nUniversal QC: reads mapping to reference genome > 80% on other loci.\nV(D)J libraries are targeted enrichments \u2014 near-100% V(D)J mapping is correct.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "V(D)J sequencing is an amplicon enrichment of immunoglobulin/TCR loci; non-VDJ\nreference mapping of <5% is the expected signature of high-quality libraries.\n"} |
| {"id": "pt_058", "accession": "GSE178279", "protocol": "bcr_vdj_low_clonotype_diversity", "universal_threshold_problem": "BCR V(D)J from tumor-infiltrating B cells: very low clonotype diversity (clonal\nexpansion). Shannon entropy < 1 (universal threshold > 3). Flagged as poor\nlibrary complexity, but clonal dominance is the expected biology in tumors.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Clonal expansion is the hallmark of antigen-driven immune responses; low\nclonotype diversity in tumor-infiltrating lymphocytes is biologically expected.\n"} |
| {"id": "pt_059", "accession": "GSE201519", "protocol": "tcr_vdj_paired_chain", "universal_threshold_problem": "TCR V(D)J paired alpha-beta sequencing: ~60-75% of cells have both chains\ndetected. Universal cell pairing threshold > 90%. Low-pairing rate in primary\ntissue is due to RNA accessibility not library failure.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "TCR chain pairing rate depends on cell activation state and RNA quality;\n60-75% pairing in primary tissue is acceptable and biologically informative.\n"} |
| {"id": "pt_060", "accession": "SYNTHETIC_DEFECT_003", "protocol": "nanopore_long_read", "universal_threshold_problem": "None \u2014 this is a genuinely defective sample", "expected_naive_action": "FAIL", "expected_correct_action": "FAIL", "ground_truth": "DEFECTIVE", "notes": "Synthetic defect: pore blockage event injected \u2014 80% of reads are <50 bp\n(failed pore runs). Both universal and protocol-aware QC should correctly FAIL.\nValidates that nanopore profile does not over-correct toward PASS.\n"} |
| {"id": "pt_061", "accession": "SYNTHETIC_DEFECT_004", "protocol": "wgbs_bisulfite", "universal_threshold_problem": "None \u2014 this is a genuinely defective sample", "expected_naive_action": "FAIL", "expected_correct_action": "FAIL", "ground_truth": "DEFECTIVE", "notes": "Synthetic defect: bisulfite conversion rate 78% (genuine under-conversion,\nnot tissue-specific). High false-methylation rate. Both thresholds should FAIL.\n"} |
| {"id": "pt_062", "accession": "SYNTHETIC_DEFECT_005", "protocol": "chip_seq_broad_mark", "universal_threshold_problem": "None \u2014 this is a genuinely defective sample", "expected_naive_action": "FAIL", "expected_correct_action": "FAIL", "ground_truth": "DEFECTIVE", "notes": "Synthetic defect: input DNA sequenced without any IP step (no enrichment).\nFRiP = 0.001. Even with broad-mark profile, this should FAIL \u2014 no signal present.\n"} |
| {"id": "pt_063", "accession": "GSE185658", "protocol": "snrna_seq_low_gene_detection", "universal_threshold_problem": "snRNA-seq from aged mouse brain: genes/nucleus 200-350 (threshold > 500).\nAged neurons show transcriptional quiescence with reduced mRNA output.\nUniversal QC removes >70% of aged neuronal nuclei.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Transcriptional quiescence in aged neurons produces genuinely low gene counts;\nWARN is appropriate \u2014 retain with annotation of reduced sensitivity.\n"} |
| {"id": "pt_064", "accession": "GSE233285", "protocol": "spatial_visium_low_spot_count", "universal_threshold_problem": "Visium on small biopsy punch: 200 spots under tissue (threshold > 1000 spots).\nSmall tissue sections are clinically common (core biopsy, surgical margin).\nSpaceRanger warns on <500 spots. Spatial patterns are valid despite small area.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Total spot count reflects tissue area not library quality; small biopsies are\nclinically important and valid for spatial analysis at reduced resolution.\n"} |
| {"id": "pt_065", "accession": "GSE243901", "protocol": "cite_seq_adt_high_background", "universal_threshold_problem": "CITE-seq from PBMCs with suboptimal wash: ADT background 30-50% of signal\n(universal signal/background threshold > 5). DSB normalization corrects for\nthis; raw ADT counts flagged but normalized values interpretable.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "High ADT background from imperfect washing is recoverable with DSB normalization;\nraw count QC failure does not necessarily mean protein data is unusable.\n"} |
| {"id": "pt_066", "accession": "GSE199318", "protocol": "bulk_atac_seq_neutrophil_low_peaks", "universal_threshold_problem": "Bulk ATAC-seq on circulating neutrophils: 3,000-8,000 peaks (threshold > 30,000).\nNeutrophils have highly condensed chromatin and limited open chromatin accessible\nto Tn5. Used for neutrophil activation studies comparing treated vs. untreated.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Granulocyte chromatin compaction produces genuinely low peak counts; the relevant\ncomparison is treated vs. untreated neutrophils, not neutrophils vs. T cells.\n"} |
| {"id": "pt_067", "accession": "GSE161918", "protocol": "smart_seq2_single_cell_low_depth_rare", "universal_threshold_problem": "Smart-seq2 on circulating tumor cells (CTCs): 30,000-60,000 reads/cell.\nCTCs are extremely rare, captured singly; budget sequencing unavoidable.\nUniversal reads/cell > 250,000 flags entire CTC dataset as FAIL.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "CTC studies often operate with low sequencing depth per cell due to rarity of\nmaterial; modest depth still captures clinically informative gene expression.\n"} |
| {"id": "pt_068", "accession": "GSE157827", "protocol": "rrbs_high_duplication", "universal_threshold_problem": "RRBS from FFPE: duplication rate 60-80% (universal threshold < 40%).\nRRBS on FFPE is a double-enrichment: restriction digest + bisulfite conversion\non degraded DNA \u2192 very limited unique molecules, high duplication expected.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "RRBS from FFPE has constrained unique molecule diversity; deduplication removes\ntoo much signal. Protocol-aware QC should flag WARN and recommend dedup-aware analysis.\n"} |
| {"id": "pt_069", "accession": "GSE178707", "protocol": "total_rna_seq_circular_rna", "universal_threshold_problem": "Total RNA-seq optimized for circular RNA detection: RNase R treatment degrades\nlinear RNA, enriching circular species. Linear RNA fraction 20-30% (from\nincomplete RNase R treatment). Standard QC flags 30% unmapped reads (back-splice\njunctions not in standard annotation) as alignment failure.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "circRNA studies require back-splice-aware aligners (STAR with chimeric options or\nfind_circ); standard junction QC is inapplicable to circRNA-enriched libraries.\n"} |
| {"id": "pt_070", "accession": "GSE169564", "protocol": "tcr_vdj_gamma_delta", "universal_threshold_problem": "gamma-delta T cell TCR V(D)J sequencing: <20% of reads have productive\nrearrangements (threshold > 60%). gamma-delta TCRs are less efficiently\ncaptured by V(D)J enrichment oligos optimized for alpha-beta TCRs.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "10x V(D)J enrichment is optimized for alpha-beta TCRs; gamma-delta TCRs have\nlower capture efficiency by design and require adjusted productivity thresholds.\n"} |
| {"id": "pt_071", "accession": "GSE158760", "protocol": "spatial_transcriptomics_visium", "universal_threshold_problem": "Universal threshold: genes/spot > 500. Visium spots on white matter or necrotic\ntissue legitimately yield 50-200 genes/spot. MultiQC flags low-coverage spots as\nlow-quality without tissue context.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "White matter is transcriptionally sparse; low gene counts per spot are biologically\nexpected, not technical failure. Tissue mask and histology context are required.\n"} |
| {"id": "pt_072", "accession": "GSE171892", "protocol": "long_read_nanopore_direct_rna", "universal_threshold_problem": "Universal threshold: mean read quality > Q20. Oxford Nanopore direct RNA reads\nyield Q8-Q12 median quality. Guppy high-accuracy basecalling improves this to\nQ12-Q15, still far below Illumina Q30 thresholds.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Direct RNA sequencing on R9.4.1 or R10.4 flowcells has inherently lower per-base\naccuracy than cDNA; Q10 is the accepted quality floor for this chemistry.\n"} |
| {"id": "pt_073", "accession": "GSE149683", "protocol": "crispr_screen_mageck", "universal_threshold_problem": "Universal threshold: PCR duplicate rate < 20%. MAGeCK CRISPR screens use PCR\namplification with degenerate primers; duplicate rates of 40-70% are normal and\nexpected in high-representation screens.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "CRISPR screen libraries are deliberately over-represented; high PCR duplicate\nrates reflect library complexity preservation, not contamination.\n"} |
| {"id": "pt_074", "accession": "GSE176525", "protocol": "hic_chromosome_conformation", "universal_threshold_problem": "Universal threshold: mapped read pairs > 80%. Hi-C libraries have 40-60%\nvalid ligation products; chimeric read pairs from non-adjacent restriction\nfragments are expected and removed by HiC-Pro at the pipeline level.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Hi-C valid pair rate of 40-60% is standard; \"unmapped\" pairs often represent\ndangling ends and self-circles that are correctly filtered downstream.\n"} |
| {"id": "pt_075", "accession": "MSV000085202", "protocol": "metabolomics_untargeted_lcms", "universal_threshold_problem": "Universal threshold: CV% < 20% across all features. Lipid species and secondary\nmetabolites in LC-MS untargeted metabolomics often show CV 30-50% due to\nbiological variation, not analytical failure.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "High CV for low-abundance metabolites is expected biologically; QC pool CV < 20%\nis the correct metric, not per-feature CV across biological samples.\n"} |
| {"id": "pt_076", "accession": "GSE163530", "protocol": "scrna_seq_ambient_rna_correction", "universal_threshold_problem": "Universal threshold: empty droplet rate > 5% = FAIL. SoupX/CellBender-corrected\n10x libraries report 10-25% ambient RNA contamination as PASSED data, which raw\nQC tools incorrectly flag before correction is applied.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Ambient RNA contamination at 10-25% is correctable; the flag should be raised\nonly if contamination exceeds 40% or if correction artifacts are present.\n"} |
| {"id": "pt_077", "accession": "GSE154600", "protocol": "tcr_bcr_vdj_clonal_expansion", "universal_threshold_problem": "Universal threshold: clonotype entropy > 3.0 (diverse). Tumor-infiltrating T cells\nin clonally expanded responses show entropy 0.5-1.5, which naive QC flags as\n\"low diversity / poor library\" rather than biologically significant clonal expansion.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Low clonotype entropy in TIL samples is a biological signal of antigen-driven\nclonal expansion, not a sequencing quality issue.\n"} |
| {"id": "pt_078", "accession": "GSE162690", "protocol": "smart_seq2_rare_cell_population", "universal_threshold_problem": "Universal threshold: reads/cell > 500,000. Rare cell populations in Smart-seq2\nat 1:1000 dilution yield 50,000-150,000 reads/cell when cell numbers are low,\nas total sequencing depth is spread across the whole plate.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Smart-seq2 plate depth varies by cell loading; rare population isolation plates\nhave fewer total cells and thus lower per-cell sequencing depth by design.\n"} |
| {"id": "pt_079", "accession": "GSE147255", "protocol": "multiome_atac_rnaseq", "universal_threshold_problem": "Universal threshold: ATAC TSS enrichment > 7. 10x Multiome ATAC+RNA co-captured\nnuclei yield TSS enrichment of 4-6, below single-assay ATAC thresholds, because\nthe GEM-X chemistry is optimized for dual-capture not maximal ATAC sensitivity.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Multiome ATAC inherently has lower TSS enrichment than dedicated ATAC-seq due to\nshared GEM volume; adjusted thresholds of 3-5 are appropriate for this assay.\n"} |
| {"id": "pt_080", "accession": "GSE143476", "protocol": "pbmc_bulk_rnaseq_ribo_depletion", "universal_threshold_problem": "Universal threshold: rRNA reads < 5%. Ribo-zero depleted PBMC total RNA-seq\nretains 8-15% rRNA in blood-derived samples due to incomplete depletion of the\nglobin-abundant erythrocyte rRNA fraction.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Ribo-depletion efficiency in blood samples is lower than in tissue RNA due to\nhigh globin and rRNA abundance; 8-15% residual rRNA is acceptable.\n"} |
| {"id": "pt_081", "accession": "GSE137912", "protocol": "chip_seq_low_cell_number", "universal_threshold_problem": "Universal threshold: library complexity (NRF) > 0.9. CUT&RUN with 5,000 cells\nyields NRF 0.6-0.75 due to limited input chromatin; standard ChIP-seq NRF\nthresholds assume millions of cells.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Low-cell CUT&RUN and CUT&TAG protocols operate on 1,000-50,000 cells; library\ncomplexity thresholds must be adjusted proportionally to input cell number.\n"} |
| {"id": "pt_082", "accession": "GSE188197", "protocol": "snrnaseq_cardiac_muscle", "universal_threshold_problem": "Universal threshold: UMIs/nucleus > 1,000. Cardiac muscle nuclei in snRNA-seq\nyield 500-800 UMIs due to compact nuclear volume and low cytoplasmic RNA\ncontamination in post-mortem heart tissue.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Post-mortem cardiac tissue has reduced RNA integrity; low UMI counts per nucleus\nare expected and do not indicate failed library preparation.\n"} |
| {"id": "pt_083", "accession": "GSE139334", "protocol": "atac_seq_ffpe_tissue", "universal_threshold_problem": "Universal threshold: fraction reads in peaks (FRiP) > 0.2. FFPE ATAC-seq\nyields FRiP 0.05-0.12 due to crosslinking-induced chromatin compaction and\nfragmentation artefacts from formalin fixation.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "FFPE ATAC is inherently noisier; FRiP 0.08-0.12 is acceptable for FFPE material\nand should trigger a WARN, not FAIL, with a note on fixation.\n"} |
| {"id": "pt_084", "accession": "GSE178707", "protocol": "pacbio_long_read_wgs", "universal_threshold_problem": "Universal threshold: mean read N50 > 10 kb. PacBio HiFi (CCS) reads average\n15-20 kb N50 by design, but raw CLR reads in older protocols average 5-8 kb\nN50, which naive QC incorrectly flags as short-insert contamination.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "CLR vs CCS protocol determines expected read length distribution; N50 thresholds\nmust be calibrated to the specific PacBio chemistry used.\n"} |
| {"id": "pt_085", "accession": "GSE152294", "protocol": "rrbs_cpg_methylation", "universal_threshold_problem": "Universal threshold: CpG coverage uniformity > 0.8. RRBS by design captures\nonly restriction-proximal CpGs; coverage uniformity across the genome is 0.2-0.4,\ncorrectly reflecting the enrichment design, not a failure.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "RRBS is an enrichment protocol; non-uniform genome coverage is the expected and\ncorrect outcome, not a QC failure. Uniformity should be measured within RRBS\ntarget regions only.\n"} |
| {"id": "pt_086", "accession": "GSE174072", "protocol": "cite_seq_protein_capture", "universal_threshold_problem": "Universal threshold: protein (ADT) UMIs/cell > 200. CITE-seq with isotype\ncontrol antibodies included in the panel shows 50-100 total ADT UMIs/cell for\nlow-abundance surface markers while isotype controls show 5-20 UMIs.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "ADT counts depend on surface marker abundance; low-expression markers like\nCD34 or CD56 in non-hematopoietic contexts legitimately yield low ADT UMIs.\n"} |
| {"id": "pt_087", "accession": "GSE165299", "protocol": "bulk_atac_seq_low_cell", "universal_threshold_problem": "Universal threshold: nucleosome-free region (NFR) fraction > 0.4. Bulk ATAC-seq\nfrom laser-capture microdissected samples (100-1000 cells) yields NFR 0.25-0.35\ndue to limited Tn5 tagmentation efficiency at low cell numbers.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Low-input bulk ATAC requires adjusted NFR thresholds; the nucleosomal periodicity\npattern is more diagnostic than the NFR absolute fraction for these samples.\n"} |
| {"id": "pt_088", "accession": "GSE177294", "protocol": "total_rnaseq_low_rin", "universal_threshold_problem": "Universal threshold: RNA integrity number (RIN) > 7. FFPE total RNA-seq\nwith ribo-depletion from archival samples routinely uses RIN 2-4 material;\nMultiQC flags these as degraded without considering FFPE protocol tolerance.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "FFPE RNA is inherently degraded; RIN 2-4 is the expected range and should\ntrigger a WARN (with degradation note) rather than hard FAIL.\n"} |
| {"id": "pt_089", "accession": "GSE155399", "protocol": "scrna_seq_doublet_rate", "universal_threshold_problem": "Universal threshold: doublet rate < 5%. 10x Genomics high-throughput loading\nat 20,000 cells/well yields 10-15% doublet rate by design; this is acceptable\nand handled by Scrublet/DoubletFinder post-processing.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "High-throughput 10x loading intentionally allows higher doublet rates in exchange\nfor more cells recovered; computational doublet removal is the expected workflow.\n"} |
| {"id": "pt_090", "accession": "GSE181605", "protocol": "gwas_imputation_quality", "universal_threshold_problem": "Universal threshold: imputation R\u00b2 > 0.3. Rare variant GWAS (MAF < 0.5%) with\narray-based imputation yields R\u00b2 0.1-0.25 for variants in low-LD regions.\nPLINK/IMPUTE2 flag these as poor quality though they may be well-imputed given\nthe sparse LD structure.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Rare variant imputation quality depends on reference panel density and regional\nLD; blanket R\u00b2 > 0.3 filters remove legitimate rare variant associations.\n"} |
| {"id": "pt_091", "accession": "GSE148396", "protocol": "phosphoproteomics_enrichment", "universal_threshold_problem": "Universal threshold: phosphopeptide enrichment efficiency > 80%. TiO\u2082 or IMAC\nenrichment of phosphopeptides from complex lysates yields 60-75% efficiency when\nstarting material is < 100 \u00b5g protein, which mass spec QC tools flag as poor.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Phospho enrichment efficiency is input-dependent; < 100 \u00b5g protein input is\nexpected to yield lower efficiency, and 60-75% is acceptable for clinical samples.\n"} |
| {"id": "pt_092", "accession": "GSE186458", "protocol": "tmt_labeled_proteomics", "universal_threshold_problem": "Universal threshold: TMT labeling efficiency > 95%. Clinical biopsy samples\nwith limited protein yield achieve 85-92% TMT labeling efficiency; standard\nproteomics QC tools flag this as incomplete labeling failure.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "TMT labeling efficiency < 95% is acceptable for limited clinical samples;\nunlabeled peptides can be identified and excluded computationally.\n"} |
| {"id": "pt_093", "accession": "GSE177131", "protocol": "long_read_amplicon_sequencing", "universal_threshold_problem": "Universal threshold: length distribution CV < 0.3 (monodisperse). Nanopore\namplicon panels from 16S/18S/ITS targets yield monodisperse length distributions\n(CV < 0.1), which some QC tools flag as \"amplicon contamination\" due to lack\nof length variation.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Targeted amplicon sequencing should produce a monodisperse length distribution;\nlow CV is the expected outcome of PCR amplification of a fixed-size target.\n"} |
| {"id": "pt_094", "accession": "GSE159929", "protocol": "scrna_seq_high_mito_neurons", "universal_threshold_problem": "Universal threshold: mitochondrial reads < 10%. Metabolically active neurons\nand cardiomyocytes legitimately express 20-35% mitochondrial reads; blanket\nmito filters remove live, healthy cells in these lineages.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Mitochondrial read fraction thresholds are cell-type-specific; neurons and\ncardiomyocytes require relaxed thresholds of 20-30% for their high energy demand.\n"} |
| {"id": "pt_095", "accession": "GSE180878", "protocol": "epigenomics_cut_and_tag", "universal_threshold_problem": "Universal threshold: peak FRIP > 0.2. CUT&TAG for active histone marks\n(H3K4me3) yields FRIP 0.5-0.8, but repressive marks (H3K27me3, H3K9me3)\nspread across large domains yield FRIP 0.05-0.15, which is flagged as failure.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Repressive marks form broad, low-signal domains; FRIP scores must be calibrated\nto the mark type, not a single universal threshold.\n"} |
| {"id": "pt_096", "accession": "GSE182571", "protocol": "multiome_spatial_visium_hd", "universal_threshold_problem": "Universal threshold: spots with > 200 genes. Visium HD at 2\u00b5m resolution\nyields many spots with 50-150 genes as the smaller spot size captures partial\ncells; this is the intended behavior of the assay.\n", "expected_naive_action": "FAIL", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Visium HD uses 2\u00d72 \u00b5m bins; each bin captures a sub-cellular area and lower\ngene counts per bin are expected and informative at this resolution.\n"} |
| {"id": "pt_097", "accession": "GSE171915", "protocol": "circadian_rnaseq_zeitgeber", "universal_threshold_problem": "Universal threshold: batch correction removes > 95% of PC1 variance. Zeitgeber\ntime-series RNA-seq data is intentionally batch-structured by collection time;\naggressive batch correction removes the biological signal of interest.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Time-series transcriptomics has structured variance by design; blind batch\ncorrection destroys the temporal signal. Zeitgeber grouping should be\ntreated as a biological factor, not a batch.\n"} |
| {"id": "pt_098", "accession": "GSE158481", "protocol": "scrna_seq_xenograft_mouse_human", "universal_threshold_problem": "Universal threshold: mapping rate to reference > 85%. Patient-derived xenograft\n(PDX) scRNA-seq aligned to human reference alone yields 50-65% mapping rate\nbecause mouse stromal cells (30-50% of total) map to the human genome poorly.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "PDX samples require dual-species alignment; low human-reference mapping rate\nis expected due to the mouse cell fraction and should trigger a WARN with\nrecommendation for species-deconvolution alignment.\n"} |
| {"id": "pt_099", "accession": "GSE175891", "protocol": "scrnaseq_low_complexity_library", "universal_threshold_problem": "Universal threshold: genes detected per cell > 200. Highly pure populations\n(e.g., erythrocytes, platelets, regulatory T cells) have genuinely restricted\ntranscriptomes with 50-180 genes expressed; blanket filters remove these\nbiologically important cell types.\n", "expected_naive_action": "FAIL", "expected_correct_action": "WARN", "ground_truth": "VALID", "notes": "Mature erythrocytes and platelets are transcriptionally quiescent; low gene\ncounts reflect their biology, not library failure. Context-aware thresholds\nbased on cell clustering are appropriate.\n"} |
| {"id": "pt_100", "accession": "GSE184369", "protocol": "nanopore_adaptive_sampling", "universal_threshold_problem": "Universal threshold: percent on-target reads > 80%. Nanopore adaptive sampling\n(ReadFish/Dorado) enriches target regions to 60-70% on-target in complex\ngenomes, rejecting off-target reads in real-time. QC tools flag 30-40% off-target\nreads as enrichment failure.\n", "expected_naive_action": "WARN", "expected_correct_action": "PASS", "ground_truth": "VALID", "notes": "Adaptive sampling enrichment efficiency of 60-70% is state-of-the-art for\nreal-time selective sequencing; 30-40% off-target reads are expected given\nthe reaction kinetics of pore-based rejection.\n"} |
|
|
