[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EBH2N5VB4XGVZ8ECM7YM1E":3,"skill-same:01M7EBH2N5VB4XGVZ8ECM7YM1E":62},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":30,"commitSha":31,"skillDirectory":32,"forksCount":14,"repositoryHtmlUrl":33,"repositoryHomepageUrl":33,"ownerHtmlUrl":34,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":35,"install":36,"resources":49},"01M7EBH2N5VB4XGVZ8ECM7YM1E","bio-atac-seq-atac-qc",null,"ATAC-seq library quality control -- TSS enrichment, FRiP, fragment-size periodicity, library complexity (NRF\u002FPBC1\u002FPBC2), mitochondrial fraction, and ENCODE 4 thresholds. Use when assessing whether an ATAC-seq library passes ENCODE acceptance criteria, diagnosing transposition artefacts, comparing Omni-ATAC vs standard prep quality, or selecting which replicates to drop before peak calling.","Mr-Nobody-Anonymous\u002FAll-skills","Mr-Nobody-Anonymous","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F184253119?v=4","01M7AEXFJHKTYQW5R0ZCQQTWFV","anonymous","01M7C00M3GQR4QKP9GZA42GFAK",0,"MIT","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-atac-seq-atac-qc\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: bio-atac-seq-atac-qc\ndescription: ATAC-seq library quality control -- TSS enrichment, FRiP, fragment-size periodicity, library complexity (NRF\u002FPBC1\u002FPBC2), mitochondrial fraction, and ENCODE 4 thresholds. Use when assessing whether an ATAC-seq library passes ENCODE acceptance criteria, diagnosing transposition artefacts, comparing Omni-ATAC vs standard prep quality, or selecting which replicates to drop before peak calling.\ntool_type: mixed\nprimary_tool: deeptools\nsource: \"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills\"\nsource_repository: \"GPTomics\u002FbioSkills\"\nsource_path: \"atac-seq\u002Fatac-qc\u002FSKILL.md\"\nlicense: \"MIT\"\nimported_at: \"2026-09-20\"\n---\n\n## Version Compatibility\n\nReference examples tested with: deepTools 3.5+, Picard 3.1+, samtools 1.19+, bedtools 2.31+, ATACseqQC 1.26+, pysam 0.22+, pyBigWig 0.3+, numpy 1.26+, pandas 2.2+, MultiQC 1.21+.\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- Python: `pip show \u003Cpackage>` then `help(module.function)` to check signatures\n- R: `packageVersion('\u003Cpkg>')` then `?function_name` to verify parameters\n- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt.\n\n# ATAC-seq Quality Control\n\n**\"Does my ATAC library pass ENCODE quality criteria?\"** -> Compute the seven canonical metrics (depth, alignment rate, mitochondrial fraction, library complexity, fragment-size periodicity, TSS enrichment, FRiP) and compare against ENCODE 4 thresholds, then diagnose failures.\n\n- CLI: `picard CollectInsertSizeMetrics`, `samtools flagstat`, `samtools idxstats`\n- CLI: `deeptools plotFingerprint`, `computeMatrix reference-point` + `plotProfile`\n- R: `ATACseqQC::TSSEscore`, `ATACseqQC::fragSizeDist`, `ATACseqQC::PTscore`\n- Python: custom NRF\u002FPBC from coordinate hash; pyBigWig for TSS enrichment\n\n## ENCODE 4 ATAC-seq Acceptance Thresholds\n\n| Metric | Definition | Ideal | Acceptable | Reject | Source |\n|--------|-----------|-------|------------|--------|--------|\n| Nuclear reads (after dedup, no chrM) | Mapped, MAPQ >= 30, non-chrM, deduped | >= 50M | 25-50M | \u003C 25M | ENCODE 4 ATAC-seq Standards |\n| Alignment rate | Mapped \u002F total reads | >= 95% | 80-95% | \u003C 80% | ENCODE 4 |\n| Mitochondrial fraction | chrM \u002F total mapped | \u003C 5% (Omni-ATAC), \u003C 20% (standard) | 20-50% | > 50% | Corces 2017 (Omni-ATAC) |\n| NRF (Non-Redundant Fraction) | Distinct positions \u002F total reads | >= 0.9 | 0.7-0.9 | \u003C 0.7 | Landt 2012 |\n| PBC1 (PCR Bottlenecking Coefficient 1) | Positions w\u002F 1 read \u002F Positions w\u002F >= 1 read | >= 0.9 | 0.7-0.9 | \u003C 0.7 | Landt 2012 |\n| PBC2 | Positions w\u002F 1 read \u002F Positions w\u002F 2 reads | >= 3.0 | 1.0-3.0 | \u003C 1.0 | Landt 2012 |\n| TSS enrichment (hg38, GENCODE v29) | Avg signal at TSS \u002F avg flanking | >= 7 | 5-7 | \u003C 5 | ENCODE 4 |\n| FRiP (Fraction Reads in Peaks) | Reads in MACS peaks \u002F total | >= 0.3 | 0.2-0.3 | \u003C 0.2 | ENCODE 4, Landt 2012 |\n| Insert-size periodicity | NFR + mono-nuc + di-nuc peaks visible | Clear 3+ peaks | NFR + mono only | Flat \u002F single peak | Buenrostro 2013 |\n\nENCODE thresholds are organism-specific. Mouse (mm10, GENCODE M21) TSS enrichment >= 5 is acceptable; non-model organisms have no published threshold (use cohort percentile rank instead). Methodology evolves; verify against the current ENCODE ATAC-seq Standards before reporting.\n\n## TSS Enrichment: ENCODE Method vs ATACseqQC Method\n\nThe two most common implementations DO NOT produce identical scores.\n\n| Method | Numerator | Denominator | Scaling |\n|--------|-----------|-------------|---------|\n| ENCODE pyTSSe \u002F Kundaje gtsse | Mean signal in 100 bp window centered at TSS | Mean signal in 100 bp window at +\u002F- 1900 to +\u002F- 2000 bp (flanks) | Per-base normalization to flanks; reported as fold-enrichment |\n| ATACseqQC TSSEscore | Sum signal in TSS +\u002F- 100 bp | Sum signal at +\u002F- 1000 bp flanking windows | Different window sizes; ratios are larger |\n| deeptools plotProfile | Visual; numeric ratio not standardized | Reference-point matrix | No standard score; for visualization only |\n\n**Trigger:** Comparing a TSS score across studies.\n\n**Mechanism:** Different normalization windows shift the absolute number; ATACseqQC's TSSEscore is typically 2-3x ENCODE's because of the wider flank.\n\n**Symptom:** Reported score 21 vs ENCODE-ideal 7 mismatch. Likely the calculator was ATACseqQC; the equivalent ENCODE score might be 8.\n\n**Fix:** State which implementation was used. For ENCODE comparisons, use `pyTSSe` (Kundaje lab) or implement the ENCODE recipe directly.\n\n```python\nimport numpy as np\nimport pyBigWig\n\ndef encode_tss_enrichment(bw_path, tss_bed, flank=2000):\n    \"\"\"ENCODE-style TSS enrichment: signal at TSS center \u002F signal at flanks.\"\"\"\n    bw = pyBigWig.open(bw_path)\n    profiles = []\n    for line in open(tss_bed):\n        chrom, start, end, *rest = line.strip().split('\\t')\n        tss = int(start)\n        strand = rest[2] if len(rest) > 2 else '+'\n        try:\n            vals = bw.values(chrom, tss - flank, tss + flank)\n            if vals is None or len(vals) != 2 * flank: continue\n            if strand == '-': vals = vals[::-1]\n            profiles.append(np.nan_to_num(vals))\n        except RuntimeError:\n            continue\n    avg = np.nanmean(profiles, axis=0)\n    flank_signal = np.mean(np.concatenate([avg[:100], avg[-100:]]))\n    center_signal = np.mean(avg[flank - 50: flank + 50])\n    return center_signal \u002F flank_signal if flank_signal > 0 else 0.0\n```\n\n## Fragment-Size Periodicity Patterns\n\n| Pattern | Visual signature | Interpretation | Action |\n|---------|-----------------|----------------|--------|\n| Strong tri-modal | NFR (~50bp) >> mono (~200bp) > di (~400bp) > tri (~600bp) peaks | Excellent transposition; well-positioned chromatin | Pass |\n| Clear bi-modal | NFR + mono only, di and tri faint | Acceptable; common in Omni-ATAC | Pass |\n| Single broad peak | Flat after NFR or no NFR | Over-transposition (too much Tn5) OR degraded chromatin | Reject; cannot distinguish nucleosomes |\n| Inverted (mono >> NFR) | Mono peak dominant, NFR weak | Under-transposition OR chromatin condensation | Caution; peak counts will be low |\n| Sharp 147 bp spike with no flanks | Tight peak at 147 bp | ChIP-seq input contamination (MNase-like) | Reject; not ATAC-grade |\n| 10.4 bp helical periodicity overlay | Sub-peaks at 50, 60, 70, 80 bp on NFR | Excellent chromatin structure resolution; helical phasing visible | Pass; high-quality |\n\nThe 10.4 bp helical periodicity is a Buenrostro 2013 hallmark: it reflects the helical pitch of B-form DNA, with Tn5 preferring outward-facing minor grooves on nucleosomal DNA. Its presence is a positive QC indicator but not required.\n\n## Per-Metric Failure Modes\n\n### Mitochondrial fraction > 50%\n\n**Trigger:** Standard ATAC-seq protocol on intact cells (no nuclear isolation), or insufficient detergent in lysis.\n\n**Mechanism:** Mitochondrial DNA is naked (no histones), so Tn5 hyperactively cuts it. Without nuclear-isolation steps (Omni-ATAC pre-spin, OR digitonin lysis with mt removal), chrM dominates the library.\n\n**Symptom:** `samtools idxstats sample.bam | awk '$1==\"chrM\"'` shows >50% of mapped reads on chrM.\n\n**Fix:** Re-prep with Omni-ATAC (Corces 2017) or fast-ATAC. Re-running QC on chrM-stripped BAM hides the underlying problem; the wasted sequencing remains. If chrM fraction is 30-50%, the library may still be salvageable via chrM removal but yield is reduced.\n\n### NRF \u002F PBC1 \u002F PBC2 below threshold\n\n**Trigger:** Over-amplified library; low input cell count combined with high PCR cycles.\n\n**Mechanism:** Each PCR cycle doubles starting fragments. With low complexity input (\u003C5000 cells) and >12 cycles, distinct fragments saturate and reads pile up at identical positions. NRF measures unique fragments \u002F total; PBC2 specifically detects multi-copy duplication.\n\n**Symptom:** NRF \u003C 0.7; PBC2 \u003C 1.0; massive duplicate-removal loss in `samtools markdup`.\n\n**Fix:** No fix post-hoc. Re-prep with more starting cells and fewer PCR cycles. Note: ATAC has *legitimate* duplicates at hyperaccessible sites (Tn5 cuts identically there), so NRF \u003C 0.9 is not by itself fatal. The combined PBC1 \u003C 0.7 + PBC2 \u003C 1.0 + visual coverage pile-ups confirm true bottlenecking.\n\n### TSS enrichment \u003C 5\n\n**Trigger:** Generic chromatin opening throughout the genome (over-transposition), OR genome build mismatch between TSS BED and BAM, OR strand-flip in TSS file.\n\n**Mechanism:** TSS enrichment requires that signal at TSSs is >> signal in genomic flanks. Over-transposition flattens the signal landscape. Strand-flipped TSSs subtract real signal because TSSs on - strand are calculated from the wrong direction.\n\n**Symptom:** TSS profile is flat or shows a slight dip at TSS center. Genome browser shows accessibility everywhere, not concentrated at promoters.\n\n**Fix:** Verify genome build (mm10 vs mm39 differ in TSS positions); verify GTF strand column; confirm signal track was generated post-deduplication. If TSS profile is genuinely flat, library is over-transposed and not recoverable; lower transposition time \u002F Tn5 concentration in next prep.\n\n### FRiP \u003C 0.2\n\n**Trigger:** Signal too diffuse to call peaks (over-transposition), low TSS enrichment, OR peak set is too narrow \u002F restrictive.\n\n**Mechanism:** FRiP correlates with TSS enrichment because both measure how concentrated the signal is. A diffuse library will have low FRiP regardless of peak count.\n\n**Symptom:** Peak count looks normal but FRiP \u003C 0.15.\n\n**Fix:** Check TSS enrichment first. If TSS is also low, the library is over-transposed. If TSS is OK but FRiP is low, the peak caller may be undercalling -- try `-p 0.01` (looser) and recalculate FRiP.\n\n### Replicate correlation \u003C 0.85\n\n**Trigger:** Batch effect, technical artefact, or cell-state drift between replicate biological collections.\n\n**Mechanism:** Pearson correlation on log-scaled binned counts (deepTools `multiBamSummary bins -bs 10000`) tracks coverage similarity. Below 0.85 indicates non-trivial divergence; ENCODE wants >= 0.9 for biological reps.\n\n**Fix:** Check PCA; if reps cluster apart from condition, drop the outlier or rerun. If the divergence aligns with batch, add batch as a covariate downstream (DiffBind `~Batch + Condition`). Do not silently merge with bad correlation.\n\n## Library Complexity (NRF, PBC1, PBC2)\n\n**Goal:** Detect over-amplification or low-input bottlenecks.\n\n**Approach:** Hash mapped read positions (or fragment 5' coordinates), tally how many positions have 1, 2, or more reads, and compute the three metrics.\n\n```python\nimport pysam\nfrom collections import Counter\n\ndef library_complexity(bam):\n    pos_counts = Counter()\n    total = 0\n    with pysam.AlignmentFile(bam, 'rb') as bf:\n        for r in bf.fetch():\n            if r.is_unmapped or r.is_secondary or r.is_supplementary:\n                continue\n            if r.is_duplicate:                          # Mark, not skip; PBC counts pre-dedup\n                pass\n            total += 1\n            key = (r.reference_name, r.reference_start, r.is_reverse)\n            pos_counts[key] += 1\n    distinct = len(pos_counts)\n    histogram = Counter(pos_counts.values())            # {1: N1, 2: N2, ...}\n    n1 = histogram.get(1, 0)\n    n2 = histogram.get(2, 0)\n    nrf = distinct \u002F total if total else 0.0\n    pbc1 = n1 \u002F distinct if distinct else 0.0\n    pbc2 = n1 \u002F n2 if n2 else float('inf')\n    return {'NRF': nrf, 'PBC1': pbc1, 'PBC2': pbc2, 'total': total, 'distinct': distinct}\n```\n\n`r.is_duplicate` is informational only here; ENCODE NRF\u002FPBC are computed pre-deduplication on the raw mapped BAM.\n\n## Cross-Replicate QC\n\n```bash\n# Spearman correlation (more robust than Pearson for ATAC)\nmultiBamSummary bins -bs 10000 -p 8 \\\n    --bamfiles rep1.bam rep2.bam rep3.bam \\\n    -o multi.npz\n\nplotCorrelation -in multi.npz \\\n    --corMethod spearman --whatToPlot heatmap --skipZeros \\\n    -o spearman_heatmap.png\n\n# Fingerprint (per-bin signal cumulative -- diagonal = no enrichment, sharp curve = good)\nplotFingerprint -p 8 -b rep1.bam rep2.bam rep3.bam \\\n    --labels rep1 rep2 rep3 \\\n    --skipZeros --numberOfSamples 50000 \\\n    -o fingerprint.png \\\n    --outQualityMetrics fingerprint_metrics.txt\n```\n\ndeepTools fingerprint quality metrics report a synthetic JS distance without a reference; the (non-synthetic) Jensen-Shannon distance column is only computed when a reference sample is supplied via `--JSDsample`. Larger values indicate stronger enrichment.\n\n## Library Complexity Extrapolation (preseq)\n\n**Goal:** Predict whether re-sequencing would rescue a low-NRF library, separating \"library is bottlenecked\" from \"we just sequenced too shallow.\"\n\n**Approach:** Fit preseq's rational-function (Pade) approximation of the Good-Toulmin power-series estimator on observed BAM read positions; extrapolate distinct-fragment yield as a function of additional sequencing depth.\n\n```bash\n# c_curve: observed complexity at current depth\npreseq c_curve -B sample.bam -o sample.ccurve.tsv -s 1e6\n\n# lc_extrap: predicted complexity at higher depth (extrapolation -e here 200M; preseq default -e is 1e10, step -s default 1M)\npreseq lc_extrap -B sample.bam -o sample.lcextrap.tsv -e 200000000 -s 5000000\n```\n\nInterpretation: if `lc_extrap` shows distinct-fragment count flattening before 100M reads, the library is bottlenecked (re-sequencing won't help; re-prep needed). If it continues to climb, re-sequencing will recover more unique reads. Use alongside NRF\u002FPBC1\u002FPBC2 to decide library re-prep vs deeper sequencing.\n\n## Sex-Chromosome QC\n\n**Trigger:** Clinical-grade ATAC; biobank-scale studies; sample-mix-up detection.\n\n**Mechanism:** chrY has minimal coverage in female samples; XIST locus (chrX) is highly accessible only in female cells (X-inactivation). Sample-swap or sex-misassignment detectable from these two loci.\n\n```bash\n# chrY read fraction\nsamtools idxstats sample.bam | awk '$1==\"chrY\"{print $3 \u002F $2}'   # reads per bp\n\n# XIST locus accessibility (chrX:73820651-73852753 in hg38)\nsamtools view -c sample.bam chrX:73820651-73852753\n```\n\nFemale: chrY reads\u002Fbp ~0; XIST count high. Male: chrY reads\u002Fbp ~male coverage; XIST count low. Discrepancy with sample metadata flags swap.\n\n## Cell-Cycle Effect on Accessibility\n\n**Trigger:** Proliferating cell lines (K562, HEK293, HeLa); samples with high S\u002FG2M signature.\n\n**Mechanism:** Replication-associated chromatin opening adds 5-15% global accessibility shift in proliferating cells; without correction, condition-specific cell-cycle differences confound differential analysis.\n\n**Detection:** Score cells\u002Fsamples for S-phase signature (Macosko 2015 cell cycle gene set adapted for chromatin: regulated origin loci, replication-stress-response genes); for bulk ATAC, compute per-sample peak intersection with replication-origin atlas (Repli-seq peaks).\n\n**Fix for differential:** Add S-phase score as covariate in DESeq2 design (`~Sphase + Condition`); for scATAC, regress on TF-IDF residuals analogous to Seurat CellCycleScoring.\n\n## Spike-in QC (Drosophila or E. coli Chromatin)\n\n**Trigger:** Studies where global accessibility shift is biological (HDAC inhibitor, DNMT inhibitor, differentiation).\n\n**Mechanism:** Per-library normalization (RPM, CPM) erases global accessibility shifts because total reads are nominally constant. Exogenous chromatin spike-in (Drosophila S2 or E. coli Tn5-naive chromatin added pre-Tn5) provides an external scaling reference.\n\n**Pipeline:** Align reads to a concatenated human + Drosophila reference; count spike-in reads per sample; normalize by spike-in (not by total reads). Reske 2020 Epigenetics Chromatin shows that normalization-method choice materially changes differential-accessibility results when a global accessibility shift is expected (ARID1A\u002FPIK3CA endometrial-epithelium case study), motivating an external reference such as a chromatin spike-in.\n\n**QC threshold:** spike-in fraction 0.5-5% of total reads is the workable range. Below 0.1% spike-in is unreliable; above 10% suggests too much spike-in (loss of cellular reads).\n\n## Comprehensive QC Aggregation\n\n**Goal:** Produce a per-sample report card with PASS\u002FFAIL flags against ENCODE thresholds.\n\n**Approach:** Compute each metric independently, compare to thresholds, write a tab-delimited report consumable by MultiQC.\n\n```python\nimport json, subprocess, sys\nfrom pathlib import Path\n\nENCODE_THRESHOLDS = {\n    'nuclear_reads_M': (25, 50),                      # (min acceptable, ideal)\n    'mt_fraction': (0.5, 0.05),                       # (max acceptable, ideal); inverted\n    'NRF': (0.7, 0.9), 'PBC1': (0.7, 0.9), 'PBC2': (1.0, 3.0),\n    'TSS_enrichment': (5.0, 7.0), 'FRiP': (0.2, 0.3),\n}\n\ndef grade(value, thr_acceptable, thr_ideal, inverted=False):\n    if inverted:\n        return 'FAIL' if value > thr_acceptable else ('PASS' if value \u003C= thr_ideal else 'WARN')\n    return 'FAIL' if value \u003C thr_acceptable else ('PASS' if value >= thr_ideal else 'WARN')\n\ndef report(metrics, out_tsv):\n    rows = []\n    for k, (acc, ideal) in ENCODE_THRESHOLDS.items():\n        if k not in metrics: continue\n        inverted = (k == 'mt_fraction')\n        flag = grade(metrics[k], acc, ideal, inverted=inverted)\n        rows.append((k, metrics[k], acc, ideal, flag))\n    with open(out_tsv, 'w') as f:\n        f.write('metric\\tvalue\\tacceptable\\tideal\\tflag\\n')\n        for r in rows: f.write('\\t'.join(map(str, r)) + '\\n')\n```\n\n## MultiQC Aggregation\n\n```bash\n# Run after generating per-sample QC outputs\nmultiqc \\\n    fastqc\u002F \\\n    picard\u002F \\\n    samtools_stats\u002F \\\n    macs2\u002F \\\n    deeptools\u002F \\\n    -o multiqc_report\n```\n\nMultiQC ingests Picard CollectInsertSizeMetrics, samtools flagstat, deepTools plotFingerprint output, and MACS peaks tables. It does NOT compute TSS enrichment or NRF; pipe a custom `_mqc.tsv` for those.\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| TSS enrichment off by 3x from expected | Wrong implementation (ENCODE vs ATACseqQC) | State the formula; convert by recomputing |\n| NRF = 1.0 exactly | BAM was already deduplicated -> all positions distinct | Compute NRF on raw mapped BAM (pre-dedup) |\n| PBC2 = inf | No positions with 2 reads | Library is too sparse; PBC2 unreliable below ~5M reads |\n| Mt fraction reported but BAM has no `chrM` | Mitochondrial chromosome named `MT`, `Mt`, or `chromosome:MT` | Match `samtools idxstats` chromosome name to the filter |\n| Insert size distribution flat after Picard | Sample is single-end | Insert size only valid for paired-end; switch to deeptools fragmentSize |\n| Replicates correlate poorly but PCA looks fine | High background dominates correlation | Use `--skipZeros`; or compute correlation on peak counts only |\n| FRiP differs by 2x between identical pipeline runs | Peak set differs (q-value cutoff drift) | Pin caller version + cutoff; FRiP is peak-set-dependent |\n| TSS enrichment lower than expected on Omni-ATAC | Used standard TSS BED on FFPE-prepped sample | FFPE TSSs are degraded; use peak-based metric instead |\n\n## References\n\n- Buenrostro JD et al 2013 Nat Methods 10:1213 (ATAC-seq protocol; fragment-size periodicity)\n- Corces MR et al 2017 Nat Methods 14:959 (Omni-ATAC; mt fraction reduction protocol)\n- Landt SG et al 2012 Genome Res 22:1813 (ENCODE\u002FmodENCODE QC framework, NRF\u002FPBC definitions; the PBC1\u002FPBC2 split is a later ENCODE-pipeline refinement)\n- ENCODE 4 ATAC-seq Data Standards (encodeproject.org\u002Fatac-seq) -- canonical thresholds\n- Ou J et al 2018 BMC Genomics 19:169 (ATACseqQC R package; TSSEscore implementation)\n- Ramirez F et al 2016 Nucleic Acids Res 44:W160 (deepTools, plotFingerprint JSD)\n- Daley T & Smith AD 2013 Nat Methods 10:325 (preseq library-complexity extrapolation model; the lc_extrap re-sequencing decision)\n\n## Related Skills\n\n- atac-seq\u002Fatac-peak-calling - FRiP requires peaks; QC drives accept\u002Freject before calling\n- atac-seq\u002Fnucleosome-positioning - Fragment-size analysis\n- atac-seq\u002Fsingle-cell-atac - per-cell QC has different thresholds\n- read-qc\u002Fquality-reports - upstream FastQC\n- alignment-files\u002Fbam-statistics - samtools flagstat \u002F idxstats\n- alignment-files\u002Fduplicate-handling - dedup before NRF\u002FPBC computation\n","## Version Compatibility\n\nReference examples tested with: deepTools 3.5+, Picard 3.1+, samtools 1.19+, bedtools 2.31+, ATACseqQC 1.26+, pysam 0.22+, pyBigWig 0.3+, numpy 1.26+, pandas 2.2+, MultiQC 1.21+.\n\nBefore using code patterns, verify installed versions match. If versions differ:\n- Python: `pip show \u003Cpackage>` then `help(module.function)` to check signatures\n- R: `packageVersion('\u003Cpkg>')` then `?function_name` to verify parameters\n- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt.\n\n# ATAC-seq Quality Control\n\n**\"Does my ATAC library pass ENCODE quality criteria?\"** -> Compute the seven canonical metrics (depth, alignment rate, mitochondrial fraction, library complexity, fragment-size periodicity, TSS enrichment, FRiP) and compare against ENCODE 4 thresholds, then diagnose failures.\n\n- CLI: `picard CollectInsertSizeMetrics`, `samtools flagstat`, `samtools idxstats`\n- CLI: `deeptools plotFingerprint`, `computeMatrix reference-point` + `plotProfile`\n- R: `ATACseqQC::TSSEscore`, `ATACseqQC::fragSizeDist`, `ATACseqQC::PTscore`\n- Python: custom NRF\u002FPBC from coordinate hash; pyBigWig for TSS enrichment\n\n## ENCODE 4 ATAC-seq Acceptance Thresholds\n\n| Metric | Definition | Ideal | Acceptable | Reject | Source |\n|--------|-----------|-------|------------|--------|--------|\n| Nuclear reads (after dedup, no chrM) | Mapped, MAPQ >= 30, non-chrM, deduped | >= 50M | 25-50M | \u003C 25M | ENCODE 4 ATAC-seq Standards |\n| Alignment rate | Mapped \u002F total reads | >= 95% | 80-95% | \u003C 80% | ENCODE 4 |\n| Mitochondrial fraction | chrM \u002F total mapped | \u003C 5% (Omni-ATAC), \u003C 20% (standard) | 20-50% | > 50% | Corces 2017 (Omni-ATAC) |\n| NRF (Non-Redundant Fraction) | Distinct positions \u002F total reads | >= 0.9 | 0.7-0.9 | \u003C 0.7 | Landt 2012 |\n| PBC1 (PCR Bottlenecking Coefficient 1) | Positions w\u002F 1 read \u002F Positions w\u002F >= 1 read | >= 0.9 | 0.7-0.9 | \u003C 0.7 | Landt 2012 |\n| PBC2 | Positions w\u002F 1 read \u002F Positions w\u002F 2 reads | >= 3.0 | 1.0-3.0 | \u003C 1.0 | Landt 2012 |\n| TSS enrichment (hg38, GENCODE v29) | Avg signal at TSS \u002F avg flanking | >= 7 | 5-7 | \u003C 5 | ENCODE 4 |\n| FRiP (Fraction Reads in Peaks) | Reads in MACS peaks \u002F total | >= 0.3 | 0.2-0.3 | \u003C 0.2 | ENCODE 4, Landt 2012 |\n| Insert-size periodicity | NFR + mono-nuc + di-nuc peaks visible | Clear 3+ peaks | NFR + mono only | Flat \u002F single peak | Buenrostro 2013 |\n\nENCODE thresholds are organism-specific. Mouse (mm10, GENCODE M21) TSS enrichment >= 5 is acceptable; non-model organisms have no published threshold (use cohort percentile rank instead). Methodology evolves; verify against the current ENCODE ATAC-seq Standards before reporting.\n\n## TSS Enrichment: ENCODE Method vs ATACseqQC Method\n\nThe two most common implementations DO NOT produce identical scores.\n\n| Method | Numerator | Denominator | Scaling |\n|--------|-----------|-------------|---------|\n| ENCODE pyTSSe \u002F Kundaje gtsse | Mean signal in 100 bp window centered at TSS | Mean signal in 100 bp window at +\u002F- 1900 to +\u002F- 2000 bp (flanks) | Per-base normalization to flanks; reported as fold-enrichment |\n| ATACseqQC TSSEscore | Sum signal in TSS +\u002F- 100 bp | Sum signal at +\u002F- 1000 bp flanking windows | Different window sizes; ratios are larger |\n| deeptools plotProfile | Visual; numeric ratio not standardized | Reference-point matrix | No standard score; for visualization only |\n\n**Trigger:** Comparing a TSS score across studies.\n\n**Mechanism:** Different normalization windows shift the absolute number; ATACseqQC's TSSEscore is typically 2-3x ENCODE's because of the wider flank.\n\n**Symptom:** Reported score 21 vs ENCODE-ideal 7 mismatch. Likely the calculator was ATACseqQC; the equivalent ENCODE score might be 8.\n\n**Fix:** State which implementation was used. For ENCODE comparisons, use `pyTSSe` (Kundaje lab) or implement the ENCODE recipe directly.\n\n```python\nimport numpy as np\nimport pyBigWig\n\ndef encode_tss_enrichment(bw_path, tss_bed, flank=2000):\n    \"\"\"ENCODE-style TSS enrichment: signal at TSS center \u002F signal at flanks.\"\"\"\n    bw = pyBigWig.open(bw_path)\n    profiles = []\n    for line in open(tss_bed):\n        chrom, start, end, *rest = line.strip().split('\\t')\n        tss = int(start)\n        strand = rest[2] if len(rest) > 2 else '+'\n        try:\n            vals = bw.values(chrom, tss - flank, tss + flank)\n            if vals is None or len(vals) != 2 * flank: continue\n            if strand == '-': vals = vals[::-1]\n            profiles.append(np.nan_to_num(vals))\n        except RuntimeError:\n            continue\n    avg = np.nanmean(profiles, axis=0)\n    flank_signal = np.mean(np.concatenate([avg[:100], avg[-100:]]))\n    center_signal = np.mean(avg[flank - 50: flank + 50])\n    return center_signal \u002F flank_signal if flank_signal > 0 else 0.0\n```\n\n## Fragment-Size Periodicity Patterns\n\n| Pattern | Visual signature | Interpretation | Action |\n|---------|-----------------|----------------|--------|\n| Strong tri-modal | NFR (~50bp) >> mono (~200bp) > di (~400bp) > tri (~600bp) peaks | Excellent transposition; well-positioned chromatin | Pass |\n| Clear bi-modal | NFR + mono only, di and tri faint | Acceptable; common in Omni-ATAC | Pass |\n| Single broad peak | Flat after NFR or no NFR | Over-transposition (too much Tn5) OR degraded chromatin | Reject; cannot distinguish nucleosomes |\n| Inverted (mono >> NFR) | Mono peak dominant, NFR weak | Under-transposition OR chromatin condensation | Caution; peak counts will be low |\n| Sharp 147 bp spike with no flanks | Tight peak at 147 bp | ChIP-seq input contamination (MNase-like) | Reject; not ATAC-grade |\n| 10.4 bp helical periodicity overlay | Sub-peaks at 50, 60, 70, 80 bp on NFR | Excellent chromatin structure resolution; helical phasing visible | Pass; high-quality |\n\nThe 10.4 bp helical periodicity is a Buenrostro 2013 hallmark: it reflects the helical pitch of B-form DNA, with Tn5 preferring outward-facing minor grooves on nucleosomal DNA. Its presence is a positive QC indicator but not required.\n\n## Per-Metric Failure Modes\n\n### Mitochondrial fraction > 50%\n\n**Trigger:** Standard ATAC-seq protocol on intact cells (no nuclear isolation), or insufficient detergent in lysis.\n\n**Mechanism:** Mitochondrial DNA is naked (no histones), so Tn5 hyperactively cuts it. Without nuclear-isolation steps (Omni-ATAC pre-spin, OR digitonin lysis with mt removal), chrM dominates the library.\n\n**Symptom:** `samtools idxstats sample.bam | awk '$1==\"chrM\"'` shows >50% of mapped reads on chrM.\n\n**Fix:** Re-prep with Omni-ATAC (Corces 2017) or fast-ATAC. Re-running QC on chrM-stripped BAM hides the underlying problem; the wasted sequencing remains. If chrM fraction is 30-50%, the library may still be salvageable via chrM removal but yield is reduced.\n\n### NRF \u002F PBC1 \u002F PBC2 below threshold\n\n**Trigger:** Over-amplified library; low input cell count combined with high PCR cycles.\n\n**Mechanism:** Each PCR cycle doubles starting fragments. With low complexity input (\u003C5000 cells) and >12 cycles, distinct fragments saturate and reads pile up at identical positions. NRF measures unique fragments \u002F total; PBC2 specifically detects multi-copy duplication.\n\n**Symptom:** NRF \u003C 0.7; PBC2 \u003C 1.0; massive duplicate-removal loss in `samtools markdup`.\n\n**Fix:** No fix post-hoc. Re-prep with more starting cells and fewer PCR cycles. Note: ATAC has *legitimate* duplicates at hyperaccessible sites (Tn5 cuts identically there), so NRF \u003C 0.9 is not by itself fatal. The combined PBC1 \u003C 0.7 + PBC2 \u003C 1.0 + visual coverage pile-ups confirm true bottlenecking.\n\n### TSS enrichment \u003C 5\n\n**Trigger:** Generic chromatin opening throughout the genome (over-transposition), OR genome build mismatch between TSS BED and BAM, OR strand-flip in TSS file.\n\n**Mechanism:** TSS enrichment requires that signal at TSSs is >> signal in genomic flanks. Over-transposition flattens the signal landscape. Strand-flipped TSSs subtract real signal because TSSs on - strand are calculated from the wrong direction.\n\n**Symptom:** TSS profile is flat or shows a slight dip at TSS center. Genome browser shows accessibility everywhere, not concentrated at promoters.\n\n**Fix:** Verify genome build (mm10 vs mm39 differ in TSS positions); verify GTF strand column; confirm signal track was generated post-deduplication. If TSS profile is genuinely flat, library is over-transposed and not recoverable; lower transposition time \u002F Tn5 concentration in next prep.\n\n### FRiP \u003C 0.2\n\n**Trigger:** Signal too diffuse to call peaks (over-transposition), low TSS enrichment, OR peak set is too narrow \u002F restrictive.\n\n**Mechanism:** FRiP correlates with TSS enrichment because both measure how concentrated the signal is. A diffuse library will have low FRiP regardless of peak count.\n\n**Symptom:** Peak count looks normal but FRiP \u003C 0.15.\n\n**Fix:** Check TSS enrichment first. If TSS is also low, the library is over-transposed. If TSS is OK but FRiP is low, the peak caller may be undercalling -- try `-p 0.01` (looser) and recalculate FRiP.\n\n### Replicate correlation \u003C 0.85\n\n**Trigger:** Batch effect, technical artefact, or cell-state drift between replicate biological collections.\n\n**Mechanism:** Pearson correlation on log-scaled binned counts (deepTools `multiBamSummary bins -bs 10000`) tracks coverage similarity. Below 0.85 indicates non-trivial divergence; ENCODE wants >= 0.9 for biological reps.\n\n**Fix:** Check PCA; if reps cluster apart from condition, drop the outlier or rerun. If the divergence aligns with batch, add batch as a covariate downstream (DiffBind `~Batch + Condition`). Do not silently merge with bad correlation.\n\n## Library Complexity (NRF, PBC1, PBC2)\n\n**Goal:** Detect over-amplification or low-input bottlenecks.\n\n**Approach:** Hash mapped read positions (or fragment 5' coordinates), tally how many positions have 1, 2, or more reads, and compute the three metrics.\n\n```python\nimport pysam\nfrom collections import Counter\n\ndef library_complexity(bam):\n    pos_counts = Counter()\n    total = 0\n    with pysam.AlignmentFile(bam, 'rb') as bf:\n        for r in bf.fetch():\n            if r.is_unmapped or r.is_secondary or r.is_supplementary:\n                continue\n            if r.is_duplicate:                          # Mark, not skip; PBC counts pre-dedup\n                pass\n            total += 1\n            key = (r.reference_name, r.reference_start, r.is_reverse)\n            pos_counts[key] += 1\n    distinct = len(pos_counts)\n    histogram = Counter(pos_counts.values())            # {1: N1, 2: N2, ...}\n    n1 = histogram.get(1, 0)\n    n2 = histogram.get(2, 0)\n    nrf = distinct \u002F total if total else 0.0\n    pbc1 = n1 \u002F distinct if distinct else 0.0\n    pbc2 = n1 \u002F n2 if n2 else float('inf')\n    return {'NRF': nrf, 'PBC1': pbc1, 'PBC2': pbc2, 'total': total, 'distinct': distinct}\n```\n\n`r.is_duplicate` is informational only here; ENCODE NRF\u002FPBC are computed pre-deduplication on the raw mapped BAM.\n\n## Cross-Replicate QC\n\n```bash\n# Spearman correlation (more robust than Pearson for ATAC)\nmultiBamSummary bins -bs 10000 -p 8 \\\n    --bamfiles rep1.bam rep2.bam rep3.bam \\\n    -o multi.npz\n\nplotCorrelation -in multi.npz \\\n    --corMethod spearman --whatToPlot heatmap --skipZeros \\\n    -o spearman_heatmap.png\n\n# Fingerprint (per-bin signal cumulative -- diagonal = no enrichment, sharp curve = good)\nplotFingerprint -p 8 -b rep1.bam rep2.bam rep3.bam \\\n    --labels rep1 rep2 rep3 \\\n    --skipZeros --numberOfSamples 50000 \\\n    -o fingerprint.png \\\n    --outQualityMetrics fingerprint_metrics.txt\n```\n\ndeepTools fingerprint quality metrics report a synthetic JS distance without a reference; the (non-synthetic) Jensen-Shannon distance column is only computed when a reference sample is supplied via `--JSDsample`. Larger values indicate stronger enrichment.\n\n## Library Complexity Extrapolation (preseq)\n\n**Goal:** Predict whether re-sequencing would rescue a low-NRF library, separating \"library is bottlenecked\" from \"we just sequenced too shallow.\"\n\n**Approach:** Fit preseq's rational-function (Pade) approximation of the Good-Toulmin power-series estimator on observed BAM read positions; extrapolate distinct-fragment yield as a function of additional sequencing depth.\n\n```bash\n# c_curve: observed complexity at current depth\npreseq c_curve -B sample.bam -o sample.ccurve.tsv -s 1e6\n\n# lc_extrap: predicted complexity at higher depth (extrapolation -e here 200M; preseq default -e is 1e10, step -s default 1M)\npreseq lc_extrap -B sample.bam -o sample.lcextrap.tsv -e 200000000 -s 5000000\n```\n\nInterpretation: if `lc_extrap` shows distinct-fragment count flattening before 100M reads, the library is bottlenecked (re-sequencing won't help; re-prep needed). If it continues to climb, re-sequencing will recover more unique reads. Use alongside NRF\u002FPBC1\u002FPBC2 to decide library re-prep vs deeper sequencing.\n\n## Sex-Chromosome QC\n\n**Trigger:** Clinical-grade ATAC; biobank-scale studies; sample-mix-up detection.\n\n**Mechanism:** chrY has minimal coverage in female samples; XIST locus (chrX) is highly accessible only in female cells (X-inactivation). Sample-swap or sex-misassignment detectable from these two loci.\n\n```bash\n# chrY read fraction\nsamtools idxstats sample.bam | awk '$1==\"chrY\"{print $3 \u002F $2}'   # reads per bp\n\n# XIST locus accessibility (chrX:73820651-73852753 in hg38)\nsamtools view -c sample.bam chrX:73820651-73852753\n```\n\nFemale: chrY reads\u002Fbp ~0; XIST count high. Male: chrY reads\u002Fbp ~male coverage; XIST count low. Discrepancy with sample metadata flags swap.\n\n## Cell-Cycle Effect on Accessibility\n\n**Trigger:** Proliferating cell lines (K562, HEK293, HeLa); samples with high S\u002FG2M signature.\n\n**Mechanism:** Replication-associated chromatin opening adds 5-15% global accessibility shift in proliferating cells; without correction, condition-specific cell-cycle differences confound differential analysis.\n\n**Detection:** Score cells\u002Fsamples for S-phase signature (Macosko 2015 cell cycle gene set adapted for chromatin: regulated origin loci, replication-stress-response genes); for bulk ATAC, compute per-sample peak intersection with replication-origin atlas (Repli-seq peaks).\n\n**Fix for differential:** Add S-phase score as covariate in DESeq2 design (`~Sphase + Condition`); for scATAC, regress on TF-IDF residuals analogous to Seurat CellCycleScoring.\n\n## Spike-in QC (Drosophila or E. coli Chromatin)\n\n**Trigger:** Studies where global accessibility shift is biological (HDAC inhibitor, DNMT inhibitor, differentiation).\n\n**Mechanism:** Per-library normalization (RPM, CPM) erases global accessibility shifts because total reads are nominally constant. Exogenous chromatin spike-in (Drosophila S2 or E. coli Tn5-naive chromatin added pre-Tn5) provides an external scaling reference.\n\n**Pipeline:** Align reads to a concatenated human + Drosophila reference; count spike-in reads per sample; normalize by spike-in (not by total reads). Reske 2020 Epigenetics Chromatin shows that normalization-method choice materially changes differential-accessibility results when a global accessibility shift is expected (ARID1A\u002FPIK3CA endometrial-epithelium case study), motivating an external reference such as a chromatin spike-in.\n\n**QC threshold:** spike-in fraction 0.5-5% of total reads is the workable range. Below 0.1% spike-in is unreliable; above 10% suggests too much spike-in (loss of cellular reads).\n\n## Comprehensive QC Aggregation\n\n**Goal:** Produce a per-sample report card with PASS\u002FFAIL flags against ENCODE thresholds.\n\n**Approach:** Compute each metric independently, compare to thresholds, write a tab-delimited report consumable by MultiQC.\n\n```python\nimport json, subprocess, sys\nfrom pathlib import Path\n\nENCODE_THRESHOLDS = {\n    'nuclear_reads_M': (25, 50),                      # (min acceptable, ideal)\n    'mt_fraction': (0.5, 0.05),                       # (max acceptable, ideal); inverted\n    'NRF': (0.7, 0.9), 'PBC1': (0.7, 0.9), 'PBC2': (1.0, 3.0),\n    'TSS_enrichment': (5.0, 7.0), 'FRiP': (0.2, 0.3),\n}\n\ndef grade(value, thr_acceptable, thr_ideal, inverted=False):\n    if inverted:\n        return 'FAIL' if value > thr_acceptable else ('PASS' if value \u003C= thr_ideal else 'WARN')\n    return 'FAIL' if value \u003C thr_acceptable else ('PASS' if value >= thr_ideal else 'WARN')\n\ndef report(metrics, out_tsv):\n    rows = []\n    for k, (acc, ideal) in ENCODE_THRESHOLDS.items():\n        if k not in metrics: continue\n        inverted = (k == 'mt_fraction')\n        flag = grade(metrics[k], acc, ideal, inverted=inverted)\n        rows.append((k, metrics[k], acc, ideal, flag))\n    with open(out_tsv, 'w') as f:\n        f.write('metric\\tvalue\\tacceptable\\tideal\\tflag\\n')\n        for r in rows: f.write('\\t'.join(map(str, r)) + '\\n')\n```\n\n## MultiQC Aggregation\n\n```bash\n# Run after generating per-sample QC outputs\nmultiqc \\\n    fastqc\u002F \\\n    picard\u002F \\\n    samtools_stats\u002F \\\n    macs2\u002F \\\n    deeptools\u002F \\\n    -o multiqc_report\n```\n\nMultiQC ingests Picard CollectInsertSizeMetrics, samtools flagstat, deepTools plotFingerprint output, and MACS peaks tables. It does NOT compute TSS enrichment or NRF; pipe a custom `_mqc.tsv` for those.\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| TSS enrichment off by 3x from expected | Wrong implementation (ENCODE vs ATACseqQC) | State the formula; convert by recomputing |\n| NRF = 1.0 exactly | BAM was already deduplicated -> all positions distinct | Compute NRF on raw mapped BAM (pre-dedup) |\n| PBC2 = inf | No positions with 2 reads | Library is too sparse; PBC2 unreliable below ~5M reads |\n| Mt fraction reported but BAM has no `chrM` | Mitochondrial chromosome named `MT`, `Mt`, or `chromosome:MT` | Match `samtools idxstats` chromosome name to the filter |\n| Insert size distribution flat after Picard | Sample is single-end | Insert size only valid for paired-end; switch to deeptools fragmentSize |\n| Replicates correlate poorly but PCA looks fine | High background dominates correlation | Use `--skipZeros`; or compute correlation on peak counts only |\n| FRiP differs by 2x between identical pipeline runs | Peak set differs (q-value cutoff drift) | Pin caller version + cutoff; FRiP is peak-set-dependent |\n| TSS enrichment lower than expected on Omni-ATAC | Used standard TSS BED on FFPE-prepped sample | FFPE TSSs are degraded; use peak-based metric instead |\n\n## References\n\n- Buenrostro JD et al 2013 Nat Methods 10:1213 (ATAC-seq protocol; fragment-size periodicity)\n- Corces MR et al 2017 Nat Methods 14:959 (Omni-ATAC; mt fraction reduction protocol)\n- Landt SG et al 2012 Genome Res 22:1813 (ENCODE\u002FmodENCODE QC framework, NRF\u002FPBC definitions; the PBC1\u002FPBC2 split is a later ENCODE-pipeline refinement)\n- ENCODE 4 ATAC-seq Data Standards (encodeproject.org\u002Fatac-seq) -- canonical thresholds\n- Ou J et al 2018 BMC Genomics 19:169 (ATACseqQC R package; TSSEscore implementation)\n- Ramirez F et al 2016 Nucleic Acids Res 44:W160 (deepTools, plotFingerprint JSD)\n- Daley T & Smith AD 2013 Nat Methods 10:325 (preseq library-complexity extrapolation model; the lc_extrap re-sequencing decision)\n\n## Related Skills\n\n- atac-seq\u002Fatac-peak-calling - FRiP requires peaks; QC drives accept\u002Freject before calling\n- atac-seq\u002Fnucleosome-positioning - Fragment-size analysis\n- atac-seq\u002Fsingle-cell-atac - per-cell QC has different thresholds\n- read-qc\u002Fquality-reports - upstream FastQC\n- alignment-files\u002Fbam-statistics - samtools flagstat \u002F idxstats\n- alignment-files\u002Fduplicate-handling - dedup before NRF\u002FPBC computation\n",{"name":5,"source":24,"license":15,"tool_type":25,"description":7,"imported_at":26,"source_path":27,"primary_tool":28,"source_repository":29},"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills","mixed","2026-09-20","atac-seq\u002Fatac-qc\u002FSKILL.md","deeptools","GPTomics\u002FbioSkills","main","2d1750717747097b602acf30bd09a7cbd4b4c6ea","awesome_skills\u002Fbiology\u002Fbio-atac-seq-atac-qc","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous","2026-11-09 11:45:36",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-atac-seq-atac-qc",{"tool":42,"command":43},"bunx","bunx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-atac-seq-atac-qc",{"tool":45,"command":46},"pnpm","pnpm dlx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-atac-seq-atac-qc","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-atac-seq-atac-qc\u002FSKILL.md",true,[50,56],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"examples\u002Fatac_qc_metrics.R","OTHER","其他",3402,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-atac-seq-atac-qc\u002Fexamples\u002Fatac_qc_metrics.R",{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"usage-guide.md","REFERENCE","参考文档",6087,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-atac-seq-atac-qc\u002Fusage-guide.md",[63,71,78,85,92,99],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":70},"01M7FFFZPC2025JQQHPHDX154J","dei-assessment","Assess diversity, equity, and inclusion with metrics. TRIGGERS - Use when user needs help with dei-assessment related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fdei-assessment\u002FSKILL.md",[],[],"2026-11-16 10:05:34",{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":70},"01M7FFFWN8H4WA7A6Y398QKB6J","dei-action-plan","Create DEI action plans with goals and initiatives. TRIGGERS - Use when user needs help with dei-action-plan related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fdei-action-plan\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":70},"01M7FFFSHTM0QZRVZ2NRJTJ0PY","culture-fit-assessment","Design culture add assessment processes. TRIGGERS - Use when user needs help with culture-fit-assessment related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fculture-fit-assessment\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":70},"01M7FFFMA12RKEQA38KHJSXXPS","cowork-to-code-bridge","Use an already-installed, independently verified cowork-to-code bridge to run narrowly approved actions on the user's own macOS, Linux, or WSL2 machine through a local file queue.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fcowork-to-code-bridge\u002FSKILL.md",[],[],{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":70},"01M7FFFH1F6M8YRGW5V487GH1K","counter-offer-response","Draft counter-offer responses for retention. TRIGGERS - Use when user needs help with counter-offer-response related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fcounter-offer-response\u002FSKILL.md",[],[],{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":103,"categories":104,"keywords":105,"publishedAt":70},"01M7FFFDGEHV8Y7J9N40HZXW52","continuous-feedback-system","Design continuous feedback systems. TRIGGERS - Use when user needs help with continuous-feedback-system related tasks.","https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fhr-people\u002Fcontinuous-feedback-system\u002FSKILL.md",[],[]]