[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EBJ60KF7PST6Z7JZJ6FHSQ":3,"skill-same:01M7EBJ60KF7PST6Z7JZJ6FHSQ":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},"01M7EBJ60KF7PST6Z7JZJ6FHSQ","bio-atac-seq-footprinting",null,"Detect transcription factor binding footprints in ATAC-seq using TOBIAS, HINT-ATAC, Wellington, or scprinter. Use when identifying bound TF sites within accessible regions, correcting Tn5 insertion bias before footprinting, choosing between cleavage-based and aggregate-based footprinters, or comparing differential TF activity between conditions.","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-footprinting\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: bio-atac-seq-footprinting\ndescription: Detect transcription factor binding footprints in ATAC-seq using TOBIAS, HINT-ATAC, Wellington, or scprinter. Use when identifying bound TF sites within accessible regions, correcting Tn5 insertion bias before footprinting, choosing between cleavage-based and aggregate-based footprinters, or comparing differential TF activity between conditions.\ntool_type: cli\nprimary_tool: tobias\nsource: \"https:\u002F\u002Fgithub.com\u002FGPTomics\u002FbioSkills\"\nsource_repository: \"GPTomics\u002FbioSkills\"\nsource_path: \"atac-seq\u002Ffootprinting\u002FSKILL.md\"\nlicense: \"MIT\"\nimported_at: \"2026-09-20\"\n---\n\n## Version Compatibility\n\nReference examples tested with: TOBIAS 0.16+, RGT HINT-ATAC 1.0.2+, Wellington (pyDNase) 0.3+, scprinter 0.1+, samtools 1.19+, bedtools 2.31+, pyBigWig 0.3+, MEME suite 5.5+.\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- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws unexpected errors, introspect the installed package and adapt rather than retrying.\n\n# TF Footprinting\n\n**\"Identify TF binding footprints in my ATAC-seq data\"** -> Detect short DNA stretches (typically 6-20 bp) of reduced Tn5 cleavage within accessible regions, where a bound TF physically protects DNA. Requires (1) Tn5 sequence-bias correction, (2) per-base footprint scoring, (3) motif-anchored detection.\n\n- CLI: `TOBIAS ATACorrect` -> `TOBIAS ScoreBigwig` (formerly `FootprintScores`) -> `TOBIAS BINDetect`\n- CLI: `rgt-hint footprinting --atac-seq` (HINT-ATAC, single-step)\n- CLI: `wellington_footprints.py` (legacy DNase, adapted for ATAC)\n- Python: `scprinter` (multi-scale, single-cell aware; Hu 2025 Nature)\n\nTn5 has a strong sequence preference (Karabacak Calviello 2019), reading approximately +\u002F- 4 bp around the insertion site. Without bias correction, \"footprints\" reflect Tn5 sequence preference rather than TF binding. This is the single most important step.\n\n## Algorithmic Taxonomy\n\n| Tool | Bias model | Scoring | Min depth | Strength | Fails when |\n|------|-----------|---------|-----------|----------|------------|\n| TOBIAS (BINDetect) | +\u002F-12 bp k-mer window (`--k_flank` 12), dinucleotide weight matrix (DWM) | Two-step: continuous footprint score then motif-anchored bound\u002Funbound classification | >= 50M nuclear reads | Mature, peer-reviewed (Bentsen 2020), differential support, modular pipeline | Below 50M reads; sequencing errors near motif inflate background |\n| HINT-ATAC | Hidden-Markov + dinucleotide bias correction | HMM emits open\u002Ffootprint\u002Fclosed states; calls ranked footprints | >= 50M | Single-step; integrates motif matching; handles DNase too | Less control over individual stages; HMM occasionally over-segments |\n| Wellington (pyDNase) | DNase-developed; ATAC adaptation by post-shift | Cleavage-rate Poisson Z-score | >= 50M (DNase >= 80M) | Original footprinting framework; well-validated for DNase | Designed for DNase II; ATAC-specific bias not corrected as carefully |\n| PIQ | Bayesian latent variable on cut sites | Genome-wide PWM scan + cleavage profile | >= 30M (lower because of model) | Per-TF posterior probabilities; works on lower depth | Outdated; not actively maintained; harder to install |\n| scprinter | Multi-scale CNN-based footprint and TF activity | Resolves footprints at multiple TF size scales (CTCF vs nuclear receptors) | >= 1M cells (sc) or 50M (bulk) | Modern ML approach; single-cell; multi-scale resolves problematic TF families | Newer tool; benchmarks evolving; GPU recommended |\n| TOBIAS + scprinter combination | TOBIAS bias correction + scprinter scoring | Two-step bridging | >= 50M | Combines the best bias model with multi-scale scoring | Manual pipeline, no single CLI |\n\nMethodology evolves; verify against the current Bentsen 2020, Karabacak Calviello 2019, and scPrinter (Hu 2025) benchmarks. ATAC footprinting power saturates above 100M nuclear reads; below 50M, weak-binding TFs (transient occupancy) cannot be reliably called.\n\n## Tn5 Bias and Why Correction Matters\n\n**Trigger:** Tn5 inserts preferentially at certain k-mers (Karabacak Calviello 2019 measured the protocol-specific 6-mer insertion-bias model used for bias correction). The preference is reproducible and biologically uninteresting.\n\n**Mechanism:** Without correction, every \"TF footprint\" near a high-bias k-mer reads as occupancy. Conversely, regions with low-bias flanks but real binding may show no footprint dip relative to corrected expectation.\n\n**Symptom:** Aggregate footprint at random GC-rich motifs shows V-shape; aggregate at AT-rich motifs shows inverse-V (peak instead of dip).\n\n**Fix:** Apply ATACorrect (TOBIAS), seqOutBias (Martins 2018), or HINT's dinucleotide model. Bias correction subtracts the Tn5-expected per-base profile from observed cleavage. After correction, V-shape is preserved only at TF-bound sites.\n\n**Goal:** Subtract Tn5 sequence-bias from the per-base cleavage signal so residual footprints reflect TF binding rather than enzyme preference.\n\n**Approach:** Run TOBIAS ATACorrect over the deduplicated BAM with the reference genome, consensus peaks, and ENCODE blacklist; it emits per-condition uncorrected, bias, expected, and corrected bigWigs for downstream scoring.\n\n```bash\n# TOBIAS ATACorrect: produces uncorrected, bias, expected, and corrected bigWigs\nTOBIAS ATACorrect \\\n    --bam sample.dedup.bam \\\n    --genome hg38.fa \\\n    --peaks consensus_peaks.bed \\\n    --blacklist hg38-blacklist.v2.bed \\\n    --outdir corrected\u002F \\\n    --cores 8\n# Output: sample_uncorrected.bw, sample_bias.bw, sample_expected.bw, sample_corrected.bw\n```\n\n## Tn5 Cut Geometry: +4 \u002F -5 Dual-Cut\n\nTn5 dimers cut both strands of DNA but with a 9 bp staggered offset. The cleavage event creates two free 5' ends: one shifted +4 bp from the binding center on the forward strand and -5 bp on the reverse strand. Footprinting tools must apply this shift before per-base counting:\n\n| Strand | Read 5' end correction |\n|--------|------------------------|\n| + strand | shift +4 bp downstream |\n| - strand | shift -5 bp upstream |\n\n**Trigger:** Computing per-base Tn5 cut signal manually.\n\n**Symptom:** Footprint aggregates show ~9 bp asymmetry (apex shifted from motif center).\n\n**Fix:** Apply +4\u002F-5 shift before counting; TOBIAS, HINT-ATAC, and scprinter handle this internally. Custom analyses must apply explicitly. deepTools provides this via `alignmentSieve --ATACshift` (which applies the canonical +4 \u002F -5 shift in one step).\n\n## Bias Correction Alternatives\n\n| Method | Approach | When to use |\n|--------|---------|-------------|\n| TOBIAS ATACorrect | +\u002F-12 bp k-mer window, dinucleotide weight matrix (DWM) | Default for most ATAC; fast |\n| chromBPNet bias model (Pampari 2024) | CNN trained on naked-DNA control or k-mer baseline | Best when sequence context complex; handles low-complexity flanks |\n| seqOutBias (Martins 2018) | Genome-wide k-mer frequency scaling (observed vs expected cut counts) | Independent of footprinting tool; works upstream |\n| HINT-ATAC dinucleotide | HMM-integrated dinucleotide bias | Built into HINT pipeline; less control |\n| Naked-DNA empirical | Sequence Tn5 on protein-free DNA | Gold standard for non-model organisms; expensive wet-lab |\n\nFor non-model organisms (no published Tn5 bias model), naked-DNA control is required. chromBPNet's bias model is the modern standard for human\u002Fmouse and outperforms TOBIAS at low-complexity sequence contexts (Pampari 2024). See atac-seq\u002Fdeep-learning-atac.\n\n## In Silico Variant Effect at Footprinted TF Motifs\n\n**Trigger:** A GWAS-fine-mapped or rare variant falls inside a TOBIAS-bound motif site.\n\n**Mechanism:** Sequence-based DL models (chromBPNet, Enformer) predict per-base accessibility at ref vs alt allele; combined with footprint evidence (TOBIAS bound site overlap), this produces a mechanistic hypothesis: \"variant disrupts binding of TF X at enhancer Y.\"\n\n**Workflow:** Run TOBIAS BINDetect to identify bound motif sites; for variants in bound sites, score with chromBPNet (atac-seq\u002Fdeep-learning-atac) for ref\u002Falt log2FC; |log2FC| > 1 supports functional disruption. Cross-reference with allele-specific accessibility (atac-seq\u002Fallele-specific-accessibility) for observed evidence.\n\n## Per-TF Footprinting Failure Modes\n\nDifferent TF families produce different footprint signatures. The same tool can report a clean V-shape for one family and noise for another.\n\n### CTCF -- The gold standard\n\n**Trigger:** Footprinting CTCF.\n\n**Mechanism:** CTCF has high ChIP-seq concordance, deep V-shaped footprint (~19-20 bp protected), strong sequence specificity. ChIP-seq overlap is typically >70%.\n\n**Symptom:** Aggregate corrected footprint shows clean ~20 bp dip with bilateral cleavage shoulders.\n\n**Verification:** Always validate footprinting output by checking CTCF first; if CTCF footprint is shallow, the bias correction or depth is the problem, not the biology.\n\n### Nuclear receptors (ER, AR, GR) -- Transient binding\n\n**Trigger:** Glucocorticoid response, hormone-stimulated systems.\n\n**Mechanism:** Steroid receptors bind transiently (residence time minutes vs hours for CTCF); average ATAC sample captures binding probability \u003C 30% per allele.\n\n**Symptom:** Aggregate footprint is shallow or absent despite ChIP-seq peaks at the same sites.\n\n**Fix:** Use scprinter's multi-scale model OR limit to ChIP-validated sites OR pool replicates for higher effective depth. Do not interpret absence of footprint as absence of binding.\n\n### Pioneer TFs (FOXA1, GATA, OCT4) -- Half-site footprint\n\n**Trigger:** Pioneer-factor binding to nucleosomal DNA.\n\n**Mechanism:** Pioneer factors bind one DNA face; the back face is on the histone octamer. Footprint is asymmetric (one side protected, other side accessible).\n\n**Symptom:** Aggregate plot shows asymmetric V; one shoulder is taller than the other.\n\n**Fix:** Use single-stranded scoring; HINT-ATAC has stranded mode. Treat asymmetric footprints as biologically meaningful, not artefactual.\n\n### AP-1 family (FOS, JUN) -- Heterodimer composite footprint\n\n**Trigger:** AP-1 enrichment; the JASPAR motif is composite of multiple heterodimer combinations.\n\n**Mechanism:** Different AP-1 dimers (FOS+JUN, FOS+JUNB, JUNB+JUNB) bind slightly different motifs. JASPAR entries are degenerate; scoring averages over all.\n\n**Fix:** Use specific HOCOMOCO motifs per heterodimer when distinguishing matters. Otherwise accept the composite call.\n\n### ZBTB family \u002F BTB-zinc finger -- Dynamic \u002F unfootprintable\n\n**Trigger:** Footprinting ZBTB16, BCL6, others.\n\n**Mechanism:** Dynamic binding kinetics + cofactor-mediated stabilization mean steady-state occupancy is highly variable. Some ZBTBs simply do not produce reliable ATAC footprints despite genuine ChIP-seq binding.\n\n**Fix:** Document the failure; use ChIP-seq for these TFs. Footprinting cannot rescue everything.\n\n### Forkhead \u002F homeobox (FOX, HOX) -- Short footprint \u003C 8 bp\n\n**Trigger:** Short-motif TFs.\n\n**Mechanism:** Footprint extent matches motif length; \u003C8 bp footprints are at the resolution limit of Tn5 (which has ~4 bp positional uncertainty).\n\n**Fix:** Multi-scale scoring (scprinter); aggregate over thousands of sites; do not rely on per-site calls for short motifs.\n\n## Decision Tree by Goal\n\n| Goal | Recommended pipeline |\n|------|---------------------|\n| Identify all TFs differentially bound between two conditions | TOBIAS ATACorrect (per condition) -> ScoreBigwig -> BINDetect with `--cond_names` |\n| Find the strongest single-TF binding (e.g., CTCF) | TOBIAS PlotAggregate over JASPAR CTCF motif sites; verify V-shape |\n| Per-cell footprinting (scATAC) | scprinter (single-cell mode); avoid TOBIAS unless pseudobulking by cluster |\n| Multi-scale TF activity (handle short and long simultaneously) | scprinter or TOBIAS + custom multi-scale | \n| Differential nuclear-receptor binding | TOBIAS pooled-replicate footprints + ChIP cross-validation; raw ATAC alone often misses transient binding |\n| Plant \u002F non-model organism | TOBIAS or HINT-ATAC with custom motifs; bias model retrained from genomic background |\n\n## TOBIAS Three-Step Pipeline\n\n**Goal:** Call bound\u002Funbound TF motif sites per condition and detect differential occupancy across two conditions.\n\n**Approach:** Run ATACorrect to subtract Tn5 bias from cleavage counts, ScoreBigwig to compute a continuous per-base footprint score, then BINDetect to anchor footprints to motif positions and produce per-TF differential bound calls with p-values.\n\n```bash\n# Step 1: Bias correction\nTOBIAS ATACorrect \\\n    --bam cond1.bam --genome hg38.fa \\\n    --peaks consensus.bed --blacklist hg38-blacklist.v2.bed \\\n    --outdir cond1_corrected\u002F --cores 16\n\n# Step 2: Per-base footprint scoring (continuous)\nTOBIAS ScoreBigwig \\\n    --signal cond1_corrected\u002Fcond1_corrected.bw \\\n    --regions consensus.bed \\\n    --output cond1_footprints.bw \\\n    --cores 16\n\n# Step 3: Motif-anchored bound\u002Funbound calls + differential\nTOBIAS BINDetect \\\n    --motifs JASPAR2024_CORE_vertebrates.pfm \\\n    --signals cond1_footprints.bw cond2_footprints.bw \\\n    --genome hg38.fa --peaks consensus.bed \\\n    --outdir bindetect\u002F \\\n    --cond_names cond1 cond2 \\\n    --cores 16\n```\n\nBINDetect output columns: `output_prefix`, motif info, condition counts (`cond1_bound`, `cond2_bound`), `cond1_mean_score`, `cond2_mean_score`, `cond1_cond2_change` (differential), `cond1_cond2_pvalue` (one-sided per direction).\n\n## Differential Reading\n\n| BINDetect output | Interpretation |\n|------------------|----------------|\n| `cond1_cond2_change` > 0, low pvalue | TF more bound in cond1 |\n| `cond1_cond2_change` \u003C 0, low pvalue | TF more bound in cond2 |\n| Both `cond1_bound` and `cond2_bound` near 0 | Motif present but no footprint either condition; TF likely not active |\n| `cond1_bound` >> `cond2_bound` but change small | High dynamic range; differential per-site rather than aggregate |\n\nThe differential score is the difference in mean footprint score across motif sites, not a fold-change. Magnitudes around 0.1-0.5 are typical for biologically relevant changes (TOBIAS BINDetect tutorials \u002F Bentsen 2020 examples; no formally published cutoff -- calibrate against positive controls in the current dataset).\n\n## Reconciling TOBIAS vs HINT-ATAC\n\n| Pattern | Likely cause | Action |\n|---------|--------------|--------|\n| TOBIAS calls binding, HINT does not | TOBIAS more sensitive; HINT's HMM filters edge calls | Trust if motif is canonical; suspect for novel\u002Fweak motifs |\n| HINT calls binding, TOBIAS does not | HINT's HMM occasionally over-segments and reports spurious | Verify by aggregate footprint at the called sites |\n| Both call same TF as differential but opposite directions | Different bias correction model; different bound\u002Funbound thresholds | Re-check ATACorrect output; one bias model may be miscalibrated |\n| Both flat | Library too shallow; chromatin too closed at motif sites | Pool replicates; consider scprinter multi-scale |\n\n**Operational rule:** For high-confidence reporting, require two-tool concordance (TOBIAS + HINT-ATAC OR TOBIAS + ChIP-seq overlap > 50%). Single-tool calls should be reported as exploratory.\n\n## NFR-Filtering Before Footprinting\n\n**Goal:** Restrict footprinting input to nucleosome-free (sub-100 bp) fragments where TF binding signal lives.\n\n**Approach:** Stream the BAM through awk, keep header lines and fragments whose insert size is between -100 and 100 bp, then re-index.\n\n```bash\n# Filter to fragments \u003C 100 bp (NFR) -- TF binding lives here, not on nucleosomes\nsamtools view -h sample.bam | \\\n    awk 'substr($0,1,1)==\"@\" || ($9 > 0 && $9 \u003C 100) || ($9 \u003C 0 && $9 > -100)' | \\\n    samtools view -b > sample.nfr.bam\nsamtools index sample.nfr.bam\n# Use this NFR BAM as input to TOBIAS ATACorrect\n```\n\nFiltering NFR strengthens footprint signal but discards di-nucleosome-borne information. Keep the unfiltered BAM for nucleosome-positioning analysis.\n\n## Motif Database Choice\n\n| Database | Coverage | Format | Notes |\n|----------|---------|--------|-------|\n| JASPAR 2024 CORE vertebrates | ~880 vertebrate non-redundant motifs (curated, experimentally derived) | JASPAR PFM, MEME, etc. | Default for vertebrate ATAC |\n| HOCOMOCO v12 | ~1443 curated motifs (v12 CORE) | JASPAR PFM | Best for resolving paralogues; provides secondary motif subtypes per TF |\n| CIS-BP 2.0 | ~80,000 motifs across 1000+ species | PWM, .meme | Broadest coverage including non-model species |\n| MEME-CHIP \u002F homer | Custom from peaks | .meme, .motif | When de novo motif needed |\n\nJASPAR motifs are conservatively curated; HOCOMOCO is comprehensive for human\u002Fmouse with quality scores per motif (A\u002FB\u002FC\u002FD); CIS-BP excels for non-model organisms.\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| Aggregate footprint inverted (peak instead of dip) | No bias correction; or wrong genome FASTA | Run ATACorrect; verify FASTA matches BAM build |\n| BINDetect reports zero bound sites | Default cutoff too stringent; or peakset too narrow | Raise `--bound-pvalue` (default 0.001) and inspect; verify peaks include where binding expected |\n| TOBIAS ATACorrect out of memory | Genome FASTA huge or many cores | Reduce `--cores`; use `samtools faidx` to confirm FASTA index exists |\n| Differential score noisy \u002F random | Per-condition bias correction inconsistent | Re-run ATACorrect with identical peakset and blacklist for each |\n| Empty motif file warning | JASPAR PFM format mismatch | Use `MEME suite` to convert; TOBIAS expects JASPAR format |\n| HINT-ATAC reports many tiny footprints | Default HMM over-segments | Use `--organism` flag explicitly; check `--region-file` is consensus, not raw peaks |\n| Wellington crashes on paired-end ATAC | Wellington was DNase-targeted, single-end model | Use TOBIAS instead, or convert paired-end to cuts-only BED |\n| Per-site footprint is V-shape but aggregate is flat | Mixing strands; some motifs on - strand | Aggregate function should handle strand; verify input motif strand column |\n\n## References\n\n- Buenrostro JD et al 2013 Nat Methods 10:1213 (ATAC-seq protocol)\n- Karabacak Calviello A et al 2019 Genome Biol 20:42 (protocol-specific Tn5\u002FDNase bias modeling for footprinting)\n- Bentsen M et al 2020 Nat Commun 11:4267 (TOBIAS framework, benchmark)\n- Li Z et al 2019 Genome Biol 20:45 (HINT-ATAC)\n- Piper J et al 2013 NAR 41:e201 (Wellington \u002F pyDNase)\n- Sherwood RI et al 2014 Nat Biotechnol 32:171 (PIQ)\n- Hu Y et al 2025 Nature 638:779 (scPrinter\u002FPRINT; multi-scale single-cell footprinting)\n- Martins AL et al 2018 NAR 46:e9 (seqOutBias bias correction alternative)\n- Castro-Mondragon JA et al 2022 NAR 50:D165 (JASPAR 2022, recently 2024)\n- Vorontsov IE et al 2024 NAR 52:D154 (HOCOMOCO v12)\n\n## Related Skills\n\n- atac-seq\u002Fatac-peak-calling - Generate input peakset (NFR-only optional)\n- atac-seq\u002Fatac-qc - Confirm depth >= 50M before footprinting\n- atac-seq\u002Fsingle-cell-atac - scprinter for single-cell footprinting\n- atac-seq\u002Fmotif-deviation - Complementary chromVAR for accessibility variability\n- atac-seq\u002Fdeep-learning-atac - chromBPNet bias correction alternative; in silico variant effect at footprints\n- atac-seq\u002Fallele-specific-accessibility - Observed allelic imbalance at TF-bound sites\n- chip-seq\u002Fpeak-annotation - Cross-validate footprints with ChIP peaks\n- sequence-manipulation\u002Fmotif-search - Underlying motif scanning patterns\n- gene-regulatory-networks\u002Fscenic-regulons - Downstream regulatory network inference\n","## Version Compatibility\n\nReference examples tested with: TOBIAS 0.16+, RGT HINT-ATAC 1.0.2+, Wellington (pyDNase) 0.3+, scprinter 0.1+, samtools 1.19+, bedtools 2.31+, pyBigWig 0.3+, MEME suite 5.5+.\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- CLI: `\u003Ctool> --version` then `\u003Ctool> --help` to confirm flags\n\nIf code throws unexpected errors, introspect the installed package and adapt rather than retrying.\n\n# TF Footprinting\n\n**\"Identify TF binding footprints in my ATAC-seq data\"** -> Detect short DNA stretches (typically 6-20 bp) of reduced Tn5 cleavage within accessible regions, where a bound TF physically protects DNA. Requires (1) Tn5 sequence-bias correction, (2) per-base footprint scoring, (3) motif-anchored detection.\n\n- CLI: `TOBIAS ATACorrect` -> `TOBIAS ScoreBigwig` (formerly `FootprintScores`) -> `TOBIAS BINDetect`\n- CLI: `rgt-hint footprinting --atac-seq` (HINT-ATAC, single-step)\n- CLI: `wellington_footprints.py` (legacy DNase, adapted for ATAC)\n- Python: `scprinter` (multi-scale, single-cell aware; Hu 2025 Nature)\n\nTn5 has a strong sequence preference (Karabacak Calviello 2019), reading approximately +\u002F- 4 bp around the insertion site. Without bias correction, \"footprints\" reflect Tn5 sequence preference rather than TF binding. This is the single most important step.\n\n## Algorithmic Taxonomy\n\n| Tool | Bias model | Scoring | Min depth | Strength | Fails when |\n|------|-----------|---------|-----------|----------|------------|\n| TOBIAS (BINDetect) | +\u002F-12 bp k-mer window (`--k_flank` 12), dinucleotide weight matrix (DWM) | Two-step: continuous footprint score then motif-anchored bound\u002Funbound classification | >= 50M nuclear reads | Mature, peer-reviewed (Bentsen 2020), differential support, modular pipeline | Below 50M reads; sequencing errors near motif inflate background |\n| HINT-ATAC | Hidden-Markov + dinucleotide bias correction | HMM emits open\u002Ffootprint\u002Fclosed states; calls ranked footprints | >= 50M | Single-step; integrates motif matching; handles DNase too | Less control over individual stages; HMM occasionally over-segments |\n| Wellington (pyDNase) | DNase-developed; ATAC adaptation by post-shift | Cleavage-rate Poisson Z-score | >= 50M (DNase >= 80M) | Original footprinting framework; well-validated for DNase | Designed for DNase II; ATAC-specific bias not corrected as carefully |\n| PIQ | Bayesian latent variable on cut sites | Genome-wide PWM scan + cleavage profile | >= 30M (lower because of model) | Per-TF posterior probabilities; works on lower depth | Outdated; not actively maintained; harder to install |\n| scprinter | Multi-scale CNN-based footprint and TF activity | Resolves footprints at multiple TF size scales (CTCF vs nuclear receptors) | >= 1M cells (sc) or 50M (bulk) | Modern ML approach; single-cell; multi-scale resolves problematic TF families | Newer tool; benchmarks evolving; GPU recommended |\n| TOBIAS + scprinter combination | TOBIAS bias correction + scprinter scoring | Two-step bridging | >= 50M | Combines the best bias model with multi-scale scoring | Manual pipeline, no single CLI |\n\nMethodology evolves; verify against the current Bentsen 2020, Karabacak Calviello 2019, and scPrinter (Hu 2025) benchmarks. ATAC footprinting power saturates above 100M nuclear reads; below 50M, weak-binding TFs (transient occupancy) cannot be reliably called.\n\n## Tn5 Bias and Why Correction Matters\n\n**Trigger:** Tn5 inserts preferentially at certain k-mers (Karabacak Calviello 2019 measured the protocol-specific 6-mer insertion-bias model used for bias correction). The preference is reproducible and biologically uninteresting.\n\n**Mechanism:** Without correction, every \"TF footprint\" near a high-bias k-mer reads as occupancy. Conversely, regions with low-bias flanks but real binding may show no footprint dip relative to corrected expectation.\n\n**Symptom:** Aggregate footprint at random GC-rich motifs shows V-shape; aggregate at AT-rich motifs shows inverse-V (peak instead of dip).\n\n**Fix:** Apply ATACorrect (TOBIAS), seqOutBias (Martins 2018), or HINT's dinucleotide model. Bias correction subtracts the Tn5-expected per-base profile from observed cleavage. After correction, V-shape is preserved only at TF-bound sites.\n\n**Goal:** Subtract Tn5 sequence-bias from the per-base cleavage signal so residual footprints reflect TF binding rather than enzyme preference.\n\n**Approach:** Run TOBIAS ATACorrect over the deduplicated BAM with the reference genome, consensus peaks, and ENCODE blacklist; it emits per-condition uncorrected, bias, expected, and corrected bigWigs for downstream scoring.\n\n```bash\n# TOBIAS ATACorrect: produces uncorrected, bias, expected, and corrected bigWigs\nTOBIAS ATACorrect \\\n    --bam sample.dedup.bam \\\n    --genome hg38.fa \\\n    --peaks consensus_peaks.bed \\\n    --blacklist hg38-blacklist.v2.bed \\\n    --outdir corrected\u002F \\\n    --cores 8\n# Output: sample_uncorrected.bw, sample_bias.bw, sample_expected.bw, sample_corrected.bw\n```\n\n## Tn5 Cut Geometry: +4 \u002F -5 Dual-Cut\n\nTn5 dimers cut both strands of DNA but with a 9 bp staggered offset. The cleavage event creates two free 5' ends: one shifted +4 bp from the binding center on the forward strand and -5 bp on the reverse strand. Footprinting tools must apply this shift before per-base counting:\n\n| Strand | Read 5' end correction |\n|--------|------------------------|\n| + strand | shift +4 bp downstream |\n| - strand | shift -5 bp upstream |\n\n**Trigger:** Computing per-base Tn5 cut signal manually.\n\n**Symptom:** Footprint aggregates show ~9 bp asymmetry (apex shifted from motif center).\n\n**Fix:** Apply +4\u002F-5 shift before counting; TOBIAS, HINT-ATAC, and scprinter handle this internally. Custom analyses must apply explicitly. deepTools provides this via `alignmentSieve --ATACshift` (which applies the canonical +4 \u002F -5 shift in one step).\n\n## Bias Correction Alternatives\n\n| Method | Approach | When to use |\n|--------|---------|-------------|\n| TOBIAS ATACorrect | +\u002F-12 bp k-mer window, dinucleotide weight matrix (DWM) | Default for most ATAC; fast |\n| chromBPNet bias model (Pampari 2024) | CNN trained on naked-DNA control or k-mer baseline | Best when sequence context complex; handles low-complexity flanks |\n| seqOutBias (Martins 2018) | Genome-wide k-mer frequency scaling (observed vs expected cut counts) | Independent of footprinting tool; works upstream |\n| HINT-ATAC dinucleotide | HMM-integrated dinucleotide bias | Built into HINT pipeline; less control |\n| Naked-DNA empirical | Sequence Tn5 on protein-free DNA | Gold standard for non-model organisms; expensive wet-lab |\n\nFor non-model organisms (no published Tn5 bias model), naked-DNA control is required. chromBPNet's bias model is the modern standard for human\u002Fmouse and outperforms TOBIAS at low-complexity sequence contexts (Pampari 2024). See atac-seq\u002Fdeep-learning-atac.\n\n## In Silico Variant Effect at Footprinted TF Motifs\n\n**Trigger:** A GWAS-fine-mapped or rare variant falls inside a TOBIAS-bound motif site.\n\n**Mechanism:** Sequence-based DL models (chromBPNet, Enformer) predict per-base accessibility at ref vs alt allele; combined with footprint evidence (TOBIAS bound site overlap), this produces a mechanistic hypothesis: \"variant disrupts binding of TF X at enhancer Y.\"\n\n**Workflow:** Run TOBIAS BINDetect to identify bound motif sites; for variants in bound sites, score with chromBPNet (atac-seq\u002Fdeep-learning-atac) for ref\u002Falt log2FC; |log2FC| > 1 supports functional disruption. Cross-reference with allele-specific accessibility (atac-seq\u002Fallele-specific-accessibility) for observed evidence.\n\n## Per-TF Footprinting Failure Modes\n\nDifferent TF families produce different footprint signatures. The same tool can report a clean V-shape for one family and noise for another.\n\n### CTCF -- The gold standard\n\n**Trigger:** Footprinting CTCF.\n\n**Mechanism:** CTCF has high ChIP-seq concordance, deep V-shaped footprint (~19-20 bp protected), strong sequence specificity. ChIP-seq overlap is typically >70%.\n\n**Symptom:** Aggregate corrected footprint shows clean ~20 bp dip with bilateral cleavage shoulders.\n\n**Verification:** Always validate footprinting output by checking CTCF first; if CTCF footprint is shallow, the bias correction or depth is the problem, not the biology.\n\n### Nuclear receptors (ER, AR, GR) -- Transient binding\n\n**Trigger:** Glucocorticoid response, hormone-stimulated systems.\n\n**Mechanism:** Steroid receptors bind transiently (residence time minutes vs hours for CTCF); average ATAC sample captures binding probability \u003C 30% per allele.\n\n**Symptom:** Aggregate footprint is shallow or absent despite ChIP-seq peaks at the same sites.\n\n**Fix:** Use scprinter's multi-scale model OR limit to ChIP-validated sites OR pool replicates for higher effective depth. Do not interpret absence of footprint as absence of binding.\n\n### Pioneer TFs (FOXA1, GATA, OCT4) -- Half-site footprint\n\n**Trigger:** Pioneer-factor binding to nucleosomal DNA.\n\n**Mechanism:** Pioneer factors bind one DNA face; the back face is on the histone octamer. Footprint is asymmetric (one side protected, other side accessible).\n\n**Symptom:** Aggregate plot shows asymmetric V; one shoulder is taller than the other.\n\n**Fix:** Use single-stranded scoring; HINT-ATAC has stranded mode. Treat asymmetric footprints as biologically meaningful, not artefactual.\n\n### AP-1 family (FOS, JUN) -- Heterodimer composite footprint\n\n**Trigger:** AP-1 enrichment; the JASPAR motif is composite of multiple heterodimer combinations.\n\n**Mechanism:** Different AP-1 dimers (FOS+JUN, FOS+JUNB, JUNB+JUNB) bind slightly different motifs. JASPAR entries are degenerate; scoring averages over all.\n\n**Fix:** Use specific HOCOMOCO motifs per heterodimer when distinguishing matters. Otherwise accept the composite call.\n\n### ZBTB family \u002F BTB-zinc finger -- Dynamic \u002F unfootprintable\n\n**Trigger:** Footprinting ZBTB16, BCL6, others.\n\n**Mechanism:** Dynamic binding kinetics + cofactor-mediated stabilization mean steady-state occupancy is highly variable. Some ZBTBs simply do not produce reliable ATAC footprints despite genuine ChIP-seq binding.\n\n**Fix:** Document the failure; use ChIP-seq for these TFs. Footprinting cannot rescue everything.\n\n### Forkhead \u002F homeobox (FOX, HOX) -- Short footprint \u003C 8 bp\n\n**Trigger:** Short-motif TFs.\n\n**Mechanism:** Footprint extent matches motif length; \u003C8 bp footprints are at the resolution limit of Tn5 (which has ~4 bp positional uncertainty).\n\n**Fix:** Multi-scale scoring (scprinter); aggregate over thousands of sites; do not rely on per-site calls for short motifs.\n\n## Decision Tree by Goal\n\n| Goal | Recommended pipeline |\n|------|---------------------|\n| Identify all TFs differentially bound between two conditions | TOBIAS ATACorrect (per condition) -> ScoreBigwig -> BINDetect with `--cond_names` |\n| Find the strongest single-TF binding (e.g., CTCF) | TOBIAS PlotAggregate over JASPAR CTCF motif sites; verify V-shape |\n| Per-cell footprinting (scATAC) | scprinter (single-cell mode); avoid TOBIAS unless pseudobulking by cluster |\n| Multi-scale TF activity (handle short and long simultaneously) | scprinter or TOBIAS + custom multi-scale | \n| Differential nuclear-receptor binding | TOBIAS pooled-replicate footprints + ChIP cross-validation; raw ATAC alone often misses transient binding |\n| Plant \u002F non-model organism | TOBIAS or HINT-ATAC with custom motifs; bias model retrained from genomic background |\n\n## TOBIAS Three-Step Pipeline\n\n**Goal:** Call bound\u002Funbound TF motif sites per condition and detect differential occupancy across two conditions.\n\n**Approach:** Run ATACorrect to subtract Tn5 bias from cleavage counts, ScoreBigwig to compute a continuous per-base footprint score, then BINDetect to anchor footprints to motif positions and produce per-TF differential bound calls with p-values.\n\n```bash\n# Step 1: Bias correction\nTOBIAS ATACorrect \\\n    --bam cond1.bam --genome hg38.fa \\\n    --peaks consensus.bed --blacklist hg38-blacklist.v2.bed \\\n    --outdir cond1_corrected\u002F --cores 16\n\n# Step 2: Per-base footprint scoring (continuous)\nTOBIAS ScoreBigwig \\\n    --signal cond1_corrected\u002Fcond1_corrected.bw \\\n    --regions consensus.bed \\\n    --output cond1_footprints.bw \\\n    --cores 16\n\n# Step 3: Motif-anchored bound\u002Funbound calls + differential\nTOBIAS BINDetect \\\n    --motifs JASPAR2024_CORE_vertebrates.pfm \\\n    --signals cond1_footprints.bw cond2_footprints.bw \\\n    --genome hg38.fa --peaks consensus.bed \\\n    --outdir bindetect\u002F \\\n    --cond_names cond1 cond2 \\\n    --cores 16\n```\n\nBINDetect output columns: `output_prefix`, motif info, condition counts (`cond1_bound`, `cond2_bound`), `cond1_mean_score`, `cond2_mean_score`, `cond1_cond2_change` (differential), `cond1_cond2_pvalue` (one-sided per direction).\n\n## Differential Reading\n\n| BINDetect output | Interpretation |\n|------------------|----------------|\n| `cond1_cond2_change` > 0, low pvalue | TF more bound in cond1 |\n| `cond1_cond2_change` \u003C 0, low pvalue | TF more bound in cond2 |\n| Both `cond1_bound` and `cond2_bound` near 0 | Motif present but no footprint either condition; TF likely not active |\n| `cond1_bound` >> `cond2_bound` but change small | High dynamic range; differential per-site rather than aggregate |\n\nThe differential score is the difference in mean footprint score across motif sites, not a fold-change. Magnitudes around 0.1-0.5 are typical for biologically relevant changes (TOBIAS BINDetect tutorials \u002F Bentsen 2020 examples; no formally published cutoff -- calibrate against positive controls in the current dataset).\n\n## Reconciling TOBIAS vs HINT-ATAC\n\n| Pattern | Likely cause | Action |\n|---------|--------------|--------|\n| TOBIAS calls binding, HINT does not | TOBIAS more sensitive; HINT's HMM filters edge calls | Trust if motif is canonical; suspect for novel\u002Fweak motifs |\n| HINT calls binding, TOBIAS does not | HINT's HMM occasionally over-segments and reports spurious | Verify by aggregate footprint at the called sites |\n| Both call same TF as differential but opposite directions | Different bias correction model; different bound\u002Funbound thresholds | Re-check ATACorrect output; one bias model may be miscalibrated |\n| Both flat | Library too shallow; chromatin too closed at motif sites | Pool replicates; consider scprinter multi-scale |\n\n**Operational rule:** For high-confidence reporting, require two-tool concordance (TOBIAS + HINT-ATAC OR TOBIAS + ChIP-seq overlap > 50%). Single-tool calls should be reported as exploratory.\n\n## NFR-Filtering Before Footprinting\n\n**Goal:** Restrict footprinting input to nucleosome-free (sub-100 bp) fragments where TF binding signal lives.\n\n**Approach:** Stream the BAM through awk, keep header lines and fragments whose insert size is between -100 and 100 bp, then re-index.\n\n```bash\n# Filter to fragments \u003C 100 bp (NFR) -- TF binding lives here, not on nucleosomes\nsamtools view -h sample.bam | \\\n    awk 'substr($0,1,1)==\"@\" || ($9 > 0 && $9 \u003C 100) || ($9 \u003C 0 && $9 > -100)' | \\\n    samtools view -b > sample.nfr.bam\nsamtools index sample.nfr.bam\n# Use this NFR BAM as input to TOBIAS ATACorrect\n```\n\nFiltering NFR strengthens footprint signal but discards di-nucleosome-borne information. Keep the unfiltered BAM for nucleosome-positioning analysis.\n\n## Motif Database Choice\n\n| Database | Coverage | Format | Notes |\n|----------|---------|--------|-------|\n| JASPAR 2024 CORE vertebrates | ~880 vertebrate non-redundant motifs (curated, experimentally derived) | JASPAR PFM, MEME, etc. | Default for vertebrate ATAC |\n| HOCOMOCO v12 | ~1443 curated motifs (v12 CORE) | JASPAR PFM | Best for resolving paralogues; provides secondary motif subtypes per TF |\n| CIS-BP 2.0 | ~80,000 motifs across 1000+ species | PWM, .meme | Broadest coverage including non-model species |\n| MEME-CHIP \u002F homer | Custom from peaks | .meme, .motif | When de novo motif needed |\n\nJASPAR motifs are conservatively curated; HOCOMOCO is comprehensive for human\u002Fmouse with quality scores per motif (A\u002FB\u002FC\u002FD); CIS-BP excels for non-model organisms.\n\n## Common Errors\n\n| Error \u002F symptom | Cause | Solution |\n|-----------------|-------|----------|\n| Aggregate footprint inverted (peak instead of dip) | No bias correction; or wrong genome FASTA | Run ATACorrect; verify FASTA matches BAM build |\n| BINDetect reports zero bound sites | Default cutoff too stringent; or peakset too narrow | Raise `--bound-pvalue` (default 0.001) and inspect; verify peaks include where binding expected |\n| TOBIAS ATACorrect out of memory | Genome FASTA huge or many cores | Reduce `--cores`; use `samtools faidx` to confirm FASTA index exists |\n| Differential score noisy \u002F random | Per-condition bias correction inconsistent | Re-run ATACorrect with identical peakset and blacklist for each |\n| Empty motif file warning | JASPAR PFM format mismatch | Use `MEME suite` to convert; TOBIAS expects JASPAR format |\n| HINT-ATAC reports many tiny footprints | Default HMM over-segments | Use `--organism` flag explicitly; check `--region-file` is consensus, not raw peaks |\n| Wellington crashes on paired-end ATAC | Wellington was DNase-targeted, single-end model | Use TOBIAS instead, or convert paired-end to cuts-only BED |\n| Per-site footprint is V-shape but aggregate is flat | Mixing strands; some motifs on - strand | Aggregate function should handle strand; verify input motif strand column |\n\n## References\n\n- Buenrostro JD et al 2013 Nat Methods 10:1213 (ATAC-seq protocol)\n- Karabacak Calviello A et al 2019 Genome Biol 20:42 (protocol-specific Tn5\u002FDNase bias modeling for footprinting)\n- Bentsen M et al 2020 Nat Commun 11:4267 (TOBIAS framework, benchmark)\n- Li Z et al 2019 Genome Biol 20:45 (HINT-ATAC)\n- Piper J et al 2013 NAR 41:e201 (Wellington \u002F pyDNase)\n- Sherwood RI et al 2014 Nat Biotechnol 32:171 (PIQ)\n- Hu Y et al 2025 Nature 638:779 (scPrinter\u002FPRINT; multi-scale single-cell footprinting)\n- Martins AL et al 2018 NAR 46:e9 (seqOutBias bias correction alternative)\n- Castro-Mondragon JA et al 2022 NAR 50:D165 (JASPAR 2022, recently 2024)\n- Vorontsov IE et al 2024 NAR 52:D154 (HOCOMOCO v12)\n\n## Related Skills\n\n- atac-seq\u002Fatac-peak-calling - Generate input peakset (NFR-only optional)\n- atac-seq\u002Fatac-qc - Confirm depth >= 50M before footprinting\n- atac-seq\u002Fsingle-cell-atac - scprinter for single-cell footprinting\n- atac-seq\u002Fmotif-deviation - Complementary chromVAR for accessibility variability\n- atac-seq\u002Fdeep-learning-atac - chromBPNet bias correction alternative; in silico variant effect at footprints\n- atac-seq\u002Fallele-specific-accessibility - Observed allelic imbalance at TF-bound sites\n- chip-seq\u002Fpeak-annotation - Cross-validate footprints with ChIP peaks\n- sequence-manipulation\u002Fmotif-search - Underlying motif scanning patterns\n- gene-regulatory-networks\u002Fscenic-regulons - Downstream regulatory network inference\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","cli","2026-09-20","atac-seq\u002Ffootprinting\u002FSKILL.md","tobias","GPTomics\u002FbioSkills","main","2d1750717747097b602acf30bd09a7cbd4b4c6ea","awesome_skills\u002Fbiology\u002Fbio-atac-seq-footprinting","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-footprinting",{"tool":42,"command":43},"bunx","bunx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-atac-seq-footprinting",{"tool":45,"command":46},"pnpm","pnpm dlx skills add Mr-Nobody-Anonymous\u002FAll-skills --skill bio-atac-seq-footprinting","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-footprinting\u002FSKILL.md",true,[50,56],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"examples\u002Frun_tobias.sh","SCRIPT","脚本",2967,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-atac-seq-footprinting\u002Fexamples\u002Frun_tobias.sh",{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"usage-guide.md","REFERENCE","参考文档",7203,"https:\u002F\u002Fgithub.com\u002FMr-Nobody-Anonymous\u002FAll-skills\u002Fblob\u002Fmain\u002Fawesome_skills\u002Fbiology\u002Fbio-atac-seq-footprinting\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",[],[]]