Skip to main content
ClaudeWave
Skill2.7k estrellas del repoactualizado 2mo ago

bio-atac-seq-atac-peak-calling

This skill performs peak calling on ATAC-seq data using MACS3 with parameters optimized for identifying open chromatin regions from aligned BAM files. Use it when analyzing accessible chromatin from ATAC-seq experiments, as it applies ATAC-specific settings like Tn5 shift correction, paired-end mode, and duplicate retention that differ from standard ChIP-seq peak calling approaches.

Instalar en Claude Code
Copiar
git clone --depth 1 https://github.com/FreedomIntelligence/OpenClaw-Medical-Skills /tmp/bio-atac-seq-atac-peak-calling && cp -r /tmp/bio-atac-seq-atac-peak-calling/skills/bio-atac-seq-atac-peak-calling ~/.claude/skills/bio-atac-seq-atac-peak-calling
Después abre una sesión nueva de Claude Code; el skill carga automáticamente.

SKILL.md

## Version Compatibility

Reference examples tested with: Bowtie2 2.5.3+, MACS3 3.0+, samtools 1.19+

Before using code patterns, verify installed versions match. If versions differ:
- CLI: `<tool> --version` then `<tool> --help` to confirm flags

If code throws ImportError, AttributeError, or TypeError, introspect the installed
package and adapt the example to match the actual API rather than retrying.

# ATAC-seq Peak Calling

**"Call peaks from my ATAC-seq data"** → Identify open chromatin regions using ATAC-specific parameters (no input control, shifted Tn5 cut sites, paired-end mode).
- CLI: `macs3 callpeak -t atac.bam -f BAMPE -g hs --nomodel --shift -75 --extsize 150`

## Basic MACS3 for ATAC-seq

**Goal:** Identify open chromatin regions from ATAC-seq data using ATAC-specific peak calling parameters.

**Approach:** Run MACS3 in paired-end mode with Tn5 shift correction, no model building, and duplicate retention since ATAC-seq generates natural duplicates at accessible sites.

```bash
# Standard ATAC-seq peak calling
macs3 callpeak \
    -t sample.bam \
    -f BAMPE \
    -g hs \
    -n sample \
    --outdir peaks/ \
    -q 0.05 \
    --nomodel \
    --shift -75 \
    --extsize 150 \
    --keep-dup all \
    -B
```

## Key ATAC-seq Parameters

```bash
# Explained parameters
macs3 callpeak \
    -t sample.bam \        # Treatment BAM
    -f BAMPE \             # Paired-end BAM (uses fragment size)
    -g hs \                # Genome size: hs (human), mm (mouse)
    -n sample \            # Output name prefix
    --nomodel \            # Don't build shifting model
    --shift -75 \          # Shift reads to center on Tn5 cut site
    --extsize 150 \        # Extend reads to this size
    --keep-dup all \       # Keep duplicates (ATAC has natural duplicates)
    -B \                   # Generate bedGraph for visualization
    --call-summits         # Call peak summits
```

## Why These Parameters?

| Parameter | Reason |
|-----------|--------|
| --nomodel | ATAC doesn't have control, can't build model |
| --shift -75 | Centers on Tn5 insertion site |
| --extsize 150 | Smooths signal around cut sites |
| --keep-dup all | Tn5 creates duplicate cuts at accessible sites |
| -f BAMPE | Uses actual fragment size from paired-end |

## Paired-End vs Single-End

```bash
# Paired-end (recommended for ATAC)
macs3 callpeak -f BAMPE -t sample.bam ...

# Single-end (less common)
macs3 callpeak -f BAM -t sample.bam \
    --nomodel --shift -75 --extsize 150 ...
```

## Call Peaks on NFR Only

**Goal:** Call peaks using only nucleosome-free fragments for sharper regulatory element detection.

**Approach:** Filter BAM to fragments <100 bp (NFR), then call peaks with adjusted shift/extsize parameters matching the shorter fragment size.

```bash
# First, filter to nucleosome-free reads (<100bp fragments)
samtools view -h sample.bam | \
    awk 'substr($0,1,1)=="@" || ($9>0 && $9<100) || ($9<0 && $9>-100)' | \
    samtools view -b > nfr.bam

# Call peaks on NFR
macs3 callpeak \
    -t nfr.bam \
    -f BAMPE \
    -g hs \
    -n sample_nfr \
    --nomodel \
    --shift -37 \
    --extsize 75 \
    --keep-dup all \
    -q 0.01
```

## Broad Peaks (Optional)

```bash
# For broader accessible regions
macs3 callpeak \
    -t sample.bam \
    -f BAMPE \
    -g hs \
    -n sample_broad \
    --nomodel \
    --shift -75 \
    --extsize 150 \
    --broad \
    --broad-cutoff 0.1
```

## Batch Processing

**Goal:** Call peaks on multiple ATAC-seq samples in one pass.

**Approach:** Loop over BAM files and run MACS3 with consistent ATAC-specific parameters for each sample.

```bash
#!/bin/bash
GENOME=hs  # hs for human, mm for mouse
OUTDIR=peaks

mkdir -p $OUTDIR

for bam in *.bam; do
    sample=$(basename $bam .bam)
    echo "Processing $sample..."

    macs3 callpeak \
        -t $bam \
        -f BAMPE \
        -g $GENOME \
        -n $sample \
        --outdir $OUTDIR \
        --nomodel \
        --shift -75 \
        --extsize 150 \
        --keep-dup all \
        -q 0.05 \
        -B \
        --call-summits
done
```

## Output Files

| File | Description |
|------|-------------|
| _peaks.narrowPeak | Peak locations (BED-like) |
| _summits.bed | Peak summit positions |
| _peaks.xls | Peak statistics (Excel format) |
| _treat_pileup.bdg | Signal track (bedGraph) |
| _control_lambda.bdg | Background (if control provided) |

## narrowPeak Format

```
chr1  100  500  peak1  500  .  10.5  50.2  45.1  200
```

Columns: chrom, start, end, name, score, strand, signalValue, pValue, qValue, summit_offset

## Convert to BigWig

```bash
# Sort bedGraph
sort -k1,1 -k2,2n sample_treat_pileup.bdg > sample.sorted.bdg

# Convert to BigWig
bedGraphToBigWig sample.sorted.bdg chrom.sizes sample.bw
```

## Merge Replicates

```bash
# Pool BAMs before peak calling (recommended for final peaks)
samtools merge -@ 8 merged.bam rep1.bam rep2.bam rep3.bam

# Call peaks on merged
macs3 callpeak -t merged.bam -f BAMPE -g hs -n merged ...
```

## IDR for Replicate Consistency

**Goal:** Identify reproducible peaks across biological replicates using the Irreproducible Discovery Rate framework.

**Approach:** Call peaks on each replicate independently, then run IDR to score peak reproducibility and filter to a high-confidence set.

```bash
# Call peaks on each replicate
macs3 callpeak -t rep1.bam -f BAMPE -g hs -n rep1 ...
macs3 callpeak -t rep2.bam -f BAMPE -g hs -n rep2 ...

# Run IDR
idr --samples rep1_peaks.narrowPeak rep2_peaks.narrowPeak \
    --input-file-type narrowPeak \
    --output-file idr_peaks.txt \
    --plot

# Filter by IDR threshold
awk '$5 >= 540' idr_peaks.txt > reproducible_peaks.bed
```

## Related Skills

- read-alignment/bowtie2-alignment - Align ATAC-seq reads
- atac-seq/atac-qc - Quality control
- chip-seq/peak-calling - ChIP-seq comparison
- genome-intervals/bed-file-basics - Work with peak files
aav-vector-design-agentSkill
adaptyvSkill

Cloud laboratory platform for automated protein testing and validation. Use when designing proteins and needing experimental validation including binding assays, expression testing, thermostability measurements, enzyme activity assays, or protein sequence optimization. Also use for submitting experiments via API, tracking experiment status, downloading results, optimizing protein sequences for better expression using computational tools (NetSolP, SoluProt, SolubleMPNN, ESM), or managing protein design workflows with wet-lab validation.

adhd-daily-plannerSkill

Time-blind friendly planning, executive function support, and daily structure for ADHD brains. Specializes in realistic time estimation, dopamine-aware task design, and building systems that

aeonSkill

This skill should be used for time series machine learning tasks including classification, regression, clustering, forecasting, anomaly detection, segmentation, and similarity search. Use when working with temporal data, sequential patterns, or time-indexed observations requiring specialized algorithms beyond standard ML approaches. Particularly suited for univariate and multivariate time series analysis with scikit-learn compatible APIs.

agent-browserSkill

Browse the web for any task — research topics, read articles, interact with web apps, fill forms, take screenshots, extract data, and test web pages. Use whenever a browser would be useful, not just when the user explicitly asks.

agentd-drug-discoverySkill
ai-analyzerSkill

AI驱动的综合健康分析系统,整合多维度健康数据、识别异常模式、预测健康风险、提供个性化建议。支持智能问答和AI健康报告生成。

alphafold-databaseSkill

Access AlphaFold's 200M+ AI-predicted protein structures. Retrieve structures by UniProt ID, download PDB/mmCIF files, analyze confidence metrics (pLDDT, PAE), for drug discovery and structural biology.