Skip to content

RastQC.md

Latest commit

 

History

History
152 lines (118 loc) · 4.68 KB

RastQC.md

File metadata and controls

152 lines (118 loc) · 4.68 KB
Raw

RastQC Agent Skill

RastQC is a fast sequencing quality control tool. Use this skill to run QC on FASTQ, BAM, or SAM files automatically.

Installation

# Install from GitHub (requires Rust 1.70+)
cargo install --git https://github.com/Huang-lab/RastQC.git

# With Nanopore Fast5/POD5 support
cargo install --git https://github.com/Huang-lab/RastQC.git --features nanopore

After installation, rastqc is available in $PATH. Verify with rastqc --version.

If cargo is not in PATH, check ~/.cargo/bin/cargo. If Rust is not installed: curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh.

Quick Reference

Basic QC (short-read Illumina data)

rastqc -o <output_dir> <input_files...>

Long-read QC (ONT / PacBio)

rastqc --long-read -o <output_dir> <input_files...>

The --long-read flag is auto-enabled for .fast5 and .pod5 files. For long-read FASTQ, you must pass it explicitly.

Common Options

Flag Purpose
-o <DIR> Output directory (default: current dir)
-t <N> Number of threads (default: all CPUs)
--long-read Enable long-read modules (N50, Quality Stratified Length, Homopolymer)
--time Show per-file timing breakdown
--nozip Write HTML report only, skip ZIP
--summary Force multi-file summary (auto for 2+ files)
--multiqc-json Output native MultiQC JSON
--exit-code Return 0=pass, 1=warn, 2=fail for pipeline gates
--serve Launch web browser to view reports
-q / --quiet Suppress progress output
--stdin Read FASTQ from stdin pipe

Input Formats

  • FASTQ (plain, .gz, .bz2)
  • BAM / SAM
  • Fast5 / POD5 (requires --features nanopore build)
  • Standard input (--stdin or pass - as filename)

Output Files

For each input file sample.fastq.gz:

File Description
sample_fastqc.html Self-contained HTML report with SVG charts
sample_fastqc.zip ZIP archive (contains HTML + data + summary)
sample_multiqc.json MultiQC JSON (only with --multiqc-json)

For multiple files:

File Description
summary.html Cross-sample dashboard with pass/warn/fail matrix
summary.tsv Machine-readable summary for scripting

Agent Workflow: Automatic QC

When asked to run QC on sequencing data, follow this workflow:

Step 1: Check rastqc is installed

which rastqc || cargo install --git https://github.com/Huang-lab/RastQC.git

Step 2: Detect input files

ls <data_dir>/*.fastq.gz <data_dir>/*.fq.gz <data_dir>/*.bam 2>/dev/null

Step 3: Determine read type

Check if files are long-read by examining a few reads:

# Peek at read lengths (if mean > 500bp, it's long-read)
gzip -dc <file> 2>/dev/null | head -400 | awk 'NR%4==2 {sum+=length($0); n++} END {printf "%.0f\n", sum/n}'

If mean read length > 500 bp, or filename contains ont, nanopore, pacbio, hifi, revio, or extension is .fast5/.pod5, add --long-read.

Step 4: Run RastQC

# Short-read
rastqc --time -o <output_dir> <files...>

# Long-read
rastqc --long-read --time -o <output_dir> <files...>

# With MultiQC JSON
rastqc --multiqc-json --time -o <output_dir> <files...>

Step 5: Interpret results

Read the stderr output for the summary, or parse:

# Quick pass/fail check
cat <output_dir>/summary.tsv

# Check specific file's module results
unzip -p <output_dir>/sample_fastqc.zip "*/summary.txt"

Step 6: Report to user

Summarize:

  • Total files processed and wall-clock time
  • Number of pass/warn/fail modules
  • Any files with failures and which modules failed
  • Open the HTML report: open <output_dir>/summary.html

Interpreting Module Results

Module WARN means FAIL means
Per Base Quality Median < Q25 at any position Median < Q20 at any position
Per Sequence Quality Mode quality <= Q27 Mode quality <= Q20
Per Base Content |A-T| or |G-C| > 10% > 20%
Per Sequence GC >15% deviation from normal >30% deviation
N Content N% > 5% at any position N% > 20%
Duplication <70% unique reads <50% unique
Overrepresented Any sequence > 0.1% > 1%
Adapter Content >5% adapter at any position >10%
Homopolymer (LR) >5% bases in runs >10% bases in runs

Pipeline Integration Example

# Nextflow-style QC gate
rastqc --exit-code -o qc_results/ sample.fastq.gz
if [ $? -eq 2 ]; then
    echo "QC FAILED - stopping pipeline"
    exit 1
fi

# Pipe from samtools
samtools fastq aligned.bam | rastqc --stdin -o qc_results/

# Batch + MultiQC
rastqc --multiqc-json -o qc_results/ data/*.fastq.gz
multiqc qc_results/