bowtie-mapping-nextflow

A Nextflow pipeline that aligns short reads to a reference genome with Bowtie2 and derives coverage tracks and, for ChIP-seq histone modification data, called peaks.

Overview

This pipeline covers the common short-read mapping and signal-generation steps used across several VEuPathDB genomic data types: origins-of-replication profiling, splice-site read mapping, and ChIP-seq (histone modification, transcription factor binding, MNase, FAIRE, and DNase experiments). Samples are aligned to a reference genome with Bowtie2, optionally deduplicated, and then compared against a paired reference/input sample with deeptools bamCompare to produce ratio coverage tracks. For ChIP-seq datasets, the pipeline additionally builds HOMER tag directories (or, for MNase data, runs DANPOS) to generate coverage bigWigs, and — for histone modification experiments with a matched input sample — calls peaks with HOMER and indexes the results as a tabix-indexed GFF for loading into VEuPathDB's genome browser and study infrastructure.

Requirements

• Nextflow (DSL2)
• Docker or Singularity/Apptainer — processes run in the veupathdb/bowtiemapping:1.0.0 image plus quay.io/biocontainers/deeptools, quay.io/biocontainers/ucsc-bedgraphtobigwig, quay.io/biocontainers/ucsc-wigtobigwig, biocontainers/danpos, and biocontainers/tabix (select the engine via the docker or singularity profile/config in conf/)

Usage

nextflow run VEuPathDB/bowtie-mapping-nextflow -r main \
  --input /path/to/input \
  --genome /path/to/genome.fasta \
  --datasetType chipSeq \
  --experimentType histonemod \
  --inputFileType fastq \
  --hasPairedReads true \
  --removePCRDuplicates true \
  --outputDir /path/to/output \
  -C conf/docker.config \
  -resume

The pipeline has a single (default) entry point, made up of two chained sub-workflows:

1. bowtieMapping — reads params.input/<samplesheetFileName>, builds a Bowtie2 index from params.genome, aligns each sample's reads with bowtie2, optionally removes PCR duplicates (samtools rmdup), and indexes the resulting BAM.
2. coverageAndPeaks — pairs each sample with the reference/input sample named in its ref field, runs deeptools bamCompare to produce a ratio bigWig per pair, and then — when datasetType is chipSeq — either runs DANPOS (for experimentType = mnase) or builds HOMER tag directories and bedGraph/bigWig coverage tracks (for histonemod, tfbinding, faire, dnase). For experimentType = histonemod, peaks are additionally called with HOMER findPeaks against the paired input sample and indexed as a tabix-compressed GFF.

Samplesheet

params.input/<samplesheetFileName> is a CSV (header row skipped) with columns: sample ID, read file 1, read file 2 (if paired), and reference/input sample ID (the ref column used to pair a sample with its control for bamCompare and peak calling). Read file paths are resolved relative to params.input unless given as absolute paths.

Key Parameters

Parameter	Default	Description
input	$launchDir/input	Directory containing the samplesheet and (relative) read files
samplesheetFileName	samplesheet.csv	Samplesheet filename, resolved under input
genome	$launchDir/input/genome.fasta	Reference genome FASTA used for the Bowtie2 index and chromosome sizes
datasetType	—	Data type being processed: originsOfReplication, spliceSites, or chipSeq
experimentType	—	For chipSeq datasets: histonemod, tfbinding, mnase, faire, or dnase
inputFileType	—	fastq or fasta, passed to bowtie2 as -q/-f
hasPairedReads	true	Whether samples have paired-end (-1/-2) or single-end (-U) reads
removePCRDuplicates	false	Whether to run samtools rmdup on alignments before downstream analysis
saveAlignments	false	Publish indexed BAM/BAI files to outputDir
saveCoverage	false	Publish bigWig coverage tracks to outputDir
profileSetName	—	Profile set name written into browser track/peak config output for the study loader
gffFileName	output.gff	Filename used for the sorted, tabix-indexed peaks GFF
outputDir	$launchDir/output	Directory published output files are written to

Output

Depending on datasetType/experimentType and the save* flags, published to outputDir:

• Indexed BAM alignments (<sample>.bam, <sample>.bam.bai) when saveAlignments is true
• Ratio coverage bigWigs from bamCompare (<sample>_vs_<ref>.bw) and their browser track config, collected into metadata_ratios
• For ChIP-seq data with saveCoverage enabled: unlogged coverage bigWigs (HOMER-derived bedGraph→bigWig for most experiment types, or DANPOS smoothed signal→bigWig for MNase), with browser track configs collected into metadata_unlogged
• For histonemod experiments: called peaks per sample (<sample>_peaks.txt), a merged tabix-indexed peaks GFF (<gffFileName>.gz / .gz.tbi), and a collected insert_study_results config file for loading peak results