Bowtie2 Alignment for ChIP-seq / ATAC-seq

Hard rules

1. No fabricated citations. Every cited work must resolve to a verifiable
2. No claim without provenance. Every quantitative or factual claim
3. No silent failure. Every script invocation, API call, or tool use must declare its exit status and what to do on non-zero. A skill that silently swallows errors is a violation.

When to use

• ChIP-seq / CUT&RUN / CUT&Tag alignment.
• ATAC-seq alignment (with --local for Tn5 soft-clipping).
• Whole-genome bisulfite-seq (with --no-mixed --no-discordant).
• Any read length 50-150 bp where speed matters.

When NOT to use

• Long reads → use minimap2.
• Variant calling on whole-genome data → use bwa-mem or bwa-mem2.
• RNA-seq splice alignment → use STAR or HISAT2.

Prerequisites

• bowtie2 ≥ 2.5
• samtools ≥ 1.19
• Reference FASTA + Bowtie2 index (bowtie2-build)

Code patterns

Index the reference

bowtie2-build --threads 8 reference/genome.fa genome_bt2
# Creates genome_bt2.{1,2,3,4,rev.1,rev.2}

For very large genomes, use the --large-index flag.

Paired-end alignment (ChIP-seq default)

bowtie2 -p 16 --no-mixed --no-discordant \
    -x genome_bt2 \
    -1 reads_R1.fq.gz -2 reads_R2.fq.gz \
    --rg-id sample1 --rg SM:sample1 --rg PL:ILLUMINA \
    --rg LB:lib1 |
  samtools sort -@ 8 -m 4G -o sample1.bam -
samtools index sample1.bam

--no-mixed --no-discordant ensures that only properly paired reads are reported; useful for fragment length analysis.

Single-end alignment (CUT&RUN, ATAC-seq)

bowtie2 -p 16 --local -x genome_bt2 -U reads.fq.gz |
  samtools sort -@ 8 -o s1.bam -
samtools index s1.bam

--local enables soft-clipping at read ends, which is essential for Tn5 transposase-cut reads in ATAC-seq.

--very-sensitive for low-input or divergent samples

bowtie2 --very-sensitive-local -p 16 -x bt2 -1 R1.fq -2 R2.fq | \
  samtools sort -@ 8 -o s.bam -

Sensitivity levels: --fast, --sensitive (default), --very-sensitive. Time increases roughly 2x between levels.

Allow unpaired reads from a paired run

Drop --no-mixed --no-discordant if you want to keep the unpaired mates:

bowtie2 -p 16 -x bt2 -1 R1.fq -2 R2.fq | samtools sort -@ 8 -o s.bam -

ATAC-seq specific: soft-clip Tn5, keep mitochondrial reads but flag them later

bowtie2 -p 16 --local --no-mixed --no-discordant -X 2000 \
    -x bt2 -1 R1.fq.gz -2 R2.fq.gz |
  samtools sort -@ 8 -o s.bam -

Then mark duplicates and remove chrM with:

samtools view -h s.bam | grep -v chrM | samtools view -b -o s.no_chrM.bam
samtools index s.no_chrM.bam

Read group in the header (for downstream tools)

bowtie2 -p 16 --rg-id s1 --rg SM:s1 --rg PL:ILLUMINA --rg LB:lib1 \
    -x bt2 -1 R1.fq -2 R2.fq | samtools sort -@ 8 -o s.bam -

Insert size distribution (for ChIP-seq fragment QC)

samtools view -f 2 s.bam | awk '{print $9}' | sort -n | uniq -c > insert_sizes.txt

nf-core integration

nf-core/chipseq and nf-core/atacseq default to Bowtie2. No CLI changes needed.

Common pitfalls

• Using --end-to-end for ATAC-seq. Tn5 transposase inserts at precise offsets and the read may start with the transposon sequence, so --local is required.
• No read group. Downstream tools (MACS2, deepTools) often require @RG headers.
• Allowing --no-mixed --no-discordant for very low input. If most of your reads are unpaired, this discards them. Drop these flags.
• Indexing the wrong reference. Bowtie2 indexes are large (.bt2 files). Keep them in a reference/ directory.
• Confusing Bowtie1 and Bowtie2. Bowtie1 is ungapped; for ChIP/ATAC, you almost always want Bowtie2.

Validation

• samtools flagstat s.bam — should show high mapping rate (≥80% for ChIP-seq, ≥95% for ATAC-seq excluding chrM).
• samtools view -c -f 2 s.bam — properly-paired count.
• Insert size histogram: ChIP-seq TF should peak at ~150-300 bp; nucleosome-depleted ChIP-seq at ~50-100 bp; ATAC-seq at sub-nucleosomal + nucleosomal modes.
• samtools view -F 4 s.bam | wc -l — mapped reads.

Open alternatives

References

• Bowtie2 paper: Langmead & Salzberg 2012, Nature Methods — 10.1038/nmeth.1923
• Bowtie2 manual: http://bowtie-bio.sourceforge.net/bowtie2/manual.shtml
• Companion: ors-bioinformatics-sequence-bwa-alignment, ors-bioinformatics-sequence-bwa-mem2-alignment.

Changelog

• 1.0.0 (2026-06-10): Initial adaptation by Pradyumna Jayaram from bio-bowtie2-alignment (bioSkills-main/read-alignment/bowtie2-alignment)."

Cross-references

Other skills in this category:

• batch-processing
• bwa-alignment
• bwa-mem2-alignment
• codon-usage
• compressed-sequence-files
• fastq-quality-scores
• filter-sequences
• format-conversion
• hisat2-alignment
• motif-search
• paired-end-fastq
• pysam-genomics
• read-write-sequences
• reverse-complement
• sam-bam-basics
• samtools-bam-processing
• seq-objects
• sequence-properties
• sequence-slicing
• sequence-statistics
• star-alignment
• transcription-translation