Bowtie2

Bowtie2 is an ultrafast and memory-efficient aligner for short DNA sequences to a reference genome. It is particularly well-suited for aligning reads that do not contain large gaps (e.g., ChIP-seq, ATAC-seq, DNA-seq) and is optimized for reads from 50 bp to several hundred bp.

https://bowtie-bio.sourceforge.net/bowtie2/
https://github.com/BenLangmead/bowtie2

Key Features: * Very fast alignment with low memory footprint * Supports gapped alignment (insertions/deletions) * Handles paired-end and single-end data * Provides multiple alignment modes (end-to-end, local) * Efficient for small genomes (bacteria, viruses) to mammalian genomes * Outputs SAM format compatible with downstream tools * Highly sensitive alignment with tunable parameters

Alignment Modes: * End-to-end mode (default): Requires the entire read to align, trimming is not allowed. Best for high-quality reads where the full read is expected to match the reference. * Local mode (--local): Allows soft-clipping of read ends that don’t match well. Better for reads with adapters or low-quality ends that weren’t properly trimmed.

Typical Workflow:

Step 1: Build index (one-time setup):

bowtie2-build reference.fasta ref_index

This creates index files (.1.bt2, .2.bt2, .3.bt2, .4.bt2, .rev.1.bt2, .rev.2.bt2).

Step 2: Align paired-end reads (end-to-end mode):

bowtie2 \
    -p 8 \                          # use 8 threads
    -x ref_index \                  # index prefix
    -1 sample_R1.fastq.gz \         # forward reads
    -2 sample_R2.fastq.gz \         # reverse reads
    --very-sensitive \              # preset: slower but more accurate
    --rg-id sample1 \               # read group ID
    --rg SM:sample1 \               # sample name
    -S sample.sam                   # output SAM file

Step 3: Convert to sorted BAM:

samtools view -bS sample.sam | samtools sort -@ 4 -o sample.sorted.bam -
samtools index sample.sorted.bam

Alignment presets: Bowtie2 provides convenient presets that adjust multiple parameters simultaneously: - --very-fast: Fastest, least sensitive - --fast: Fast, moderately sensitive - --sensitive (default): Balanced speed and sensitivity - --very-sensitive: Slower but more thorough search - --very-sensitive-local: For local alignment mode

When to use Bowtie2 vs BWA-MEM: - Bowtie2: Best for DNA sequencing without long insertions/deletions (ChIP-seq, ATAC-seq, small genomes, metagenomics). Faster and uses less memory than BWA-MEM. - BWA-MEM: Better for WGS/WES with structural variants, longer reads (>100 bp), and when you need split-read support. More robust to sequencing errors.

Reporting options: - -k N: Report up to N alignments per read (default: search for 1) - --no-unal: Suppress unaligned reads in output - --no-mixed: For paired-end, only align when both reads align - --no-discordant: Only report concordant paired-end alignments

Fragment length constraints for paired-end: The -I and -X parameters set the minimum and maximum fragment length (insert size):

bowtie2 -x ref -1 R1.fq -2 R2.fq -I 100 -X 500  # expect 100-500 bp fragments

This prevents spurious alignments where reads map too close or too far apart.

--- title: "Bowtie2" --- Bowtie2 is an ultrafast and memory-efficient aligner for short DNA sequences to a reference genome. It is particularly well-suited for aligning reads that do not contain large gaps (e.g., ChIP-seq, ATAC-seq, DNA-seq) and is optimized for reads from 50 bp to several hundred bp. - https://bowtie-bio.sourceforge.net/bowtie2/ - https://github.com/BenLangmead/bowtie2 **Key Features:** * Very fast alignment with low memory footprint * Supports gapped alignment (insertions/deletions) * Handles paired-end and single-end data * Provides multiple alignment modes (end-to-end, local) * Efficient for small genomes (bacteria, viruses) to mammalian genomes * Outputs SAM format compatible with downstream tools * Highly sensitive alignment with tunable parameters **Alignment Modes:** * **End-to-end mode (default)**: Requires the entire read to align, trimming is not allowed. Best for high-quality reads where the full read is expected to match the reference. * **Local mode (`--local`)**: Allows soft-clipping of read ends that don't match well. Better for reads with adapters or low-quality ends that weren't properly trimmed. **Typical Workflow:** **Step 1: Build index (one-time setup):** ```bash bowtie2-build reference.fasta ref_index ``` This creates index files (.1.bt2, .2.bt2, .3.bt2, .4.bt2, .rev.1.bt2, .rev.2.bt2). **Step 2: Align paired-end reads (end-to-end mode):** ```bash bowtie2 \ -p 8 \ # use 8 threads -x ref_index \ # index prefix -1 sample_R1.fastq.gz \ # forward reads -2 sample_R2.fastq.gz \ # reverse reads --very-sensitive \ # preset: slower but more accurate --rg-id sample1 \ # read group ID --rg SM:sample1 \ # sample name -S sample.sam # output SAM file ``` **Step 3: Convert to sorted BAM:** ```bash samtools view -bS sample.sam | samtools sort -@ 4 -o sample.sorted.bam - samtools index sample.sorted.bam ``` **Alignment presets:** Bowtie2 provides convenient presets that adjust multiple parameters simultaneously: - `--very-fast`: Fastest, least sensitive - `--fast`: Fast, moderately sensitive - `--sensitive` (default): Balanced speed and sensitivity - `--very-sensitive`: Slower but more thorough search - `--very-sensitive-local`: For local alignment mode **When to use Bowtie2 vs BWA-MEM:** - **Bowtie2**: Best for DNA sequencing without long insertions/deletions (ChIP-seq, ATAC-seq, small genomes, metagenomics). Faster and uses less memory than BWA-MEM. - **BWA-MEM**: Better for WGS/WES with structural variants, longer reads (>100 bp), and when you need split-read support. More robust to sequencing errors. **Reporting options:** - `-k N`: Report up to N alignments per read (default: search for 1) - `--no-unal`: Suppress unaligned reads in output - `--no-mixed`: For paired-end, only align when both reads align - `--no-discordant`: Only report concordant paired-end alignments **Fragment length constraints for paired-end:** The `-I` and `-X` parameters set the minimum and maximum fragment length (insert size): ```bash bowtie2 -x ref -1 R1.fq -2 R2.fq -I 100 -X 500 # expect 100-500 bp fragments ``` This prevents spurious alignments where reads map too close or too far apart.