05. Alignment

← Previous: Adapter trimming · Tutorial contents · Next: BAM filtering and QC →

Prepare the reference

Use an authoritative FASTA and an aligner index built from exactly that FASTA. Record the organism, assembly, source, release, checksum, included contigs, and chromosome naming convention. Annotation and blacklist resources used later must be compatible with the same assembly.

Build the Bowtie2 index once for a newly obtained reference. Do not rebuild it for every sample:

set -euo pipefail

REFERENCE_FASTA="/path/to/reference.fa"
INDEX_PREFIX="reference/bowtie2/genome"

test -s "${REFERENCE_FASTA}"
mkdir -p reference/bowtie2 logs/reference

bowtie2-build \
  --threads 2 \
  "${REFERENCE_FASTA}" "${INDEX_PREFIX}" \
  2>&1 | tee logs/reference/bowtie2_build.log

Copy-and-adapt alignment commands for one sample

set -euo pipefail

SAMPLE_ID="sample_01"
INDEX_PREFIX="reference/bowtie2/genome"
R1="results/trimmed/${SAMPLE_ID}_R1.trimmed.fastq.gz"
R2="results/trimmed/${SAMPLE_ID}_R2.trimmed.fastq.gz"

test -s "${R1}"
test -s "${R2}"
mkdir -p results/alignment logs/alignment

bowtie2 \
  --threads 2 \
  -x "${INDEX_PREFIX}" \
  -1 "${R1}" \
  -2 "${R2}" \
  -S "results/alignment/${SAMPLE_ID}.sam" \
  2> "logs/alignment/${SAMPLE_ID}.bowtie2.log"

test -s "results/alignment/${SAMPLE_ID}.sam"

Bowtie2 alignment presets and fragment-length assumptions should match the library. Do not treat an example command as a validated protocol for a new dataset. If reads were intentionally left untrimmed, change R1 and R2 to the validated raw FASTQ paths.

Alignment gate

• exit status is zero and the log is preserved;
• the input read-pair count agrees with the alignment summary;
• headers identify the expected reference;
• overall and concordant mapping rates are reviewed across all samples;
• no sample is accepted or excluded solely by comparison with an undocumented threshold.

---

← Previous: Adapter trimming · Tutorial contents · Next: BAM filtering and QC →