Index a FASTA and extract a region with samtools faidx

Build a .fai index for a FASTA file with samtools faidx, then pull any sub-sequence by chromosome and coordinate.

SSSudipta SardarUpdated July 21, 2026

Prerequisites

samtools installed (`samtools faidx`)
A FASTA file (e.g. reference.fasta)

1Index the FASTA

Build the .fai index next to your reference. This writes reference.fasta.fai and prints nothing on success. Each line is tab-separated: sequence name, length, byte offset, bases per line, and line width; run `cut -f1,2 reference.fasta.fai` to read chromosome sizes.

bash

samtools faidx reference.fasta
cut -f1,2 reference.fasta.fai

Expected output

chr1	248956422
chr2	242193529
chr3	198295559

2Extract a region by coordinate

Pass a `name:start-end` region and faidx prints the header plus the sub-sequence to stdout. The interval is 1-based and inclusive, so chr1:100-200 returns 101 bases. Redirect to a file with `> region.fasta` when you want to keep it.

bash

samtools faidx reference.fasta chr1:100-200 > region.fasta

Expected output

>chr1:100-200
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN

Troubleshooting

[faidx] Failed to fetch sequence in chr1:100-200

The name before the colon must match the FASTA header exactly, up to the first space. List the valid names with `grep '>' reference.fasta` or read the first column of reference.fasta.fai.

[E::fai_build_core] Different line length in sequence 'chr1' at line 42

faidx requires a uniform line width within each record. Re-wrap the file to a fixed width first: `seqkit seq -w 60 in.fasta > wrapped.fasta`, then index wrapped.fasta.

References

samtools faidx documentation