Compress and index a VCF with bgzip and tabix

Block-compress a VCF with bgzip and build a tabix index so you can pull variants from any genomic region in milliseconds.

SDSomenath DuttaUpdated July 21, 2026

Prerequisites

htslib installed (`bgzip`, `tabix`)
A VCF file with records sorted by chromosome and position

1Block-compress the VCF

bgzip compresses the VCF in place, replacing calls.vcf with calls.vcf.gz. Unlike plain gzip, the BGZF format writes independent compressed blocks that tabix can seek into. The command prints nothing on success.

bash

bgzip calls.vcf

Expected output

# bgzip prints nothing on success
# calls.vcf is replaced by calls.vcf.gz

2Build the tabix index

Index the compressed file with the VCF preset. This is also silent and writes calls.vcf.gz.tbi next to your data. The index maps genomic coordinates to byte offsets inside the BGZF blocks.

bash

tabix -p vcf calls.vcf.gz

Expected output

# tabix prints nothing on success
# creates the index calls.vcf.gz.tbi

3Query a genomic region

Ask for any region as chrom:start-end. Tabix seeks straight to the matching blocks and streams the overlapping records instead of scanning the whole file. Add -h if you also want the VCF header lines prepended.

bash

tabix calls.vcf.gz chr1:1000000-2000000

Expected output

chr1	1000920	rs6054257	G	A	29.0	PASS	NS=3;DP=14;AF=0.5	GT:GQ:DP	0|1:48:8
chr1	1234567	.	T	C	52.0	PASS	NS=3;DP=22;AF=0.33	GT:GQ:DP	1|0:45:12
chr1	1876543	rs62636508	C	G	41.0	PASS	NS=2;DP=18;AF=0.25	GT:GQ:DP	0|0:39:10

Troubleshooting

[tabix] the file out of order at line ...

The VCF must be position-sorted before indexing. Sort and compress in one pass with `bcftools sort calls.vcf -Oz -o calls.sorted.vcf.gz`, then run `tabix -p vcf calls.sorted.vcf.gz`.

"not a BGZF file" or tabix fails to index

The file was compressed with plain gzip, not bgzip. Decompress it with `gunzip calls.vcf.gz` and re-run `bgzip calls.vcf`.

References

tabix documentation (htslib)