VCF format

For the full specification, see the VCF format specification.

VCF (Variant Call Format) is the standard text format for storing genetic variants: the places where a sample’s sequence differs from a reference. This includes single-nucleotide polymorphisms (SNPs), insertions and deletions (indels), and larger structural variants.

Format

A VCF file has a header section followed by one data line per variant. Header lines begin with ## and describe the file and the meaning of the fields used below; the single line beginning with #CHROM names the columns. Each variant record has eight required, tab-separated fields:

• CHROM — the chromosome the variant is on.

• POS — the 1-based position of the variant on the chromosome.

• ID — an identifier for the variant (for example a dbSNP rs number), or . if none.

• REF — the reference allele (the base(s) in the reference).

• ALT — the alternate allele(s) observed in the sample.

• QUAL — a Phred-scaled quality score for the assertion that a variant is present.

• FILTER — PASS if the variant passed all filters, or a list of the filters it failed.

• INFO — a semicolon-separated list of additional annotations.

The figure below shows an example from the specification:

An example VCF file with its header lines and several variant records. Source: VCF format specification.

Software that use VCF format

VCF is produced and consumed by most variant-focused tools:

• GATK — variant calling and filtering.

• Samtools / BCFtools — variant calling and VCF manipulation.

• SnpEff — annotating variants with their predicted effects.

• VCFtools — filtering and summarizing VCF files.

• dbSNP — a public database of known variants.

How are these files generated?

VCF files are the output of a variant-calling pipeline: reads are aligned to a reference, the alignments are processed, and a variant caller compares the sample to the reference to produce the list of differences. For a full worked example of such a pipeline, see the whole genome sequencing walkthrough.