VCF / BCF

Overview

VCF (Variant Call Format) is the standard file format for describing genetic variants – SNPs, insertions, deletions, and structural variants – relative to a reference genome. It is produced by every major variant caller (GATK HaplotypeCaller, DeepVariant, bcftools, FreeBayes) and consumed by annotation tools, filtering pipelines, and population-genetic analyses.

BCF is the binary, BGZF-compressed equivalent of VCF. It is significantly faster to parse and is the recommended format for large cohorts. The relationship between VCF and BCF mirrors that of SAM and BAM.

VCF files use the extension .vcf (plain text) or .vcf.gz (block-compressed with bgzip). BCF files use .bcf. Both compressed formats support tabix or CSI indexing for fast regional queries.

Structure

A VCF file has three parts: meta-information lines, a header line, and data lines.

##fileformat=VCFv4.3
##INFO=<ID=DP,Number=1,Type=Integer,Description="Total Depth">
##INFO=<ID=AF,Number=A,Type=Float,Description="Allele Frequency">
##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
##FORMAT=<ID=DP,Number=1,Type=Integer,Description="Read Depth">
##FORMAT=<ID=GQ,Number=1,Type=Integer,Description="Genotype Quality">
#CHROM  POS     ID        REF  ALT  QUAL  FILTER  INFO          FORMAT     SAMPLE1
chr1    12345   rs123456  A    G    99    PASS    DP=50;AF=0.5  GT:DP:GQ   0/1:50:99
chr1    67890   .         CT   C    85    PASS    DP=40;AF=0.3  GT:DP:GQ   0/1:40:85

Meta-information lines

Lines starting with ## define the schema for INFO, FORMAT, FILTER, and contig fields. They are machine-readable and describe every annotation that appears in the data section.

Header line

The single line starting with #CHROM lists the eight mandatory columns followed by one column per sample.

Data columns

Col	Field	Description
1	CHROM	Chromosome or contig name.
2	POS	1-based position of the variant on the reference.
3	ID	Variant identifier (e.g. dbSNP rsID) or `.` if unknown.
4	REF	Reference allele(s).
5	ALT	Alternate allele(s), comma-separated for multi-allelic sites.
6	QUAL	Phred-scaled quality score for the variant call.
7	FILTER	`PASS` if the site passed all filters; otherwise a semicolon-separated list of filter names.
8	INFO	Semicolon-separated key-value pairs with site-level annotations.
9	FORMAT	Colon-separated keys defining the per-sample fields.
10+	SAMPLE	Per-sample genotype and annotations matching the FORMAT keys.

Genotype encoding

The GT (genotype) field uses indices into the REF/ALT alleles:

Genotype	Meaning
`0/0`	Homozygous reference
`0/1`	Heterozygous (one REF, one ALT)
`1/1`	Homozygous alternate
`1/2`	Heterozygous with two different ALT alleles
`./.`	Missing genotype (no call)
`0\|1`	Phased heterozygous (`\|` indicates phase is known)

Variant types

Type	REF	ALT	Example
SNP	`A`	`G`	Single nucleotide change
Insertion	`A`	`ATG`	Two bases inserted after position
Deletion	`CT`	`C`	One base deleted
MNP	`AT`	`GC`	Multi-nucleotide polymorphism
Complex	`ACT`	`GA`	Simultaneous insertion and deletion

Working With

Compressing and indexing

# Block-compress with bgzip (NOT gzip)
bgzip variants.vcf

# Create a tabix index
tabix -p vcf variants.vcf.gz

Viewing and querying

# View VCF header
bcftools view -h variants.vcf.gz

# Query a specific region
bcftools view variants.vcf.gz chr1:10000-50000

# View only PASS variants
bcftools view -f PASS variants.vcf.gz

Filtering variants

# Keep only biallelic SNPs with QUAL >= 30 and depth >= 10
bcftools view -m2 -M2 -v snps variants.vcf.gz \
  | bcftools filter -i 'QUAL>=30 && INFO/DP>=10' -o filtered.vcf.gz -Oz

# Exclude variants that failed filters
bcftools view -f PASS variants.vcf.gz -o pass_only.vcf.gz -Oz

Statistics and summary

# Generate variant statistics
bcftools stats variants.vcf.gz > stats.txt

# Count variants by type
bcftools stats variants.vcf.gz | grep '^SN'

Converting VCF to BCF and back

# VCF to BCF
bcftools view variants.vcf.gz -Ob -o variants.bcf
bcftools index variants.bcf

# BCF to VCF
bcftools view variants.bcf -Oz -o variants.vcf.gz
tabix -p vcf variants.vcf.gz

Merging and concatenating

# Merge VCFs from different samples (same sites)
bcftools merge sample1.vcf.gz sample2.vcf.gz -Oz -o merged.vcf.gz

# Concatenate VCFs from different regions (same samples)
bcftools concat chr1.vcf.gz chr2.vcf.gz -Oz -o combined.vcf.gz

Extracting fields

# Extract specific fields as a tab-separated table
bcftools query -f '%CHROM\t%POS\t%REF\t%ALT\t%QUAL\t[%GT]\n' \
  variants.vcf.gz > variants.tsv