author: Erik Garrison erik.garrison@bc.edu
The Variant Call Format (VCF) is a flat-file, tab-delimited textual format intended to concisely describe reference-indexed variations between individuals. VCF provides a common interchange format for the description of variation in individuals and populations of samples, and has become the defacto standard reporting format for a wide array of genomic variant detectors.
vcflib provides methods to manipulate and interpret sequence variation as it can be described by VCF. It is both:
- an API for parsing and operating on records of genomic variation as it can be described by the VCF format,
- and a collection of command-line utilities for executing complex manipulations on VCF files.
The API itself provides a quick and extremely permissive method to read and write VCF files. Extensions and applications of the library provided in the included utilities (*.cpp) comprise the vast bulk of the library's utility for most users.
vcflib provides a variety of functions for VCF manipulation:
- Generate haplotype-aware intersections (vcfintersect -i), unions (vcfintersect -u), and complements (vcfintersect -v -i).
- Overlay-merge multiple VCF files together, using provided order as precedence (vcfoverlay).
- Validate the integrity and identity of the VCF by verifying that the VCF record's REF matches a given reference file (vcfcheck).
- Convert a VCF file into a per-allele or per-genotype tab-separated (.tsv) file (vcf2tsv).
- Store a VCF file in an SQLite3 database (vcf2sqlite.py).
- Make a BED file from the intervals in a VCF file (vcf2bed.py).
- Filter variants and genotypes using arbitrary expressions based on values in the INFO and sample fields (vcffilter).
- Randomly sample a subset of records from a VCF file, given a rate (vcfrandomsample).
- Select variants of a certain type (vcfsnps, vcfbiallelic, vcfindels, vcfcomplex, etc.)
- Annotate one VCF file with fields from the INFO column of another, based on position (vcfaddinfo, vcfintersect).
- Incorporate annotations or targets provided by a BED file (vcfannotate, vcfintersect).
- Examine genotype correspondence between two VCF files by annotating samples in one file with genotypes from another (vcfannotategenotypes).
- Annotate variants with the distance to the nearest variant (vcfdistance).
- Count the number of alternate alleles represented in samples at each variant record (vcfaltcount).
- Subset INFO fields to decrease file size and processing time (vcfkeepinfo).
- Lighten up VCF files by keeping only a subset of per-sample information (vcfkeepgeno).
- Quickly obtain the list of samples in a given VCF file (vcfsamplenames).
- Remove samples from a VCF file (vcfkeepsamples, vcfremovesamples).
- Sort variants by genome coordinate (vcfstreamsort).
- Remove duplicate variants in vcfstreamsort'ed files according to their REF and ALT fields (vcfuniq).
- Break multiallelic records into multiple records (vcfbreakmulti), retaining allele-specific INFO fields.
- Combine overlapping biallelic records into a single record (vcfcreatemulti).
- Decompose complex variants into a canonical SNP and indel representation (vcfallelicprimitives), generating phased genotypes for available samples.
- Reconstitute complex variants provided a phased VCF with samples (vcfgeno2haplo).
- Set genotypes in a VCF file provided genotype likelihoods in the GL field (vcfglxgt).
- Establish putative somatic variants using reported differences between germline and somatic samples (vcfsamplediff).
- Remove samples for which the reported genotype (GT) and observation counts disagree (AO, RO) (vcfremoveaberrantgenotypes).
- Obtain aggregate statistics about VCF files (vcfstats).
- Print the receiver-operating characteristic (ROC) of one VCF given a truth set (vcfroc).
- Annotate VCF records with the Shannon entropy of flanking sequence (vcfentropy).
- Calculate the heterozygosity rate (vcfhetcount).
- Generate potential primers from VCF records (vcfprimers), to check for genome uniqueness.
- Convert the numerical represenation of genotypes provided by the GT field to a human-readable genotype format (vcfgenotypes).
- Observe how different alignment parameters, including context and entropy-dependent ones, influence variant classification and interpretation (vcfremap).
- Classify variants by annotations in the INFO field using a self-organizing map (vcfsom); re-estimate their quality given known variants.
A number of "helper" perl and python scripts (e.g. vcf2bed.py, vcfbiallelic) further extend functionality.
In practice, users are encouraged to drive the utilities in the library in a streaming fashion, using pipes, to fully utilize resources on multi-core systems during interactive work. Piping provides a convenient method to interface with other libraries (vcf-tools, BedTools, GATK, htslib, bcftools, freebayes) which interface via VCF files, allowing the composition of an immense variety of processing functions.
See src/vcfecho.cpp for basic usage. src/Variant.h and src/Variant.cpp describe methods available in the API. vcflib is incorporated into several projects, such as freebayes, which may provide a point of reference for prospective developers. Additionally, developers should be aware of that vcflib contains submodules (git repositories) comprising its dependencies (outside of lzib and a *nix environment).
vcflib includes submodules, so to obtain vcflib you have to use:
% git clone --recursive git://github.com/ekg/vcflib.git
or
% git clone --recursive https://github.com/ekg/vcflib.git
To build, use Make:
% cd vcflib
% make
Executables are built into the ./bin directory in the repository. A number of shell, perl, python, and R scripts already reside there. This makes installation easy, as users can add vcflib/bin to their path, or copy the contained executables to a directory already in their path.
usage: vcf2tsv [-n null_string] [vcf file]
Converts stdin or given VCF file to tab-delimited format, using null string to replace empty
values in the table.
usage: vcfaddinfo <vcf file> <vcf file>
Adds info fields from the second file which are not present in the first vcf file.
Uses allele frequencies in the AF info column to estimate phylogeny at multiallelic sites.
usage: vcfallelicprimitives [options] [file]
options:
-m, --use-mnps Retain MNPs as separate events (default: false)
-t, --tag-parsed FLAG Tag records which are split apart of a complex allele
with this flag
If multiple alleleic primitives (gaps or mismatches) are specified in
a single VCF record, split the record into multiple lines, but drop all
INFO fields. "Pure" MNPs are split into multiple SNPs unless the -m
flag is provided. Genotypes are phased where complex alleles have been
decomposed, provided genotypes in the input.
Counts the number of alternate alleles in the record.
usage: vcfannotate [options] [<vcf file>]
options:
-b, --bed use annotations provided by this BED file
-k, --key use this INFO field key for the annotations
-d, --default use this INFO field key for records without annotations
Intersect the records in the VCF file with targets provided in a BED file.
Intersections are done on the reference sequences in the VCF file.
If no VCF filename is specified on the command line (last argument) the VCF
read from stdin.
usage: vcfannotategenotypes <annotation-tag> <vcf file> <vcf file>
annotates genotypes in the first file with genotypes in the second adding the
genotype as another flag to each sample filed in the first file.
annotation-tag is the name of the sample flag which is added to store the
annotation. also adds a 'has\_variant' flag for sites where the second file has
a variant.
usage: vcfbreakmulti [options] [file]
If multiple alleles are specified in a single record, break the record into
multiple lines, preserving allele-specific INFO fields.
usage: vcfcheck [options] <vcf file>
options: -f, --fasta-reference FASTA reference file to use to obtain
primer sequences
Verifies that the VCF REF field matches the reference as described.
Removes reference-matching sequence from complex alleles and adjusts records to reflect positional change.
usage: vcfcommonsamples <vcf file> <vcf file> outputs each record in the
first file, removing samples not present in the second
Counts the total number of alleles in the input.
If overlapping alleles are represented across multiple records, merge them into a single record.
Adds a value to each VCF record indicating the distance to the nearest variant in the file.
usage: vcfentropy [options] <vcf file>
options: -f, --fasta-reference FASTA reference file to use to obtain
primer sequences -w, --window-size Size of the window over which to calculate entropy
Anotates the output VCF file with, for each record, EntropyLeft,
EntropyRight, EntropyCenter, which are the entropies of the sequence of the given window size to the left, right, and center of the record.
usage: vcffilter [options] <vcf file>
options:
-f, --info-filter specifies a filter to apply to the info fields of records,
removes alleles which do not pass the filter
-g, --genotype-filter specifies a filter to apply to the genotype fields of records
-s, --filter-sites filter entire records, not just alleles
-t, --tag-pass tag vcf records as positively filtered with this tag, print all records
-F, --tag-fail tag vcf records as negatively filtered with this tag, print all records
-A, --append-filter append the existing filter tag, don't just replace it
-a, --allele-tag apply -t on a per-allele basis. adds or sets the corresponding INFO field tag
-v, --invert inverts the filter, e.g. grep -v
-o, --or use logical OR instead of AND to combine filters
-r, --region specify a region on which to target the filtering, requires a BGZF
compressed file which has been indexed with tabix. any number of
regions may be specified.
Filter the specified vcf file using the set of filters.
Filters are specified in the form "<ID> <operator> <value>:
-f "DP > 10" # for info fields
-g "GT = 1|1" # for genotype fields
-f "CpG" # for 'flag' fields
Operators can be any of: =, !, <, >, |, &
Any number of filters may be specified. They are combined via logical AND
unless --or is specified on the command line. Obtain logical negation through
the use of parentheses, e.g. ! "( DP = 10 )"
For convenience, you can specify "QUAL" to refer to the quality of the site, even
though it does not appear in the INFO fields.
Count the allele frequencies across alleles present in each record in the VCF file. (Similar to vcftools --freq.)
Uses genotypes from the VCF file to correct AC (alternate allele count), AF (alternate allele frequency), NS (number of called), in the VCF records. For example:
% vcfkeepsamples file.vcf NA12878 | vcffixup - | vcffilter -f "AC > 0"
Would downsample file.vcf to only NA12878, removing sites for which the sample was not called as polymorphic.
usage: vcfflatten [file]
Removes multi-allelic sites by picking the most common alternate. Requires
allele frequency specification 'AF' and use of 'G' and 'A' to specify the
fields which vary according to the Allele or Genotype. VCF file may be
specified on the command line or piped as stdin.
usage: vcfgeno2haplo [options] [<vcf file>]
options:
-w, --window-size N compare records up to this many bp away (default 30)
-r, --reference FILE FASTA reference file, required with -i and -u
Convert genotype-based phased alleles within --window-size into haplotype alleles.
usage: vcfgenotypecompare <other-genotype-tag> <vcf file>
adds statistics to the INFO field of the vcf file describing the
amount of discrepancy between the genotypes (GT) in the vcf file and the
genotypes reported in the <other-genotype-tag>. use this after
vcfannotategenotypes to get correspondence statistics for two vcfs.
Converts numerical representation of genotypes (standard in GT field) to the alleles provided in the call's ALT/REF fields.
usage: vcfglxgt [options] <vcf file>
options:
-n, --fix-null-genotypes only apply to null and partly-null genotypes
Set genotypes using the maximum genotype likelihood for each sample.
Count the number of heterozygotes in the input VCF.
Provides the ratio between heterozygotes and homozygotes.
usage: vcfintersect [options] [<vcf file>]
options:
-b, --bed FILE use intervals provided by this BED file
-v, --invert invert the selection, printing only records which would
not have been printed out
-i, --intersect-vcf FILE use this VCF for set intersection generation
-u, --union-vcf FILE use this VCF for set union generation
-w, --window-size N compare records up to this many bp away (default 30)
-r, --reference FILE FASTA reference file, required with -i and -u
-l, --loci output whole loci when one alternate allele matches
-m, --ref-match intersect on the basis of record REF string
-t, --tag TAG attach TAG to each record's info field if it would intersect
-V, --tag-value VAL use this value to indicate that the allele is passing
'.' will be used otherwise. default: 'PASS'
-M, --merge-from FROM-TAG
-T, --merge-to TO-TAG merge from FROM-TAG used in the -i file, setting TO-TAG
in the current file.
For bed-vcf intersection, alleles which fall into the targets are retained.
For vcf-vcf intersection and union, unify on equivalent alleles within window-size bp
as determined by haplotype comparison alleles.
usage: vcfkeepgeno <vcf file> [FIELD1] [FIELD2] ...
outputs each record in the vcf file, removing FORMAT fields not listed
on the command line from sample specifications in the output
usage: vcfkeepinfo <vcf file> [FIELD1] [FIELD2] ...
outputs each record in the vcf file, removing INFO fields not listed on the command line
usage: vcfkeepsamples <vcf file> [SAMPLE1] [SAMPLE2] ...
outputs each record in the vcf file, removing samples not listed on the command line
Adds the length of the variant record (in [-/+]) relative to the reference allele to each VCF record.
Annotates the VCF stream on stdin with the number of alternate alleles at the site.
usage: vcfoverlay [options] [<vcf file> ...]
options:
-h, --help this dialog
Overlays records in the input vcf files in the order in which they appear.
Demonstration of alternate allele parsing method. This method uses pairwise alignment of REF and ALTs to determine component allelic primitives for each alternate allele.
Use vcfallelicprimitives to decompose records while preserving format.
usage: vcfprimers [options] <vcf file>
options:
-f, --fasta-reference FASTA reference file to use to obtain primer sequences
-l, --primer-length The length of the primer sequences on each side of the variant
For each VCF record, extract the flanking sequences, and write them to stdout as FASTA
records suitable for alignment. This tool is intended for use in designing validation
experiments. Primers extracted which would flank all of the alleles at multi-allelic
sites. The name of the FASTA "reads" indicates the VCF record which they apply to.
The form is >CHROM_POS_LEFT for the 3' primer and >CHROM_POS_RIGHT for the 5' primer,
for example:
>20_233255_LEFT
CCATTGTATATATAGACCATAATTTCTTTATCCAATCATCTGTTGATGGA
>20_233255_RIGHT
ACTCAGTTGATTCCATACCTTTGCCATCATGAATCATGTTGTAATAAACA
usage: vcfrandomsample [options] [<vcf file>]
options:
-r, --rate RATE base sampling probability per locus
-s, --scale-by KEY scale sampling likelihood by this Float info field
-p, --random-seed N use this random seed
Randomly sample sites from an input VCF file, which may be provided as stdin.
Scale the sampling probability by the field specified in KEY. This may be
used to provide uniform sampling across allele frequencies, for instance.
usage: vcfremap [options] [<vcf file>]
options:
-w, --ref-window-size N align using this many bases flanking each side of the reference allele
-s, --alt-window-size N align using this many flanking bases from the reference around each alternate allele
-r, --reference FILE FASTA reference file, required with -i and -u
-m, --match-score N match score for SW algorithm
-x, --mismatch-score N mismatch score for SW algorithm
-o, --gap-open-penalty N gap open penalty for SW algorithm
-e, --gap-extend-penalty N gap extension penalty for SW algorithm
-z, --entropy-gap-open use entropy scaling for the gap open penalty
-R, --repeat-gap-extend N penalize non-repeat-unit gaps in repeat sequence
-a, --adjust-vcf TAG supply a new cigar as TAG in the output VCF
For each alternate allele, attempt to realign against the reference with lowered gap open penalty.
If realignment is possible, adjust the cigar and reference/alternate alleles.
Strips genotypes which are homozygous but have observations implying heterozygosity. Requires RA (reference allele observation) and AA (alternate allele observation) for each genotype.
usage: vcfremovesamples <vcf file> [SAMPLE1] [SAMPLE2] ...
outputs each record in the vcf file, removing samples listed on the command line
usage: vcfroc [options] [<vcf file>]
options:
-t, --truth-vcf FILE use this VCF as ground truth for ROC generation
-w, --window-size N compare records up to this many bp away (default 30)
-r, --reference FILE FASTA reference file
Generates a pseudo-ROC curve using sensitivity and specificity estimated against
a putative truth set. Thresholding is provided by successive QUAL cutoffs.
usage: vcfsamplediff <tag> <sample> <sample> [ <sample> ... ] <vcf file>
tags each record where the listed sample genotypes differ with <tag>
The first sample is assumed to be germline, the second somatic.
Each record is tagged with <tag>={germline,somatic,loh} to specify the type of
variant given the genotype difference between the two samples.
Prints the names of the samples in the VCF file.
usage: vcfsom [options] [vcf file]
training:
vcfsom -s output.som -f "AF DP ABP" training.vcf
application:
vcfsom -a output.som -f "AF DP ABP" test.vcf >results.vcf
vcfsomtrains and/or applies a self-organizing map to the input VCF data
on stdin, adding two columns for the x and y coordinates of the winning
neuron in the network and an optional euclidean distance from a given
node (--center).
If a map is provided via --apply, map will be applied to input without
training. Automated filtering to an estimated FP rate is
options:
-h, --help this dialog
training:
-f, --fields "FIELD ..." INFO fields to provide to the SOM
-a, --apply FILE apply the saved map to input data to FILE
-s, --save FILE train on input data and save the map to FILE
-t, --print-training-results
print results of SOM on training input
(you can also just use --apply on the same input)
-x, --width X width in columns of the output array
-y, --height Y height in columns of the output array
-i, --iterations N number of training iterations or epochs
-d, --debug print timing information
recalibration:
-c, --center X,Y annotate with euclidean distance from center
-p, --paint-true VCF use VCF file to annotate true variants (multiple)
-f, --paint-false VCF use VCF file to annotate false variants (multiple)
-R, --paint-tag TAG provide estimated FDR% in TAG in variant INFO
-N, --false-negative replace FDR% (false detection) with FNR% (false negative)
usage: vcfstats [options] <vcf file>
-r, --region specify a region on which to target the stats, requires a BGZF
compressed file which has been indexed with tabix. any number of
regions may be specified.
-a, --add-info add the statistics intermediate information to the VCF file,
writing out VCF records instead of summary statistics
-l, --no-length-frequency don't out the indel and mnp length-frequency spectra
-m, --match-score N match score for SW algorithm
-x, --mismatch-score N mismatch score for SW algorithm
-o, --gap-open-penalty N gap open penalty for SW algorithm
-e, --gap-extend-penalty N gap extension penalty for SW algorithm
Prints statistics about variants in the input VCF file.
Reads VCF on stdin and guarantees that the positional order is correct provided out-of-order variants are no more than 100 positions in the VCF file apart.
Like GNU uniq, but for VCF records. Remove records which have the same positon, ref, and alt as the previous record.
For each record, remove any duplicate alternate alleles that may have resulted from merging separate VCF files.