Package {VCFArrow}

Title:	Fast and Memory-Efficient Manipulation and Conversion of VCF Files
Version:	0.9.0
Description:	Reads Variant Call Format (VCF) files into an S4 object whose genotypes are stored on disk as an 'Apache Arrow' dataset and loaded lazily, allowing large datasets to be processed with little memory. Provides functions to filter variants and samples, subset and merge datasets, compute summary statistics and diagnostic plots, write VCF files, and convert data to input formats of population genetic and phylogenetic software such as 'STRUCTURE', 'ADMIXTURE', 'PLINK', 'EIGENSTRAT', 'Arlequin', 'GENEPOP', 'BayeScan', 'TreeMix', 'SNAPP' and 'adegenet'.
URL:	https://legallab.github.io/VCFArrow/, https://github.com/legalLab/VCFArrow
BugReports:	https://github.com/legalLab/VCFArrow/issues
Language:	en-US
License:	MIT + file LICENSE
Encoding:	UTF-8
Depends:	R (≥ 4.1.0)
LinkingTo:	Rcpp
Imports:	Rcpp, arrow, cli, dplyr, ggplot2, methods, parallel, stats, stringr, tibble, utils, vctrs
Suggests:	adegenet, knitr, rmarkdown, svglite
VignetteBuilder:	knitr
SystemRequirements:	zlib
Config/roxygen2/version:	8.1.0
NeedsCompilation:	yes
Packaged:	2026-10-07 03:07:51 UTC; tomas
Author:	Tomas Hrbek [aut, cre, cph]
Maintainer:	Tomas Hrbek <hrbek@evoamazon.net>
Repository:	CRAN
Date/Publication:	2026-10-16 13:20:38 UTC

VCFArrow: Fast and Memory-Efficient Manipulation and Conversion of VCF Files

Description

Reads Variant Call Format (VCF) files into an S4 object whose genotypes are stored on disk as an 'Apache Arrow' dataset and loaded lazily, allowing large datasets to be processed with little memory. Provides functions to filter variants and samples, subset and merge datasets, compute summary statistics and diagnostic plots, write VCF files, and convert data to input formats of population genetic and phylogenetic software such as 'STRUCTURE', 'ADMIXTURE', 'PLINK', 'EIGENSTRAT', 'Arlequin', 'GENEPOP', 'BayeScan', 'TreeMix', 'SNAPP' and 'adegenet'.

Author(s)

Maintainer: Tomas Hrbek hrbek@evoamazon.net (ORCID) [copyright holder]

Authors:

• Tomas Hrbek hrbek@evoamazon.net (ORCID) [copyright holder]

See Also

Useful links:

• https://legallab.github.io/VCFArrow/

• https://github.com/legalLab/VCFArrow

• Report bugs at https://github.com/legalLab/VCFArrow/issues

VCFArrow S4 class

Description

S4 class holding VCF data. Variant metadata are kept in memory while genotypes are stored on disk as an Apache Arrow dataset and loaded lazily. Objects are created with read_vcf().

Slots

header

Complete VCF header (character).

info

INFO field of each variant (character).

format

FORMAT field of each variant (data.frame).

variants

CHROM, POS, ID, REF, ALT, QUAL, FILTER and precalculated per-variant metrics (data.frame).

gt

Arrow dataset of genotypes in long format.

samples

Sample names (character).

groups

Group assignment of each sample (character).

path

Location of the on-disk Arrow dataset (character).

finalizer_env

Environment used to clean up path on garbage collection.

loci

CHROM, POS, REF and ALT of every variant of the source VCF, row i describing .row_id i; never filtered, so variants can be identified (e.g. by vcf_bind()) even after they were removed from variants (data.frame).

invariant_removed

Variants removed as invariant, with their metadata (data.frame).

Author(s)

Tomas Hrbek April 2026

Subset method for VCFArrow

Description

Subset a VCFArrow object by variants (rows) and samples (columns)

Usage

## S4 method for signature 'VCFArrow'
x[i, j, ..., drop = FALSE]

Arguments

Details

This function is a method of the VCFArrow S4 class Method to subset by row and column of GT

Value

A new VCFArrow object containing the selected variants and samples

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f)
vcf[1:100, 1:5]

assess_vcf_coverage

Description

Quantifying read depth of all samples in VCF

Usage

assess_vcf_coverage(
  vcf_arrow,
  res_path,
  species,
  project,
  details = TRUE,
  max_points_per_sample = 5000L
)

Arguments

Details

This function generates a violin plot of read depths per sample.

Value

violin plot of read depth for each sample in VCF

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
assess_vcf_coverage(vcf, res_path = tempdir(),
                    species = "Phyllomedusa vaillantii",
                    project = "vaillantii")

assess_vcf_missing_data

Description

Quantifying missing data of all samples in VCF

Usage

assess_vcf_missing_data(vcf_arrow, res_path, species, project, details = TRUE)

Arguments

Details

This function generates a dataframe of absolute and relative missing data per sample, and a plot of relative missing data % per sample.

Value

dataframe and plot of missing data for each sample in VCF

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
assess_vcf_missing_data(vcf, res_path = tempdir(),
                        species = "Phyllomedusa vaillantii",
                        project = "vaillantii")

read_vcf

Description

Read VCF file and store its content within a VCFArrow object

Usage

read_vcf(vcf_file, chunk_size = 50000)

Arguments

Details

This function read a VCF file into an S4 class object in chunks, returning a VCFArrow object. It accepts both uncompressed and gz compressed files. The GT field is stored as an Apache Arrow in Long format. Various metrics are precalculated for fast and easy filtering. GT slot content stored in a TEMP directory for lazy loading.

Value

VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f)
vcf

set_vcf_groups

Description

Associate samples with group info based on "strt" file

Usage

set_vcf_groups(vcf_arrow, data_path, strt = "strata")

Arguments

Details

This function creates list of factors of sample-to-group assignments and groups based on a strata and group file. The strt is a tsv file with 2+ columns, with 2 columns named 'id' and 'pop'. By default strt is named 'strata' and found in the datapath where the VCF file is found.

Value

vcf_arrow object with groups slot filled

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- set_vcf_groups(read_vcf(f), data_path = dirname(f),
                      strt = "strata")
table(vcf@groups)

Show method for VCFArrow

Description

Print a summary of a VCFArrow object

Usage

## S4 method for signature 'VCFArrow'
show(object)

Arguments

Details

This function is a method of the VCFArrow S4 class Method to show object content summary

Value

Invisibly returns object; called for its side effect of printing a summary.

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f)
show(vcf)

vcf2admixture

Description

Converts a VCFArrow object to a PLINK .bed format infile plus ADMIXTURE-specific .pop file

Usage

vcf2admixture(
  vcf_arrow,
  out_file,
  keep_groups = NULL,
  sex = NULL,
  pheno = NULL,
  chrom_code = c("zero", "auto", "index", "keep"),
  supervised = FALSE,
  reference_groups = NULL
)

Arguments

Details

This function converts a VCFArrow object to an external PLINK .bed formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups. Sex and phenotype vectors are optional. If not defined sex = 0, pheno = -9. For supervised analyses, the supervised flag needs to be true to generate .pop file needed for ADMIXTURE's supervised mode. The reference_groups vector specifies ancestral information of individuals. When NULL, or for any missing individual, ancestry is inferred.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2admixture(vcf, out_file = file.path(tempdir(), "admixture_in"))
vcf2admixture(vcf, out_file = file.path(tempdir(), "admixture_sup"),
              supervised = TRUE,
              reference_groups = c("GS", "BS", "WA"))

vcf2apparent

Description

Converts a VCFArrow object to an Apparent format infile

Usage

vcf2apparent(vcf_arrow, out_file, keep_groups = NULL, key = "All")

Arguments

Details

This function converts a VCFArrow object to an external SmartSNP formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups. Possible relationships defined by the parameter 'kee' are All, Pa, Mo, Fa, Off.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2apparent(vcf, out_file = tempfile(fileext = ".txt"))

vcf2arlequin

Description

Converts a VCFArrow object to an Arlequin infile

Usage

vcf2arlequin(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external Arlequin formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2arlequin(vcf, out_file = tempfile(fileext = ".arp"))

vcf2bayesass

Description

Converts a VCFArrow object to BayesAss format infile

Usage

vcf2bayesass(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external BayesAss formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2bayesass(vcf, out_file = tempfile(fileext = ".immanc"))

vcf2bayescan

Description

Converts a VCFArrow object to Bayescan format infile

Usage

vcf2bayescan(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external BayesAss formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2bayescan(vcf, out_file = tempfile(fileext = ".txt"))

vcf2eigenstrat

Description

Converts a VCFArrow object to Eigenstrat format infile

Usage

vcf2eigenstrat(vcf_arrow, out_file, keep_groups = NULL, sex = NULL)

Arguments

Details

This function converts a VCFArrow object to an external BayesAss formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2eigenstrat(vcf, out_file = file.path(tempdir(), "eigenstrat_in"))

vcf2fasta

Description

Converts a VCFArrow object to a FASTA format infile

Usage

vcf2fasta(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external FASTA formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2fasta(vcf, out_file = tempfile(fileext = ".fas"))

vcf2fineradstructure

Description

Converts a VCFArrow object to FineRadStructure format infile

Usage

vcf2fineradstructure(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external Treemix formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_linked <- vcf_filter_multiSNV(vcf)
vcf2fineradstructure(vcf_linked, out_file = tempfile(fileext = ".txt"))

vcf2genepop

Description

Converts a VCFArrow object to a Genepop format infile

Usage

vcf2genepop(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external Genpop formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2genepop(vcf, out_file = tempfile(fileext = ".gen"))

vcf2genlight

Description

Converts a VCFArrow object to Genlight format infile

Usage

vcf2genlight(
  vcf_arrow,
  out_file = NULL,
  keep_groups = NULL,
  ploidy = 2L,
  save = FALSE
)

Arguments

Details

This function converts a VCFArrow object to an adegenet Genlight object. It requires the 'adegenet' package, which VCFArrow only suggests: install it with install.packages("adegenet") to use this function. Genotypes are read in chunks whose size is determined by the read_vcf() function. If no groups are defined, the default behavior is to use all groups. Genlight objects can encode polyploid genomes, by default diploid genomes are assumed. Genlight objects are in-memory S4 objects and are always returned; if save = TRUE, the object is also written to out_file with saveRDS(). Unlike the other exporters, out_file is optional because nothing is written unless save = TRUE.

Value

An adegenet genlight object.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
gl <- vcf2genlight(vcf)
gl
gl <- vcf2genlight(vcf, out_file = tempfile(fileext = ".rds"),
                   save = TRUE)

vcf2gt_long

Description

Converts a VCFArrow object to tidy long format infile

Usage

vcf2gt_long(
  vcf_arrow,
  out_file,
  keep_groups = NULL,
  format = c("feather", "parquet", "csv"),
  col_select = NULL
)

Arguments

Details

This function converts a VCFArrow object to an external SmartSNP formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups. The tidy data can be saved in either Arrow, Parquet or CSV formats. File extension is added automatically if missing. Optionally, specific columns can be saved, by default all columns are saved. The gt long slot contains pre-calculated metrics in addition to just genotypes.

Value

Invisibly returns the path of the written file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2gt_long(vcf, out_file = tempfile(), format = "csv")

vcf2migrate

Description

Converts a VCFArrow object to a MIGRATE-N format infile

Usage

vcf2migrate(
  vcf_arrow,
  out_file,
  keep_groups = NULL,
  block_size = 100L,
  method = "S"
)

Arguments

Details

This function converts a VCFArrow object to an external MIGRATE-N formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups. The function implements two different SNP formats, new (un)linked SNPs (S) - fixed blocks, and new (un)linked SNPs (N) - chromosome intervals. Alleles are sequence data when SNPs are mapped to a reference. The S option generates blocks of a specific number of SNPs and treats them as linked; this is appropriate if SNPs are extracted without a reference. The N format extracts all SNPs within a chromosome, within the block size and treats them as linked; this is appropriate if SNPs are mapped against a reference. The size of the linked block is determined by the block_size parameter. See the Migrate-N documentation (https://peterbeerli.com/migrate-html5/) for format detail.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2migrate(vcf, out_file = tempfile(fileext = ".txt"))

vcf2nexus

Description

Converts a VCFArrow object to a NEXUS format infile

Usage

vcf2nexus(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external NEXUS formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2nexus(vcf, out_file = tempfile(fileext = ".nex"))

vcf2plink_bed

Description

Converts a VCFArrow object to a PLINK .bed format infile

Usage

vcf2plink_bed(
  vcf_arrow,
  out_file,
  keep_groups = NULL,
  sex = NULL,
  pheno = NULL,
  chrom_code = c("auto", "index", "zero", "keep")
)

Arguments

Details

This function converts a VCFArrow object to an external PLINK .bed formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups. Sex and phenotype vectors are optional. If not defined sex = 0, pheno = -9.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2plink_bed(vcf, out_file = file.path(tempdir(), "plink_out"))

vcf2plink_ped

Description

Converts a VCFArrow object to a PLINK .ped format infile

Usage

vcf2plink_ped(
  vcf_arrow,
  out_file,
  keep_groups = NULL,
  sex = NULL,
  pheno = NULL,
  chrom_code = c("auto", "index", "zero", "keep")
)

Arguments

Details

This function converts a VCFArrow object to an external PLINK .ped formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups. Sex and phenotype vectors are optional. If not defined sex = 0, pheno = -9.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2plink_ped(vcf, out_file = file.path(tempdir(), "plink_out"))

vcf2related

Description

Converts a VCFArrow object to Related format infile

Usage

vcf2related(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external Related formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2related(vcf, out_file = tempfile(fileext = ".txt"))

vcf2smartsnp

Description

Converts a VCFArrow object to smartsnp tabular format infile

Usage

vcf2smartsnp(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external SmartSNP formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2smartsnp(vcf, out_file = tempfile(fileext = ".txt"))

vcf2snapp

Description

Converts a VCFArrow object to a SNAPP encoded NEXUS format infile

Usage

vcf2snapp(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external SNAPP/SNAPPER formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2snapp(vcf, out_file = tempfile(fileext = ".nex"))

vcf2snmf

Description

Converts a VCFArrow object to an sNMF format infile

Usage

vcf2snmf(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external sNMF formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2snmf(vcf, out_file = tempfile(fileext = ".geno"))

vcf2structure

Description

Converts a VCFArrow object to a Structure or FastStructure infile

Usage

vcf2structure(vcf_arrow, out_file, keep_groups = NULL, method = "S")

Arguments

Details

This function converts a VCFArrow object to an external Structure or FastStructure formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups. The flag parameter controls whether Structure (flag = 'S') or FastStructure (flag = 'F') formatted output is written out.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2structure(vcf, out_file = tempfile(fileext = ".str"))
vcf2structure(vcf, out_file = tempfile(fileext = ".fstr"),
              method = "F")

vcf2treemix

Description

Converts a VCFArrow object to Treemix format infile

Usage

vcf2treemix(vcf_arrow, out_file, keep_groups = NULL)

Arguments

Details

This function converts a VCFArrow object to an external Treemix formatted file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. If no groups are defined, the default behavior is to use all groups.

Value

Invisibly returns the input VCFArrow object; called for its side effect of writing out_file.

Author(s)

Tomas Hrbek May 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf2treemix(vcf, out_file = tempfile(fileext = ".txt"))

vcf_bind

Description

Bind two or more VCFArrow objects into a new VCFArrow object

Usage

vcf_bind(
  ...,
  mode = c("intersect", "union"),
  absent_as = c("missing", "hom_ref"),
  outgroup = FALSE,
  recover_loci = FALSE,
  .check = NULL
)

Arguments

Details

This function binds two or more VCFArrow objects with different sets of individuals, returning a new VCFArrow object. Sample names must be unique across objects. The objects need not have the same variants: mode = "intersect" keeps the variants present in all objects, mode = "union" keeps the variants present in any object and codes the genotypes an object lacks according to absent_as. For objects with the same variants both modes give the same result. In mode = "union", variants an object removed stay removed for it: a variant that one object removed by filtering but that another object contains is included, with the genotypes of the object that removed it coded according to absent_as. Binding in steps and binding all objects in one call then give the same result. With recover_loci = TRUE, the real genotypes of such variants are restored instead, read from the genotype files of the object that removed them (which keep every variant the object had when its files were written). This suits joining separately filtered datasets; binding in steps can then differ from binding in one call.

outgroup = TRUE is meant for adding outgroups to an ingroup that was filtered on its own (so that properties of the outgroup, such as its missingness, cannot remove ingroup variants): the result keeps exactly the variants of the first object, and variants removed from it stay removed. The other objects contribute their genotypes for these variants, read from their genotype files wherever present (including variants they removed, for example as invariant); genotypes they do not have are coded according to absent_as.

Variants are matched by CHROM, POS, REF and ALT, so the objects may come from different VCF files (e.g. separate sequencing runs); REF and ALT must be written the same way in all of them (including the order of alternative alleles). This is assured when all datasets were called against the same reference genome assembly, which gives the same CHROM, POS and REF for the same locus. It is generally not possible for de novo (reference-free) SNP calling, for example with DiscoSnp, where locus names and positions are generated anew in each calling run, so the same locus will usually not match between separately called datasets; such datasets should be called together in a single run instead. Multi-allelic sites can list their ALT alleles in a different order in different files, so it is best to normalize the VCFs first, for example by splitting multi-allelic sites with ⁠bcftools norm -m -any⁠. Variants keep the .row_id of the first object where its source VCF has them, other variants get new ids; the variant metadata of the first object and, in mode = "intersect", its record of removed invariant variants are kept, so bound objects can be bound again.

Value

VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
ingroup <- vcf_extract_groups(vcf, c("GS", "BS", "WA"))
outgroup <- vcf_extract_groups(vcf, "OG")
vcf_bind(ingroup, outgroup, mode = "intersect")
vcf_bind(ingroup, outgroup, mode = "union", absent_as = "missing")
vcf_bind(vcf_filter_missingness(ingroup, 0.2), outgroup, outgroup = TRUE,
         absent_as = "missing")

vcf_bind_sparse

Description

Deprecated: use vcf_bind(), which now binds objects with different variants (see its mode and absent_as arguments). vcf_bind_sparse(...) is vcf_bind(..., recover_loci = TRUE).

Usage

vcf_bind_sparse(
  ...,
  mode = c("intersect", "union"),
  absent_as = c("missing", "hom_ref")
)

Arguments

Value

VCFArrow object

vcf_copy

Description

Copy a VCFArrow object

Usage

vcf_copy(vcf_arrow)

Arguments

Details

This function copies a VCFArrow object. It makes an actual physical copy of the object.

Value

VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_copy(vcf_filter_maf(vcf, 0.05))

vcf_extract_groups

Description

Extract groups of samples from a VCFArrow object

Usage

vcf_extract_groups(
  vcf_arrow,
  groups,
  keep = TRUE,
  f_invar = TRUE,
  verbose = TRUE
)

Arguments

Details

This function removes groups of samples from a VCFArrow object, returning a new VCFArrow object. It uses the 'keep' flag to either keep or drop the groups in the list. By default will remove any loci that may have become invariant as the result of the removal of samples. By default will report removed samples, final % missing data, and number of retained samples after sample filtering.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_extract_groups(vcf, groups = c("GS", "BS"))
vcf_extract_groups(vcf, groups = "OG", keep = FALSE)

vcf_extract_samples

Description

Extract samples from a VCFArrow object

Usage

vcf_extract_samples(
  vcf_arrow,
  samples,
  keep = TRUE,
  f_invar = TRUE,
  verbose = TRUE
)

Arguments

Details

This function removes samples from a VCFArrow object, returning a new VCFArrow object. It uses the 'keep' flag to either keep or drop the samples in the list. By default will remove any loci that may have become invariant as the result of the removal of samples. By default will report removed samples, final % missing data, and number of retained samples after sample filtering.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_extract_samples(vcf, samples = c("Pv14", "Pv27", "Pv28"))
vcf_extract_samples(vcf, samples = c("Pb2Jp", "Pb2Scx", "Pb1Rd"),
                    keep = FALSE)

vcf_filter_adr

Description

Correct or remove genotypes with > % ALT/REF ratio from a VCFArrow object

Usage

vcf_filter_adr(
  vcf_arrow,
  mode = c("correct", "remove"),
  threshold = 0.1,
  f_invar = TRUE
)

Arguments

Details

This function either changes genotypes or makes genotypes missing in a VCFArrow object if they have above/below threshold normalized ADR ratio, returning a new VCFArrow object. The ADR is calculated as ADR = ALT / (REF + ALT). If ADR > threshold, the genotype becomes ALT homozygous. If ADR < threshold, the genotype becomes REF homozygous. By default will remove any loci that may have become invariant as the result of the change/removal of genotypes

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_adr(vcf, mode = "correct")

vcf_filter_biallelic

Description

Remove non-biallelic loci from a VCFArrow object

Usage

vcf_filter_biallelic(vcf_arrow)

Arguments

Details

This function removes non-biallelic loci from a VCFArrow object, returning a new VCFArrow object.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_biallelic(vcf)

vcf_filter_coverage

Description

Remove genotypes below read DP threshold from a VCFArrow object

Usage

vcf_filter_coverage(vcf_arrow, threshold = 10)

Arguments

Details

This function removes genotypes below a DP threshold from a VCFArrow object, returning a new VCFArrow object.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_coverage(vcf, threshold = 10)

vcf_filter_hets

Description

Remove loci above a heterozigosity threshold from a VCFArrow object

Usage

vcf_filter_hets(vcf_arrow, threshold = 0.5)

Arguments

Details

This function removes loci above a heterozigosity threshold from a VCFArrow object, returning a new VCFArrow object. High heterozigosities are indicative of potential paralogs.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_hets(vcf, threshold = 0.5)

vcf_filter_indels

Description

Remove indels from a VCFArrow object

Usage

vcf_filter_indels(vcf_arrow)

Arguments

Details

This function removes indel loci from a VCFArrow object, returning a new VCFArrow object.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_indels(vcf)

vcf_filter_invariant

Description

Remove invariant loci from a VCFArrow object

Usage

vcf_filter_invariant(vcf_arrow)

Arguments

Details

This function removes invariant loci from a VCFArrow object, returning a new VCFArrow object. This might be desirable after subsetting a VCFArrow object by individuals.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_invariant(vcf)

vcf_filter_maf

Description

Remove loci below MAF threshold from a VCFArrow object

Usage

vcf_filter_maf(vcf_arrow, threshold = 0.05)

Arguments

Details

This function removes loci below MAF threshold from a VCFArrow object, returning a new VCFArrow object.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_maf(vcf, threshold = 0.05)

vcf_filter_missing

Description

Remove samples with > % missing data from a VCFArrow object

Usage

vcf_filter_missing(vcf_arrow, threshold = 0.5, f_invar = TRUE, verbose = TRUE)

Arguments

Details

This function removes samples from a VCFArrow object if they have above threshold missing loci, returning a new VCFArrow object. By default will remove any loci that may have become invariant as the result of the removal of samples. By default will report removed samples, final % missing data, and number of retained samples after sample filtering.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf |>
  vcf_filter_missingness(threshold = 0.2) |>
  vcf_filter_missing(threshold = 0.3)

vcf_filter_missingness

Description

Remove loci above missingness threshold from a VCFArrow object

Usage

vcf_filter_missingness(vcf_arrow, threshold = 0.1, verbose = TRUE)

Arguments

Details

This function removes loci above missingness threshold from a VCFArrow object, returning a new VCFArrow object. Missingness is a locus focused metric, i.e. missing data per locus.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_missingness(vcf, threshold = 0.2)

vcf_filter_multiSNV

Description

Subset a VCFArrow object keeping only loci with 2+ SNVs per locus

Usage

vcf_filter_multiSNV(vcf_arrow, block_size = 10000, minSNV = 2, maxSNV = 5)

Arguments

Details

This function subsets a VCFArrow object keeping only loci with between min and max # of SNVs per locus, returning a new VCFArrow object. Default min = 2 and max = 5 SNVs per locus (recommended as input for fineRADstructure analyses). Locus is defined as a different chromosome or a block of the 'block_size' parameter value within a chromosome.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_multiSNV(vcf, block_size = 10000, minSNV = 2, maxSNV = 5)

vcf_filter_oneSNV

Description

Subset a VCFArrow object keeping only 1 SNV per locus

Usage

vcf_filter_oneSNV(vcf_arrow, block_size = 10000)

Arguments

Details

This function subsets a VCFArrow object keeping only 1 SNV per locus, returning a new VCFArrow object. Locus is defined as a different chromosome or a block of the 'block_size' parameter value within a chromosome. The first SNV independent of quality is taken (may modify this in the future).

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_oneSNV(vcf)

vcf_filter_paralogs

Description

Remove loci above REPEAT and below RANK threshold from a VCFArrow object

Usage

vcf_filter_paralogs(
  vcf_arrow,
  threshold_rk = 0.2,
  threshold_rpt = 0.5,
  keep_na = FALSE
)

Arguments

Details

This function removes loci below the RANK threshold and above the REPEAT threshold from a VCFArrow object, returning a new VCFArrow object. It applies vcf_filter_rank() and then vcf_filter_rpt(). RANK and REPEAT metrics focus on different paralog signals, and used jointly maximize true positive and minimize false positive detections. Simulations indicate that removing loci with Rk < 0.2 or RPT > 0.5 gives the best results, with Rk < 0.4 or RPT > 0.5 being slightly more conservative.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek September 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_paralogs(vcf, threshold_rk = 0.2, threshold_rpt = 0.5)

vcf_filter_pass

Description

Remove loci that did not PASS FILTER in a VCFArrow object

Usage

vcf_filter_pass(vcf_arrow)

Arguments

Details

This function removes loci that did not PASS the FILTER in a VCFArrow object, returning a new VCFArrow object. PASS indicates that a variant has successfully passed all applied quality control filters. When no filters were applied and the FILTER field is '.' (Dot), vcf_filter_pass() defaults to passing the locus.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_pass(vcf)

vcf_filter_quality

Description

Remove loci below QUAL threshold from a VCFArrow object

Usage

vcf_filter_quality(vcf_arrow, threshold = 30)

Arguments

Details

This function removes loci below QUAL threshold from a VCFArrow object, returning a new VCFArrow object. QUAL is a Phred-scaled quality metric assessing whether any variant exists at that position across all samples. QUAL is different from PL which represents the confidence in a genotype call of a specific sample. When QUAL not reported '.' (Dot), QUAL is assumed to be above threshold.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_quality(vcf, threshold = 30)

vcf_filter_rank

Description

Remove loci below RANK threshold from a VCFArrow object

Usage

vcf_filter_rank(vcf_arrow, threshold = 0.4, keep_na = FALSE)

Arguments

Details

This function removes loci below RANK threshold from a VCFArrow object, returning a new VCFArrow object. RANK is calculated by kissreads2 in DiscoSNP-RAD (Gauthier et. al. 2020) and registered as Rk in INFO. RANK is calculated as sqrt(chi-sqr/n) of allele read counts, and used for paralog detection -> very low rank values (<0.4) are indicative of paralogs.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_rank(vcf, threshold = 0.5)

vcf_filter_rpt

Description

Remove loci above REPEAT threshold from a VCFArrow object

Usage

vcf_filter_rpt(vcf_arrow, threshold = 0.5, keep_na = FALSE)

Arguments

Details

This function removes loci above REPEAT threshold from a VCFArrow object, returning a new VCFArrow object. REPEAT is calculated by disco_haplotypes in DiscoSNP-RAD (Gauthier et. al. 2020) and registered as RPT in INFO. REPEAT is calculated as max of overlap, depth score and is 1 if in a giant-component bubble, and is used for paralog detection -> high repeat values (>0.5) are indicative of paralogs.

Value

subsetted VCFArrow object

Author(s)

Tomas Hrbek September 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_filter_rpt(vcf, threshold = 0.5)

Garbage-collect VCFArrow temp directories

Description

Triggers R's garbage collector (three full passes to handle the finalizer -> pending-queue -> unlink chain), then flushes any directories whose reference count has already reached zero.

Usage

vcf_gc(force = FALSE, verbose = TRUE)

Arguments

Value

Invisibly returns NULL; called for its side effect of deleting temporary directories.

Typical workflow

  rm(vcf1, vcf2, vcf3)
  vcf_gc()              # usually sufficient
  vcf_gc(force = TRUE)  # if directories remain after rm()

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f)
rm(vcf)
vcf_gc()

Estimate RAM required for a VCFArrow export operation

Description

Estimate RAM required for a VCFArrow export operation

Usage

vcf_memory_estimate(
  vcf_arrow,
  keep_groups = NULL,
  format = c("individual", "pop", "chunk"),
  chunk_size = 100000L,
  lowmem = FALSE
)

Arguments

Value

Invisibly, a list with elements:

• n_var: number of variants

• n_samples: number of samples retained

• n_pops: number of populations

• chunk_arrow_bytes: Arrow memory per chunk read

• matrix_bytes: size of the accumulation matrices

• peak_bytes: estimated peak memory

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_memory_estimate(vcf, format = "individual")

vcf_set_workers

Description

Set the number of parallel worker processes used by VCFArrow

Usage

vcf_set_workers(workers = 1L)

Arguments

Details

Functions that process the on-disk genotype chunks independently distribute the chunks over workers background R processes: read_vcf(), the genotype-based filters, sample and group extraction, vcf_stats() and the assess functions, vcf_bind(), write_vcf(), the vcf2*() exporters and vcf2gt_long() with CSV output (with feather or Parquet output it runs serially). Results do not depend on the number of workers. The workers are started once and reused until vcf_set_workers(1) is called or the session ends. Each worker holds one chunk in memory at a time, so memory use grows with the number of workers (see vcf_memory_estimate()). Useful values are up to the number of physical CPU cores; disk speed limits the gain beyond that.

Value

Invisibly, the previous number of workers.

Author(s)

Tomas Hrbek September 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
# 2 workers (CRAN examples may use at most 2 cores)
vcf_set_workers(2)
vcf <- read_vcf(f)
vcf_set_workers(1)

vcf_stats

Description

Calculates basic stats of each samples from VCFArrow format data. Includes average read depth per individual, missing data per individual, Watterson's theta and pi.

Usage

vcf_stats(vcf_arrow, res_path, project, theta = FALSE)

Arguments

Details

This function calculates average read depth, heterozygosity number of heterozygotes, number of reference and alternative homozygotes, missing data and total number SNPs of each sample in an VCFArrow object. Optionally calls vcf_theta() to get total and group Watterson's theta and pi.

Value

table of statistics

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_stats(vcf, res_path = tempdir(), project = "vaillantii",
          theta = TRUE)

vcf_sub_loci

Description

Randomly subsets SNVs from a VCFArrow object

Usage

vcf_sub_SNVs(vcf_arrow, n_SNVs = 10000, seed = NULL)

Arguments

Details

This function subsets a VCFArrow object to specific number of SNVs, returning new VCFArrow object. The seed for random number generator is automatically generated unless otherwise specified.

Value

VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_sub_SNVs(vcf, n_SNVs = 1000, seed = 42)

vcf_sub_loci_stratified

Description

Randomly subsets SNVs from a VCFArrow object

Usage

vcf_sub_SNVs_stratified(vcf_arrow, n_SNVs = 1000, seed = NULL)

Arguments

Details

This function subsets a VCFArrow object to specific number of SNVs, returning new VCFArrow object. The subsampling is stratified, i.e. the same proportion of SNVs per CHROM. The seed for random number generator is automatically generated unless specified.

Value

VCFArrow object

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_sub_SNVs_stratified(vcf, n_SNVs = 1000, seed = 42)

Suggest a chunk_size for read_vcf() given available RAM

Description

Suggest a chunk_size for read_vcf() given available RAM

Usage

vcf_suggest_chunk_size(available_gb, n_samples, n_columns = 5L)

Arguments

Value

Invisibly, the suggested chunk size (numeric).

Examples

vcf_suggest_chunk_size(available_gb = 8, n_samples = 100)

vcf_theta

Description

Calculates basic Watterson's theta and pi for all samples and for sample groups from VCFArrow format data.

Usage

vcf_theta(vcf_arrow, keep_groups = NULL)

Arguments

Details

This function calculates Watterson's theta and pi for the entire VCFArrow object, and for groups of individuals whose grouping is indicated by the groups slot in the VCFArrow object. The statistics are accumulated chunk by chunk as running sums per group, so memory use does not grow with the number of variants, and chunks are processed in parallel when workers are set with vcf_set_workers().

Value

list of statistics

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
vcf_theta(vcf)

write_vcf

Description

Write a VCFArrow object to an external VCF file

Usage

write_vcf(vcf_arrow, out_file, gzip = FALSE)

Arguments

Details

This function writes a VCFArrow object to an external VCF file. Writing occurs in chunks whose size is determined by the read_vcf() function. Larger chunks result in faster writing speeds. It writes both uncompressed and gz compressed files. Compressing increases writing time be about 50%. For large files, it is recommended to output an uncompressed VCF file, and then compress with GZIP or PIGZ.

Value

Invisibly returns the path of the written file.

Author(s)

Tomas Hrbek April 2026

Examples

f <- system.file("extdata", "vaillantii_discosnp_sub.vcf.gz",
                 package = "VCFArrow")
vcf <- read_vcf(f) |> set_vcf_groups(dirname(f))
write_vcf(vcf, out_file = tempfile(fileext = ".vcf"))
write_vcf(vcf, out_file = tempfile(fileext = ".vcf.gz"),
          gzip = TRUE)