VariantAnnotation

Bioconductor package for reading, writing, and annotating genetic variants in VCF format. Full documentation at https://vjcitn.github.io/VariantAnnotation/.

Installation

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("VariantAnnotation")

Function categories

VCF I/O

Read, write, scan, and filter VCF files. ScanVcfParam controls which chromosomal regions, samples, and fields are loaded.

vcf <- readVcf("variants.vcf.gz", genome = "hg19")
writeVcf(vcf, "output.vcf")
filterVcf("variants.vcf.gz", genome = "hg19", destination = "filtered.vcf.gz",
          filters = FilterRules(list(minQ = function(x) !is.na(fixed(x)$QUAL) & fixed(x)$QUAL > 100)))

Key functions: readVcf, writeVcf, scanVcf, scanVcfHeader, filterVcf, indexVcf, VcfFile, ScanVcfParam

---

VCF data structures and accessors

CollapsedVCF (one row per variant, multi-allelic ALT stored as list) and ExpandedVCF (one row per allele) form the core data model. Standard accessors retrieve each VCF component.

ref(vcf)        # reference alleles
alt(vcf)        # alternate alleles (DNAStringSetList for CollapsedVCF)
info(vcf)       # INFO fields as a DataFrame
geno(vcf)$GT    # genotype matrix
header(vcf)     # VCFHeader with meta, contig, and format info
evcf <- VariantAnnotation::expand(vcf)  # CollapsedVCF -> ExpandedVCF

Key functions/classes: VCF, CollapsedVCF, ExpandedVCF, VCFHeader, fixed, ref, alt, qual, filt, info, geno, header, meta, contig, vcfFields, expand

---

Variant classification

Classify variants by mutation type. All functions operate on VCF or VRanges objects and return logical vectors.

isSNV(vcf)          # single nucleotide variants
isIndel(vcf)        # insertions and deletions
isTransition(vcf)   # transitions (A<->G, C<->T)
isSubstitution(vcf) # substitutions (same-length alleles)

Key functions: isSNV, isInsertion, isDeletion, isIndel, isDelins, isTransition, isSubstitution; VariantType class hierarchy (CodingVariants, SpliceSiteVariants, IntronVariants, FiveUTRVariants, ThreeUTRVariants, IntergenicVariants, PromoterVariants)

---

Functional annotation

Map variants to transcripts and predict protein-level consequences.

loc <- locateVariants(vcf, txdb, region = CodingVariants())
coding <- predictCoding(vcf, txdb, seqSource = Hsapiens)
summ <- summarizeVariants(vcf, txdb, mode = CodingVariants())

Key functions: locateVariants, predictCoding, summarizeVariants, getTranscriptSeqs

---

Genotype analysis

Convert VCF genotype calls to a SnpMatrix for downstream association testing (e.g. with the snpStats package).

library(snpStats)
sm <- genotypeToSnpMatrix(vcf)
col.summary(sm$genotypes)

Key functions: genotypeToSnpMatrix, GLtoGP, probabilityToSnpMatrix

---

Variant ranges (VRanges)

VRanges extends GRanges with per-variant fields: alleles, read depth, and soft/hard filter results. Useful for variant calling pipelines.

vr <- VRanges(seqnames = "chr1", ranges = IRanges(1000, 1000),
              ref = "A", alt = "T",
              totalDepth = 30, altDepth = 15,
              sampleNames = "sample1")
altFraction(vr)
asVCF(vr)

Key functions/classes: VRanges, VRangesList, makeVRangesFromGRanges, totalDepth, altDepth, refDepth, altFraction, refFraction, softFilterMatrix, hardFilters, called, asVCF, stackSamples

---

External annotation databases

Interfaces to prebuilt SQLite databases for SIFT, PolyPhen-2, and PROVEAN pathogenicity predictions.

library(SIFT.Hsapiens.dbSNP132)
db <- SIFT.Hsapiens.dbSNP132
SIFTDbColumns(db)
select(db, keys = "rs17970171", columns = SIFTDbColumns(db))

Key classes: SIFTDb, PolyPhenDb, PROVEANDb

---

Ensembl VEP REST API

Query the Ensembl Variant Effect Predictor via its REST endpoint for human variants (hg38 / GRCh38).

resp <- post_Hs_region(chr = "7", pos = 155800001,
                       id = "myVar", ref = "A", alt = "T")
fl <- system.file("extdata", "chr22.vcf.gz", package = "VariantAnnotation")
vcf <- readVcf(fl)
hits <- vep_by_region(vcf[1:10], snv_only = FALSE, chk_max = FALSE)

Key functions: post_Hs_region, vep_by_region