filterVcfBasic: Basic VCF filter function

In PureCN: Copy number calling and SNV classification using targeted short read sequencing

Description

Function to remove artifacts and low confidence/quality variant calls.

Usage

filterVcfBasic(
  vcf,
  tumor.id.in.vcf = NULL,
  use.somatic.status = TRUE,
  snp.blacklist = NULL,
  af.range = c(0.03, 0.97),
  contamination.range = c(0.01, 0.075),
  min.coverage = 15,
  min.base.quality = 25,
  min.supporting.reads = NULL,
  error = 0.001,
  target.granges = NULL,
  remove.off.target.snvs = TRUE,
  model.homozygous = FALSE,
  interval.padding = 50,
  DB.info.flag = "DB"
)

Arguments

vcf	CollapsedVCF object, read in with the readVcf function from the VariantAnnotation package.
tumor.id.in.vcf	The tumor id in the CollapsedVCF (optional).
use.somatic.status	If somatic status and germline data is available, then use this information to remove non-heterozygous germline SNPs or germline SNPS with biased allelic fractions.
snp.blacklist	A file with blacklisted genomic regions. Must be parsable by import from rtracklayer, for a example a BED file with file extension ‘.bed’.
af.range	Exclude variants with allelic fraction smaller or greater than the two values, respectively. The higher value removes homozygous SNPs, which potentially have allelic fractions smaller than 1 due to artifacts or contamination. If a matched normal is available, this value is ignored, because homozygosity can be confirmed in the normal.
contamination.range	Count variants in dbSNP with allelic fraction in the specified range. If the number of these putative contamination variants exceeds an expected value and if they are found on almost all chromosomes, the sample is flagged as potentially contaminated and extra contamination estimation steps will be performed later on.
min.coverage	Minimum coverage in tumor. Variants with lower coverage are ignored.
min.base.quality	Minimim base quality in tumor. Requires a BQ genotype field in the VCF.
min.supporting.reads	Minimum number of reads supporting the alt allele. If NULL, calculate based on coverage and assuming sequencing error of 10^-3.
error	Estimated sequencing error rate. Used to calculate minimum number of supporting reads using calculatePowerDetectSomatic.
target.granges	GenomicRanges object specifiying the target postions. Used to remove off-target reads. If NULL, do not check whether variants are on or off-target.
remove.off.target.snvs	If set to a true value, will remove all SNVs outside the covered regions.
model.homozygous	If set to TRUE, does not remove homozygous variants. Ignored in case a matched normal is provided in the VCF.
interval.padding	Include variants in the interval flanking regions of the specified size in bp. Requires target.granges.
DB.info.flag	Flag in INFO of VCF that marks presence in common germline databases. Defaults to DB that may contain somatic variants if it is from an unfiltered dbSNP VCF.

Value

A list with elements

vcf	The filtered CollapsedVCF object.
flag	A flag (logical(1)) if problems were identified.
flag_comment	A comment describing the flagging.

Author(s)

Markus Riester

See Also

calculatePowerDetectSomatic

Examples

# This function is typically only called by runAbsolute via 
# fun.filterVcf and args.filterVcf.
vcf.file <- system.file("extdata", "example.vcf.gz", package="PureCN")
vcf <- readVcf(vcf.file, "hg19")
vcf.filtered <- filterVcfBasic(vcf)        

PureCN documentation built on Nov. 8, 2020, 5:37 p.m.