VariantCaller

Calls genomic sites of variation using the specified caller. Implemented callers are (see 'caller'):

Samtools caller uses the mpileup, bcftools and vcfutils commands piped to call single-sample variants with simple depth filter. Samtools is simple yet powerful.
Genome Analysis Toolkit (GATK) caller uses the UnifiedGenotyper to call single- or multi-sample variants. Multi-sample calling enables simultaneous calling of numerous samples for improved results using information from each sample.
VarScan caller uses samtools mpileup followed by VarScan to call single- or paired-sample variants. Paired calling enables comparison of normal and tumor sample and outputs the somatic mutations.

The caller should be specified using the parameter, as well as the path to the software unless environment variables are used (GATK_HOME, VARSCAN_HOME pointing to installation directories). Samtools executables "samtools", "bcftools" and "vcfutils.pl" must be in PATH and made executable (vcfutils.pl).

The options parameter can be used to add additional software options in the software specific format. For samtools caller the options are added to only the mpileup command. The following options are hard implemented:

Samtools: varFilter with depth cap at 150
GATK: nothing
VarScan: strand filter is turned on; always output VCF (non-paired); only use reads with mapping quality larger than zero; in somatic (paired) mode process the outputs to separate high quality and low quality somatic, germline and LOH variants.

Here is a list of caller specific options that should be considered depending on the data:

Samtools: -q and -Q for mapping and base quality caps (smaller than specified are skipped)
GATK: --annotation for specific annotation or --group for annotation group; --min_base_quality_score; -stand_call_conf for calling threshold (=variant) and -stand_emit_conf for emitting threshold (=possible variant but filtered)
VarScan non-paired: --min-coverage; --p-value for calling threshold; --min-avg-qual for quality threshold; --min-reads2 for minimum reads supporting variant
VarScan paired: --min-coverage; --somatic-p-value for threshold for somatic calling

The output is always an array with one or more files. Here follows a description of the files specific for the caller:

Samtools: key "vcf" gives the calls.
GATK: key "vcf" gives the calls, "idx" gives the VCF index file and "metrics" gives additional calling metrics file.
VarScan non-paired: key "vcf" gives the calls.
VarScan paired: keys "snp" and "indel" give the variants and "validation" gives all confident sites (when 'variantsOnly' is true)

Complete documentation:

Version	1.0
Bundle	sequencing
Categories	VariationAnalysis
Authors	Rony Lindell (rony.lindell@helsinki.fi), Riku Louhimo (Riku.Louhimo@Helsinki.FI)
Issue tracker	View/Report issues
Source files	component.xml function.scala
Usage	Example with default values VariantCaller( reference = FASTA, //bam1 = BAM, //bam2 = BAM, //bams = BAMList, //intervals = BED, //dbsnp = VCF, callIndels = true, caller = "gatk", exclude = "none", gatk = "", memory = "4g", options = "", region = "all", samOptions4VS = "", threads = 1, variantsOnly = true, varscan = "" )

Inputs

Name	Type	Mandatory	Description
reference	FASTA	Mandatory	The reference genome together with possible auxiliary files.
bam1	BAM	Optional	Input bam file of normal or tumor sample. The BAM index file (.bai) should be located in the same directory when using GATK.
bam2	BAM	Optional	Tumor bam file to call from in VarScan comparison calling. The variants from 'bam1' and 'bam2' will be compared in order to separate germline and somatic variants. [varscan only]
bams	BAMList	Optional	File containing newline separated paths to bam files. When using this input GATK usage will be forced and multiple sample variant calling using the UnifiedGenotyper will be executed. One multi-sample VCF file will be produced. [gatk only] The extension of the file must be `.list`.
intervals	BED	Optional	File with genomic intervals to operate on. Only calls hitting these areas will be output. Note: This can be used to mask out introns in exome sequencing by providing an BED file containing all exonic regions.
dbsnp	VCF	Optional	File with known SNPs, usually from latest dbSNP distribution. The file is used in GATK to improve calling and for annotation. [gatk only] The file must be pointed to by the key 'vcf'.

Outputs

Name	Type	Description
snp	VCF	Output snp (and indel) calls. When indels are written into the same file (GATK, Samtools), this file will contain both snps and indels.
indel	VCF	Output indel calls. VarScan writes snps and indels into separate files.
metrics	TextFile	Calling metrics. GATK will produce a file with some information about the calling results.

Parameters

Name	Type	Default	Description
callIndels	boolean	true	If false, indel calling will be skipped and only snps are called.
caller	string	"gatk"	Software/algorithm to use to call variants. Possible values: {samtools, gatk, varscan}.
exclude	string	"none"	Genomic region from which to exclude reads (i.e., skip) in the same format as 'region'. [gatk only]
gatk	string	""	Path to GATK jar file, e.g. "/opt/gatk/GenomeAnalysisTK.jar". If empty string is given (default), GATK_HOME environment variable is assumed to point to the GATK directory where GenomeAnalysisTK.jar is located.
memory	string	"4g"	The amount of java-heap memory being allocated to GATK, given in the format "4g" for 4 gigabytes or "2560m" for 2560 megabytes (2,5g) etc. For optimal performance this should be a multiple of the threads used (see `threads`). [gatk only]
options	string	""	This string will be added to the command and can include any number of options in the software specific format, e.g. "-q 1" will skip zero-quality alignments or "-stand_emit_conf 10.0" to emit calls with quality higher than 10. See software specific documentation for more information.
region	string	"all"	Genomic region from which to select reads, e.g. "chr1" or "chr2:1-20000". The region can also be a comma separated list, e.g. "chr1,chr2:1-20000,chrX:1-4000". A more extensive list of regions can be defined as a file using the 'intervals' input. [gatk only]
samOptions4VS	string	""	This string will be added to the samtools command when VarScan is used and can include any number of options in the software specific format. If string is empty, "-q 1" will be applied to skip zero-quality alignments. See software specific documentation for more information.
threads	int	1	Number of threads allocated. Preferably allocate kINT amount of memory to accompany the threads, e.g. `14=4` gb of memory for 4 threads, or `2*8=16` gb for 8 threads. [gatk only]
variantsOnly	boolean	true	If true, only variant sites are called. When false, all confident sites are called, even those which are equal to the reference allele. Assigning this to `false` might be indicated for normal samples in normal-tumor comparison calculations.
varscan	string	""	Path to the VarScan jar file (typically VarScan.v2.X.Y.jar varying with the version). If empty string is given (default), VARSCAN_HOME environment variable is assumed to point to the VarScan directory, where VarScan.jar is the program file or a link pointing to it.

Test cases

Test case	Parameters▼	IN reference	IN bam1	IN bam2	IN bams	IN intervals	IN dbsnp	OUT snp	OUT indel	OUT metrics
case1_samtools	properties	reference	bam1	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)
	# Simple testcase for the samtools caller, caller=samtools, variantsOnly=false
case2_gatk	properties	reference	bam1	(missing)	(missing)	(missing)	dbsnp	(missing)	(missing)	(missing)
	# Simple testcase for the gatk caller, caller=gatk, callIndels=false, variantsOnly=false, memory=1g, threads=1, options=-stand_emit_conf 1 -stand_call_conf 1
case3_varscan	properties	reference	bam1	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)
	# Simple testcase for the varscan germline caller, caller=varscan, callIndels=false, options=--min-coverage 1 --min-reads2 1 --min-avg-qual 1 --p-value 0.05
case4_varscan_paired	properties	reference	bam1	bam2	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)
	# Simple testcase for the varscan somatic caller, caller=varscan, options=--min-coverage 1 --p-value 0.95 --somatic-p-value 0.10
case5_varscan_paired	properties	reference	bam1	bam2	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)
	# Simple testcase for the varscan somatic caller, caller=varscan, options=--min-coverage 1 --p-value 0.95 --somatic-p-value 0.10, samOptions4VS=-q 1 -Q 10

Generated 2019-02-07 07:42:35 by Anduril 2.0.0