VariantEval

General-purpose tool for variant evaluation (% in dbSNP, genotype concordance, Ti/Tv ratios, and a lot more)

Overview

Given a variant callset, it is common to calculate various quality control metrics. These metrics include the number of raw or filtered SNP counts; ratio of transition mutations to transversions; concordance of a particular sample's calls to a genotyping chip; number of singletons per sample; etc. Furthermore, it is often useful to stratify these metrics by various criteria like functional class (missense, nonsense, silent), whether the site is CpG site, the amino acid degeneracy of the site, etc. VariantEval facilitates these calculations in two ways: by providing several built-in evaluation and stratification modules, and by providing a framework that permits the easy development of new evaluation and stratification modules.

Input

One or more variant sets to evaluate plus any number of comparison sets.

Output

Evaluation tables detailing the results of the eval modules which were applied. For example:

 output.eval.grp:
##:GATKReport.v0.1 CountVariants : Counts different classes of variants in the sample
CountVariants  CompRod   CpG      EvalRod  JexlExpression  Novelty  nProcessedLoci  nCalledLoci  nRefLoci  nVariantLoci  variantRate ...
CountVariants  dbsnp     CpG      eval     none            all      65900028        135770       0         135770        0.00206024  ...
CountVariants  dbsnp     CpG      eval     none            known    65900028        47068        0         47068         0.00071423  ...
CountVariants  dbsnp     CpG      eval     none            novel    65900028        88702        0         88702         0.00134601  ...
CountVariants  dbsnp     all      eval     none            all      65900028        330818       0         330818        0.00502000  ...
CountVariants  dbsnp     all      eval     none            known    65900028        120685       0         120685        0.00183133  ...
CountVariants  dbsnp     all      eval     none            novel    65900028        210133       0         210133        0.00318866  ...
CountVariants  dbsnp     non_CpG  eval     none            all      65900028        195048       0         195048        0.00295976  ...
CountVariants  dbsnp     non_CpG  eval     none            known    65900028        73617        0         73617         0.00111710  ...
CountVariants  dbsnp     non_CpG  eval     none            novel    65900028        121431       0         121431        0.00184265  ...
...


Usage examples

 java -jar GenomeAnalysisTK.jar \
-T VariantEval \
-R reference.fasta \
-o output.eval.grp \
--eval:set1 set1.vcf \
--eval:set2 set2.vcf \
[--comp comp.vcf]

Count Mendelian violations for each family in a callset with multiple families (and provided pedigree)
 Java -jar GenomeAnalysisTK.jar \
-T VariantEval \
-R reference.fasta \
-o output.MVs.byFamily.table \
--eval multiFamilyCallset.vcf \
-noEV -noST \
-ST Family \
-EV MendelianViolationEvaluator


Caveat

Some stratifications and evaluators are incompatible with each other due to their respective memory requirements, such as AlleleCount and VariantSummary, or Sample and VariantSummary. If you specify such a combination, the program will output an error message and ask you to disable one of these options. We do not currently provide an exhaustive list of incompatible combinations, so we recommend trying out combinations that you are interested in on a dummy command line, to rapidly ascertain whether it will work or not.

These Read Filters are automatically applied to the data by the Engine before processing by VariantEval.

Parallelism options

This tool can be run in multi-threaded mode using this option.

Window size

This tool uses a sliding window on the reference.

• Window start: -50 bp before the locus
• Window stop: 50 bp after the locus

Command-line Arguments

Engine arguments

All tools inherit arguments from the GATK Engine' "CommandLineGATK" argument collection, which can be used to modify various aspects of the tool's function. For example, the -L argument directs the GATK engine to restrict processing to specific genomic intervals; or the -rf argument allows you to apply certain read filters to exclude some of the data from the analysis.

VariantEval specific arguments

This table summarizes the command-line arguments that are specific to this tool. For more details on each argument, see the list further down below the table or click on an argument name to jump directly to that entry in the list.

Argument name(s) Default value Summary
Required Inputs
--eval
NA Input evaluation file(s)
Optional Inputs
--comp
[] Input comparison file(s)
--dbsnp
-D
none dbSNP file
--goldStandard
-gold
none Evaluations that count calls at sites of true variation (e.g., indel calls) will use this argument as their gold standard for comparison
--knownCNVs
NA File containing tribble-readable features describing a known list of copy number variants
--stratIntervals
NA File containing tribble-readable features for the IntervalStratificiation
Optional Outputs
--out
-o
stdout An output file created by the walker. Will overwrite contents if file exists
Optional Parameters
--ancestralAlignments
-aa
NA Fasta file with ancestral alleles
--evalModule
-EV
[] One or more specific eval modules to apply to the eval track(s) (in addition to the standard modules, unless -noEV is specified)
--known_names
-knownName
[] Name of ROD bindings containing variant sites that should be treated as known when splitting eval rods into known and novel subsets
--mendelianViolationQualThreshold
-mvq
50.0 Minimum genotype QUAL score for each trio member required to accept a site as a violation. Default is 50.
--minPhaseQuality
-mpq
10.0 Minimum phasing quality
--sample
-sn
NA Derive eval and comp contexts using only these sample genotypes, when genotypes are available in the original context
--samplePloidy
-ploidy
2 Per-sample ploidy (number of chromosomes per sample)
--select_exps
-select
[] One or more stratifications to use when evaluating the data
--select_names
-selectName
[] Names to use for the list of stratifications (must be a 1-to-1 mapping)
--stratificationModule
-ST
[] One or more specific stratification modules to apply to the eval track(s) (in addition to the standard stratifications, unless -noS is specified)
Optional Flags
--doNotUseAllStandardModules
-noEV
false Do not use the standard modules by default (instead, only those that are specified with the -EV option)
--doNotUseAllStandardStratifications
-noST
false Do not use the standard stratification modules by default (instead, only those that are specified with the -S option)
--keepAC0
false If provided, modules that track polymorphic sites will not require that a site have AC > 0 when the input eval has genotypes
--list
-ls
false List the available eval modules and exit
--mergeEvals
false If provided, all -eval tracks will be merged into a single eval track
--requireStrictAlleleMatch
-strict
false If provided only comp and eval tracks with exactly matching reference and alternate alleles will be counted as overlapping

Argument details

Arguments in this list are specific to this tool. Keep in mind that other arguments are available that are shared with other tools (e.g. command-line GATK arguments); see Inherited arguments above.

--ancestralAlignments / -aa

Fasta file with ancestral alleles

File  NA

--comp / -comp

Input comparison file(s)
The variant file(s) to compare against.

This argument supports reference-ordered data (ROD) files in the following formats: BCF2, VCF, VCF3

List[RodBinding[VariantContext]]  []

--dbsnp / -D

dbSNP file
dbSNP comparison VCF. By default, the dbSNP file is used to specify the set of "known" variants. Other sets can be specified with the -knownName (--known_names) argument.

This argument supports reference-ordered data (ROD) files in the following formats: BCF2, VCF, VCF3

RodBinding[VariantContext]  none

--doNotUseAllStandardModules / -noEV

Do not use the standard modules by default (instead, only those that are specified with the -EV option)

Boolean  false

--doNotUseAllStandardStratifications / -noST

Do not use the standard stratification modules by default (instead, only those that are specified with the -S option)

Boolean  false

--eval / -eval

Input evaluation file(s)
The variant file(s) to evaluate.

This argument supports reference-ordered data (ROD) files in the following formats: BCF2, VCF, VCF3

R List[RodBinding[VariantContext]]  NA

--evalModule / -EV

One or more specific eval modules to apply to the eval track(s) (in addition to the standard modules, unless -noEV is specified)
See the -list argument to view available modules.

String[]  []

--goldStandard / -gold

Evaluations that count calls at sites of true variation (e.g., indel calls) will use this argument as their gold standard for comparison
Some analyses want to count overlap not with dbSNP (which is in general very open) but actually want to itemize their overlap specifically with a set of gold standard sites such as HapMap, OMNI, or the gold standard indels. This argument provides a mechanism for communicating which file to use

This argument supports reference-ordered data (ROD) files in the following formats: BCF2, VCF, VCF3

RodBinding[VariantContext]  none

--keepAC0 / -keepAC0

If provided, modules that track polymorphic sites will not require that a site have AC > 0 when the input eval has genotypes

boolean  false

--known_names / -knownName

Name of ROD bindings containing variant sites that should be treated as known when splitting eval rods into known and novel subsets
List of rod tracks to be used for specifying "known" variants other than dbSNP.

HashSet[String]  []

--knownCNVs / -knownCNVs

File containing tribble-readable features describing a known list of copy number variants
File containing tribble-readable features containing known CNVs. For use with VariantSummary table.

IntervalBinding[Feature]  NA

--list / -ls

List the available eval modules and exit
Note that the --list argument requires a fully resolved and correct command-line to work.

Boolean  false

--mendelianViolationQualThreshold / -mvq

Minimum genotype QUAL score for each trio member required to accept a site as a violation. Default is 50.

double  50.0  [ [ -∞  ∞ ] ]

--mergeEvals / -mergeEvals

If provided, all -eval tracks will be merged into a single eval track
If true, VariantEval will treat -eval 1 -eval 2 as separate tracks from the same underlying variant set, and evaluate the union of the results. Useful when you want to do -eval chr1.vcf -eval chr2.vcf etc.

boolean  false

--minPhaseQuality / -mpq

Minimum phasing quality

double  10.0  [ [ -∞  ∞ ] ]

--out / -o

An output file created by the walker. Will overwrite contents if file exists

PrintStream  stdout

--requireStrictAlleleMatch / -strict

If provided only comp and eval tracks with exactly matching reference and alternate alleles will be counted as overlapping

boolean  false

--sample / -sn

Derive eval and comp contexts using only these sample genotypes, when genotypes are available in the original context

Set[String]  NA

--samplePloidy / -ploidy

Per-sample ploidy (number of chromosomes per sample)

int  2  [ [ -∞  ∞ ] ]

--select_exps / -select

One or more stratifications to use when evaluating the data

ArrayList[String]  []

--select_names / -selectName

Names to use for the list of stratifications (must be a 1-to-1 mapping)

ArrayList[String]  []

--stratificationModule / -ST

One or more specific stratification modules to apply to the eval track(s) (in addition to the standard stratifications, unless -noS is specified)

String[]  []

--stratIntervals / -stratIntervals

File containing tribble-readable features for the IntervalStratificiation
File containing tribble-readable features for the IntervalStratificiation

IntervalBinding[Feature]  NA