Variant Evaluation-Problem with FP,FN

[Mahtab]

Hi all

I'm trying to evaluate variants called by GATK UnifiedGenotyper, FreeBayes and mpileup/bcftools. I'm using GATK VariantEval but although the VCF files are showing different concordance, FP and FN fields in the GATK report are zero. I've tried VariantEval providing dbSNP and not providing it. The dataset I'm using is from Broad NA12878.

Any suggestions is appreciated.

Thanks
Mahtab

[alonie]

Hi Mahtab

I think the GATK VariantEval module uses the switch --goldStandard to specify the truth set for FP/FN evaluation (rather than the --comp switch, which is for concordance evaluation)

see http://www.broadinstitute.org/gatk/g...riantEval.html for more details

Andrew

[Mahtab]

Thanks Andrew. This option does not seem to be available in Galaxy.