How to Create a Truth & training set for Variant Recalibrator

Question

[yoh1242]

Hello! I need a step to step Guide on how to Create a Truth and Training set for VariantRecalibrator GATK v4.1.4. The Guide available on GATK website was not helpful enough and I have yet to make progress.

I have downloaded a list of known indels BQSR file from 1000bullgenome website and plan to use it for the -resource part of the command line. Kindly mention the steps necessary to turn this file into a suitable resource file for variant recalibrator step of GATK.

[Ram]

was not helpful enough

Where did you face problems? Which guide are you referring to and how far did you get?

[yoh1242]

I have performed HaplotypeCaller but can not proceed with VariantRecalibrator due to lack of resource file.

I have downloaded 1000bullgenome.vcf and performed VariantCalling and VariantFIltration on the file but when I run the Commandline.

java -Xmx8g -jar /home/zafar/miniconda3/share/gatk4-4.1.4.1-0/gatk-package-4.1.4.1-local.jar VariantRecalibrator -R UMD3.1_chromosomes.fa -V H-BQSRSunny2.vcf -resource:1000bullgenome,known=true,training=true,truth=true,prior=6.0 ARSUMD.Filtered.vcf -an QD -an MQ -an MQRankSum -an ReadPosRankSum -an FS -an SOR -mode SNP -tranche 100.0 -tranche 99.9 -tranche 99.0 -tranche 90.0 -O Sunny2VR.recal -tranches-file Sunny2VR.tranches -rscript-file Sunny2VR.plots.R

I get the following Error:

A USER ERROR has occurred: Bad input: Values for QD annotation not detected for ANY training variant in the input callset. VariantAnnotator may be used to add these annotations.
A USER ERROR has occurred: Bad input: Values for QD annotation not detected for ANY training variant in the input callset. VariantAnnotator may be used to add these annotations.

I have tried both GATK Variant Annotator and snpEff on the resource vcf file but nothing is working.

[Ram]

You've still not pointed us to the "guide available on GATK website" you refer to.

[yoh1242]

I cant seem to find the link on the new GATK forum. Regardless, The steps were as follows: Perform SelectVariants, followed by VariantFiltration and finish it with Variant Annotation on the known SNP and Indel vcf file to create a Training Set resource file. Despite following this step I have yet to yield any result

[GenoMax]

Hello yoh1242!

• Where Can I Download Truth, Training sets for the bovine genome.

Is this a follow-up question for above thread or are you basically asking the same question again?

[yoh1242]

DIfferent but same question.

I am asking for sites where I can download -resource vcf files for this very question mentioned here