GATK Base Quality Recalibration

[ashuchawla]

Could any one please tell me what is the role of the dbsnp vcf file in the GATK Base Quality Score recalibration program -> http://www.broadinstitute.org/gsa/wi..._recalibration

[nickloman]

It's to mask out known SNPs so those positions don't get counted as mismatches / sequence errors.

[ashuchawla]

It's to mask out known SNPs so those positions don't get counted as mismatches / sequence errors.

Great thanks....I have another confusion, I have noticed the presence of similar snp's multiple times in the dbsnp file, do you know if this makes any difference to the recalibration process? Also, could you please if possible direct me to the actual code(java packages) where this dbsnp file is being used so that I can understand its functionality??

[nickloman]

Have a dig around here

http://www.broadinstitute.org/gsa/wiki/index.php/Base_quality_score_recalibration

and

https://github.com/broadgsa/gatk/tree/master/public/java/src/org/broadinstitute/sting

Similar SNPs or the same SNP? I don't think the presence of the same SNP would make any difference to the result, the call is either ignored or it isn't.

[ashuchawla]

Have a dig around here

http://www.broadinstitute.org/gsa/wiki/index.php/Base_quality_score_recalibration

and

https://github.com/broadgsa/gatk/tree/master/public/java/src/org/broadinstitute/sting

Similar SNPs or the same SNP? I don't think the presence of the same SNP would make any difference to the result, the call is either ignored or it isn't.

I beleive same snp's, even I think it wont make any difference but i needed confirmation. And I have gone through most of these packages u posted the link for....was unable to find the one that I need.

[nickloman]

java\src\org\broadinstitute\sting\gatk\walkers\recalibration\CountCovariatesWalker.java seems to be the relevant file...