overlapping read pairs in GATK 4.0.1.2

Hi,

I came across this post regarding overlapping read pair: https://gatkforums.broadinstitute.org/gatk/discussion/1745/does-unifiedgenotyper-count-overlapping-paired-end-reads-as-two-separate-reads

I could not interpret the last comment by @ebanks. Could you explain me in terms of the new Haplotypecaller?

Best Answer

Answers

  • SheilaSheila Broad InstituteMember, Broadie, Moderator

    @prasundutta87
    Hi,

    I am not sure which part you are confused about, but I will assume

    if they agree, to cap them each at Q20 (which is half of the expected PCR error rate).

    Basically, if the overlapping bases (from the two mates) agree (are the same base), set both base qualities to 20. I guess PCR error rate is 1/50.

    if they disagree, to set them both to Q0.

    If both bases are different from each other, set the base qualities of both to 0.

    I hope this helps.

    -Sheila

  • prasundutta87prasundutta87 EdinburghMember

    Hi,

    My confusion was about what happens when you set both the base qualities to 20, how does it help in the bayes likelihood model. Sorry, just trying to understand it in layman terms instead of going too mathematical about it.

Sign In or Register to comment.