Hi all,
I have a question about the genotype likelihoods calculation of GATK (see the source code annotation org.broadinstitute.sting.gatk.walkers.DiploidSNPGenotypeLikelihoods.java)
For heterozygous genotypes and if bi is in G, I do not konw why P(bi | G) = 1 - P(error | q1) / 2 + P(error | q1) / 6, I think P(bi | G)=1/2 (1- P(error | q1) + P(error | q1)/3)=1/2 -P(error | q1)/3.
Take G=AC for example, P(bi | AC)=1/2(P(bi | A)+P(bi | C)). If bi=A, then P(bi | G)=1/2 (1- P(error | q1) + P(error | q1)/3).
Could anyone explain it for me?
I have a question about the genotype likelihoods calculation of GATK (see the source code annotation org.broadinstitute.sting.gatk.walkers.DiploidSNPGenotypeLikelihoods.java)
For heterozygous genotypes and if bi is in G, I do not konw why P(bi | G) = 1 - P(error | q1) / 2 + P(error | q1) / 6, I think P(bi | G)=1/2 (1- P(error | q1) + P(error | q1)/3)=1/2 -P(error | q1)/3.
Take G=AC for example, P(bi | AC)=1/2(P(bi | A)+P(bi | C)). If bi=A, then P(bi | G)=1/2 (1- P(error | q1) + P(error | q1)/3).
Could anyone explain it for me?