How To Calculate Genotype Likelihood Using Quality Scores For Reads

Question

[Mcmahanl]

After google search and reading some articles, I still have no idea how to calculate genotype likelihood for a locus using quality scores for reads using simple probability. The probability math symbols in those articles are hard for me to understand how to do the calculation.

For example, for a locus, the reference sequence has T. There are 6 reads, 4 reads have T and 2 reads have G (with quality score = 10) at that locus.

If the true genotype of this locus is [T,T], the question is how does one calculate the probability of this genotype at this locus, i.e. P(D | [T, T])? D for the given data of reads.

Answer 1

[Pierre Lindenbaum]

see Heng Li's "Mathematical Notes on SAMtools Algorithms / 4.4 Likelihood of data given genotype " in http://www.broadinstitute.org/gatk/media/docs/Samtools.pdf

[stat.1405]

Thank you so much Pierre Lindenbaum, it is really helpful