It is possible to convert PL values to genotype probabilities using the phred-scaled likelihoods. The process involves applying the phred-scaled likelihoods to the Hardy-Weinberg equilibrium equation, which is a mathematical model used to describe the distribution of genotypes in a population. The equation is as follows:

p^2 = p(AA) 2pq = p(Aa) q^2 = p(aa)

where p is the frequency of the A allele and q is the frequency of the non-A allele. The genotype probabilities can be obtained by solving this equation for p(AA), p(Aa), and p(aa) using the PL values as input.

One way to do this is to use the package "VCFR" which provide functions to read and manipulate VCF files in R.

Another way is to use the package "GenotypeIO" which provides utility functions for reading and writing genotype data in R.

You can also use the package "SNPRelate" which has a function "gtl2gprob()" that converts genotype likelihoods to genotype probabilities.

be aware of organising the files for each package. you may need to adjust accordingly.

Here is some code to read in a vcf file into R and convert PL to locus-specific GL scaled between 0-2. You can then use these GL values to calculate an allele frequency for each population by sum(gl)/2*(number of individuals in pop). Sorry it is not well annotated. You can compare the GL to the hard genotype calls in the vcf to prove to yourself it's correct. Hope this helps!

library(data.table)

library(vcfR)

#### read in vcf ####

vcf

@Rami Yannam Thank you for your answer, I did it in VCFR using Trevor's script and it worked.