recommendations for lcWGS derived genotype likelihoods

[laurahspencer]

Hello! Is it possible to use tensorqtl to perform QTL analyses with genotype likelihoods derived from low-coverage whole genome sequencing data, instead of hard-called genotypes?

Thanks for any guidance you can provide!

[francois-a]

Hi, definitely. The genotype input is flexible as long as it's in a compatible format. If the likelihoods are stored as dosages in PLINK pgen you can pass the --dosages flag to read those instead of the hard calls.