Hello,
I have imputed genotype data from the Michigan Imputation Server and I would like to use this for cis-eQTL mapping with MatrixeQTL.
My question is about whether to use the genotype data from the GT slot or the DS slot.
From my understanding, the GT slot contains the estimated most likely genotype and the DS slot contains the estimated alternate allele dosage.
The GT slot takes the values "0|0", "1|0", "0|1", "1|1". For eQTL mapping these could be coerced as follows:
"0|0" = 0
"1|0" = 1
"0|1" = 1
"1|1" = 2
The DS slot contains continuous values from 0-2, i.e. mostly 0, 1 and 2 but with some decimals, e.g. 0.5, reflecting uncertainty in the genotype calls.
Does anyone have any advice regarding best practices about which data to use for eQTL mapping? e.g. the "best guess" genotypes from the GT slot, the continuous values from the DS slot, or the DS values rounded to the nearest integer? or something else?
Thank you.
