Manual corrections for allele mismatches

Question

[Sakshi]

I am doing PRS for the first time. I wanted to understand if anyone has a comprehensive pipeline for managing the differences in alleles between base and target. In my mind the logic was for the following scenarios (maf is for eff, MAF_c is for A1)

A1      A2      eff     ref     maf     MAF_c

T       G       T       G       0.26    0.2947 <- perfect leave as is

T       C       T       C       0.32    0.3939 <- perfect leave as is

C       T       T       C       0.93    0.09869 <- mismatch, but since maf for A1~ 1-maf for eff we can flip beta

G       A       A       G       0.98    0.01034 <- mismatch, but since maf for A1~ 1-maf for eff we can flip beta

T       G       T       G       0.09    0.05242 <- perfect

C       T       T       C       0.97    0.01663 <-mismatch, but since maf for A1~ 1-maf for eff we can flip beta

C       G       C       G       0.11    0.07325<- perfect

T       C       T       C       0.06    0.07478<-perfect

T       C       T       C       0.1     0.05731<-perfect

What abou a scenario where

C       T       T       C       0.93    0.90 

where the alleles are swapped but maf for A1 is very similar to maf for eff? but the alleles are appearing swapped? what do we do here?

Would appreciate if someone could layout their approach to all possible scenarios they've encountered in their analysis

Answer 1

[curious]

take a look at thomas winkler's stuff: https://www.nature.com/articles/nprot.2014.071