Polygenic risk scores

Question

[namkungho87]

Hi all,

I am using the PRSice software for evaluating polygenic risk scores for my psych-chip data. After running the software, I got polygenic risk scores for each subject with SZ GWAS base data.

I found that almost all the polygenic scores for my subject were negative values. For example, -0.000972588 -0.000166018 -0.000792577 -0.00120931 -0.00104668, etc.

I don't understand what negative polygenic risk scores mean. Is is fine for me to get negative values? Or, does it seem that there were some program errors?

Please advise me. Thanks!

[WouterDeCoster]

I'm unaware of the software you are using, but intuitively to me this means that you a) selected protective variants or b) used your phenotype groups in the wrong order (compared controls vs patients instead of patients vs controls).

[zx8754]

In addition to @WouterDeCoster's guesses, it means subjects carry mostly "protective" alleles.

Answer 1

[namkungho87]

Thanks for your reply. After changing the discovery data set from the mentioned one to another set from PGC, I could found positive values for most SZ patients, meaning that polygenic risk scores for my subjects are strongly dependent on the selection of discovery data sets. I any case, thanks!