search

choishingwan / PRSice

A software package for calculating, applying, evaluating and plotting the results of polygenic risk scores
http://prsice.info
GNU General Public License v3.0
180 stars 85 forks source link

All *.all.scores files are identical #71

Closed wiedenhoeft closed 6 years ago

wiedenhoeft commented 6 years ago

I ran PRSice 2.1.3.beta on four quantitative phenotype columns, which I confirmed are all different. However, using the --all-score option, the resulting *.all.score files are byte-by-byte identical (according to fdupes). While its possible that the values are similar for strongly related phenotypes, the fact that they don't differ by a single digit makes me think that this is a bug.

choishingwan commented 6 years ago

Could you please send me the full log file?

choishingwan commented 6 years ago

Oh, wait, I should've read more carefully.

You are using different phenotype for the same samples, right? In that case, the all score should be the same because

  1. You have the same base file, therefore the input effect sizes are the same
  2. You have the same genotype, as you are only changing the phenotype, your PRS should be the same

While your *.all.score files should be the same ( I should've only generated one), your .prsice and .summary file should differ.

wiedenhoeft commented 6 years ago

Oh yes, my bad, these are the actual PRS, I was confusing those with the *.prsice files. But I agree, it is confusing to have one for each phenotype.