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

scores #342

Closed mfazel closed 11 months ago

mfazel commented 12 months ago

Hi, I have a probably simple question which I tried to answer by reading docs and previous issues but could not find an answer to it. I could not find anywhere in the documentations and --help what your definition for --a1 (A1) and --a2 (A2) are. If I assume you are using plink output and PRSice A1/A2 and plink's are the same, then A1 is usually the minor allele and the one that is tested. It's also mentioned in the documentation that PRSice2 takes care of flipping the OR/BETA direction and converting 0/1/2 to 2/1/0 if necessary to count the risk alleles for each individual. Now my question is, why when I run any dataset including PRSice toy example, there are negative scores in the toy.best file? Am I missing something or how do I interpret the negative scores? should user manually modify sumstat and plink file before running PRSice? here are the first 10 values from toy.best

<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns="http://www.w3.org/TR/REC-html40">

FID | IID | In_Regression | PRS -- | -- | -- | -- CAS_1 | CAS_1 | Yes | -0.006 CAS_2 | CAS_2 | Yes | -0.00631 CAS_3 | CAS_3 | Yes | -0.00227 CAS_4 | CAS_4 | Yes | -0.00204 CAS_5 | CAS_5 | Yes | -0.00083 CAS_6 | CAS_6 | Yes | -0.00225 CAS_7 | CAS_7 | Yes | -0.00069 CAS_8 | CAS_8 | Yes | -0.00413 CAS_9 | CAS_9 | Yes | 0.002567 CAS_10 | CAS_10 | Yes | 0.001863

Thanks. Mehdi

choishingwan commented 11 months ago

score can be negative. The way to understand that is to think that a variant with a negative effect size is reducing the risk of disease relative to the non-effective allele.