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
187 stars 89 forks source link

Issue with the output files (.summary and .prsice files) #350

Closed WeiChenPan closed 9 months ago

WeiChenPan commented 9 months ago

Hi,

First of all, thank you for your excellent tool.

I'm currently running PRSice2 on UK Biobank and having some issues with the output files (.summary and .prsice files). In .summary file, the threshold and P-value of the best-fit model PRSice2 found are 1 and 0, respectively. But I find it strange that the P-value is 0 because it implies non-significance. Screenshot from 2024-01-31 11-31-11

So I checked the .prsice file and found that most P-vales of any threshold are 0. Screenshot from 2024-01-31 11-42-13

Below is the additional command line arguments (extra_options) I added for running my project: --snp Name --chr Chr --bp Pos --A1 Alt --A2 Ref --stat Effect --pvalue Pval --or --pheno-col Hematologic_malignancy_cc --cov-file Hematologic_malignancy_imputation.phe.pheno --cov-col sex,age,ethnic_background,alcohol_status,hypertension,ever_smoked,SBP,BMI,NLR,CRP

I'm not sure which step went wrong or if it's just a problem with the parameters. Did anyone have the same problem before?

Best, Wei-Chen

choishingwan commented 9 months ago

p-value of 0 means it is very significant. In your case I will check if there are sample overlap between your GWAS and your target, as that is the most likely reason for such significant results.