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

Set-based Clumping #347

Closed JinhanZhu1 closed 9 months ago

JinhanZhu1 commented 1 year ago

Hello,

I have some questions regards to the set-based clumping used in PRset. Based on the documentation, it said "to maximize signal within each gene set, we must perform clumping for each gene sets separately". My questions are:

  1. Suppose I have 3 gene sets and one background set, will the clumping happened in 4 sets separately? If yes, will it leads to same SNP has different effect size in different sets due to various LD structure in each set? (Which can potentially biased the P value)
  2. During random sampling m snps for N times from background set in competitive-p computation, does the random sampling only include SNPs that has GWAS stats?

Really appreciate your time and help!!

choishingwan commented 1 year ago

  1. Yes, it happens 4 times. No, prsice don't modify effect sizes as we only do clumping and thresholding, but not penalization

  2. Yes. Only snps with summary stat will be in the analysis set (as we use post clump snps)

On Thu, Oct 26, 2023, 6:28 PM Zhujh @.***> wrote:

Hello,

I have some questions regards to the set-based clumping used in PRset. Based on the documentation, it said "to maximize signal within each gene set, we must perform clumping for each gene sets separately". My questions are:

  1. Suppose I have 3 gene sets and one background set, will the clumping happened in 4 sets separately? If yes, will it leads to same SNP has different effect size in different sets due to various LD structure in each set? (Which can potentially biased the P value)
  2. During random sampling m snps for N times from background set in competitive-p computation, does the random sampling only include SNPs that has GWAS stats?

Really appreciate your time and help!!

— Reply to this email directly, view it on GitHub https://github.com/choishingwan/PRSice/issues/347, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJTRYTYENKQHVU5LL6HZKDYBLPZ7AVCNFSM6AAAAAA6R5LR46VHI2DSMVQWIX3LMV43ASLTON2WKOZRHE3DINBWGA3TINA . You are receiving this because you are subscribed to this thread.Message ID: @.***>