Closed ruthchia closed 4 years ago
Hi Ruth,
For step1, the plink file is used for constructing the genetic relationship matrix. It is ok to specify a plink file with one chromosome if it fits your research interest, but usually we recommend using a genome-wide plink file, so the sample relatedness can be well captured. BTW, this plink file usually contains genotyped variants.
For step 2, you may definitely specify files for dosages to test by chromosome.
Thanks, Wei
On Thu, Sep 27, 2018 at 1:24 PM Ruth Chia notifications@github.com wrote:
Hi, I am wondering if it would be possible to run SAIGE per chromosome when LOCO is set to FALSE? I am asking because my genotype and imputed dosage files are already broken down per chromosome and to have to merge them seems a little cumbersome and the files are huge. Can you please clarify? Thanks! Ruth
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/weizhouUMICH/SAIGE/issues/52, or mute the thread https://github.com/notifications/unsubscribe-auth/AJfdL880UnkM8Bsrd9h5hb1DANUapft4ks5ufQnHgaJpZM4W9HKL .
Hi Wei, Thanks for the clarification! Appreciate it! Ruth
From: weizhouUMICH notifications@github.com Reply-To: weizhouUMICH/SAIGE reply@reply.github.com Date: Thursday, September 27, 2018 at 1:34 PM To: weizhouUMICH/SAIGE SAIGE@noreply.github.com Cc: "Chia, Ruth (NIH/NIA/IRP) [E]" ruth.chia@nih.gov, Author author@noreply.github.com Subject: Re: [weizhouUMICH/SAIGE] running SAIGE per chromosome (#52)
Hi Ruth,
For step1, the plink file is used for constructing the genetic relationship matrix. It is ok to specify a plink file with one chromosome if it fits your research interest, but usually we recommend using a genome-wide plink file, so the sample relatedness can be well captured. BTW, this plink file usually contains genotyped variants.
For step 2, you may definitely specify files for dosages to test by chromosome.
Thanks, Wei
On Thu, Sep 27, 2018 at 1:24 PM Ruth Chia notifications@github.com wrote:
Hi, I am wondering if it would be possible to run SAIGE per chromosome when LOCO is set to FALSE? I am asking because my genotype and imputed dosage files are already broken down per chromosome and to have to merge them seems a little cumbersome and the files are huge. Can you please clarify? Thanks! Ruth
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/weizhouUMICH/SAIGE/issues/52, or mute the thread https://github.com/notifications/unsubscribe-auth/AJfdL880UnkM8Bsrd9h5hb1DANUapft4ks5ufQnHgaJpZM4W9HKL .
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/weizhouUMICH/SAIGE/issues/52#issuecomment-425178380, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ATxQ4ahVdb1Pp74DxhOxm7VeD2bLS3baks5ufQwPgaJpZM4W9HKL.
For the plink genotype file used for Step 1, is there a standard preparation recommended? For example, is LD pruning and a MAF cutoff recommended?
Hi @jjfarrell,
Sorry for my late reply!! We usually use LD-pruned plink genotype file for GRM with common markers. Since version 0.36.1 , an option minMAFforGRM is added to specify a min MAF cutoff for markers used for GRM from the plink file. The default value is minMAFforGRM = 0.01
Please reopen if you still encounter this issue. Thanks! Wei
Hi, I am wondering if it would be possible to run SAIGE per chromosome when LOCO is set to FALSE? I am asking because my genotype and imputed dosage files are already broken down per chromosome and to have to merge them seems a little cumbersome and the files are huge. Can you please clarify? Thanks! Ruth