search

xihaoli / STAARpipeline-Tutorial

The tutorial for performing single-/multi-trait association analysis of whole-genome/whole-exome sequencing (WGS/WES) studies using FAVORannotator, STAARpipeline and STAARpipelineSummary
GNU General Public License v3.0
21 stars 17 forks source link

Question about selecting variants to calculate sparse GRM #3

Closed ikarus97 closed 1 year ago

ikarus97 commented 1 year ago

Dear authors,

I'm trying to use your pipeline to our association study, but got questions about the way to build GRM. Is it correct to use the common (MAF > 5%) variants to build GRM, so that we can capture population structure among individuals? Or should I use the rare variants (those to be used in the main analysis), so we capture the similarity between individuals as will be observed in the main analysis?

Thank you, In-Hee

xihaoli commented 1 year ago

Hi In-Hee,

Thanks for your question. To our knowledge, the standard pipelines to build GRM (e.g. GENESIS and GCTA) are based on KING to infer sample relatedness. KING recommends filtering out rare variants when generating GRM for WGS studies. More details can be found in the KING Tutorial.

Please feel free to let me know if you have any additional questions.

Best, Xihao