search

stephenslab / mvsusieR

A new model and algorithm for multivariate Bayesian variable selection regression.
https://stephenslab.github.io/mvsusieR
Other
21 stars 7 forks source link

Trait-specific sample sizes? #52

Open harryyiheyang opened 2 months ago

harryyiheyang commented 2 months ago

Hi there, When I used the mvsusie_rss(), a universal sample size was required for all traits. However, this rarely happens in practice, e.g., the following numbers are the sample sizes of GWAS summary data: BMI DBP HDL LDL SBP SMK T2D TG 747649.9 969641.0 1232760.0 1222125.6 946218.1 801386.2 747649.9 1242643.5 Is it available to allow trait-specific sample sizes? Thank you very much Yihe

gaow commented 2 months ago

@zouyuxin do we take a vector of sample sizes ordered same way as the phenotypes?

stephens999 commented 2 months ago

It's strongly advised to use the same individuals for all traits, using imputation where necessary to achieve this. When you use different individuals the results may be unreliable.

On Thu, Aug 22, 2024, 18:41 gaow @.***> wrote:

@zouyuxin https://github.com/zouyuxin do we take a vector of sample sizes ordered same way as the phenotypes?

— Reply to this email directly, view it on GitHub https://github.com/stephenslab/mvsusieR/issues/52#issuecomment-2305923094, or unsubscribe https://github.com/notifications/unsubscribe-auth/AANXRRKJY6OBK3AVGFVVZBLZSZZKBAVCNFSM6AAAAABM64BALOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMBVHEZDGMBZGQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>

harryyiheyang commented 2 months ago

Hi @stephens999 and @gaow I appreciate your clarification! Yihe