search

stephenslab / susieR

R package for "sum of single effects" regression.
https://stephenslab.github.io/susieR
Other
169 stars 42 forks source link

Run SuSiE using individual level genetic data #221

Closed mkoromina closed 3 months ago

mkoromina commented 3 months ago

Hi,

I was wondering whether there is a tutorial/vignette describing how to run SuSiE using individual level genetic data. In that case, which SuSiE R functions would be the best to use?

Thank you!

pcarbo commented 3 months ago

@mkoromina Thanks for your interest in susieR.

You could start here: https://stephenslab.github.io/susieR/articles/finemapping.html

More vignettes can be found here: https://stephenslab.github.io/susieR/articles/index.html

Let us know if you have any other questions.

mkoromina commented 3 months ago

Thank you @pcarbo very much for this!

A few more questions that I have are related to: (i) how the genotype shall be formatted in order to run susieR, ie will it be a genotype matrix where each row will be a variants and the columns are the genotypes for the different individuals? , and, (ii) can we run this vignette in the scenario where Y variable is binary/categorical?

pcarbo commented 3 months ago

Hi @mkoromina, the first vignette ("Fine-mapping example") should answer question (i).

For (ii), the answer is more complicated, but there is some work supporting the use of linear regression models for fine-mapping binary traits (with individual-level data — not as much summary data, e.g., susie_rss). This is briefly discussed in the FINEMAP paper by Benner et al. Also see papers by Matti Pirinen (Annals of Applied Statistics, Nature Genetics).

mkoromina commented 3 months ago

thank you very much @pcarbo for all this information!