search

junyangq / snpnet

snpnet: Fast and scalable lasso/elastic-net solver for large SNP data
32 stars 15 forks source link

Both FID and IID should be used for identifier for individuals #6

Closed yk-tanigawa closed 5 years ago

yk-tanigawa commented 5 years ago

Hi Junyang,

As we talked last time, I strongly recommend you to use both FID and IID as the identifier for individuals. https://github.com/junyangq/snpnet/blob/4b0687cbe092ba1936f16f1b2059532b57faa5b4/R/snpnet.R#L75

I think you can mimic the behavior of BEDMatrix::BEDMatrix(), meaning that you may concatenate the two IDs with an underscore, like FID_IID. https://github.com/junyangq/snpnet/blob/4b0687cbe092ba1936f16f1b2059532b57faa5b4/R/snpnet.R#L81

Also, I would recommend changing the separator in .phe file. I think the current package assume it is a 'csv' file, but the default behavior of PLINK bed/bim/fam file is a tab-separated file.

junyangq commented 5 years ago

This has been resolved. Thanks for the suggestion @yk-tanigawa