jhmarcus / gwass Goto Github PK

This is very useful. Here are some things that would be nice

1. Allow for unused columns in the csv file (this is more a pipeline issue but I'll put it here so everything is together)
2. In the output, the "referenece_allele" column should be renamed to "effect_allele". The alt_allele column should be renamed to other_allele
3. In the reference csv, the "regresion_type" column should be called something like "log_transforme_beta" and then should be yes if the raw data report OR and no if it is a linear regression or the raw data report log(OR).

Right now it is hardcoded for the raw sum stat files to be '.tsv' so we have to preprocess files that aren't. It would be simple and nice to support other common file formats. We ran into issues b/c some files had mixed sep but adding some other standards would be useful to avoid manual preprocessing.