search

precimed / mostest

Multivariate Omnibus Statistical Test
GNU General Public License v3.0
22 stars 2 forks source link

NA values in phenotype/.fam file #9

Open mbytc1 opened 3 years ago

mbytc1 commented 3 years ago

If possible, please could you provide a fix so that is is possible to provide NA values for participants/samples across the phenotype file (e.g. if have removed due to being relatedness or population stratification)? (NB i have made sure that they are consistent participants between phenotype and bfiles). Currently, if I try to run it with NA values for any participant, the output all looks correct aside from the univariate sumstats output, where no BETA and SE values are provided. I know that I could instead run an extra stage before MOStest of making new bfiles using plink with limiting to those participants with no NA values, but if possible, it would be great and much simpler if there is a fix within MOSTest to allow it. Thank you

zqsha commented 3 years ago

Yes, completely agree. I met the same issue. Hopefully, the developers can fix this issue. Or already updated?

denvdm commented 2 years ago

Hi, sorry for late reply. While this generated an (unnecessary) error with the earliest versions of MOSTest, it should run smoothly now, if you set missing values to 'NaN'.