How to deal with NA column in MASH analysis

Open rfeng2023 opened this issue 2 years ago • 1 comments

When I was trying to Compute MASH posteriors in MASH analysis, there is a NA column in bhat and sbhat matrix in file /mnt/vast/hpc/csg/rf2872/Work/MASH_csg.q/output/RDS/ALL_Ast_End_Exc_Inh_Mic_OPC_Oli.log2cpm.bed.processed_phenotype.per_chrom_ALL_Ast_End_Exc_Inh_Mic_OPC_Oli.log2cpm.ALL_Ast_End_Exc_Inh_Mic_OPC_Oli.covariate.pca.resid.PEER.cov.chr9.norminal.cis_long_table.RMI1.rds, which is a split rds file form sumstat merger vcf (maybe more than 1 NA columns in other files).

I tried to delete the NA column, so the ncol of above matrices is 5, which is different from the mash model, and would cause the job crash in downstream analysis mash_compute_posterior_matrices. Is there any proper way to deal with this issue?

Mar 13 '23 16:03 rfeng2023

Thanks for documenting @rfeng2023! as discussed in person this PM, please post update on this ticket to summarize what we concluded how to move forward.

Mar 14 '23 21:03 gaow