search

xiaolei-lab / rMVP

:postbox: A Memory-efficient, Visualization-enhanced, and Parallel-accelerated Tool For Genome-Wide Association Study
Apache License 2.0
255 stars 69 forks source link

Different output from GAPIT Farmcpu and rMVP Farmcpu #57

Open shameem356 opened 3 years ago

shameem356 commented 3 years ago

Hello,

I was using same set of data in GAPIT Farmcpu and rMVP Farmcpu (the genotype file in not imputed ).We have got very significant result from GAPIT Farmcpu. But rMVP Farmcpu is showing different result compared to GAPIT Farmcpu. Is it possible ?

XiaoleiLiuBio commented 3 years ago

Yes, this is possible if the genotype data was not fully imputed, and also the default sets of FarmCPU function-related parameters are different between GAPIT and rMVP.

shameem356 commented 3 years ago

Hello @XiaoleiLiuBio , Thank you so much for the reply. Can we use genotype file in rMVP Farmcpu which is in not imputed, ? If yes, how much reliable the answer ? my rMVP results SNPs are functionally very good as compared to GAPIT. So please can you clarify this case ?

XiaoleiLiuBio commented 3 years ago

All functions of the association tests, e.g., GLM, MLM, and FarmCPU don't accept genotype data with missing values. Please impute the genotype data by professional imputation tools before using rMVP. If only a small proportion of genotype values are missing, you can use rMVP to impute the missing genotypic values by major alleles.