Post date: Nov 02, 2014 9:23:19 PM
I decided to try piMass in place of the gemma BSLMM. This basically drops the true polygenic term which could alleviate any weirdness with the kinship matrix.
I ran three replicate chains for each modal and mean genotype data set, here is an example:
pimass-lin -g fix_geno_timemaC.txt -p pheno_samplesC.txt -w 1000000 -s 3000000 -num 100 -o pimass_timemaCrep0 -cc
Not the fix_geno* part just adds an number to each locus name so that each is unique.