I am using JMP Genomics to model a multi-location experiment with approx 400 cultivars. With a MAF = 0.05 around 10.000 SNPs are retained. The analyses without accounting for relatedness among the cultivars, seem to be OK. However, I know that some of these cultivars are related. And I would like to account for that in the modelling. In particular, I would like to include the Q-K matrices when I do analyses across locations, i.e., when all locations are analysed simultanously.
If a tutorial describing this already exist, I would be happy to be adviced to it. If that is not the case, however, I would appreciate feedback on my problem.
Great question. This gets a little complicated, so I'd suggest emailing firstname.lastname@example.org to open a Tech Support track so we can discuss the details further. The general idea is that your Q and K matrices will be composed of identical blocks repeated for each location. You can use the full data set and estimate Q and K all at once, or you can estimate them at a single location and then replicate them manually for the other locations (more efficient).