Post date: Nov 22, 2013 12:3:30 AM
I quantified locus effects for the combined sets, Ac, Ms, GLA, SLA, and all. Several loci have non-zero posterior inclusion probabilities. For example, 16 loci have pip >= 0.05 in one or more sets (some in multiple). For example, the pip for s36p248085 is 0.69 in GLA and 0.23 in SLA (but interestingly only 0.02 for all). The new figures and effect size text files are in projects/lycaeides_hostplant/melGemma. Here are a few thoughts on how I should eventually present these,
Figures with pip or effect along scaffolds, the larger scaffolds could be labelled whereas smaller ones could be combined.
Perhaps a list of the top loci, with maf, nature of variant, and annotation information if possible.
Perhaps sliding window for pips, but it still doesn't seem very relevant as the main signal is for individual snps not scaffolds or runs of snps.
One key thing for me to remember is that even though only host specific loci will be useful for downstream analyses, all performance associated snps are relevant for describing the genetic basis of variation in performance.
Note, the relevant R, perl and result files have 'comb' in their names.