Post date: Oct 10, 2013 3:21:44 PM
I checked whether we have evidence that any of the top five SNPs from Aaron's initial GWA were effected by selection in the within-generation transplant experiment. Here are the five SNPs,
# trait scaffold position
1 lateralhue 1067 217427
2 dorsal 3448 47421
3 dorsal 2100 34576
4 dorsal 3448 26532
5 dorsal 6255 2808
Note, we have only mapped GBS data to the genome reference sequence at the resolution level of GBS contig not individual SNP, but I don't think doing things differently would change the results below (and a 100 bp resolution is probably pretty appropriate for comparing the GWA and transplant experiment results). These analyses use files in projects/timema_mortality_experiment/
We have experimental data for the four SNPs in the GBS contig of Aaron's 2nd SNP, three SNPs in the GBS contig of his 3rd SNP, and one in the GBS contig of his 4th SNP (we don't have anything for Aaaron's 1st or 5th SNP). None of these eight experiemntal SNPs showed exceptional allele frequency change in the Caenothus treatment (the highest one has P(exceptional) = 0.796). Moreover, none of our 114 exceptional change SNPs are on the same scaffolds as these 8 SNPs. The only possible evidence we have that these SNPs were affected by selection in the experiment is the 7 of the 8 have negative point estimates for selection coefficients (note, these selection coefficients are not in the current version of the paper because we only report selection coefficents for exceptional change loci). This could be kind of interesting but I need to look at it more (and maybe it isn't even worth pursuing).
One other thought, we could run some sort test of an association between stripe phenotype and genotype for the survivors of this experiment that only considers the exceptional change loci (this greatly reduces multiple testing issues). We could then ask whether, at least for survivors, the genotypes at any of these loci are associated with the stripe phenotype (we could even consider there aggregate association).