Environmental associations at individual loci
After filtering out the intergenic SNPs that might result in false
positives, we were left with 927,740 (22.3%) SNPs in or near genes.
These were then used for the association analyses. This is similar to
the approach used by Jordan et al. (2017) for Eucalyptus . After
the running of LFMM2 (p < 10-5) for GEA, we
found 1,458 significant associations with the five selected
environmental variables (Table 1). PCK4 (April 1stsnowpack) had the most associations, with TMIN (minimum winter
temperature) having the following highest number. Few SNPs were
associated with more than one climatic variable, with the highest degree
of overlap between PCK4 and TMIN (64 SNPs) and between CWD and TMIN (17
SNPs) (Fig. 2).
For PCK4 and TMIN, there were roughly similar numbers of associated SNPs
in upstream and downstream regions versus the gene itself, with 14% of
associated SNPs being missense (non-synonymous) mutations (Table 2).
SNPs associated with CWD were also roughly evenly split between flanking
regions and the main gene sequence, but only 3% were missense
mutations. A higher proportion of SNPs associated with TMAX (maximum
summer temperature) were within the gene, with 22% being missense
mutations, while PPTW (winter precipitation) showed the opposite
pattern, with 69% of SNPs being in the flanking regions.