Data from: Modularity of genes involved in local adaptation to climate despite physical linkage

Background: Linkage among genes experiencing different selection pressures
can make natural selection less efficient. Theory predicts that when local
adaptation is driven by complex and non-covarying stresses, increased
linkage is favored for alleles with similar pleiotropic effects, with
increased recombination favored among alleles with contrasting pleiotropic
effects. Here, we introduce a framework to test these predictions with a
co-association network analysis, which clusters loci based on differing
associations. We use this framework to study the genetic architecture of
local adaptation to climate in lodgepole pine, Pinus contorta, based on
associations with environments. Results: We identify many clusters of
candidate genes and SNPs associated with distinct environments, including
aspects of aridity and freezing, and discover low recombination rates
among some candidate genes in different clusters. Only a few genes contain
SNPs with effects on more than one distinct aspect of climate. There is
limited correspondence between co-association networks and gene regulatory
networks. We further show how associations with environmental principal
components can lead to misinterpretation. Finally, simulations illustrate
both benefits and caveats of co-association networks. Conclusions: Our
results support the prediction that different selection pressures favor
the evolution of distinct groups of genes, each associating with a
different aspect of climate. But our results went against the prediction
that loci experiencing different sources of selection would have high
recombination among them. These results give new insight into evolutionary
debates about the extent of modularity, pleiotropy, and linkage in the
evolution of genetic architectures.