Multilocus analyses reveal little evidence for lineage‐wide adaptive evolution within major clades of soft pines (Pinus subgenus Strobus)

Abstract

Estimates from molecular data for the fraction of new nonsynonymous mutations that are adaptive vary strongly across plant species. Much of this variation is due to differences in life history strategies as they influence the effective population size (Ne ). Ample variation for these estimates, however, remains even when comparisons are made across species with similar values of Ne . An open question thus remains as to why the large disparity for estimates of adaptive evolution exists among plant species. Here, we have estimated the distribution of deleterious fitness effects (DFE) and the fraction of adaptive nonsynonymous substitutions (α) for 11 species of soft pines (subgenus Strobus) using DNA sequence data from 167 orthologous nuclear gene fragments. Most newly arising nonsynonymous mutations were inferred to be so strongly deleterious that they would rarely become fixed. Little evidence for long-term adaptive evolution was detected, as all 11 estimates for α were not significantly different from zero. Nucleotide diversity at synonymous sites, moreover, was strongly correlated with attributes of the DFE across species, thus illustrating a strong consistency with the expectations from the Nearly Neutral Theory of molecular evolution. Application of these patterns to genome-wide expectations for these species, however, was difficult as the loci chosen for the analysis were a biased set of conserved loci, which greatly influenced the estimates of the DFE and α. This implies that genome-wide parameter estimates will need truly genome-wide data, so that many of the existing patterns documented previously for forest trees (e.g. little evidence for signature of selection) may need revision.

The challenge of separating signatures of local adaptation from those of isolation by distance and colonization history: The case of two white pines

“…Briefly, an initial set of 118 gene sequences from the White Pine Resequencing Project (WHISP, http://dendrome.ucdavis.edu/wpgp; Eckert et al., 2013), randomly distributed across the genome, was selected. We also included 23 candidate genes for growth, phenology, and cold hardiness in Picea glauca , 24 candidate genes for wood formation in P. glauca , one candidate gene for adaptation to aridity in Pinus taeda , and two gene sequences available from GenBank (see Nadeau et al., 2015 for more details).…”

“…Forty‐one genes (51 SNPs) and 48 genes (52 SNPs) contained SNPs only in P. strobus and P. monticola , respectively. We deduced SNP annotations (i.e., noncoding, synonymous, nonsynonymous) for 71 fully annotated genes from the WHISP dataset (Eckert et al., 2013). For the other gene sets, the Picea glauca gene catalog was used to deduce coding regions and SNP annotations.…”

The Evolutionary Genetics of the Genes Underlying Phenotypic Associations for Loblolly Pine (Pinus taeda, Pinaceae)

“…With that said, however, the genome-wide estimate of a using sugar pine as the outgroup remained significantly positive (a = 0.10, 95% C.I. : 0.01-0.20), with the difference likely attributed to the increase of conserved amplicons in this set (Kern 2009;Eckert et al 2013). A more thorough analysis, however, was not performed, as only 950 amplicons were available, and many of the amplicons associated with phenotypes were not in this set.…”

“…Depending on the method used to estimate a, population structure would need to be pronounced, the sampling relatively even across diverged populations, and magnitudes of population structure would need to differ between categories of amplicons due to demographic history (see Eckert et al 2013). Further work based on population-level sampling, however, would be needed to quantify any effects of population structure on our inferences.…”

“…When sugar pine was used as an outgroup, however, these quantitative patterns largely disappeared. This is likely a function of the decreased number of amplicons available for analysis (3484 vs. 950), as well as the bias imposed by use of PCR primers from diverged taxa (Kern 2009;Eckert et al 2013). …”