Genome scans for divergent selection in natural populations of the widespread hardwood species Eucalyptus grandis (Myrtaceae) using microsatellites

Abstract

Identification of loci or genes under natural selection is important for both understanding the genetic basis of local adaptation and practical applications, and genome scans provide a powerful means for such identification purposes. In this study, genome-wide simple sequence repeats markers (SSRs) were used to scan for molecular footprints of divergent selection in Eucalyptus grandis, a hardwood species occurring widely in costal areas from 32° S to 16° S in Australia. High population diversity levels and weak population structure were detected with putatively neutral genomic SSRs. Using three F_ST outlier detection methods, a total of 58 outlying SSRs were collectively identified as loci under divergent selection against three non-correlated climatic variables, namely, mean annual temperature, isothermality and annual precipitation. Using a spatial analysis method, nine significant associations were revealed between F_ST outlier allele frequencies and climatic variables, involving seven alleles from five SSR loci. Of the five significant SSRs, two (EUCeSSR1044 and Embra394) contained alleles of putative genes with known functional importance for response to climatic factors. Our study presents critical information on the population diversity and structure of the important woody species E. grandis and provides insight into the adaptive responses of perennial trees to climatic variations.

Figure 1. Geographic distribution of the 16 Eucalyptus grandis populations studied.

The map was generated using software ArcGIS 10.0 (http://www.esri.com/software/arcgis/). Full description of the populations can be found in Table 1. SF, state forest; ACT, Australian Capital Territory.

Figure 2. Genetic structure of 16 Eucalyptus grandis populations based on 31 putatively neutral gSSR loci.

Full description of the populations can be found in Table 1. (a) Principal coordinates analysis (PCoA). (b) Unweighted pair group method with arithmetic mean (UPGMA) dendrogram. (c) Individual proportion and population membership to each of the clusters inferred in STRUCTURE analysis (K = 2).

Figure 3. Linear regression for three significant associations between F_ST outlier allele frequencies and climatic variables.

Each dot represents a group of homogeneous populations in K-means climatic partition. (a) The 120 bp allele of locus Embra180 associated with mean annual temperature. (b,c) The 276 bp allele of EUCeSSR0755 associated with mean annual temperature and isothermality, respectively.