- Estimates were obtained from ancestral recombination graphs (ARGs) along chromosome 8. Gene density as the proportion of the window containing coding sequence. The estimated TMRCAs in thousands of yr ago (ka) were obtained assuming a 10 yr generation time. N0, ancestral effective population size; Nc, current effective population size.
Correlates and you can hotspots out-of recombination
Correlations calculated based on 6122 windows of 100 kb revealed that ? is significantly positively correlated, albeit in decreasing magnitude, with nucleotide diversity, gene density, Lancaster top hookup sites distance from the centromere and GC content (Table 5). g. chromosomes 1, 3, 4 and 6), corroborated by the higher positive correlations (Table S2), although more centrally located peaks were observed on chromosomes 7 and 10 for which correlations were negative (Table S2; Fig. 5). Again no significant impact of the localized assembly inconsistencies of the E. grandis genome version 1.1 was seen on the genome-wide profiles of population recombination rate (Fig. S4) and the chromosome-specific estimates of ? (Fig. S5). The profiles of gene density also showed higher values toward the tips of the chromosomes and significant positive correlations with the distance from the centromere were seen for all chromosome except no. 9 (Table S2). In opposite to those spiky patterns, very homogeneous degrees of nucleotide diversity were seen across the entire genome and, as expected, ?w was strongly negatively correlated with gene density and positively correlated with GC content. A total of 179 hotspots of recombination were detected in the E. grandis genome. When compared with the rest of the genome, the hotspot regions were found to be significantly enriched for GO terms associated with chemical reactions and pathways that are part of the normal metabolic processes, while terms associated with regulation, signal transduction and response were significantly underrepresented (Table 6).
- Significance values (P-values in parentheses) were obtained using a permutation test to account for the spatial autocorrelation among the 100 kb windows. ?, population recombination rate; ?w, Watterson nucleotide diversity.
- Wade, gene ontology.
Discussion
A combination of whole-genome pooled resequencing and high-density SNP genotyping was used to report the first genome-wide examination of key features of the Eucalyptus genome that have a profound impact on the understanding of fundamental evolutionary issues and on the practice of molecular breeding: nucleotide diversity, population recombination and extent of LD. We corroborated previous results showing that E. grandis displays one of the highest nucleotide diversities in plants in general, and forest trees in particular (?w = 0.022), consistent with its exceptional genetic and phenotypic variation (Grattapaglia et al., 2012 ), resulting from a high estimated genome-wide mutation rate (4.9 ? 10 ?8 bp ?1 per generation). The genomic landscape of recombination showed that the 11 Eucalyptus chromosomes display a relatively similar rate of recombination, which in turn was found to be significantly positively correlated with nucleotide diversity, gene density, distance from the centromere and GC content. This picture was consistent with the fairly constant recombination rates seen along the linkage maps. Both Infinium and sequence-based SNP data converged to essentially equivalent measures of genome-wide population recombination rate (? = 0.53–1.47 ? 10 ?3 bp ?1 ) using alternative estimation methods. Thus, at the genome-wide scale, mutation was found to be c. 1.5 times more important than recombination in shaping the genomic diversity of E. grandis. When a closer look was taken at the larger variance associated with the genome-wide estimate of ? obtained from sequence data, a surprisingly variable pattern emerged. A progressive decrease of ? by two orders of magnitude was seen when going from the short (0–100 bp) pairwise SNP distance range to the genome-wide scale (0–50 kb). Perhaps not so surprisingly, given recent reports showing variable extents of LD in other forest trees, we provide compelling evidence that the extent of LD in Eucalyptus, when assessed at the genome-wide scale, decays considerably more slowly (c. 4–6 kb) when compared with previous reports based on the analysis of short sequence stretches in candidate genes.