Genomic evidence of introgression and adaptation in a model subtropical tree species, Eucalyptus grandis

Abstract

The genetic consequences of adaptation to changing environments can be deciphered using population genomics, which may help predict species' responses to global climate change. Towards this, we used genome‐wide SNP marker analysis to determine population structure and patterns of genetic differentiation in terms of neutral and adaptive genetic variation in the natural range of Eucalyptus grandis, a widely cultivated subtropical and temperate species, serving as genomic reference for the genus. We analysed introgression patterns at subchromosomal resolution using a modified ancestry mapping approach and identified provenances with extensive interspecific introgression in response to increased aridity. Furthermore, we describe potentially adaptive genetic variation as explained by environment‐associated SNP markers, which also led to the discovery of what is likely a large structural variant. Finally, we show that genes linked to these markers are enriched for biotic and abiotic stress responses.

Genomic consequences of artificial selection during early domestication of a wood fibre crop

“…Population differentiation patterns were investigated and compared among breeding populations, and between breeding populations, wild E. grandis (including 362 individuals from three subpopulations; Mostert‐O'Neill et al ., 2021) and other Latoangulatae species as published by Silva‐Junior et al . (2015) using four approaches: principal component analysis (PCA) with normalization to each marker’s standard deviation in S VS 8; sparse nonnegative matrix factorization (sNMF) using the lea R package (Frichot et al ., 2014; Frichot & François, 2015) – the values for K tested were K = 2 to K = 10 with five repetitions of each value and the minimum cross‐entropy (CE) was determined for each value of K and visualized; discriminant analysis of principal components (DAPC) using the adegenet R package (Jombart, 2008; Jombart et al ., 2010) with Bayesian information criterion (BIC) used to determine the most probable cluster number in the data set with K = 1 to K = 15 tested; and the extent of differentiation among breeding populations, and between breeding and wild E. grandis populations was quantified as F ‐statistics, F ST , as described by Weir & Cockerham (1984), with 95% confidence intervals in S VS 8.…”

“…(2019a) and Mostert‐O’Neill et al . (2021). Two sets of genes, within 2 and 6 kb, were analysed based on the lower and upper estimates of LD decay as determined by Silva‐Junior & Grattapaglia (2015), to account for large variations in genome‐wide LD patterns in the breeding populations.…”

“…Suspected introgression was further tested by interrogating individual genotypes for the presence of genomic segments not originating from E. grandis by ancestry mapping using the Efficient Inference of Local Ancestry ( eila ) R package (Yang et al ., 2013) as described by Mostert‐O’Neill et al . (2021) with the breakpoint penalty λ = 30. Briefly, probable ancestry was calculated for each SNP using the same three reference populations as described by Mostert‐O'Neill et al .…”

“…Next, the cumulative probability estimates of all SNPs on a chromosomal segment were used to assign each segment to one of the three ancestral populations. The penalty for allowing breakpoints (λ), where ancestry switches from one ancestral population to another along a chromosomal segment, was previously optimized based on ancestry mapping conducted in pure species and hybrid individuals for which the ancestry was known (Mostert‐O’Neill et al ., 2021). This approach allowed for the identification of even small introgressed genomic segments, which could confound the detection of genomic regions selected during early domestication.…”

“…We also investigate the possibility that interspecific hybridization and recent infusions from unimproved, wild material have contributed to the genetic diversity in South African breeding populations, as is suggested to have occurred in the domestication of other crops (He et al ., 2011; Myles et al ., 2011; Cornille et al ., 2012; Wu et al ., 2014; Baute et al ., 2015). This is done by elucidating the population structure and genetic differentiation of breeding populations relative to wild E. grandis populations (Mostert‐O’Neill et al ., 2021) and species with which E. grandis could have hybridized ex situ (Silva‐Junior et al ., 2015). Next, we define the core E. grandis breeding germplasm, representative of the advanced‐generation population that has been under selection for a century.…”