Forest tree species adaptation to climate across biomes: Building on the legacy of ecological genetics to anticipate responses to climate change

Intraspecific variation plays a critical role in extant and future forest responses to climate change. Forest tree species with wide climatic niches rely on the intraspecific variation resulting from genetic adaptation and phenotypic plasticity to accommodate spatial and temporal climate variability. A centuries-old legacy of forest ecological genetics and provenance trials has provided a strong foundation upon which to continue building on this knowledge, which is critical to maintain climate-adapted forests. Our overall objective is to understand forest trees intraspecific responses to climate across species and biomes, while our specific objectives are to describe ecological genetics models used to build our foundational knowledge, summarize modeling approaches that have expanded the traditional toolset, and extensively review the literature from 1994 to 2021 to highlight the main contributions of this legacy and the new analyzes of provenance trials. We reviewed 103 studies comprising at least three common gardens, which covered 58 forest tree species, 28 of them with range-wide studies. Although studies using provenance trial data cover mostly commercially important forest tree species from temperate and boreal biomes, this synthesis provides a global overview of forest tree species adaptation to climate. We found that evidence for genetic adaptation to local climate is commonly present in the species studied (79%), being more common in conifers (87.5%) than in broadleaf species (67%). In 57% of the species, clines in fitness-related traits were associated with temperature variables, in 14% of the species with precipitation, and in 25% of the species with both. Evidence of adaptation lags was found in 50% of the species with range-wide studies. We conclude that ecological genetics models and analysis of provenance trial data provide excellent insights on intraspecific genetic variation, whereas the role and limits of phenotypic plasticity, which will likely determine the fate of extant forests, is vastly understudied.

Reduced within-population quantitative genetic variation is associated with climate harshness in maritime pine

How evolutionary forces interact to maintain genetic variation within populations has been a matter of extensive theoretical debates. While mutation and exogenous gene flow increase genetic variation, stabilizing selection and genetic drift are expected to deplete it. To date, levels of genetic variation observed in natural populations are hard to predict without accounting for other processes, such as balancing selection in heterogeneous environments. We aimed to empirically test three hypotheses: (i) admixed populations have higher quantitative genetic variation due to introgression from other gene pools, (ii) quantitative genetic variation is lower in populations from harsher environments (i.e., experiencing stronger selection), and (iii) quantitative genetic variation is higher in populations from heterogeneous environments. Using growth, phenological and functional trait data from three clonal common gardens and 33 populations (522 clones) of maritime pine (Pinus pinaster Aiton), we estimated the association between the population-specific total genetic variances (i.e., among-clone variances) for these traits and ten population-specific indices related to admixture levels (estimated based on 5165 SNPs), environmental temporal and spatial heterogeneity and climate harshness. Populations experiencing colder winters showed consistently lower genetic variation for early height growth (a fitness-related trait in forest trees) in the three common gardens. Within-population quantitative genetic variation was not associated with environmental heterogeneity or population admixture for any trait. Our results provide empirical support for the potential role of natural selection in reducing genetic variation for early height growth within populations, which indirectly gives insight into the adaptive potential of populations to changing environments.

Crop genomic selection with deep learning and environmental data: A survey

Machine learning techniques for crop genomic selections, especially for single-environment plants, are well-developed. These machine learning models, which use dense genome-wide markers to predict phenotype, routinely perform well on single-environment datasets, especially for complex traits affected by multiple markers. On the other hand, machine learning models for predicting crop phenotype, especially deep learning models, using datasets that span different environmental conditions, have only recently emerged. Models that can accept heterogeneous data sources, such as temperature, soil conditions and precipitation, are natural choices for modeling GxE in multi-environment prediction. Here, we review emerging deep learning techniques that incorporate environmental data directly into genomic selection models.