The scales and signatures of climate adaptation by the Arabidopsis transcriptome

Abstract

AbstractClines in allele frequency and trait variation can be highly informative for understanding how populations have historically adapted to climate variation across landscapes. However, as a consequence of the many complexities inherent to this process, these climate-associated differentiation patterns can be confounded, misleading, or obscured. Molecular phenotypes like gene expression levels are a potentially valuable means for resolving these complexities. Their intermediate position between genomes and organismal traits and their interrelatedness structured by gene regulatory networks can help parse how different climatic factors contribute to unique components of range-wide or region-specific diversity patterns. Here, we demonstrate these explanatory values of gene expression variation through integrative analyses of transcriptomic data from 665 Arabidopsis thaliana accessions. Differentiation of co-expressed genes is often associated with source site climate. Although some patterns hold range-wide, many other gene expression clines are specific to particular ancestry groups, reflecting how broad-scale and local combinations of selective agents differentially resolve functional interrelationships between plant defense, drought tolerance, and life history traits. We also extend these analyses to parse how different factors explain climate-associated variation in flowering time and its plasticity. Expression of key regulators FLC and SOC1 strongly predicts time to flower, consistent with previous work, but our results also highlight novel relationships that indicate as yet unexplored climate-related connections between defense signaling and flowering. Finally, we show that integrative models combining genotype and gene expression information predict variation in flowering time under ecologically realistic conditions more accurately than models based on either source alone.Significance StatementPopulations often adapt to local conditions along climate gradients, and associations between climate parameters and traits or alleles often indicate a history of adaptive differentiation. However, such signals can be obscured or misleading due to the complex genetics underlying trait variation or other historical processes, frustrating our capacity to reveal how populations adapt to diverse climates. As a molecular intermediate between genetic polymorphisms and their impact on organismal traits, gene expression variation is a useful readout for addressing several of these difficulties. Here, we leverage transcriptomic data from hundreds of Arabidopsis thaliana accessions to reveal continental and region-specific patterns of climate-associated differentiation as well as investigate how gene expression adaptation at both scales shapes flowering time variation along climate gradients.