Taking genomics outdoors: linking local adaptation, trait variation, and gene expression in grass ecotypes across a rainfall gradient

Abstract With increasing droughts, understanding local adaptation and drought tolerance in ecologically dominant species is crucial for enhancing ecosystem resilience. We leveraged a long-term reciprocal garden to assess local adaptation and drought responses in Andropogon gerardi , a foundation grassland species in the US Great Plains. Our objectives were to identify adaptive traits, explore gene expression responses across a rainfall gradient and under experimental drought and integrate trait-based analyses with gene expression profiles to test for local adaptation. Reciprocal gardens, established a decade ago, are composed of different A. gerardi ecotypes sourced along a rainfall gradient (MAP 480–1167mm yr -1 ) and sown as ecological communities. Rainout shelters imposed experimental drought. We hypothesized that ecotypes should perform best in their homesite, reflecting local adaptation. The dry ecotype should perform best under rainouts and exhibit traits and expression profiles favoring drought tolerance; the wet ecotype should favor traits and gene expression enhancing growth and resource acquisition. We found ecotypes had highest biomass and cover in their homesite, confirming local adaptation. Under experimental drought, the dry ecotype demonstrated improved performance at the wet site, confirming its adaptive value under water limitation. The dry ecotype showed stress-tolerance strategies (shorter, more water-efficient, upregulated drought-response genes), while the wet ecotype emphasized growth strategies (taller, higher biomass, upregulation of growth hormone gibberellin). Using co-expression networks, gene clusters linked adaptive traits, revealing genetic mechanisms of adaptation. Results advance our understanding of adaptation by linking gene expression and trait variation to drought responses in locally-adapted ecotypes, informing ecotype climate-matching under drought.