2020 ESA Annual Meeting (August 3 - 6)

OOS 52 Abstract - Using current and historical climate to assess adaptive genomics across multiple species in the Atlantic Forest, Brazil

Monday, August 3, 2020: 3:45 PM
Laura Bertola1, Mariana Vasconcellos2,3, Roberta Damasceno4, Ivan Prates3, Marcelo Reginato2, Nick Steiner3, Ashfaq Khan3, Natalia Quinteros3, Andressa Nuss5, Fabio Raposo do Amaral6, Ana Beatriz6, Cristina Miyaki4, Karina Silva-Brandão7, Luiza Magaldi7, Andre V. L. Freitas7, Miguel Rodrigues4, Fabián Michelangeli2, Kyle McDonald3, Ana Carnaval3 and Michael Hickerson3, (1)The City University of New York, (2)The New York Botanical Garden, New York, NY, (3)City College of New York, New York, NY, (4)Universidade de São Paulo, São Paulo, Brazil, (5)Federal University of Rio Grande do Sul, Porto Alegra, Brazil, (6)Universidade Federal de São Paulo, São Paulo, Brazil, (7)Universidade Estadual de Campinas, Campinas, Brazil
Background/Question/Methods

The South American Atlantic Forest domain is a biodiversity hotspot characterized by steep climatic gradients and high environmental heterogeneity. Studies of the spatial distribution of neutral (non-adaptive) genetic diversity - informed mostly by mitochondrial DNA - demonstrated that the patterns of diversity have been influenced by both past and current climate, however different parts of the forest have been impacted in contrasting ways. Present-day climate constrains the spatial distribution of neutral genetic diversity in species that occupy the southern, montane, cool-associated portion of the forest. In contrast, historical climates strongly impacted the distribution of neutral genetic diversity in lineages that live in the warmer (mostly northern) lowlands. Here, we use genome-wide RADseq data collected from 25 taxa distributed throughout the Atlantic Forest domain, including plants, amphibians, lizards, birds and butterflies, to investigate how shared climatic gradients correlate with genomic adaptation in individual taxa.

To quantify the current climatic features, we use three datasets: remote sensing data (MODIS Land Surface Temperature and CHIRPS precipitation), bioclimatic variables derived from remote sensing data, and bioclimatic variable derived from weather station data (CHELSA). Since we hypothesize that certain patterns of adaptation are the result of historical climate, we also include projections of historical climate from four different time periods: 1) Last Interglacial (~130 ka), 2), Last Glacial Maximum (~21ka), 3) Heinrich Stadial 1 (~17.0-14.7ka), 4) Late-Holocene/Meghalayan (4.2-0.3ka). We used latent factor mixed models (LFMM) to detect signatures of selection while accounting for taxon-specific population structure.

Results/Conclusions

Depending on the taxon and its distribution, our approach flagged 1-15% of the sampled SNPs as potentially under selection. First results indicate that species which are restricted to the (cooler) southern part of the forest, tend to have a higher number of loci flagged associated with present-day climate. Lowland, mostly northern, species have relatively higher numbers of putatively adaptive loci in association with past climates. While these patterns recapitulate those previously reported for neutral diversity, in some species they are less apparent; it remains to be tested if these non-conforming species have changed their distributions more dramatically in response to climatic oscillations over time, therefore losing specific signatures of past climate legacies.

Data collected in this study 1) give us insight into community-level responses to shared environmental gradients and shifts; 2) provide new views on the evolutionary history of the species in the Atlantic Forest; and 3) give us guidelines on how to preserve regional biodiversity.