Population genetics of the coral Acropora millepora: toward genomic prediction of bleaching

INTRODUCTION Coral reefs worldwide are suffering losses at an alarming rate as a result of anthropogenic climate change. Increased seawater temperatures, even only slightly above long-term maxima, can induce bleaching—the breakdown of the symbiotic relationship between coral hosts and their intracellular photosynthetic dinoflagellates from the family Symbiodiniaceae. Because these symbionts provide the majority of energy required by the coral host, prolonged periods of bleaching can eventually lead to the death of the colony. In the face of rapidly increasing temperatures, new conservation strategies are urgently needed to prevent future mass losses of coral cover, and these benefit from an understanding of the genetic basis of bleaching. RATIONALE Bleaching responses vary within and among coral species; in the reef-building coral Acropora millepora, a commonly distributed species across the Indo-Pacific, these differences have been shown to be at least partly heritable. In principle, therefore, interindividual differences in bleaching should be predictable from genomic data. Here, we demonstrate the feasibility of using a genomics-based approach to predict individual bleaching responses and suggest ways in which this can inform new strategies for coral conservation. RESULTS We first generated a chromosome-scale genome assembly as well as whole-genome sequences for 237 samples collected at 12 reefs distributed across the central Great Barrier Reef during peak bleaching in 2017. We showed that we can reliably impute genotypes in low-coverage sequencing data with a modestly sized reference haplotype panel, demonstrating a cost-effective approach for future large-scale whole-genome sequencing efforts. Very little population structure was detected across the sampled reefs, which was likely the result of the broadcast spawning mode of reproduction in A. millepora. Against this genomic background, we detected unusually old variation at the heat-shock co-chaperone sacsin, which is consistent with long-term balancing selection acting on this gene. Our genomic sequencing approach simultaneously provides a quantitative measure of bleaching and identifies the composition of symbiont species present within individual coral hosts. Testing more than 6.8 million variants for associations with three different measures of bleaching response, no single site reached genome-wide significance, indicating that variation in bleaching response is not due to common loci of large effect. However, a model that incorporates genetic effects estimated from the genome-wide association data, genomic data on relative symbiont species composition, and environmental variables is predictive of individual bleaching phenotypes. CONCLUSION Understanding the genetics of heat and bleaching tolerance will be critical to predict coral adaptation and the future of coral reef ecosystems under climate change. This knowledge also supports both conventional management approaches and the development of new interventions. Our work provides insight into the genetic architecture of bleaching response and serves as a proof of principle for the use of genomic approaches in conservation efforts. We show that a model based on environmental factors, genomic data from the symbiont, and genome-wide association data in the coral host can help distinguish individuals most tolerant to bleaching from those that are most susceptible. These results thus build a foundation toward a genomic predictor of bleaching response in A. millepora and other coral species