Genomic Prediction of Disease Resistance Provides a Path to Marker Assisted Restoration in a Wetland Foundation Tree...

Tree species worldwide are under threat from non-native pathogens that impact forests and the ecosystem services they provide. Myrtle rust, caused by Austropuccinia psidii, is one example, first detected in Australia in 2010. This fungal pathogen infects immature tissue from a wide range of Myrtaceae hosts, including the wetland foundation species Melaleuca quinquenervia. Durable restoration action for this species would preferentially incorporate disease-resistant individuals. Our aim for this study was to identify genetic markers and develop a genomic prediction model that could assist with selection for resistance. We conducted artificial inoculation of a panel of seedlings and measured their immune responses to myrtle rust. We then performed whole genome sequencing (3.2 M common SNPs) and conducted a Genome Wide Association Study (N = 492), which revealed clusters of significantly associated SNPs in three chromosomal regions, including clusters of putative R genes. Associated SNPs were filtered to a panel of 1049 for a highly accurate genomic prediction model (R = 0.83, precision = 0.65 for predicting which plants were in the top third for resistance). This provides a relatively inexpensive approach to identifying resistant individuals or seed lots for restoration and a template for managing myrtle rust impacts while maintaining population genetic diversity.