Worldwide Population Structure of the Coffee Rust Fungus Is Strongly Shaped by Local Adaptation and Breeding History.

The devastating disease coffee leaf rust, caused by , has been a major constraint to worldwide coffee production. Recently, populations were shown to be structured into three divergent genetic lineages with marked host specialization (C1, C2, and C3). However, there is yet no overall understanding of the population dynamics and adaptation of the most widespread and epidemiological relevant group (C3). We used restriction site-associated DNA sequencing to generate 13,804 single nucleotide polymorphisms (SNPs) across a worldwide collection of 99 isolates. Phylogenetic analyses uncovered a well-supported structuring within C3, with three main subgroups (SGs; SGI, SGII, and SGIII), which seem to reflect the historical distribution, breeding, and exchange of coffee cultivars. SGI shows a ladder-like diversification pattern and occurs across all four continents sampled, SGII is mainly restricted to Africa, and SGIII is observed only in Timor, revealing a higher genetic differentiation. Outlier and association tests globally identified 112 SNPs under putative positive selection, which impacted population structure. In particular, 29 overlapping SNPs per se seemed to have an extremely strong effect on population divergence. We also found exclusive and fixed alleles associated with the SGs supporting local adaptation. Functional annotation revealed that transposable elements may play a role in host adaptation. Our study provides a higher-resolution perspective on the evolutionary history of on cultivated coffee, showing its strong ability to adapt and the strength of the selective force imposed by coffee hosts, which should be taken into account when designing strategies for pathogen dissemination control and selective breeding.