A Family-Wide Analysis of Disease Resistance Genes in the Rosaceae Using a Custom Bioinformatic Pipeline

Efficiently targeted mining of genomic data benefits plant breeding. We developed a Python-based bioinformatics pipeline for the discovery, characterization, and phylogenetic analysis of disease resistance (R) genes from public genome databases. We conducted a comparative analysis of R genes from the high-quality genome sequences of eleven species within the Rosaceae, including important fruit crops like apple, pear, peach, and strawberry. From these eleven genomes, a total of 2,489 putative R genes encoding full-length NB-ARC domains were discovered. The pipeline classified 793 putative R genes as TIR, and 1,696 as non-TIR subtypes. The pipeline additionally clustered putative R protein sequences into the RosaR80 Framework, a set of 651 homology groups based on an 80% amino acid identity threshold for the NB-ARC domain. Maximum Likelihood analysis of RosaR80 homology groups revealed idiosyncratic patterns of diversification. Groups present in all genomes represented highly conserved R gene lineages, with preserved gene copy number and syntenic chromosomal positions between Rosaceous species. RosaR80 homology groups derived from a single species represented R gene lineages with species-specific patterns of diversification. For example, in Malus × domestica, cloned and candidate genes affording resistance to apple scab (Venturia inaequalis) were derived from large, highly diversified, species-specific lineages, often located in chromosomal gene clusters, suggesting a recent co-evolutionary arms race with the pathogen. In addition to being a tool for generating new evolutionary hypotheses, our pipeline facilitates the identification of agronomically useful R gene lineages for plant improvement, and may inform evaluation of gene bank collections for targeted allele mining.