Abstract
Recombination is one of the forces that helps to shape the genetic diversity of populations by facilitating crossover events, in which homologous chromosomal sequences are exchanged. This process generates novel allelic combinations, enabling populations to adapt to selection pressures. Understanding the factors influencing crossover placement is vital for breeders, as it allows for the targeted transfer of specific traits or genes to offspring. In this study, we explore three different types of genome features, such as k-mers, expression elements and repetitive elements, and their relationships with recombination in ten intraspecific populations of Brassica oleracea, and one interspecific cross between two tomato species; Solanum lycopersicum and Solanum pimpinellifolium. Our results reveal that specific AT-rich k-mers, expression elements from gene annotation, and certain repetitive elements are positively associated with meiotic recombination. In contrast, CG-rich k-mers and other repetitive elements, such as some LTR retrotransposon families, show negative associations. These features were subsequently used to train regression-based machine learning models capable of predicting recombination patterns along chromosomes. Our findings suggest that plant genomes contain sufficient information to infer recombination landscapes along chromosomes.