Bulding a SNP-based Multipoint Sugarcane Linkage Map Using Loci With All Possible Dosages

Sugarcane is an autopolyploid species with high number of chromossomes and frequent aneuploidy. Due to its genome complexity, it does not fully benefit from modern molecular markers technologies. The vast majority of available genetic maps are based on simplified approaches, using only loci that are present in single copies on the genome and thus segregate in a 1:1 fashion. Even in the few studies that included markers with higher dosages, multipoint analysis using statistical methods such as hidden markov models were not used. Moreover, these studies generally assumed that sugarcane is octoploid and used marker systems with limited information (e. g. SSR). We will show that the usage of modern marker systems, such as SNPs (that are codominant for sugarcane), combined with a new statistical model developed for building linkage maps for autopolyploids with any ploidy level, could result in more saturated maps including loci with all available dosages. The assumptions of our model are more realistic, and homology groups with any ploidy level could be estimated. By using multipoint analysis based on hidden markov models, information from all markers in each homology group is used simultaneously, increasing the statistical power and resulting in more saturated maps. Simulated and real sugarcane results will be presented and discussed. Implications of this new approach for genetic studies in polyployd species, and also for QTL mapping, will be addressed.