Fast and Deterministic Linkage Mapping of Polyploids Using an Approximate Calculation of Recombination Frequencies

Until now, linkage mapping of polyploids was restricted to tetraploid crops and specific markers. We developed a new approach that produces fast linkage maps independent of the level of ploidy. The increased number of chromosome copies lead to more possibilities of recombination and double reduction. Instead of using computationally expensive approaches like the expectation-maximization (EM) algorithm to determine pairwise recombination frequencies for all markers, we apply a fast heuristic that approximates the value. The sorting of markers within the linkage group is solved efficiently by hierarchical clustering combined with a leaf ordering algorithm. Compared to sliding window approaches of other open source tools, our solution is faster and leads to better chromosome wide results. For non-deterministic marker orders, we extended the sum of adjacent recombination frequencies (SARF) criteria to neighboring markers, until the order is solved unambiguously. Our algorithm was applied to both simulated and real data of different ploidy levels. In both cases we obtained linkage maps with the expected numbers of linkage groups and orders of markers. Our approach enables scientist and breeders of horti- and agriculturally important polyploid crops to use the full genetic potential of their lines.