Characterization of Linkage Maps Produced with a Minimum Spanning Tree Algorithm

The Flipper deletion mapper, which implements a variant of Kruskal’s minimum spanning tree algorithm, was used to produce linkage maps in recombinant inbred populations of wheat, cotton, and 39 families of the soybean NAM (nested association mapping population), as well as the entire NAM population itself. Flipper had been improved to be able to deduplicate its input markers, i.e., to exclude markers that did not recombine with some other marker, and then to allow insertion of duplicate markers once the recombining markers had been ordered. About half of the markers mapped in single families could not be mapped in the combined NAM population because of too-high minimum recombination fraction, which is consistent with facile SNP appearance and disappearance by independent mutations in repetitive sequence. All but the whole-NAM map exhibited frequent runs of recombining or cosegregating markers, and this impression was reinforced by simulation results and chi-square tests based on the directly calculated probability that a marker recombines with a neighbor in the map. The probability that any marker recombined with the same number of neighbors in any two maps was greater than expected by chance alone, but fell far short of consistency among the maps. The whole-NAM map had very few perfectly cosegregating markers. The NAM markers came from a finished genome assembly, so it was possible to compare the distribution of recombinations to nucleotide distances in the assembly, to address whether the uneven recombination in the family maps was related to nucleotide distance, regionally (for example, gene) enhanced recombination, or sequence-specific enhancement of recombination (hotspots). Preliminary results indicate that nucleotide distance and regional variation in recombination frequency on the scale of chromosome arms can fully account for the unevenness of recombination across the map, although it is difficult to eliminate a role for sequence specificity.