Le lundi 21 septembre 2015 à 15:00 - CIRAD - amphi Jacques AlliotJoão Ricardo Bachega
Genetic linkage maps have traditionally been constructed for outbred populations using two-point approaches. Although these approaches have provided advances on ordering and positioning of the molecular markers in genetic linkage maps, they do not consider the information of all the markers simultaneously to find the most likely order and position into linkage groups. The multipoint approach based on hidden Markov models (HMM) is a powerful strategy to use the information of all the markers simultaneously, obtaining more precise and accurate estimates of the orders, positions and distances between markers. HMM for linkage map construction is implemented in free OneMap package available in R/CRAN, which was initially developed for analyzing outcrossing species. Currently, OneMap allows the construction of genetic linkage maps from different experimental crosses, such as full-sib, F2, backcrosses and recombinant inbred lines (RILs). For all these cases, molecular markers with different segregation patterns are used simultaneously for constructing genetic linkage map, providing with the HMM a powerful approach to represent the genome of the species and detect quantitative trait loci (QTLs) for important traits in subsequent linkage analysis studies. OneMap is under development to support large data sets, which are already available for many crop species from different genotyping-by-sequencing strategies. Moreover, a novel HMM-based model is under development to allow the construction of genetic linkage maps for autopolyploid species, which was initially motivated from single nucleotide polymorphism (SNP) data of the highly polyploid and complex sugarcane (Saccharum spp.) genome.