OptiMAS: a decision support tool to conduct Marker Assisted Selection programs

With the increasing use of markers in breeding programs, it is important to develop decision support tools to help breeders in implementing their Marker-Assisted Recurrent Selection (MARS) project. OptiMAS has been developed with the possibility to consider a multi-allelic context, which opens new prospects to further accelerate genetic gain by assembling favorable alleles issued from diverse parental lines. Algorithms have been deployed to trace parental QTL alleles identified as favorable throughout selection generations, using information given by markers located in the vicinity of the estimated QTL positions. Using these results, probabilities of allele transmission are computed in different MAS schemes and mating designs (intercrossing, selfing, backcrossing, double haploids, RIL). Then, strategies are proposed to select the best plants and to efficiently intermate them. OptiMAS includes three different modules, corresponding to the different steps of the selection program:

- Computation of genotypic probabilities - Estimation of genetic values (molecular score) for each individual (at all/each QTL).
- Selection of individuals via a truncation selection based on the molecular score, or on a weighted molecular score giving more or less importance to the different QTL and a utility criterion which aims at selecting candidates based on the possibility of obtaining superior genotype in their progeny. Manual selection and complementarities between candidates regarding the favorable alleles they carry were also taken into account.
- Identification of crosses to be made among selected individuals: half-diallel between selected candidates or "better-half" strategy which consists of avoiding crosses between the worst selected individuals.