Hotel Estoril Eden, Monte Estoril,
5-8 October 2005



Text Box: Participants
NextText Box: Programme

se of Bayesian Inference for Genetic Mapping

A.M.Mayoral, J.Morales, E.Carbonell, M.J.Asíns and M.Alacreu
Centro de Investigación Operativa, Depto. Estadística, Matemática e Informática, Universidad Miguel Hernández, Spain

We develop a unified Bayesian model to estimate recombination fractions
among all markers within a linkage group and also an algorithm to obtain the genetic map from the estimates. We consider here codominant markers and several family designs: Back-Cross, Duplo Haploids, F2, F3, ... ; Fn. Our approach allows for:

  •  accounting prior uncertainty about the recombination fractions, r, known to be between 0 and 0.5 for markers within the same linkage group;
  • a unified model for the observed data of descendants, based on the frequencies for every configuration between pair of markers;
  • a very intuitive way to communicate results about estimates and the
    ordering of markers, based on probabilities of occurrence, much more informative than p-values;
  • assessing the certainty degree obtained on different maps;
  • assessing the certainty degree obtained on the location of each marker in the map.

The proposed Bayesian model for estimation uses objective prior distributions. Its simplicity and flexibility to adapt to any population of the cited above is relevant. Monte Carlo integration and MCMC simulation methods are necessary for obtaining results. Then, conclusions are easily interpreted. The model and correct performance of the ordering algorithm are illustrated on several simulated examples.