| Summary: | Genetic mapping contributes to the understanding of functional mechanisms that underlie the constitution of living organisms and their physiology. For example, genetic mapping can be used in conceiving new treatments of congenital or infectious diseases and in selecting plants and animals that have a higher and better production. The most common approaches of genetic mapping exploit the allelic segregation in a pedigree during only a few number of generations and, consequently, they do not have a sufficient resolution to allow an effective gene isolation and cloning. An alternative to these approaches is to study allelic associations along the history of a population. This requires accurate models of population demography, genetic inheritance and allelic associations.
This thesis contributes to the modelling of allelic associations (linkage disequilibrium, LD) and to the assessment of the effects of selection and inbreeding. In a simulation framework, we fitted the multimarker-multiallelic LD with an exponential function characterised by two parameters: the distance (R) at which LD is independent of the genetic distance and the LD reached at this distance (residual LD, rs). As an application of this approach, the LD was estimated in five populations of pigs. We observed a long range LD (>10cM) that was explained by the random drift. Moreover, significantly increased LD was found in regions harbouring selected QTL (quantitative trait loci), suggesting an effect of selection. Fitting LD with the exponential model proposed in simulations indicated that mapping methods using LD (LDM) can achieve a resolution of ~3cM in the populations of pigs we have studied and can be powerful with a marker spacing of 5-10cM.
As illustrated with these data from pigs, the model that we used to fit LD offers opportunities to characterise allelic association in populations, estimate the required marker density for genome-wide LD studies and determine the expected resolution of LDM. It is also shown that the proposed model can help overcoming the assumptions of asymptotic linkage equilibrium and independence between markers that are commonly made in LDM but are not always fulfilled.
|
|---|
