Estimation of genetic parameters using linkage between a marker gene and a locus underlying a quantitative character in F2 populations

A maximum likelihood algorithm is developed for estimating the recombination frequency in a segregating population (F2), between a marker gene and a locus affecting a quantitative trait as well as estimating the means and variances of the three genotypes of the quantitative trait. The maximum likelihood estimates are compared with the moment estimates of these parameters obtained from the algorithm described by Luo & Kearsey in 1989. It is concluded from computer simulation results that the maximum likelihood algorithm provides more accurate estimates and is more robust to changes in the value of the recombination frequency than the moment solutions, particularly with heterogenous variances. The difference between the genetic model considered here and by Luo & Kearsey and that by Darwasi & Weller, in 1992, is also discussed. Both methods for estimating r and gene effects become biased for high values of r and low values of heritability, but the results are better for data with complete dominance than for additive data.