Population improvement in lucerne (Medicago sativa L.): genetic analyses in original and improved populations

Most quantitative genetics analyses are limited to the first (mean) and sec ond (variance) degree statistics and their derivatives. Analyses based on t hird (skewness) and fourth (kurtosis) degree statistics can be useful espec ially for detecting and characterising the nature of gene interactions. Thi rd and fourth degree statistics were analysed and used to interpret differe nces in forage yield among S-1 families of lucerne derived from double-cros s populations that were synthesised before (OGDC) and after (AGDC) improvem ent via inbreeding and selection. Higher levels of genetic load (deleteriou s alleles) were revealed in the OGDC population compared with the improved population. The analyses also revealed the importance of gene interaction f or forage yield in lucerne. In the unselected OGDC population, interaction between alleles in repulsion phase linkages was more important, whereas, in the selected AGDC population, interaction between alleles linked in coupli ng phase assumed greater importance. The above results suggest that inbreed ing and selection in lucerne can accumulate favourable alleles over generat ions of selection and result in population improvement. Skewness and kurtos is are relatively easy to compute and interpret and should serve as valuabl e tools in tetraploid quantitative genetics analyses.