New Methodology for Mapping Resistance Trait Loci in Crop Cultivar Population
|School||Nanjing Agricultural College|
|Course||Crop Genetics and Breeding|
|Keywords||cultivar populations resistance trait QTL mapping association analysis multi-RTL association study|
Many important taits in animal, plant and human are resistance traits, the traits are controlled by polygenes and the phenotypes are modified by environmental factors. As a consequence, the phenotypic values usually show quantitative variation. The phenotypes of such traits, however, are often measured in discrete but orderd categoties. The detection of resistance trait loci (RTL) is helpful to crop genetic improvement for the traits. However, less phenotypic information for the traits increases the difficulty in the genetic dissection of the traits, especially in the detection of RTL in crop cultivar population.In this study there were two aspects in the detection of RTL in crop cultivar population. First, association study for binary and ordered categorical traits was used to estimate the effect and position of RTL and the critical value of liability under threshold model using maximum-likelihood-based expectation and conditional maximization (ECM) algorithm. Second, multi-RTL association study for binary trait was used to estimate the effect and position of RTL and the critical value of liability under threshold model using empirical-Bayes-based Newton-Raphson iteration approach. Here the RTL effect was estimated by the posterior median. All the approaches were validated by Monte Carlo simulation studies. The main results were as follows.1) The association study for resistance traits was compared with chi-squared test approach. Although there were no difference for the power in the detection of RTL between the above two approaches, the former could estimate RTL effect and its proportion of phenotypic variation explained by RTL. Results from Monte Carlo simulation studies showed that suitable sample sizes for the detection of RTL, with 5% heritability and 2 to 4 alleles, for binary trait were from 500 to 450; those with 10% heritability were 300 and those with 15% heritability were 200. When the number of categorical trait was 3, the corresponding sample sizes were 450 to 400; 300 to 250 and 250 to 200.We have investigated the effect of rare allele with the 2.5%,5%,10%and 15% proportion at the RTL on the association study. Results from Monte Carlo simulation studies showed that the rare allele, with less than 10%, might be randomly changed into the other alleles; the rare allele, with less than 5%, might be changed into one of the other alleles as well; and the rare allele, with larger than 10%, might be not changed into the other alleles.2) Results from multi-RTL association study were similar to those from the above association study. In addition, the false positive rate for the former was less that that for the latter.