مدل سازي ساختار كوواريانس آثار نشانگر هاي DNA در بين صفات همبسته با استفاده از يك مدل تحليل عاملي تصادفي

Modeling covariance structure of DNA marker effects for correlated traits employing a random factor analysis model

The objective of conducting this study was to develop a multi-trait analysis approach for modeling pleiotropic effects of quantitative trait loci (QTL). In the first part of the research a statistical method was developed to allow choosing appropriate model for the gene effect. The proposed method was based on linear mixed models in which the effects of DNA markers for correlated traits were considered as random and the covariance structure for these effects was a factor analysis model. Covariance of the residual effects among traits was modeled using an unstructured model. Phenotypic and molecular data were simulated and analyzed to investigate the efficiency of the proposed method. The effects of marker density and sample size on the power of factor analysis model to detect pleiotropic QTL were studied. The power of the model increased as the sample size increased. Increasing the marker density from one marker per 20 cM to one marker per 5 cM led to increase in the power of the model. However, increasing marker density to one marker per cM compared with one marker per 5 cM had a negligible effect on the power of the model. The results indicated that the random factor analysis model is a powerful model for modeling covariance structure of DNA markers effects among correlated traits and locating pleiotropic QTL.