Quantitative Trait Loci Mapping for Trait in Categorical Scale
Afendi, Farit Mochamad
MetadataShow full item record
Genes or regions on chromosome underlying a quantitative trait are called Quantitative Trait Loci (QTL). Characterizing genes controlling quantitative trait on their position in chromosome and their effect on trait is through a process called QTL mapping. In estimating the QTL position and its effect, QTL mapping basically utilize the association between QTL and DNA markers. However, many important traits are obtained in categorical scale, such as resistance from certain disease. From a theoretical point of view, QTL mapping method assuming continuous trait could not be applied to categorical trait. This research was focusing on the assessment of the performance of Maximum Likelihood (ML) and Regression (REG) approach employed in QTL mapping as well as the performance of Lander and Botstein (LB) and Piepho method in determining critical value in testing the existence of QTL for binary and ordinal trait by means of simulation study. The simulation study to evaluate the performance of ML and REG approach was conducted by taking into account several factors that may affecting the performance of both approaches. The factors are: (1) marker density; (2) QTL effect; (3) sample size; (4) shape of phenotypic distribution; (5) number of categories; and (6) number of QTL. Moreover, the simulation study for evaluating LB and Piepho method in determining critical value was conducted by generating distribution of the test statistic under null hypothesis. From simulation study, it was obtained that LB and Piepho method showing similar performance in determining critical value in testing the existence of QTL for binary and ordinal trait. The simulation study also indicating that both methods could be used in determining critical value in QTL mapping analysis for binary trait as well as for ordinal trait if the REG approach is used but not if ML approach is used due to their poor performance. In assessing the performance of ML and REG approach in QTL mapping analysis for binary trait, the two approaches showing comparable performance; whereas for ordinal trait REG approach showing poor performance compared with ML approach in estimating thresholds. As a result, in QTL mapping analysis, ML and REG approach could be used when dealing with binary trait, whereas ML approach is suggested when dealing with ordinal trait.