Application of neural network to linkage analysis of quantitative traits

• Main Authors: Min-Shiuann Yang, 楊民鉉
• Other Authors: 戴政
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/46624290848206087276

碩士 === 國立臺灣大學 === 流行病學研究所 === 89 === There are many methods for detecting the position of disease loci for complex traits. In this thesis the back-propagation neural network approach is used to simultaneously detect multiple loci for complex traits. Neural networks are set up by iteratively training the transmission paths from the input (markers) to the output (disease state) individual by individual. The markers that have the highest contributed path values are selected as the candidate gene for a trait. Simulated data are generated for investigating the performance of this method. Our result indicates that the neural networks method is useful in detection of QTL. It can be considered as an alternative method in analysis of complex traits.