Development of a Quantiative Structure Activity Relationship Model to Predict Mutagenicity of Atomatic Nitro Based on Hierarchical Support Vector Regression

• Main Authors: You-Chen Lu, 呂侑宸
• Other Authors: Max K. Leong
• Format: Others
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/64415124179721863380

碩士 === 國立東華大學 === 化學系 === 101 === Aromatic nitro compounds are commonly seen in industries, food, and environment, and they are closely related to mutagenicity or even carcinogenicity in some cases. As such, it is of necessity to develop an in silico model to predict their mutagenicity. The objective of this study was to construct a predictive model using 15 descriptors and the hierarchical support vector regression (HSVR) scheme based on Salmonella typhimurium TA98‒S9 mutagenicity data compiled from the literature. Various statistical analyses were employed to ensure its accuracy and predictivity. The predictions by the HSVR model are in good agreement with the observed values for those molecules in the training set (n = 226, r2= 0.909, = 0.901, RMSE = 0.627, s = 0.375), test set (n = 56, r2= 0.827, RMSE = 0.732, s = 0.441), and outlier set (n = 8, r2 = 0.845, RMSE = 0.315, s = 0.200). The results indicate that the HSVR model performed better than published models statistically and can be employed as a tool to predict mutagenicity of new aromatic nitro compounds.