Categorical QSAR Models for Skin Sensitization based upon Local Lymph Node Assay Classification Measures Part 2: 4D-Fingerprint Three-State and Two-2-State Logistic Regression Models

• DOI: 10.1093/toxsci/kfm185
• Publication Date: 2007-08-04T00:40:40Z
• Authors: Yi Li; Dahua Pan; Jianzhong Liu; Petra S. Kern; G. Frank Gerberick; Anton J. Hopfinger; Yufeng J. Tseng
• Keywords: Categorical variable; Data set

Three and four state categorical quantitative structure–activity relationship (QSAR) models for skin sensitization have been constructed using data from the murine Local Lymph Node Assay studies. These are same we previously used to build two-state (sensitizer, nonsensitizer) QSAR (Li et al., 2007, Chem. Res. Toxicol. 20, 114–128). 4D-fingerprint descriptors derived 4D-molecular similarity paradigm generate these models. A training set of 196 a test 22 structurally diverse compounds were in this study. Logistic regression, partial least square coupled logistic regression The three-state model gives classification accuracy 73.4% 63.6% set, while random average value any is 33.3%. two-2-state [four categories total] 83.2% 54.6% 25%. An analysis skin-sensitization developed study, as well our previous analysis, suggests that "moderate" sensitizers may be main source limited accuracy.