Gene Expression Programming for Prediction of Acute Toxicity of Aldehydes

A novel way for building quantitative structure activity relationship (QSAR) model to predict the toxicity of aldehydes for rat was presented. A novel algorithm Gene Expression Programming (GEP) was utilized. GEP uses character linear chromosomes composed of genes structurally organized in a head and a tail. The chromosomes function as a genome and are subjected to modification by means of different operators. The chromosomes encode expression trees which are the object of selection. The rat lethal dose 50% (LD50) of 88 diverse aldehydes was modeled using the descriptors calculated from the molecular structure along using a quantitative structure-activity relationship technique. Gene Expression programming (GEP) was utilized to construct the linear and nonlinear prediction models, leading to a good cross-validation correlation coefficient (R-2) of 0.947 and root means square errors (S-2) of 0.037. The prediction results of the nonlinear GEP model seem to be better than those of support vector machine (SVM). The GEP may offer new possibilities for solving more complex technological and scientific problems.