CORAL (‘CORrelation And Logic’) is freeware available on the Internet www.insilico.eu/coral The aim of this program is to establish a correlation between an endpoint and descriptors calculated with a simplified molecular input line entry system (SMILES). Three models calculated by CORAL for toxicity towards rat (-pLD50) of inorganic substances (three random splits) have shown that CORAL could be a good tool to model this endpoint. The average statistical characteristics for the CORAL models are the following: n=38, r2=0.8461, q2=0.8298, s=0.273, F=198 (subtraining set); n=37, r2=0.8144, s=0.322, F=154 (calibration set); and n=10, r2=0.8004, Rm (test)2 =0.7815, s=0.240, F=32 (validation set). [...]
Optimal descriptors calculated with simplified molecular input line entry system (SMILES) have been examined as a tool for prediction of anxiolytic activity. Descriptors calculated with SMILES (a) of keto-isomers; (b) of enol-isomers; and (c) of both keto-isomers together with enol-isomers have been studied. Three approaches have been compared: 1. classic’ training-test’ system 2. balance of correlations and 3. balance of correlations with ideal slopes. The best statistical characteristics for the external validation set took place for optimal descriptors calculated with SMILES of both keto-form and enol-form (i.e., molecular structure was represented in the format: ’sMILES of keto-form. SMILES of enol-form’) by means of balance of correlations with ideal slopes. The predictive potential of this model was checked with three random splits. [...]
JavaScript is turned off in your web browser. Turn it on to take full advantage of this site, then refresh the page.