* Leave-one-out cross-validation (LOO CV) procedure is performed using the whole PASS training set for validation of prediction quality. The prediction result is compared with known experimental data for the studied compound. The procedure is repeated for all compounds from the PASS training set; then the average Invariant Accuracy of Prediction (IAP=1-IEP) values are calculated for each biological activity and for all biological activites.
IAP equals numerically to ROC AUC

* Leave-one-out cross-validation (LOO CV) procedure is performed using the whole PASS training set for validation of prediction quality. The prediction result is compared with known experimental data for the studied compound. The procedure is repeated for all compounds from the PASS training set; then the average Invariant Accuracy of Prediction (IAP=1-IEP) values are calculated for each biological activity and for all biological activites.
IAP equals numerically to ROC AUC

Training set

Training sets on cytotoxicity of chemical compounds built on experimental data from CHEMBLdb were used to train PASS for “structure-cell line cytotoxicity” relationship prediction. The average prediction accuracy calculated by leave-one-out cross-validation procedure is approximately 93% for cytotoxicity prediction for cancer cell lines and non-tumor cell lines.

The outcome of the collection of experimental data was the training set of 59,882 structures of compounds, which reflects current knowledge about the cytotoxic substances in relation to 943 human cell lines.