Summary of results for human liver microsomal models generated with very large training sets and different molecular descriptors

HLM Model with CDK and SMARTS KeysHLM Model with MOE2D and SMARTS Keys
No. of descriptors578818
No. of training set compounds193,650193,930
Cross-validation results38,730 compounds38,786 compounds
Training R20.790.77
20% test set R20.690.69
Blind data set (2310 compounds)
Continuous and categorical
    Positive predictive value0.800.823
Time (s/compound)0.2520.303