We defined the best precision threshold as the P-value cut-off that achieved prediction of a maximal proportion of true positives while requiring at least one false positive gene.
Disregarding low-confidence predictions, YLoc can achieve prediction accuracies of over 90%.
Similarly, the GEMS-SVM classifier with the 6 predictive genes achieved prediction accuracies of 88.9%, 87.5% and 89.8% respectively for three acetaminophen data sets from NCT008, NTP and NCT informatics challenge.
The orthology based method achieved prediction accuracy of 83.86%, and among the novel predictions, it only covered 56.35% of our novel discoveries.
After applying the RD threshold of 3.03 Å to discriminate the exposed (< = 3.03 Å) and buried (>3.03 Å) residues in the current dataset, we achieved prediction accuracies of 74.1% and 82.9% for exposed and buried residues, respectively, with the overall prediction accuracy of 78.2%.
It is then possible to examine if the achieved prediction results of the original data set are significantly different than random predictions.
The achieved prediction ability was 76.1% of cross-validated grouped cases correctly classified.
They achieved prediction error 23% on average of internal target positions based on the clinical data observed between external marker and internal marker motions.
Moreover, the possibility to achieve better prediction of rat oral acute toxicity by combining the predictions from multiple QSAR models was explored.
