The conformation generation for test set compounds was carried out in a similar way like training set compounds using conformation analysis algorithm.
In addition, in the GPCR, and PI data set compounds were described by their physicochemical properties.
Figure 11 Visualizing the selected test set compounds of the Caco-2 dataset.
The structures of the training set compounds were shown in Fig. 2.
By recalculating all scores for the training set scaffolds, using only activity data for training set compounds, we simulated the situation where test set compounds are new and unknown.
The conformation generation for the test set compounds was carried out in a similar way, as the training set compounds using the BEST conformational analysis algorithm, implemented within LigandScout.
The best model is able to predict the experimental toxicity of the test set compounds with an accuracy of 90%.
The average similarity to the nearest neighbour in the training set for the test set compounds was 0.53.
Plotted are acd_logd values (x-axis) versus the predicted logD values (y-axis) for 100,000 test set compounds.
Compounds B7, B8, B10, D2-D4, E2, F2, H8, J4, and J8 were selected as test set compounds.
