Fig 1.
Diversity within anticancer drugs (A, left side) and diversity within natural products database (B, right side).
Fig 2.
Flowcharts for the modeling process (2a), and the ISE engine (2b).
Fig 3.
Physicochemical properties distribution of anticancer drugs (A) Molecular weight distribution, (B) Log P values, (C) Number of H-bond acceptors [lip_acc], (D) Number of H-bond donors [lip_don], (E) Number of rigid bonds, (F) number of rotatable bonds, (G) Number of aromatic atoms.
Fig 4.
Violation distribution of anticancer drugs to Lipinski rule of 5 for drug-likeness (left side) and Oprea rule for lead-likeness (right side).
Table 1.
Three filters out of the 29 filters used for producing the anticancer indexing model.
The Matthews correlation coefficients (MCCs), the true positive (TP) percentages, the true negative (TN) percentages, and the descriptors' ranges are shown.
Table 2.
Descriptors' redundancy.
Fig 5.
Redundancy of descriptors in the 29 filters used to produce the anticancer indexing model.
The picture was constructed by using WORDLE module.
Fig 6.
Indexing model for anticancer potential activity: True/false positives percentage (left Y-axis) and Matthews's correlation coefficient (MCC, right Y-axis) illustrated against molecular bioactivity index threshold (MBI, X-axis).
Fig 7.
Enrichment plot of the anticancer potential activity-indexing model of natural products.
Fig 8.
A receiver operating characteristic (ROC) curve showing the performance of the anticancer bioactivity-indexing model.
Fig 9.
Twelve of the natural products that are scored highly as potential anticancer drug candidates according to our ISE-based anticancer indexing model.