Artificial Intelligence in Drug Formulation and Delivery: Benefits, Trends, and Future Perspectives

Aurora-kinase is a key regulator of centrosome function during mitosis and meiosis and is involved in many mitotic events in the cell cycle.Aurora kinases, specifically Aurora-A are extensively expressed in many tumors.As a result, targeting Aurora-A kinase is established to be an important target in the treatment of cancer.In our study, the 3D-QSAR pharmacophore model generation using the Hypogen algorithm was performed to generate a valid predictive pharmacophore model using a data set of 35 reported pyrazole and Furano-pyrimidine analogs of Aurora-A inhibitors.The pharmacophore model selected had a cost difference of 87.029 a correlation coefficient of 0.97 and an RMS of 0.83, thus it was proven to be statistically significant.The pharmacophore model showed one hydrogen bond donor, one hydrophobic, and two ring aromatic features.This model was selected to virtually screen different databases (SPECS, and scPDB databases).Fit value was used to filter the screened ligands.Three top hits of this screening were further subjected to docking studies in the binding site of the Aurora-A kinase receptor (PDB ID: 4JBO).Docking results of the top three hits had a high binding affinity to the Aurora-A kinase receptor and showed a similar pattern of binding interactions to the reference (IBPR001).Subsequently, the top hits are predicted to be potential Aurora-A kinase inhibitors.As a result, this research reveals potential promising Aurora-A kinase hits which can be further optimized for novel Aurora-A kinase inhibitors with higher efficacy and safety profile.