Molecular docking-based classification and systematic QSAR analysis of indoles as Pim kinase inhibitors

• Published in: SAR and QSAR in environmental research 2020-05, Vol.31 (5), p.399-419
• Main Authors: Kalaki, Z., Asadollahi-Baboli, M.
• Format: Article
• Language: English
• Subjects: adaptive neuro fuzzy inference system; Cancer; Colon; discriminant analysis; docking based descriptors; Enzymes; Hydrogen bonding; Hydrogen bonds; Indoles; Inhibitors; Kinases; Learning algorithms; Leukemia; Machine learning; Molecular docking; Pim kinase enzyme; Prostate; QSAR; Residues; Structure-activity relationships
• ISSN: 1062-936X; 1029-046X
• DOI: 10.1080/1062936X.2020.1751277

Pim kinase enzyme has an essential role in the treatment of prostate, colon and acute myeloid leukaemia cancers. The indoles inhibitors were docked in the enzyme's active pocket in order to survey the inhibition mechanism and extract the ligands' conformations. The docking outcome shows that the active inhibitors have strong van der Waals interactions with residues of Ile185, Leu44, Leu120 and Leu174, hydrogen bonds with residues of Asp128, Arg122 and Glu171 and π-π interaction with the residue of Phe49. The sum of these interactions is ~80 kcal mol −1 contributing ~90% of total binding free energies. Using docking-based molecular descriptors, the unsupervised and supervised classifications were successfully carried out with the accuracy of 0.82 and 0.95, respectively, to categorize the active/inactive Pim kinase inhibitors. The vigorous quantitative assessment was performed using different machine learning techniques. The constructed QSAR model [(r 2 cal , r 2 p , r 2 m and Q 2 LOO ) > 0.80 and (SE cal , SE p and SE LOO )