Abstract

QSAR study was performed on a series of aryl carboxylic acid amide derivatives (62 analogs) to establish structural features required for human dihydroorotate dehydrogenase (hDHODH) inhibition. Statistical significant QSAR models were developed for the prediction of hDHODH inhibitory activity by applying MLR analysis (r2 = 0.851 and q2 = 0.795), PCR analysis (r2 = 0.713 and q2 = 0.667) and PLS analysis (r2 = 0.848 and q2 = 0.802). QSAR study emphasized the importance of topological, estate number, hydrophobic and alignment independent descriptors for the prediction of hDHODH inhibitory activity. SaasCcount descriptor suggested the presence of carbon atoms in five member aryl ring system. Positive impact of alignment independent descriptors reveals the presence of carbonyl oxygen and chloro group in this series of compounds. DistTopo signifies basic connectivity of atoms in the molecules. High degree of predictability of the proposed QSAR models offers a great potential for the design and development of potent hDHODH inhibitors.

Keywords: Dihydroorotate dehydrogenase (DHODH), QSAR, Multiple linear regression (MLR), Principle component regression (PCR), Partial least square (PLS) analysis, aryl carboxylic acid amide derivatives