ABSTRACT
The multivariate image analysis (MIA) applied to quantitative structure–activity relationship (QSAR) approach has been shown to be a promising tool in the modeling of bioactivities of chemical compounds. This chapter provides a comparative analysis of the traditional MIA-QSAR, aug-MIA-QSAR, and aug-MIA-QSARcolor approaches in the prediction of the Michaelis constant (K m) and substrate cleavage rate (K cat) of modified peptides against the dengue virus type 2 (DENV). The aug-MIA-QSARcolor method demonstrates superior predictive power compared to the traditional MIA-QSAR and aug-MIA-QSAR models, respectively, suggesting that the incorporation of color schemes 172to discriminate the different atom types in chemical structures and the modification of atomic sizes to achieve proportionality to the van der Waals radius improve the predictive ability of the MIA-QSAR method. Structural interpretation of the aug-MIA-QSARcolor models reveals that small, unbranched amino acids in the A1 and A4 positions are associated with greater substrate binding affinity (K m) and a higher substrate cleavage rate (K cat) when present in the positions A1 and A2, while acidic amino acids in the positions A2 and A4 are related with low substrate affinity and low substrate cleavage rates when in the A1 position. These findings provide relevant chemical information on the structural characteristics critical in the planning of novel target compounds against the DENV in a quicker and inexpensive way.
