Identification of Electronic and Structural Descriptors of Adenosine Analogues Related to Inhibition of Leishmanial Glyceraldehyde-3-Phosphate Dehydrogenase

Abstract

Quantitative structure–activity relationship (QSAR) studies were performed in order to identify molecular features responsible for the antileishmanial activity of 61 adenosine analogues acting as inhibitors of the enzyme glyceraldehyde 3-phosphate dehydrogenase of Leishmania mexicana (LmGAPDH). Density functional theory (DFT) was employed to calculate quantum-chemical descriptors, while several structural descriptors were generated with Dragon 5.4. Variable selection was undertaken with the ordered predictor selection (OPS) algorithm, which provided a set with the most relevant descriptors to perform PLS, PCR and MLR regressions. Reliable and predictive models were obtained, as attested by their high correlation coefficients, as well as the agreement between predicted and experimental values for an external test set. Additional validation procedures were carried out, demonstrating that robust models were developed, providing helpful tools for the optimization of the antileishmanial activity of adenosine compounds.

Pattern Recognition Techniques Applied to the Study of Leishmanial Glyceraldehyde-3-Phosphate Dehydrogenase Inhibition

“…The descriptor that provides the greatest modeling power is Polarizability, which means that this is the variable that best describes the data set for the two classes observed in the training set. It is interesting to note that the results obtained in this work are in agreement with previous studies which have indicated the importance of steric and electrostatic properties in explaining the affinity of adenosines to Lm GAPDH [ 21 , 22 ].…”

“…The aim of this study, involving adenosine derivatives, their affinities to Lm GAPDH and pattern recognition techniques, is to understand the fundamental effects involved in the interaction between the bioactive ligands and the biological target. The computational procedure employed here has enabled discrimination of the studied compounds, with higher (Class 1) and lower (Class 2) affinities to Lm GAPDH, through molecular descriptors obtained by quantum chemical calculations (E LUMO , QR 2 , QR 4 , Volume and Polarizability), differently from previous studies where more complex calculations were required [ 21 ] or only topological descriptors were able to provide statistically validated QSAR models [ 22 ]. All pattern recognition models obtained in the present work have shown internal consistency and were externally validated with a set of test compounds.…”

A structure–activity relationship study of the toxicity of ionic liquids using an adapted Ferreira–Kiralj hydrophobicity parameter

“…27 The final reduction of variables was carried out in this program. [29][30][31] This method uses partial least squares (PLS), a regression method which reduces the size of the data by transforming them into mutually orthogonal latent variables (LVs), 31,32 to build models by rearranging the columns of the matrix in such a way that the most important descriptors, classified according to an informative vector (correlation vector, regression vector, and their product), are placed in the first column. [29][30][31] This method uses partial least squares (PLS), a regression method which reduces the size of the data by transforming them into mutually orthogonal latent variables (LVs), 31,32 to build models by rearranging the columns of the matrix in such a way that the most important descriptors, classified according to an informative vector (correlation vector, regression vector, and their product), are placed in the first column.…”

“…Matrices of descriptors were subjected to the method of selection of variables called ordered predictors selection (OPS), 28 an iterative algorithm for building QSAR models. [29][30][31] This method uses partial least squares (PLS), a regression method which reduces the size of the data by transforming them into mutually orthogonal latent variables (LVs), 31,32 to build models by rearranging the columns of the matrix in such a way that the most important descriptors, classified according to an informative vector (correlation vector, regression vector, and their product), are placed in the first column. In this study, the three vectors were used simultaneously.…”

Exploring the Potential of Spherical Harmonics and PCVM for Compounds Activity Prediction

“…For instance, classical QSAR (Quantitative Structure-Activity Relationships) methodologies [19] have given their contribution [20,21,22,23]. Lozano et al identified molecular features responsible for the antileishmanial activity of 61 adenosine analogues acting as inhibitors of the enzyme glyceraldehyde 3-phosphate dehydrogenase of Leishmania mexicana (LmGAPDH) [24]. Adeniji et al made a great effort to develop a model that relates the structures of 50 compounds to their activities against M. tuberculosis [25].…”