Facundo Pérez-Giménez

Discrimination and selection of new potential antibacterial compounds using simple topological descriptors

The aim of the work was to discriminate between antibacterial and non-antibacterial drugs by topological methods and to select new potential antibacterial agents from among new structures. The method used for antibacterial activity selection was a linear discriminant analysis (LDA). It is possible to obtain a QSAR interpretation of the information contained in the discriminant function. We make use of the pharmacological distribution diagrams (PDDs) as a visualizing technique for the identification and selection of new antibacterial agents.

Artificial neural networks and linear discriminant analysis: a valuable combination in the selection of new antibacterial compounds.

A set of topological descriptors has been used to discriminate between antibacterial and nonantibacterial drugs. Topological descriptors are simple integers calculated from the molecular structure represented in SMILES format. The methods used for antibacterial activity discrimination were linear discriminant analysis (LDA) and artificial neural networks of a multilayer perceptron (MLP) type. The following plot frequency distribution diagrams were used: a function of the number of drugs within a value interval of the discriminant function and the output value of the neural network versus these values. Pharmacological distribution diagrams (PDD) were used as a visualizing technique for the identification of antibacterial agents. The results confirmed the discriminative capacity of the topological descriptors proposed. The combined use of LDA and MLP in the guided search and the selection of new structures with theoretical antibacterial activity proved highly effective, as shown by the in vitro activity and toxicity assays conducted.

Drugs and nondrugs: an effective discrimination with topological methods and artificial neural networks.

A set of topological and structural descriptors has been used to discriminate general pharmacological activity. To that end, we selected a group of molecules with proven pharmacological activity including different therapeutic categories, and another molecule group without any activity. As a method for pharmacological activity discrimination, an artificial neural network was used, dividing molecules into active and inactive, to train the network and externally validate it. The following plot frequency distribution diagrams were used: a function of the number of drugs within a value interval, and the output value of the neural network versus these values. Pharmacological distribution diagrams (PDD) were used as a visualizing technique for the identification of drug and nondrug molecules. The results confirmed the discriminative capacity of the topological descriptors proposed.

Artificial neural network applied to the discrimination of antibacterial activity by topological methods

Abstract A new topological method that makes it possible to discriminate the active and inactive molecules on the basis of their chemical structures is applied in the present study to the antibacterial agents. This method uses neural networks in which training algorithms are used as well as different concepts and methods of artificial intelligence with a suitable set of topological descriptors. It is possible to obtain a QSAR interpretation of the information contained in the network after the training has been carried out.

Atom- and Bond-Based 2D TOMOCOMD-CARDD Approach and Ligand-Based Virtual Screening for the Drug Discovery of New Tyrosinase Inhibitors

Two-dimensional atom- and bond-based TOMOCOMD-CARDD descriptors and linear discriminant analysis (LDA) are used in this report to perform a quantitative structure-activity relationship (QSAR) study of tyrosinase-inhibitory activity. A database of inhibitors of the enzyme is collected for this study, within 246 highly dissimilar molecules presenting antityrosinase activity. In total, 7 discriminant functions are obtained by using the whole set of atom- and bond-based 2D indices. All the LDA-based QSAR models show accuracies above 90% in the training set and values of the Matthews correlation coefficient (C) varying from 0.85 to 0.90. The external validation set shows globally good classifications between 89% and 91% and C values ranging from 0.75 to 0.81. Finally, QSAR models are used in the selection/identification of the 20 new dicoumarins subset to search for tyrosinase inhibitory activity. Theoretical and experimental results show good correspondence between one another. It is important to remark that most compounds in this series exhibit a more potent inhibitory activity against the mushroom tyrosinase enzyme than the reference compound, Kojic acid (IC50 = 16.67 μM), resulting in a novel nucleus base (lead) with antityrosinase activity, and this could serve as a starting point for the drug discovery of novel tyrosinase inhibitor lead compounds. ( Journal of Biomolecular Screening 2008:1014-1024)

Computational discovery of novel trypanosomicidal drug-like chemicals by using bond-based non-stochastic and stochastic quadratic maps and linear discriminant analysis

Herein we present results of a quantitative structure-activity relationship (QSAR) studies to classify and design, in a rational way, new antitrypanosomal compounds by using non-stochastic and stochastic bond-based quadratic indices. A data set of 440 organic chemicals, 143 with antitrypanosomal activity and 297 having other clinical uses, is used to develop QSAR models based on linear discriminant analysis (LDA). Non-stochastic model correctly classifies more than 93% and 95% of chemicals in both training and external prediction groups, respectively. On the other hand, the stochastic model shows an accuracy of about the 87% for both series. As an experiment of virtual lead generation, the present approach is finally satisfactorily applied to the virtual evaluation of 9 already synthesized in house compounds. The in vitro antitrypanosomal activity of this series against epimastigote forms of Trypanosoma cruzi is assayed. The model is able to predict correctly the behaviour for the majority of these compounds. Four compounds (FER16, FER32, FER33 and FER 132) showed more than 70% of epimastigote inhibition at a concentration of 100 microg/mL (86.74%, 78.12%, 88.85% and 72.10%, respectively) and two of these chemicals, FER16 (78.22% of AE) and FER33 (81.31% of AE), also showed good activity at a concentration of 10 microg/mL. At the same concentration, compound FER16 showed lower value of cytotoxicity (15.44%), and compound FER33 showed very low value of 1.37%. Taking into account all these results, we can say that these three compounds can be optimized in forthcoming works, but we consider that compound FER33 is the best candidate. Even though none of them resulted more active than Nifurtimox, the current results constitute a step forward in the search for efficient ways to discover new lead antitrypanosomals.

Prediction of chromatographic properties for a group of natural phenolic derivatives by molecular topology

A study was made of the relationship between the RM values obtained by thin-layer chromatography for a group of phenols and connectivity indices proposed by Kier and Hall. By using multivariate regression the corresponding connectivity functions were obtained, which were selected based on their respective statistical parameters. Regression analysis of the connectivity functions showed a correct prediction of the experimental elution sequence for this group of molecules using silica gel stationary phases and mobile phases of different polarity. Random and stability studies of the different prediction models selected were carried out, and good stability and null randomness were obtained in all cases.

IMMAN: free software for information theory-based chemometric analysis

The features and theoretical background of a new and free computational program for chemometric analysis denominated IMMAN (acronym for Information theory-based CheMoMetrics ANalysis) are presented. This is multi-platform software developed in the Java programming language, designed with a remarkably user-friendly graphical interface for the computation of a collection of information-theoretic functions adapted for rank-based unsupervised and supervised feature selection tasks. A total of 20 feature selection parameters are presented, with the unsupervised and supervised frameworks represented by 10 approaches in each case. Several information-theoretic parameters traditionally used as molecular descriptors (MDs) are adapted for use as unsupervised rank-based feature selection methods. On the other hand, a generalization scheme for the previously defined differential Shannon’s entropy is discussed, as well as the introduction of Jeffreys information measure for supervised feature selection. Moreover, well-known information-theoretic feature selection parameters, such as information gain, gain ratio, and symmetrical uncertainty are incorporated to the IMMAN software (http://mobiosd-hub.com/imman-soft/), following an equal-interval discretization approach. IMMAN offers data pre-processing functionalities, such as missing values processing, dataset partitioning, and browsing. Moreover, single parameter or ensemble (multi-criteria) ranking options are provided. Consequently, this software is suitable for tasks like dimensionality reduction, feature ranking, as well as comparative diversity analysis of data matrices. Simple examples of applications performed with this program are presented. A comparative study between IMMAN and WEKA feature selection tools using the Arcene dataset was performed, demonstrating similar behavior. In addition, it is revealed that the use of IMMAN unsupervised feature selection methods improves the performance of both IMMAN and WEKA supervised algorithms.Graphical abstractGraphic representation for Shannon’s distribution of MD calculating software.

Trends in information theory-based chemical structure codification

This report offers a chronological review of the most relevant applications of information theory in the codification of chemical structure information, through the so-called information indices. Basically, these are derived from the analysis of the statistical patterns of molecular structure representations, which include primitive global chemical formulae, chemical graphs, or matrix representations. Finally, new approaches that attempt to go “back to the roots” of information theory, in order to integrate other information-theoretic measures in chemical structure coding are discussed.

Shannon's, mutual, conditional and joint entropy information indices: generalization of global indices defined from local vertex invariants.

A new mathematical approach is proposed in the definition of molecular descriptors (MDs) based on the application of information theory concepts. This approach stems from a new matrix representation of a molecular graph (G) which is derived from the generalization of an incidence matrix whose row entries correspond to connected subgraphs of a given G, and the calculation of the Shannons entropy, the negentropy and the standardized information content, plus for the first time, the mutual, conditional and joint entropy-based MDs associated with G. We also define strategies that generalize the definition of global or local invariants from atomic contributions (local vertex invariants, LOVIs), introducing related metrics (norms), means and statistical invariants. These invariants are applied to a vector whose components express the atomic information content calculated using the Shannons, mutual, conditional and joint entropybased atomic information indices. The novel information indices (IFIs) are implemented in the program TOMOCOMDCARDD. A principal component analysis reveals that the novel IFIs are capable of capturing structural information not codified by IFIs implemented in the software DRAGON. A comparative study of the different parameters (e.g. subgraph orders and/or types, invariants and class of MDs) used in the definition of these IFIs reveals several interesting results. The mutual entropy-based indices give the best correlation results in modeling of a physicochemical property, namely the partition coefficient of the 34 derivatives of 2-furylethylenes, among the classes of indices investigated in this study. In a comparison with classical MDs it is demonstrated that the new IFIs give good results for various QSPR models.