Ligand and structure-based discovery of phosphorus-containing compounds as potential metalloproteinase inhibitors.

In this study, a methodology is proposed, combining ligand- and structure-based virtual screening tools, for the identification of phosphorus-containing compounds as inhibitors of zinc metalloproteases. First, we use Dragon molecular descriptors to develop a Linear Discriminant Analysis classification model, which is widely validated according to the OECD principles. This model is simple, robust, stable and has good discriminating power. Furthermore, it has a defined applicability domain and it is used for virtual screening of the DrugBank database. Second, docking experiments are carried out on the identified compounds that showed good binding energies to the enzyme thermolysin. Considering the potential toxicity of phosphorus-containing compounds, their toxicological profile is evaluated according to Protox II. Of the five molecules evaluated, two show carcinogenic and mutagenic potential at small LD50, not recommended as drugs, while three of them are classified as non-toxic, and could constitute a starting point for the development of new vasoactive metalloprotease inhibitor drugs. According to molecular dynamics simulation, two of them show stable interactions with the active site maintaining coordination with the metal. A high agreement is evident between QSAR, docking and molecular dynamics results, demonstrating the potentialities of the combination of these tools.

Ligand-based discovery of new potential acetylcholinesterase inhibitors for Alzheimer's disease treatment.

The enzyme acetylcholinesterase (AChE) is currently a therapeutic target for the treatment of neurodegenerative diseases. These diseases have highly variable causes but irreversible evolutions. Although the treatments are palliative, they help relieve symptoms and allow a better quality of life, so the search for new therapeutic alternatives is the focus of many scientists worldwide. In this study, a QSAR-SVM classification model was developed by using the MATLAB numerical computation system and the molecular descriptors implemented in the Dragon software. The obtained parameters are adequate with accuracy of 88.63% for training set, 81.13% for cross-validation experiment and 81.15% for prediction set. In addition, its application domain was determined to guarantee the reliability of the predictions. Finally, the model was used to predict AChE inhibition by a group of quinazolinones and benzothiadiazine 1,1-dioxides obtained by chemical synthesis, resulting in 14 drug candidates with in silico activity comparable to acetylcholine.

Computational identification of chemical compounds with potential anti-Chagas activity using a classification tree.

Chagas disease is endemic to 21 Latin American countries and is a great public health problem in that region. Current chemotherapy remains unsatisfactory; consequently the need to search for new drugs persists. Here we present a new approach to identify novel compounds with potential anti-chagasic action. A large dataset of 584 compounds, obtained from the Drugs for Neglected Diseases initiative, was selected to develop the computational model. Dragon software was used to calculate the molecular descriptors and WEKA software to obtain the classification tree. The best model shows accuracy greater than 93.4% for the training set; the tree was also validated using a 10-fold cross-validation procedure and through a test set, achieving accuracy values over 90.5% and 92.2%, correspondingly. The values of sensitivity and specificity were around 90% in all series; also the false alarm rate values were under 10.5% for all sets. In addition, a simulated ligand-based virtual screening for several compounds recently reported as promising anti-chagasic agents was carried out, yielding good agreement between predictions and experimental results. Finally, the present work constitutes an example of how this rational computer-based method can help reduce the cost and increase the rate in which novel compounds are developed against Chagas disease.

Thorough evaluation of OECD principles in modelling of 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine derivatives using QSARINS.

The human immunodeficiency virus is a lethal pathology considered as a worldwide problem. The search for new strategies for the treatment of this disease continues to be a great challenge in the scientific community. In this study, a series of 107 derivatives of 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine, previously evaluated experimentally against HIV-I reverse transcriptase, was used to model antiretroviral activity. A model of linear regression, implemented in the QSARINS software, was developed with a genetic algorithm for variable selection. The fit of its parameters was good and exhaustive validation, according to the OECD regulatory principles, was performed. Also, the applicability domain was established. In addition, its robustness (r 2 = 0.84), stability (Q 2 LOO = 0.81; Q 2 LMO = 0.80) and good predictive power (r 2 EXT = 0.85) is proved. So, it was used to predict the antiretroviral activity of eight compounds obtained by rational drug design. Finally, it can be affirmed that the proposed tools allow the rapid and economic identification of potential antiretroviral drugs.

Machine learning-based models to predict modes of toxic action of phenols to Tetrahymena pyriformis.

The phenols are structurally heterogeneous pollutants and they present a variety of modes of toxic action (MOA), including polar narcotics, weak acid respiratory uncouplers, pro-electrophiles, and soft electrophiles. Because it is often difficult to determine correctly the mechanism of action of a compound, quantitative structure-activity relationship (QSAR) methods, which have proved their interest in toxicity prediction, can be used. In this work, several QSAR models for the prediction of MOA of 221 phenols to the ciliated protozoan Tetrahymena pyriformis, using Chemistry Development Kit descriptors, are reported. Four machine learning techniques (ML), k-nearest neighbours, support vector machine, classification trees, and artificial neural networks, have been used to develop several models with higher accuracies and predictive capabilities for distinguishing between four MOAs. They showed global accuracy values between 95.9% and 97.7% and area under Receiver Operator Curve values between 0.978 and 0.998; additionally, false alarm rate values were below 8.2% for training set. In order to validate our models, cross-validation (10-folds-out) and external test-set were performed with good behaviour in all cases. These models, obtained with ML techniques, were compared with others previously reported by other researchers, and the improvement was significant.

QuBiLs-MAS method in early drug discovery and rational drug identification of antifungal agents.

The QuBiLs-MAS approach is used for the in silico modelling of the antifungal activity of organic molecules. To this effect, non-stochastic (NS) and simple-stochastic (SS) atom-based quadratic indices are used to codify chemical information for a comprehensive dataset of 2478 compounds having a great structural variability, with 1087 of them being antifungal agents, covering the broadest antifungal mechanisms of action known so far. The NS and SS index-based antifungal activity classification models obtained using linear discriminant analysis (LDA) yield correct classification percentages of 90.73% and 92.47%, respectively, for the training set. Additionally, these models are able to correctly classify 92.16% and 87.56% of 706 compounds in an external test set. A comparison of the statistical parameters of the QuBiLs-MAS LDA-based models with those for models reported in the literature reveals comparable to superior performance, although the latter were built over much smaller and less diverse datasets, representing fewer mechanisms of action. It may therefore be inferred that the QuBiLs-MAS method constitutes a valuable tool useful in the design and/or selection of new and broad spectrum agents against life-threatening fungal infections.

Comparative study to predict toxic modes of action of phenols from molecular structures.

Quantitative structure-activity relationship models for the prediction of mode of toxic action (MOA) of 221 phenols to the ciliated protozoan Tetrahymena pyriformis using atom-based quadratic indices are reported. The phenols represent a variety of MOAs including polar narcotics, weak acid respiratory uncouplers, pro-electrophiles and soft electrophiles. Linear discriminant analysis (LDA), and four machine learning techniques (ML), namely k-nearest neighbours (k-NN), support vector machine (SVM), classification trees (CTs) and artificial neural networks (ANNs), have been used to develop several models with higher accuracies and predictive capabilities for distinguishing between four MOAs. Most of them showed global accuracy of over 90%, and false alarm rate values were below 2.9% for the training set. Cross-validation, complementary subsets and external test set were performed, with good behaviour in all cases. Our models compare favourably with other previously published models, and in general the models obtained with ML techniques show better results than those developed with linear techniques. We developed unsupervised and supervised consensus, and these results were better than our ML models, the results of rule-based approach and other ensemble models previously published. This investigation highlights the merits of ML-based techniques as an alternative to other more traditional methods for modelling MOA.

Event-based criteria in GT-STAF information indices: theory, exploratory diversity analysis and QSPR applications.

Versatile event-based approaches for the definition of novel information theory-based indices (IFIs) are presented. An event in this context is the criterion followed in the "discovery" of molecular substructures, which in turn serve as basis for the construction of the generalized incidence and relations frequency matrices, Q and F, respectively. From the resultant F, Shannon's, mutual, conditional and joint entropy-based IFIs are computed. In previous reports, an event named connected subgraphs was presented. The present study is an extension of this notion, in which we introduce other events, namely: terminal paths, vertex path incidence, quantum subgraphs, walks of length k, Sach's subgraphs, MACCs, E-state and substructure fingerprints and, finally, Ghose and Crippen atom-types for hydrophobicity and refractivity. Moreover, we define magnitude-based IFIs, introducing the use of the magnitude criterion in the definition of mutual, conditional and joint entropy-based IFIs. We also discuss the use of information-theoretic parameters as a measure of the dissimilarity of codified structural information of molecules. Finally, a comparison of the statistics for QSPR models obtained with the proposed IFIs and DRAGON's molecular descriptors for two physicochemical properties log P and log K of 34 derivatives of 2-furylethylenes demonstrates similar to better predictive ability than the latter.

Improvement of charge-transfer indices for multifunctional amino acids: application to lysozyme.

Valence topological Charge-Transfer (CT) indices are applied to the calculation of pH at the isoelectric point (pI). The model is generalized for molecules with heteroatoms. The ability of the indices for the description of molecular charge distribution is established by comparing them with the pI of 21 amino acids. Linear correlation models are obtained. The CT indices improve multivariable regression equations for pI. The variance decreases by 95%. No superimposition of the corresponding G(k)-J(k) and G(k)(V)-J(k)(V) pairs is observed in most fits, which diminishes the risk of collinearity. The inclusion of heteroatoms in pi-electron system is beneficial for the description of pI, the because of either the role of the additional p orbitals provided by heteroatom or the role of steric factors in pi-electron conjugation. The use of CT/valence CT indices {G(k), J(k), G(k)(V), J(k)(V)} gives limited results for modelling amino acid pI. The inclusion of the number of acidic/basic groups improves the models. The effect is specially noticeable for amino acids with more than two functional groups. The fitting line for 21 amino acids is used to estimate lysozyme pI replacing (1 + Delta n/n(T)) with (M + Delta n)/n(T). The results for lysozyme fragments can estimate the pI of the whole protein with 1-13% error.

Structural, chemical topological, electrotopological and electronic structure hypotheses.

The first important hypothesis in the prediction of properties of synthesized molecules is the structural hypothesis. In the study of drug-receptor interactions, the case where the three-dimensional structure of the receptor is known allows the application of molecular simulation and energy calculations to estimate the binding affinity for a proposed series of compounds. The chemical topological hypothesis permits the description of molecular structures without using concepts such as force or energy. These notions would not be as dominant as supposed since they should be able to be deduced from topology. Although topological descriptors are able to describe specific physicochemical properties, there is no any mechanistic interpretation for topological descriptors, but they can be considered as essential magnitudes, able to describe molecular structure as an alternative and independent approach, since they allow an algebraic description of the structure itself. The hypothesis is based on the ability of topological descriptors for the prediction of virtually any kind of structure-related macroscopic property, and to design directly new compounds showing predetermined properties. It is possible to design new lead drugs ignoring the explicit mechanism of action, but only by topological similarity with other active compounds. This is a basic difference with all the conventional methods based on the use of physical variables, which need to know the specific drug-receptor interaction. The electrotopological hypothesis may be considered a synthesis of both previous hypotheses. Finally, the electronic structure hypothesis is formulated.

Binding of vinyl polymers to anionic model membranes.

The association of poly2-vinylpyridine (P2VPy) and poly4-vinylpyridine (P4VPy) to dimyristoylphosphatidic acid (DMPA) small unilamellar vesicles (SUVs) was studied as a function of pH, ionic strength (I), polymer concentration and temperature using spectrofluorimetry. Poly(vinylpyridine) (PVPy) data were transformed into association isotherms and analyzed in terms of binding and partition models. In the case of polyions, the inclusion of the activity coefficient in both models was essential. Moreover, a relating equation was proposed to compare parameters based on both theoretical approaches. On the basis of the results obtained, a model was developed to analyze polymer adsorption at the surface level, in which the length of the hydrophobic chain and the position of the N atom in the pyridinium ring play an important role. Transition temperature (Tc) for DMPA (ca. 55 degrees C) is decreased between 15 degrees C-19 degrees C in the presence of PVPy. Van't Hoff isochore showed that the binding constant (KA) accounted for average PVPy-DMPA two-dimensional solid and liquid interactions. KA decreased with I in the presence of both polymers, but was more sensitive to I in the case of P2VPy. Likewise, the number of phospholipid heads (N) involved in the binding process decreased with I in the presence of PVPy. The influence of I was more significant on N than on KA.

A new topological index to elucidate apolar hydrocarbons.

A new topological charge-transfer index is applied to elucidate the polar character of hydrocarbons. The dipole moments calculated by algebraic and vector semisums of charge-transfer indices are defined in this work. The combination of the charge-transfer indices allows the estimation of the molecular dipole moments. The ability of the indices for the description of the molecular charge distribution is established by comparing them with the dipole moment of a heterogeneous set of 57 polar and 53 apolar hydrocarbons. Linear and nonlinear correlation models are obtained. The new charge-transfer index, mu(vec), elucidates whether hydrocarbons are apolar, and improves the multivariable nonlinear regression equation for the dipole moment. When comparing with previous results, smaller superpositions of the corresponding pairs Gk, Jk are observed in our fits. This diminishes the risk of colinearity.

Calculation of partition coefficient and hydrophobic moment of the secondary structure of lysozyme.

A method that permits a semiquantitative estimate of the partitioning of any solute between any two media is presented. As an example, the partition coefficients and hydrophobic moment of the secondary structure of lysozyme are calculated. Program GSCAP is written as a version of Pascal's solvent-dependent conformational analysis (SCAP) program. The dipole moments calculated for the helices are trebled with respect to that for the sheet. For helices, the main contribution to the water-accessible surface area is the hydrophobic term, while the hydrophilic part dominates in the sheet. Molecular globularity and the three studied partition coefficients differentiate between helices and sheet.

Universal organic solvent-water partition coefficient model

A method that permits a semiquantitative estimate of the partitioning of any solute between any two media is presented. As an example, the organic solvent-water partition coefficients P are calculated. Program GSCAP is written as a version of Pascal's SCAP program. The only needed parameters are the dielectric constant and molecular volume of the organic solvent. The log P results are compared with the Pomona database. The average absolute deviation is 1.48 log units and the standard deviation is 1.66 log units.

Characterizing cavities in model inclusion molecules: a comparative study.

We have selected fullerene-60 and -70 cavities as model systems in order to test several methods for characterizing inclusion molecules. The methods are based on different technical foundations such as a square and triangular tessellation of the molecule taken as a unitary sphere, spherical tessellation of the molecular surface, numerical integration of the atomic volumes and surfaces, triangular tessellation of the molecular surface, and a cubic lattice approach to a molecular space. Accurate measures of the molecular volume and surface area have been performed with the pseudo-random Monte Carlo (MCVS) and uniform Monte Carlo (UMCVS) methods. These calculations serve as a reference for the rest of the methods. The SURMO2 and MS methods have not recognized the cavities and may not be convenient for intercalation compounds. The programs that have detected the cavities never exceed 5% deviation relative to the reference values for molecular volume and surface area. The GEPOL algorithm, alone or combined with TOPO, shows results in good agreement with those of the UMCVS reference. The uniform random number generator provides the fastest convergence for UMCVS and a correct estimate of the standard deviations. The effect of the internal cavity on the accessible surfaces has been calculated.

Torsional effects on the molecular polarizabilities of the benzothiazole (A)-benzobisthiazole (B) oligomer A-B13-A.

We outline a method for the calculation of multipole moments and molecular dipole-dipole (alpha), dipole-quadrupole (A), and quadrupole-quadrupole (C) polarizabilities, which we have successfully applied to benzothiazole (A)-benzobisthiazole (B) oligomer A-B13-A. Three model rotational isomers have been characterized: (1) the fully planar (000) rotational isomer; (2) a conformation with each unit rotated 10 degrees in the alternate direction (+(-)+), and (3) a rotational isomer with each unit rotated 10 degrees in the same direction (+3). The dipole moment, mu, is smaller for isomers 000 and +(-)+ than for isomer +3. The calculation of alpha A, and C has been performed by use of the interacting induced dipoles polarization model, which calculates tensor effective anisotropic point polarizabilities (method of Applequist). The values of alpha, A, and C are in the same order of magnitude as reference calculations (PAPID) program). The values of A are rather sensitive to mu, which varies under rotation, explaining the greatest value of magnitude of Ax,xx for polar isomer +3. This rotational isomer has the maximum hydrophilic accessible surface, which would improve solubility in water. It is found that small torsional changes can enhance solubility by increasing the hydrophilic accessible surface without too much affecting the values of alpha and C. However, the torsion of the oligomer can vary the value of mu and so modify A.

Improved AMYR program: an algorithm for the theoretical simulation of molecular associations, including geometrical and topological characterization of the dimers.

Program AMYR, originally written by S. Fraga (University of Alberta, Canada), allows for the calculation of molecular associations using a pair-wise atom-atom potential. The interaction energy is evaluated through a 1/R expansion. Our improved version includes a dispersion energy term in the potential corrected by damping functions, the possibility of carrying out energy minimizations through variable metric methods, as well as the optional calculation of geometrical and topological indices. Program AMYR has been adapted also for high-performance computing and vectorization. An interactive version of the program carries out real-time molecular graphics showing simultaneously the energy profile of the calculations.