The optimal selection of chemical features (molecular descriptors) is an essential pre-processing step for the efficient application of computational intelligence techniques in virtual screening for identification of bioactive molecules in drug discovery. The selection of molecular descriptors has key influence in the accuracy of affinity prediction. In order to improve this prediction, we examined a Random Forest (RF)-based approach to automatically select molecular descriptors of training data for ligands of kinases, nuclear hormone receptors, and other enzymes. The reduction of features to use during prediction dramatically reduces the computing time over existing approaches and consequently permits the exploration of much larger sets of experimental data. To test the validity of the method, we compared the results of our approach with the ones obtained using manual feature selection in our previous study (Perez-Sanchez et al., 2014).The main novelty of this work in the field of drug discovery is the use of RF in two different ways: feature ranking and dimensionality reduction, and classification * Corresponding author: Tel.: +34 610488989; fax: +34 965903681Email addresses: gcano@dtic.ua.es (Gaspar Cano), jgr@ua.es (Jose Garcia-Rodriguez), agarcia@dtic.ua.es (Alberto Garcia-Garcia), hperez@ucam.edu (Horacio Perez-Sanchez), benedikt@hi.is (Jón Atli Benediktsson), anilth@hi.is (Anil Thapa), A.Barr1@westminster.ac.uk (Alastair Barr)

Abstract. Virtual Screening (VS) methods can considerably aid clinical research, predicting how ligands interact with drug targets. Most VS methods suppose a unique binding site for the target, but it has been demonstrated that diverse ligands interact with unrelated parts of the target and many VS methods do not take into account this relevant fact. This problem is circumvented by a novel VS methodology named BINDSURF that scans the whole protein surface in order to find new hotspots, where ligands might potentially interact with, and which is implemented in last generation massively parallel GPU hardware, allowing fast processing of large ligand databases. BINDSURF can thus be used in drug discovery, drug design, drug repurposing and therefore helps considerably in clinical research. However, the accuracy of most VS methods and concretely BINDSURF is constrained by limitations in the scoring function that describes biomolecular interactions, and even nowadays these uncertainties are not completely understood. In order to improve accuracy of the scoring functions used in BINDSURF we propose a hybrid novel approach where neural networks (NNET) and support vector machines (SVM) methods are trained with databases of known active (drugs) and inactive compounds, being this information exploited afterwards to improve BINDSURF VS predictions.

ABSTRACT:Virtual Screening (VS) methods can considerably aid clinical research, predicting how ligands interact with drug targets. However, the accuracy of most VS methods is constrained by limitations in the scoring function that describes biomolecular interactions, and even nowadays these uncertainties are not completely understood. In order to improve accuracy of scoring functions used in most VS methods we propose a hybrid novel approach where neural networks (NNET) and support vector machines (SVM) methods are trained with databases of known active (drugs) and inactive compounds, being this information exploited afterwards to improve VS predictions.

Adenosine, a widespread and endogenous nucleoside that acts as a powerful neuromodulator in the nervous system, is a promising therapeutic target in a wide range of conditions. The structural similarity between xanthine derivatives and neurotransmitter adenosine has led to the derivatives of the heterocyclic ring being among the most abundant chemical classes of ligand antagonists of adenosine receptor subtypes. Small changes in the xanthine scaffold have resulted in a wide array of adenosine receptor antagonists. In this work, we developed a QSAR model for the [Formula: see text] subtype, which is, as yet, not well characterized, with two purposes in mind: to predict adenosine [Formula: see text] antagonist activity and to offer a substructural interpretation of this group of xanthines. The QSAR model provided good classifications of both the test and external sets. In addition, most of the contributions to adenosine [Formula: see text] receptor affinity derived by subfragmentation of the molecules in the training set agree with the relationships observed in the literature. These two factors mean that this QSAR ensemble could be used as a model to predict future adenosine [Formula: see text] antagonist candidates.

No abstract