Classifier Ensemble Based on Feature Selection and Diversity Measures for Predicting the Affinity of A2B Adenosine Receptor Antagonists

A(2B) adenosine receptor antagonists may be beneficial in treating diseases like asthma, diabetes, diabetic retinopathy, and certain cancers. This has stimulated research for the development of potent ligands for this subtype, based on quantitative structure-affinity relationships. In this work, a new ensemble machine learning algorithm is proposed for classification and prediction of the ligand-binding affinity of A(2B) adenosine receptor antagonists. This algorithm is based on the training of different classifier models with multiple training sets (composed of the same compounds but represented by diverse features). The k-nearest neighbor, decision trees, neural networks, and support vector machines were used as single classifiers. To select the base classifiers for combining into the ensemble, several diversity measures were employed. The final multiclassifier prediction results were computed from the output obtained by using a combination of selected base classifiers output, by utilizing different mathematical functions including the following: majority vote, maximum and average probability. In this work, 10-fold cross- and external validation were used. The strategy led to the following results: i) the single classifiers, together with previous features selections, resulted in good overall accuracy, ii) a comparison between single classifiers, and their combinations in the multiclassifier model, showed that using our ensemble gave a better performance than the single classifier model, and iii) our multiclassifier model performed better than the most widely used multiclassifier models in the literature. The results and statistical analysis demonstrated the supremacy of our multiclassifier approach for predicting the affinity of A(2B) adenosine receptor antagonists, and it can be used to develop other QSAR models.

[1]  Giampiero Spalluto,et al.  Progress in the pursuit of therapeutic adenosine receptor antagonists , 2006, Medicinal research reviews.

[2]  Ludmila I. Kuncheva,et al.  Measures of Diversity in Classifier Ensembles and Their Relationship with the Ensemble Accuracy , 2003, Machine Learning.

[3]  Roberto Todeschini,et al.  Structure/Response Correlations and Similarity/Diversity Analysis by GETAWAY Descriptors, 1. Theory of the Novel 3D Molecular Descriptors , 2002, J. Chem. Inf. Comput. Sci..

[4]  Nikolaos M. Avouris,et al.  EVALUATION OF CLASSIFIERS FOR AN UNEVEN CLASS DISTRIBUTION PROBLEM , 2006, Appl. Artif. Intell..

[5]  Aliuska Duardo-Sanchez,et al.  From QSAR models of drugs to complex networks: state-of-art review and introduction of new Markov-spectral moments indices. , 2012, Current topics in medicinal chemistry.

[6]  Ludmila I. Kuncheva,et al.  Combining Pattern Classifiers: Methods and Algorithms , 2004 .

[7]  Ting Chen,et al.  Ensemble Feature Selection: Consistent Descriptor Subsets for Multiple QSAR Models , 2007, J. Chem. Inf. Model..

[8]  Juan José Rodríguez Diez,et al.  Rotation Forest: A New Classifier Ensemble Method , 2006, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[9]  K. Varani,et al.  Recent developments in the field of A2A and A3 adenosine receptor antagonists. , 2003, European journal of medicinal chemistry.

[10]  K. Varani,et al.  Design, synthesis, and biological evaluation of new 8-heterocyclic xanthine derivatives as highly potent and selective human A2B adenosine receptor antagonists. , 2004, Journal of medicinal chemistry.

[11]  Alexander Tropsha,et al.  Best Practices for QSAR Model Development, Validation, and Exploitation , 2010, Molecular informatics.

[12]  M. P. González,et al.  Search for new antagonist ligands for adenosine receptors from QSAR point of view. How close are we? , 2008, Medicinal research reviews.

[13]  B. Fredholm,et al.  International Union of Pharmacology. XXV. Nomenclature and classification of adenosine receptors. , 2001, Pharmacological reviews.

[14]  S. Moro,et al.  Fluorosulfonyl- and bis-(beta-chloroethyl)amino-phenylamino functionalized pyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidine derivatives: irreversible antagonists at the human A3 adenosine receptor and molecular modeling studies. , 2001, Journal of medicinal chemistry.

[15]  Manuela Pavan,et al.  A distance measure between models: a tool for similarity/diversity analysis of model populations , 2004 .

[16]  Christian Lemmen,et al.  Using Ensembles to Classify Compounds for Drug Discovery , 2003, J. Chem. Inf. Comput. Sci..

[17]  Ann Nowé,et al.  GA(M)E-QSAR: A Novel, Fully Automatic Genetic-Algorithm-(Meta)-Ensembles Approach for Binary Classification in Ligand-Based Drug Design , 2012, J. Chem. Inf. Model..

[18]  Y. Kurogi,et al.  1,2,4-Triazolo[5,1-i]purine derivatives as highly potent and selective human adenosine A(3) receptor ligands. , 2002, Journal of medicinal chemistry.

[19]  Igor V Tetko,et al.  A comparison of different QSAR approaches to modeling CYP450 1A2 inhibition , 2011, J. Chem. Inf. Model..

[20]  K. Jacobson,et al.  Emerging adenosine receptor agonists , 2007, Expert opinion on emerging drugs.

[21]  Pier Andrea Borea,et al.  Design, synthesis, and biological evaluation of C9- and C2-substituted pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidines as new A2A and A3 adenosine receptors antagonists. , 2003, Journal of medicinal chemistry.

[22]  G. Spalluto,et al.  7-Substituted 5-amino-2-(2-furyl)pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidines as A2A adenosine receptor antagonists: a study on the importance of modifications at the side chain on the activity and solubility. , 2002, Journal of medicinal chemistry.

[23]  Rafael Bello,et al.  ANN-QSAR model for selection of anticancer leads from structurally heterogeneous series of compounds. , 2007, European journal of medicinal chemistry.

[24]  Dimitar Hristozov,et al.  Exploring Potency and Selectivity Receptor Antagonist Profiles Using a Multilabel Classification Approach: The Human Adenosine Receptors as a Key Study , 2009, J. Chem. Inf. Model..

[25]  Paola Gramatica,et al.  Structure/Response Correlations and Similarity/Diversity Analysis by GETAWAY Descriptors, 2. Application of the Novel 3D Molecular Descriptors to QSAR/QSPR Studies , 2002, J. Chem. Inf. Comput. Sci..

[26]  Lourdes Santana,et al.  Medicinal chemistry and bioinformatics--current trends in drugs discovery with networks topological indices. , 2007, Current topics in medicinal chemistry.

[27]  K. Klotz,et al.  Pyrazolo[4,3‐e]1,2,4‐triazolo[1,5‐c]pyrimidine derivatives as adenosine receptor ligands: A starting point for searching A2B adenosine receptor antagonists , 2001 .

[28]  Fernanda Borges,et al.  Combining QSAR classification models for predictive modeling of human monoamine oxidase inhibitors. , 2013, European journal of medicinal chemistry.

[29]  Charles X. Ling,et al.  Using AUC and accuracy in evaluating learning algorithms , 2005, IEEE Transactions on Knowledge and Data Engineering.

[30]  K. Klotz,et al.  Medicinal chemistry and pharmacology of A2B adenosine receptors. , 2003, Current topics in medicinal chemistry.

[31]  Yoav Freund,et al.  A decision-theoretic generalization of on-line learning and an application to boosting , 1997, EuroCOLT.

[32]  F. Sanz,et al.  1-, 3- and 8-substituted-9-deazaxanthines as potent and selective antagonists at the human A2B adenosine receptor. , 2008, Bioorganic & medicinal chemistry.

[33]  K. Varani,et al.  New strategies for the synthesis of A3 adenosine receptor antagonists. , 2003, Bioorganic & medicinal chemistry.

[34]  A. Cavalli,et al.  Novel 1,3-dipropyl-8-(3-benzimidazol-2-yl-methoxy-1-methylpyrazol-5-yl)xanthines as potent and selective A₂B adenosine receptor antagonists. , 2012, Journal of medicinal chemistry.

[35]  S. Gessi,et al.  Adenosine receptor antagonists: translating medicinal chemistry and pharmacology into clinical utility. , 2008, Chemical reviews.

[36]  8-Substituted-9-deazaxanthines as adenosine receptor ligands: design, synthesis and structure-affinity relationships at A2B. , 2004, European journal of medicinal chemistry.

[37]  David J. Hand,et al.  A Simple Generalisation of the Area Under the ROC Curve for Multiple Class Classification Problems , 2001, Machine Learning.

[38]  M. Loza,et al.  1,3-Dialkyl-8-(hetero)aryl-9-OH-9-deazaxanthines as potent A2B adenosine receptor antagonists: design, synthesis, structure-affinity and structure-selectivity relationships. , 2008, Bioorganic & medicinal chemistry.

[39]  Ernesto Estrada,et al.  Spectral Moments of the Edge-Adjacency Matrix of Molecular Graphs, 2. Molecules Containing Heteroatoms and QSAR Applications , 1997, J. Chem. Inf. Comput. Sci..

[40]  Yuqing Song,et al.  Structural Predictions of Adenosine 2B Antagonist Affinity Using Molecular Field Analysis , 2001 .

[41]  Alexander Golbraikh,et al.  Predictive QSAR modeling workflow, model applicability domains, and virtual screening. , 2007, Current pharmaceutical design.

[42]  Alexander Golbraikh,et al.  Combinatorial QSAR of Ambergris Fragrance Compounds , 2004, J. Chem. Inf. Model..

[43]  Ernesto Estrada,et al.  Spectral Moments of the Edge Adjacency Matrix in Molecular Graphs. 3. Molecules Containing Cycles , 1998, J. Chem. Inf. Comput. Sci..

[44]  K. Varani,et al.  Recent developments in the field of A3 adenosine receptor antagonists , 2003 .

[45]  S. Moro,et al.  Pyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidine derivatives as highly potent and selective human A(3) adenosine receptor antagonists: influence of the chain at the N(8) pyrazole nitrogen. , 2000, Journal of medicinal chemistry.

[46]  Barbara Cacciari,et al.  Synthesis, biological activity, and molecular modeling investigation of new pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine derivatives as human A(3) adenosine receptor antagonists. , 2002, Journal of medicinal chemistry.

[47]  R. Quinn,et al.  Adenosine receptors as potential therapeutic targets. , 1999, Drug discovery today.

[48]  Pedro Larrañaga,et al.  A review of feature selection techniques in bioinformatics , 2007, Bioinform..

[49]  Ting Wang,et al.  Boosting: An Ensemble Learning Tool for Compound Classification and QSAR Modeling , 2005, J. Chem. Inf. Model..

[50]  Olivier Taboureau,et al.  Classification of Cytochrome P450 1A2 Inhibitors and Noninhibitors by Machine Learning Techniques , 2009, Drug Metabolism and Disposition.

[51]  Ernesto Estrada,et al.  Spectral Moments of the Edge Adjacency Matrix in Molecular Graphs, 1. Definition and Applications to the Prediction of Physical Properties of Alkanes , 1996, J. Chem. Inf. Comput. Sci..

[52]  S. Moro,et al.  Pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine derivatives as adenosine receptor antagonists. Influence of the N5 substituent on the affinity at the human A 3 and A 2B adenosine receptor subtypes: a molecular modeling investigation. , 2003, Journal of medicinal chemistry.

[53]  Benjamin A. Ellingson,et al.  Conformer Generation with OMEGA: Algorithm and Validation Using High Quality Structures from the Protein Databank and Cambridge Structural Database , 2010, J. Chem. Inf. Model..

[54]  Alexander Golbraikh,et al.  Combinatorial QSAR Modeling of P-Glycoprotein Substrates , 2006, J. Chem. Inf. Model..