Flemming Steen Jørgensen

hERG Classification Model Based on a Combination of Support Vector Machine Method and GRIND Descriptors

The human Ether-a-go-go Related Gene (hERG) potassium channel is one of the major critical factors associated with QT interval prolongation and development of arrhythmia called Torsades de Pointes (TdP). It has become a growing concern of both regulatory agencies and pharmaceutical industries who invest substantial effort in the assessment of cardiac toxicity of drugs. The development of in silico tools to filter out potential hERG channel inhibitors in early stages of the drug discovery process is of considerable interest. Here, we describe binary classification models based on a large and diverse library of 495 compounds. The models combine pharmacophore-based GRIND descriptors with a support vector machine (SVM) classifier in order to discriminate between hERG blockers and nonblockers. Our models were applied at different thresholds from 1 to 40 microm and achieved an overall accuracy up to 94% with a Matthews coefficient correlation (MCC) of 0.86 ( F-measure of 0.90 for blockers and 0.95 for nonblockers). The model at a 40 microm threshold showed the best performance and was validated internally (MCC of 0.40 and F-measure of 0.57 for blockers and 0.81 for nonblockers, using a leave-one-out cross-validation). On an external set of 66 compounds, 72% of the set was correctly predicted ( F-measure of 0.86 and 0.34 for blockers and nonblockers, respectively). Finally, the model was also tested on a large set of hERG bioassay data recently made publicly available on PubChem ( http://pubchem.ncbi.nlm.nih.gov/assay/assay.cgi?aid=376) to achieve about 73% accuracy ( F-measure of 0.30 and 0.83 for blockers and nonblockers, respectively). Even if there is still some limitation in the assessment of hERG blockers, the performance of our model shows an improvement between 10% and 20% in the prediction of blockers compared to other methods, which can be useful in the filtering of potential hERG channel inhibitors.

Prediction methods and databases within chemoinformatics: emphasis on drugs and drug candidates

MOTIVATION To gather information about available databases and chemoinformatics methods for prediction of properties relevant to the drug discovery and optimization process. RESULTS We present an overview of the most important databases with 2-dimensional and 3-dimensional structural information about drugs and drug candidates, and of databases with relevant properties. Access to experimental data and numerical methods for selecting and utilizing these data is crucial for developing accurate predictive in silico models. Many interesting predictive methods for classifying the suitability of chemical compounds as potential drugs, as well as for predicting their physico-chemical and ADMET properties have been proposed in recent years. These methods are discussed, and some possible future directions in this rapidly developing field are described.

Mutational Mapping and Modeling of the Binding Site for (S)-Citalopram in the Human Serotonin Transporter

The serotonin transporter (SERT) regulates extracellular levels of the neurotransmitter serotonin (5-hydroxytryptamine) in the brain by facilitating uptake of released 5-hydroxytryptamine into neuronal cells. SERT is the target for widely used antidepressant drugs, including imipramine, fluoxetine, and (S)-citalopram, which are competitive inhibitors of the transport function. Knowledge of the molecular details of the antidepressant binding sites in SERT has been limited due to lack of structural data on SERT. Here, we present a characterization of the (S)-citalopram binding pocket in human SERT (hSERT) using mutational and computational approaches. Comparative modeling and ligand docking reveal that (S)-citalopram fits into the hSERT substrate binding pocket, where (S)-citalopram can adopt a number of different binding orientations. We find, however, that only one of these binding modes is functionally relevant from studying the effects of 64 point mutations around the putative substrate binding site. The mutational mapping also identify novel hSERT residues that are crucial for (S)-citalopram binding. The model defines the molecular determinants for (S)-citalopram binding to hSERT and demonstrates that the antidepressant binding site overlaps with the substrate binding site.

Classification of Cytochrome P450 1A2 Inhibitors and Noninhibitors by Machine Learning Techniques

The cytochrome P450 (P450) superfamily plays an important role in the metabolism of drug compounds, and it is therefore highly desirable to have models that can predict whether a compound interacts with a specific isoform of the P450s. In this work, we provide in silico models for classification of CYP1A2 inhibitors and noninhibitors. Training and test sets consisted of approximately 400 and 7000 compounds, respectively. Various machine learning techniques, such as binary quantitative structure activity relationship, support vector machine (SVM), random forest, kappa nearest neighbor (kNN), and decision tree methods were used to develop in silico models, based on Volsurf and Molecular Operating Environment descriptors. The best models were obtained using the SVM, random forest, and kNN methods in combination with the BestFirst variable selection method, resulting in models with 73 to 76% of accuracy on the test set prediction (Matthews correlation coefficients of 0.51 and 0.52). Finally, a decision tree model based on Lipinskis Rule-of-Five descriptors was also developed. This model predicts 67% of the compounds correctly and gives a simple and interesting insight into the issue of classification. All of the models developed in this work are fast and precise enough to be applicable for virtual screening of CYP1A2 inhibitors or noninhibitors or can be used as simple filters in the drug discovery process.

Prediction of pH-dependent aqueous solubility of druglike molecules

In the present work, the Henderson-Hasselbalch (HH) equation has been employed for the development of a tool for the prediction of pH-dependent aqueous solubility of drugs and drug candidates. A new prediction method for the intrinsic solubility was developed, based on artificial neural networks that have been trained on a druglike PHYSPROP subset of 4548 compounds. For the prediction of acid/base dissociation coefficients, the commercial tool Marvin has been used, following validation on a data set of 467 molecules from the PHYSPROP database. The best performing network for intrinsic solubility predictions has a cross-validated root mean square error (RMSE) of 0.70 log S-units, while the Marvin pKa plug-in has an RMSE of 0.71 pH-units. A data set of 27 drugs with experimentally determined pH-solubility curves was assembled from the literature for the validation of the combined pH-dependent model, giving a mean RMSE of 0.79 log S-units. Finally, the combined model has been applied on profiling the solubility space at low pH of five large vendor libraries.

Virtual screening and prediction of site of metabolism for cytochrome P450 1A2 ligands.

With the availability of an increasing number of high resolution 3D structures of human cytochrome P450 enzymes, structure-based modeling tools are more readily used. In this study we explore the possibilities of using docking and scoring experiments on cytochrome P450 1A2. Three different questions have been addressed: 1. Binding orientations and conformations were successfully predicted for various substrates. 2. A virtual screen was performed with satisfying enrichment rates. 3. A classification of individual compounds into active and inactive was performed. It was found that while docking can be used successfully to address the first two questions, it seems to be more difficult to perform the classification. Different scoring functions were included, and the well-characterized water molecule in the active site was included in various ways. Results are compared to experimental data and earlier classification data using machine learning methods. The possibilities and limitations of using structure-based drug design tools for cytochrome P450 1A2 come to light and are discussed.

A projector formulation for the Van Vleck transformation: I. Degenerate case

The Van Vleck transformation as previously described is generalized so that it can isolate several levels simultaneously. It is shown that the contact transformation involves projectors in an implicit way and that it can be considered as a special case of the generalized Van Vleck transformation. We discuss to what extent the recursion relations (for further transformation of the Nth transform) given in part I can be carried over to the generalized Van Vleck transformation and to the contact transformation. An effective hamiltonian for a particular ‘cluster’ of zeroth-order levels can be obtained in several ways. We compare the results of the usual Van Vleck transformation, Soliverezs transformation and the contact transformation.

Human peptide transporters: therapeutic applications

Peptide transporters are epithelial solute carriers. Their functional role has been characterised in the small intestine and proximal tubules, where they are involved in absorption of dietary peptides and peptide reabsorption, respectively. Currently, two peptide transporters, PepT1 and PepT2, which possess transport activity, have been identified. The transporters are not drug targets per se, but due to uniquely broad substrate specificity they have proven to be relevant in drug therapy at the level of drug transport. Therapeutic agents such as orally active β-lactam antibiotics, bestatin, prodrugs of acyclovir and gancyclovir have oral bioavailabilities, which are largely a result of their interaction with PepT1. The transporters have therefore received considerable attention in relation to drug delivery. The aim of the present review is to highlight structural requirements for binding to peptide transporters, as well as their role in drug delivery and in potential future drug design and targeted tissue delivery of peptides and peptidomimetics.

Towards an understanding of drug resistance in malaria: Three-dimensional structure of Plasmodium falciparum dihydrofolate reductase by homology building

A three-dimensional (3-D) model of dihydrofolate reductase (DHFR) from Plasmodium falciparum has been constructed by homology building. The model building has been based on a structural alignment of five X-ray structures of DHFR from different species. The 3-D model of the plasmodial DHFR was obtained by amino acid substitution in the human DHFR, which was chosen as template, modification of four loops (two insertions, two deletions) and subsequent energy minimization. The active site of P. falciparum DHFR was analyzed and compared to human DHFR with respect to sequence variations and structural differences. Based on this analysis the molecular consequences of point mutations known to be involved in drug resistance were discussed. The significance of the most important point mutation causing resistance, S108N, could be explained by the model, whereas the point mutations associated with enhanced resistance, N51I and C59R, seem to have a more indirect effect on inhibitor binding.

Prediction of cytochrome P450 mediated metabolism.

Cytochrome P450 enzymes (CYPs) form one of the most important enzyme families involved in the metabolism of xenobiotics. CYPs comprise many isoforms, which catalyze a wide variety of reactions, and potentially, a large number of different metabolites can be formed. However, it is often hard to rationalize what metabolites these enzymes generate. In recent years, many different in silico approaches have been developed to predict binding or regioselective product formation for the different CYP isoforms. These comprise ligand-based methods that are trained on experimental CYP data and structure-based methods that consider how the substrate is oriented in the active site or/and how reactive the part of the substrate that is accessible to the heme group is. We will review key aspects for various approaches that are available to predict binding and site of metabolism (SOM), what outcome can be expected from the predictions, and how they could potentially be improved.