Virtual Screening Experiments Research Articles

The influence of negative training set size on machine learning-based virtual screening.

BackgroundThe paper presents a thorough analysis of the influence of the number of negative training examples on the performance of machine learning methods.ResultsThe impact of this rather neglected aspect of machine learning methods application was examined for sets containing a fixed number of positive and a varying number of negative examples randomly selected from the ZINC database. An increase in the ratio of positive to negative training instances was found to greatly influence most of the investigated evaluating parameters of ML methods in simulated virtual screening experiments. In a majority of cases, substantial increases in precision and MCC were observed in conjunction with some decreases in hit recall. The analysis of dynamics of those variations let us recommend an optimal composition of training data. The study was performed on several protein targets, 5 machine learning algorithms (SMO, Naïve Bayes, Ibk, J48 and Random Forest) and 2 types of molecular fingerprints (MACCS and CDK FP). The most effective classification was provided by the combination of CDK FP with SMO or Random Forest algorithms. The Naïve Bayes models appeared to be hardly sensitive to changes in the number of negative instances in the training set.ConclusionsIn conclusion, the ratio of positive to negative training instances should be taken into account during the preparation of machine learning experiments, as it might significantly influence the performance of particular classifier. What is more, the optimization of negative training set size can be applied as a boosting-like approach in machine learning-based virtual screening.

Discovery of novel inhibitors of vascular endothelial growth factor-A–Neuropilin-1 interaction by structure-based virtual screening

Neuropilin-1 (NRP-1), one of the most important co-receptors of vascular endothelial growth factor-A (VEGF-A), increases its angiogenic action in several chronic diseases including cancer by increasing the activity of associated tyrosine kinase receptors, VEGFR1 and VEGFR2. Binding of VEGF-A to NRP-1 plays a critical role in pathological angiogenesis and tumor progression. Today, targeting this interaction is a validated approach to fight against angiogenesis-dependent diseases. Only anti-NRP-1 antibodies, peptide and peptidomimetic drug-candidates or hits have been developed thus far. In order to identify potent orally active small organic molecules various experimental and in silico approaches can be used. Here we report, novel promising small drug-like molecules disrupting the binding of VEGF-A165 to NRP-1. We carried out structure-based virtual screening experiments using the ChemBridge compound collection on the VEGF-A165 binding pocket of NRP-1. After docking and two rounds of similarity search computations, we identified 4 compounds that inhibit the biotinylated VEGF-A165 binding to recombinant NRP-1 with Ki of about 10μM. These compounds contain a common chlorobenzyloxy alkyloxy halogenobenzyl amine scaffold that can serve as a base for further development of new NRP-1 inhibitors.

Benzene-1,3-dicarboxylic acid 2,5-dimethylpyrrole derivatives as multiple inhibitors of bacterial Mur ligases (MurC–MurF)

Enzymes catalyzing the biosynthesis of bacterial peptidoglycan represent traditionally a collection of highly selective targets for novel antibacterial drug design. Four members of the bacterial Mur ligase family—MurC, MurD, MurE and MurF—are involved in the intracellular steps of peptidoglycan biosynthesis, catalyzing the synthesis of the peptide moiety of the Park’s nucleotide. In our previous virtual screening campaign, a chemical class of benzene-1,3-dicarboxylic acid 2,5-dimethylpyrrole derivatives exhibiting dual MurD/MurE inhibition properties was discovered. In the present study we further investigated this class of compounds by performing inhibition assays on all four Mur ligases (MurC–MurF). Furthermore, molecular dynamics (MD) simulation studies of one of the initially discovered compound 1 were performed to explore its geometry as well as its energetic behavior based on the Linear Interaction Energy (LIE) method. Further in silico virtual screening (VS) experiments based on the parent active compound 1 were conducted to optimize the discovered series. Selected hits were assayed against all Escherichia coli MurC–MurF enzymes in biochemical inhibition assays and molecules 10–14 containing benzene-1,3-dicarboxylic acid 2,5-dimethylpyrrole coupled with five member-ring rhodanine moiety were found to be multiple inhibitors of the whole MurC–MurF cascade of bacterial enzymes in the micromolar range. Steady-state kinetics studies suggested this class to act as competitive inhibitors of the MurD enzyme towards d-Glu. These compounds represent novel valuable starting point in the development of novel antibacterial agents.

The Discovery of a Novel and Selective Inhibitor of PTP1B Over TCPTP: 3D QSAR Pharmacophore Modeling, Virtual Screening, Synthesis, and Biological Evaluation

Given the special role of insulin and leptin signaling in various biological responses, protein-tyrosine phosphatase-1B (PTP1B) was regarded as a novel therapeutic target for treating type 2 diabetes and obesity. However, owing to the highly conserved (sequence identity of about 74%) in active pocket, targeting PTP1B for drug discovery is a great challenge. In this study, we employed the software package Discovery Studio to develop 3D QSAR pharmacophore models for PTP1B and TCPTP inhibitors. It was further validated by three methods (cost analysis, test set prediction, and Fisher's test) to show that the models can be used to predict the biological activities of compounds without costly and time-consuming synthesis. The criteria for virtual screening were also validated by testing the selective PTP1B inhibitors. Virtual screening experiments and subsequent in vitro evaluation of promising hits revealed a novel and selective inhibitor of PTP1B over TCPTP. After that, a most likely binding mode was proposed. Thus, the findings reported here may provide a new strategy in discovering selective PTP1B inhibitors.

Improvement of Virtual Screening Results by Docking Data Feature Analysis

In this study, we propose a novel approach to evaluate virtual screening (VS) experiments based on the analysis of docking output data. This approach, which we refer to as docking data feature analysis (DDFA), consists of two steps. First, a set of features derived from the docking output data is computed and assigned to each molecule in the virtually screened library. Second, an artificial neural network (ANN) analyzes the molecule's docking features and estimates its activity. Given the simple architecture of the ANN, DDFA can be easily adapted to deal with information from several docking programs simultaneously. We tested our approach on the Directory of Useful Decoys (DUD), a well-established and highly accepted VS benchmark. Outstanding results were obtained by DDFA not only in comparison with the conventional rankings of the docking programs used in this work but also with respect to other methods found in the literature. Our approach performs with similar good results as the best available methods, which, however, also require substantially more computing time, economic resources, and/or expert intervention. Taken together, DDFA represents an automatic and highly attractive methodology for VS.

A rotation-translation invariant molecular descriptor of partial charges and its use in ligand-based virtual screening.

BackgroundMeasures of similarity for chemical molecules have been developed since the dawn of chemoinformatics. Molecular similarity has been measured by a variety of methods including molecular descriptor based similarity, common molecular fragments, graph matching and 3D methods such as shape matching. Similarity measures are widespread in practice and have proven to be useful in drug discovery. Because of our interest in electrostatics and high throughput ligand-based virtual screening, we sought to exploit the information contained in atomic coordinates and partial charges of a molecule.ResultsA new molecular descriptor based on partial charges is proposed. It uses the autocorrelation function and linear binning to encode all atoms of a molecule into two rotation-translation invariant vectors. Combined with a scoring function, the descriptor allows to rank-order a database of compounds versus a query molecule. The proposed implementation is called ACPC (AutoCorrelation of Partial Charges) and released in open source. Extensive retrospective ligand-based virtual screening experiments were performed and other methods were compared with in order to validate the method and associated protocol.ConclusionsWhile it is a simple method, it performed remarkably well in experiments. At an average speed of 1649 molecules per second, it reached an average median area under the curve of 0.81 on 40 different targets; hence validating the proposed protocol and implementation.

Condorcet and borda count fusion method for ligand-based virtual screening.

BackgroundIt is known that any individual similarity measure will not always give the best recall of active molecule structure for all types of activity classes. Recently, the effectiveness of ligand-based virtual screening approaches can be enhanced by using data fusion. Data fusion can be implemented using two different approaches: group fusion and similarity fusion. Similarity fusion involves searching using multiple similarity measures. The similarity scores, or ranking, for each similarity measure are combined to obtain the final ranking of the compounds in the database.ResultsThe Condorcet fusion method was examined. This approach combines the outputs of similarity searches from eleven association and distance similarity coefficients, and then the winner measure for each class of molecules, based on Condorcet fusion, was chosen to be the best method of searching. The recall of retrieved active molecules at top 5% and significant test are used to evaluate our proposed method. The MDL drug data report (MDDR), maximum unbiased validation (MUV) and Directory of Useful Decoys (DUD) data sets were used for experiments and were represented by 2D fingerprints.ConclusionsSimulated virtual screening experiments with the standard two data sets show that the use of Condorcet fusion provides a very simple way of improving the ligand-based virtual screening, especially when the active molecules being sought have a lowest degree of structural heterogeneity. However, the effectiveness of the Condorcet fusion was increased slightly when structural sets of high diversity activities were being sought.

Toward the computer-aided discovery of FabH inhibitors. Do predictive QSAR models ensure high quality virtual screening performance?

Antibiotic resistance has increased over the past two decades. New approaches for the discovery of novel antibacterials are required and innovative strategies will be necessary to identify novel and effective candidates. Related to this problem, the exploration of bacterial targets that remain unexploited by the current antibiotics in clinical use is required. One of such targets is the β-ketoacyl-acyl carrier protein synthase III (FabH). Here, we report a ligand-based modeling methodology for the virtual-screening of large collections of chemical compounds in the search of potential FabH inhibitors. QSAR models are developed for a diverse dataset of 296 FabH inhibitors using an in-house modeling framework. All models showed high fitting, robustness, and generalization capabilities. We further investigated the performance of the developed models in a virtual screening scenario. To carry out this investigation, we implemented a desirability-based algorithm for decoys selection that was shown effective in the selection of high quality decoys sets. Once the QSAR models were validated in the context of a virtual screening experiment their limitations arise. For this reason, we explored the potential of ensemble modeling to overcome the limitations associated to the use of single classifiers. Through a detailed evaluation of the virtual screening performance of ensemble models it was evidenced, for the first time to our knowledge, the benefits of this approach in a virtual screening scenario. From all the obtained results, we could arrive to a significant main conclusion: at least for FabH inhibitors, virtual screening performance is not guaranteed by predictive QSAR models.

Toward activated homology models of the human M1 muscarinic acetylcholine receptor

Structure-based virtual screening offers a good opportunity for the discovery of selective M1 muscarinic acetylcholine receptor (mAChR) agonists for the treatment of Alzheimer's disease. However, no 3-D structure of an M1 mAChR is yet available and the homology models that have been previously reported are only able to identify antagonists in virtual screening experiments. In this study, we generated a homology model of the human M1 mAChR, based on the crystal structure of an M3 mAChR as the template. This initial model was modified, using the agonist-bound crystal structure of a β2-adrenergic receptor as a guide, to give two possible activated structures. The T192 side chain was adjusted in both structures and one of the structures also had the whole of transmembrane (TM) 5 rotated and tilted toward the inner channel of the transmembrane region. The binding sites of all three structures were then refined by induced-fit docking (IFD) with acetylcholine. Virtual screening experiments showed that all three refined models could efficiently differentiate agonists from decoy molecules, with the TM5-modified models also giving good agonist/antagonist selectivity. The whole range of agonists and antagonists was observed to bind within the orthosteric site of the structure obtained by IFD refinement alone, implying that it has inactive state character. In contrast, the two TM5-modified structures were unable to accommodate the antagonists, supporting the proposition that they possess activated state character.

Application of Docking and QM/MM-GBSA Rescoring to Screen for Novel Myt1 Kinase Inhibitors

Identification of compounds that can bind to a target protein with high affinity is a nontrivial task in structure-based drug design. Several approaches ranging from simple scoring methods to more computationally demanding methods are usually applied for this purpose. In the current work, we used ligand docking in combination with QM/MM-GBSA, MM-GBSA, and MM-PBSA rescoring to discriminate between active and inactive Myt1 kinase inhibitors. Results show that QM/MM-GBSA rescoring performs better than normal docking scores or MM-GBSA rescoring in classifying active and inactive inhibitors. We also applied QM/MM-GBSA rescoring to estimate the binding affinities of compounds from different virtual screening runs. To prove our approach and to confirm its predictive power, a few compounds which were predicted to be active were purchased and experimentally tested. Among the five selected compounds, three showed significant inhibition of recombinant Myt1. PD-173952, which yielded a favorable QM/MM-GBSA binding free energy, showed a K(i) value of 8.1 nM. In addition, two compounds, PD-180970 and saracatinib, showed inhibition at the low micromolar level. Thus, the developed protocol might be useful for further virtual screening experiments to better discriminate between active and inactive compounds and to further optimize the identified hits.

An infrastructure to mine molecular descriptors for ligand selection on virtual screening.

The receptor-ligand interaction evaluation is one important step in rational drug design. The databases that provide the structures of the ligands are growing on a daily basis. This makes it impossible to test all the ligands for a target receptor. Hence, a ligand selection before testing the ligands is needed. One possible approach is to evaluate a set of molecular descriptors. With the aim of describing the characteristics of promising compounds for a specific receptor we introduce a data warehouse-based infrastructure to mine molecular descriptors for virtual screening (VS). We performed experiments that consider as target the receptor HIV-1 protease and different compounds for this protein. A set of 9 molecular descriptors are taken as the predictive attributes and the free energy of binding is taken as a target attribute. By applying the J48 algorithm over the data we obtain decision tree models that achieved up to 84% of accuracy. The models indicate which molecular descriptors and their respective values are relevant to influence good FEB results. Using their rules we performed ligand selection on ZINC database. Our results show important reduction in ligands selection to be applied in VS experiments; for instance, the best selection model picked only 0.21% of the total amount of drug-like ligands.

A new graph-based molecular descriptor using the canonical representation of the molecule.

Molecular similarity is a pervasive concept in drug design. The basic idea underlying molecular similarity is the similar property principle, which states that structurally similar molecules will exhibit similar physicochemical and biological properties. In this paper, a new graph-based molecular descriptor (GBMD) is introduced. The GBMD is a new method of obtaining a rough description of 2D molecular structure in textual form based on the canonical representations of the molecule outline shape and it allows rigorous structure specification using small and natural grammars. Simulated virtual screening experiments with the MDDR database show clearly the superiority of the graph-based descriptor compared to many standard descriptors (ALOGP, MACCS, EPFP4, CDKFP, PCFP, and SMILE) using the Tanimoto coefficient (TAN) and the basic local alignment search tool (BLAST) when searches were carried.

Improving structural similarity based virtual screening using background knowledge.

BackgroundVirtual screening in the form of similarity rankings is often applied in the early drug discovery process to rank and prioritize compounds from a database. This similarity ranking can be achieved with structural similarity measures. However, their general nature can lead to insufficient performance in some application cases. In this paper, we provide a link between ranking-based virtual screening and fragment-based data mining methods. The inclusion of binding-relevant background knowledge into a structural similarity measure improves the quality of the similarity rankings. This background knowledge in the form of binding relevant substructures can either be derived by hand selection or by automated fragment-based data mining methods.ResultsIn virtual screening experiments we show that our approach clearly improves enrichment factors with both applied variants of our approach: the extension of the structural similarity measure with background knowledge in the form of a hand-selected relevant substructure or the extension of the similarity measure with background knowledge derived with data mining methods.ConclusionOur study shows that adding binding relevant background knowledge can lead to significantly improved similarity rankings in virtual screening and that even basic data mining approaches can lead to competitive results making hand-selection of the background knowledge less crucial. This is especially important in drug discovery and development projects where no receptor structure is available or more frequently no verified binding mode is known and mostly ligand based approaches can be applied to generate hit compounds.

Large‐scale virtual screening experiments on Windows Azure‐based cloud resources

SUMMARYMolecular docking simulations have high potential to contribute to a wide area of molecular and biomedical research in various disciplines including molecular biology, drug design, environmental studies and psychology. Conducting large‐scale molecular docking experiments requires a vast amount of computing resources. Several types of distributed computing infrastructures have been investigated and utilized recently to conduct such simulations, including service and desktop grid systems or local clusters. This paper investigates and analyses how Windows Azure‐based cloud resources can be applied for this purpose. A virtual screening experiment framework has been implemented on a Windows Azure‐based cloud using the generic worker concept. Virtual machines can be instantiated in the cloud on demand scaling up the simulations based on the volume of molecules to be docked and the available financial resources. Bioscientists are able to execute the simulations and visualise the results from a high‐level user interface. The paper describes the experiences when implementing the molecular docking application on this novel platform and provides the first benchmarking experiments to evaluate the suitability of the infrastructure for computation intensive simulations. Copyright © 2013 John Wiley & Sons, Ltd.

DRUG TARGET PRIORITIZATION IN PLASMODIUM FALCIPARUM THROUGH METABOLIC NETWORK ANALYSIS, AND INHIBITOR DESIGNING USING VIRTUAL SCREENING AND DOCKING APPROACH

The genome sequence of Plasmodium falciparum reveals that many metabolic pathways are unique as compared to its human host. Metabolic Network Analysis was carried out to find the essential enzymes critical for the survival of the pathogen. In the present study, choke point and load point analysis was used to locate putative targets. The identified targets were further checked to confirm that no alternate pathway or human homolog exists. Among the top 15 enzymes obtained from this analysis, we have selected P. falciparum orotidine-5'-monophosphate decarboxylase (PfODCase) enzyme as it is sequentially and structurally different from that of humans, for searching novel inhibitors. A five-point 3D pharmacophore was generated for the crystal structure of PfODCase complexes with uridine-5'-monophosphate (U5P). The binding site environment shows three H-bond acceptors, one H-bond donor and one negative ionizable feature. This pharmacophore model was used as a 3D query to perform virtual screening experiments against 2,664,779 standard lead compounds obtained from the freely available ZINC database. Top 10 hits obtained from virtual screening were selected for molecular docking experiments against PfODCase in order to verify their results and to have a better insight into their binding modes. Here, docking of U5P with PfODCase is used as a control. We have identified six compounds, among them, few are U5P analogs and others are novel ones with diverse scaffolds. The key residues: Lys42, Asp20, Lys72, Ser127, Ala184, Gln185 and Arg203 at the main binding pocket of PfODCase are responsible for better stability of diverse ligands. These compounds according to their free energy of binding could serve as potent leads for designing novel inhibitors against malarial ODCase enzyme.

Structure- and ligand-based drug design of novel p38-alpha MAPK inhibitors in the fight against the Alzheimer’s disease

Alzheimer’s disease (AD) is characterized microscopically by the presence of amyloid plaques, which are accumulations of beta-amyloid protein inter-neurons, and neurofibrillary tangles formed predominantly by highly phosphorylated forms of the microtubule-associated protein, tau, which form tangled masses that consume neuronal cell body, possibly leading to neuronal dysfunction and ultimately death. p38α mitogen-activated protein kinase (MAPK) has been implicated in both events associated with AD, tau phosphorylation and inflammation. p38α MAPK pathway is activated by a dual phosphorylation at Thr180 and Tyr182 residues. Drug design of p38α MAPK inhibitors is mainly focused on small molecules that compete for Adenosine triphosphate in the catalytic site. Here, we used different approaches of structure- and ligand-based drug design and medicinal chemistry strategies based on a selected p38α MAPK structure deposited in the Protein Data Bank in complex with inhibitor, as well as others reported in literature. As a result of the virtual screening experiments performed here, as well as molecular dynamics, molecular interaction fields studies, shape and electrostatic similarities, activity and toxicity predictions, and pharmacokinetic and physicochemical properties, we have selected 13 compounds that meet the criteria of low or no toxicity potential, good pharmacotherapeutic profile, predicted activities, and calculated values ​​comparable with those obtained for the reference compounds, while maintaining the main interactions observed for the most potent inhibitors.

Optimization and visualization of the edge weights in optimal assignment methods for virtual screening.

BackgroundLigand‐based virtual screening plays a fundamental part in the early drug discovery stage. In a virtual screening, a chemical library is searched for molecules with similar properties to a query molecule by means of a similarity function. The optimal assignment of chemical graphs has proven to be a valuable similarity function for many cheminformatic tasks, such as virtual screening. The optimal assignment assumes all atoms of a query molecule to be equally important, which is not realistic depending on the binding mode of a ligand. The importance of a query molecule’s atoms can be integrated in the optimal assignment by weighting the assignment edges. We optimized the edge weights with respect to the virtual screening performance by means of evolutionary algorithms. Furthermore, we propose a visualization approach for the interpretation of the edge weights.ResultsWe evaluated two different evolutionary algorithms, differential evolution and particle swarm optimization, for their suitability for optimizing the assignment edge weights. The results showed that both optimization methods are suited to optimize the edge weights. Furthermore, we compared our approach to the optimal assignment with equal edge weights and two literature similarity functions on a subset of the Directory of Useful Decoys using sophisticated virtual screening performance metrics. Our approach achieved a considerably better overall and early enrichment performance. The visualization of the edge weights enables the identification of substructures that are important for a good retrieval of ligands and for the binding to the protein target.ConclusionsThe optimization of the edge weights in optimal assignment methods is a valuable approach for ligand‐based virtual screening experiments. The approach can be applied to any similarity function that employs the optimal assignment method, which includes a variety of similarity measures that have proven to be valuable in various cheminformatic tasks. The proposed visualization helps to get a better understanding of the binding mode of the analyzed query molecule.

Using graph-based consensus clustering for combining K-means clustering of heterogeneous chemical structures

Consensus clustering methods are motivated by the success of combining multiple classifiers in many areas. In this paper, graph-based consensus clustering is used to improve the quality of chemical compound clustering by enhancing the robustness, novelty, consistency and stability of individual clusterings. For this purpose, Hyper-Graph Partitioning Algorithm (HGPA) [1], was applied. The clustering is evaluated based on the ability to separate actives from inactives molecules in each cluster and the results were compared with the Ward's clustering method. The chemical dataset MDL Drug Data Report (MDDR) database has been used for experiments. The MDL Drug Data Report (MDDR) database consists of 102516 molecules. For the experiments, the dataset DS1 was chosen from the MDDR database. This dataset has been used for many virtual screening experiments [2-4]. The dataset DS1contains 10 heterogeneous activity classes (8568 molecules). For the clustering experiments, two 2D fingerprint descriptors will be used which are developed by Scitegic's Pipeline Pilot [5]. These are 120-bit ALOGP and 1024-bit extended connectivity fingerprints (ECFP_4). The results were evaluated based on the effectiveness of the methods to separate actives from non-actives molecules using QPI- (for quality partition index) measure, which was devised by Varin et al. [6]. As defined by [7], an active cluster as a non-singleton cluster for which the percentage of active molecules in the cluster is greater than the percentage of active molecules in the dataset as a whole. Let p be the number of actives in active clusters, q the number of inactives in active clusters, r the number of actives in inactive clusters (i.e., clusters that are not active clusters) and s the number of singleton actives. The high value occurs when the actives are clustered tightly together and separated from the inactive molecules. Then the quality partition index, QPI, is defined to be: QPI=pp+q+r+s (1) Then, the results will be compared with Ward's individual clustering method, the standard clustering method for chemoinformatics applications. The generation process has been done by multiple run of K-means algorithms, each with random initialization of cluster centroids. The number of partitions generated in this step is ranged between n = 5 to n = 50, with 5-times step. Then, all the generated partitions were combined using HGPA to obtain the consensus partition. This process is done for each fingerprint (ALOGP and ECFP_4). The mean of QPI values are averaged over the ten activity classes of the datasets. Tables ​Tables1,1, ​,22 show the effectiveness of MDDR dataset clustering using ALOGP and ECFP_4 fingerprints. The best PQI value of consensus clustering methods for each column has been bold-faced for ease of reference. Table 1 Effectivenss of clustering of high diverse MDDR dataset: ALOGP Fingerprint. Table 2 Effectivenss of clustering of high diverse MDDR dataset: ECFP_4 Fingerprint. Visual inspection of the results enables comparisons to be made between the effectiveness of clustering of MDDR datasets and Ward's method, the best of choice clustering method for chemoinformatics applications. In addition, ten times of consensus clustering, for each fingerprint were observed in order to study the effectiveness of consensus clustering with different ensemble sizes. The results show that HGPA consensus clustering gives robust and novel result when K-means algorithm is run 20-50 times using ALOGP. The performance of consensus clustering outperforms the Wards' method. For consensus of dataset which represented by ECFP_4 fingerprint, the best QPI values of consensus clustering are obtained from ensembles of size n = 20-50. The performance of consensus clustering gives robust results which are better than overall performance of individual clusterings. The values of QPI in both datasets for consensus clustering are close to the Wards method. The consensus clustering, HGPA, provide stable clusters by decreasing the sensitivity to noise and outliers. The average percentages of singleton clusters of individual clusterings compared with consensus clustering for both fingerprints. The results show that consensus clustering partition the datasets with average percentage of singleton equal to zero, which is much better than individual clusterings and Wards' method. For example, 16.72% of molecules of DS1 are clustered as singletons when Wards method is applied on ALOGP fingerprint with number of clusters equal to 1000 clusters. Finally we conclude that graph-based consensus clustering can improve the effectiveness of chemical compounds clustering. The performance of consensus clustering is more robust, novel, stable, consistent, and out-perform Wards' method in case of using ALOGP fingerprint. By using ECFP_4 fingerprint, consensus clustering methods provide more robust, stable, and consistent clustering and close to the Wards clustering results. The experiments reported here suggest that graph-based consensus clustering can improve the quality of individual clustering by using the efficient algorithm, K-means algorithm, to generate the ensemble with size (20-50) for both structurally diverse chemical datasets.

How to pick a winning team: approaches towards the selection of computationally derived protein structures for ensemble-based virtual screening

The necessity of treating receptor flexibility in protein-ligand docking has been widely acknowledged and is the subject of extensive research in the field of drug discovery [1]. The use of multiple discrete protein conformations, so-called ensemble docking, has been proven to be a valid concept to mimic target plasticity in docking experiments [2,3]. Using molecular dynamics (MD) the number of different conformations that can be generated is practically unlimited. Not all of these conformations can be included in the pose prediction or virtual screening process for reasons of computational cost. Moreover, some of them will be more suitable for docking purposes than others. The question arises if and how adequate protein conformations can be selected systematically a priori based on quantifiable structural features. For neuraminidase and cyclin-dependent kinase II, snapshots of molecular dynamics simulation trajectories have been clustered and structurally assessed by applying a variety of methods. Extensive cross docking and virtual screening experiments show that relatively large differences in docking performance are caused by only very subtle conformational changes within the protein which cannot be captured by the currently applied characterisation methods. As an alternative, cross docking capability can be used as a reliable indicator towards the selection of suitable conformations for ensemble-based virtual screening. In combination with short minimisations of docked poses in the binding site, virtual screening performance can be further improved and ensembles of MD snapshots can be built which perform as well as the generally superior holo crystal structures.

Exploring Distinct Binding Site Regions of ß2 Adrenergic Receptor via Coarse-Grained Molecular Dynamics Simulations

\beta_2-Adrenergic receptor (\beta_2AR) is a G protein- coupled receptor that is highly flexible and able to recognize a wide range of ligands through its conformational variations. Active and inactive conformations revealed by recent crystallographic experiments do not provide a complete dynamic picture of the receptor, especially in the binding site. In this study, molecular dynamics (MD) simulation through a residue-based coarse-grained model is used as an alternative and efficient method to explore a wider conformational search space. The system was composed of \beta_2AR embedded into a 1-palmitoyl-2-oleoyl-phosphatidylcholine membrane bilayer with surrounding water. A total of 6 \mu s of simulation at constant NPT was performed for a system of 6868 coarse-grained beads. The system reached equilibrium at around 0.1 \mu s. The overall 3-dimensional structure was well preserved throughout the simulation. Local residue-based fluctuations were in good agreement with fully atomistic MD simulations. Four distinct snapshots were selected and reverse-mapped to all-atom representations with around 65,000 atoms. Each reverse-mapped system was later subjected to 100 ns of MD simulation for equilibration. Root mean square deviation clustering analysis yielded distinct receptor conformers for the binding site regions, which were suggested to be alternative representations of the binding pocket and thus were proposed as plausible targets in docking-based virtual screening experiments for the discovery of novel antagonists.