Potential and limitations of ensemble docking

A major problem in structure-based virtual screening applications is the appropriate selection of a single or even multiple protein structures to be used in the virtual screening process. A priori it is unknown which protein structure(s) will perform best in a virtual screening experiment. We investigated the performance of ensemble docking, as a function of ensemble size, for eight targets of pharmaceutical interest. Starting from single protein structure docking results, for each ensemble size up to 500,000 combinations of protein structures were generated, and, for each ensemble, pose prediction and virtual screening results were derived. Comparison of single to multiple protein structure results suggests improvements when looking at the performance of the worst and the average over all single protein structures to the performance of the worst and average over all protein ensembles of size two or greater, respectively. We identified several key factors affecting ensemble docking performance, including the sampling accuracy of the docking algorithm, the choice of the scoring function, and the similarity of database ligands to the cocrystallized ligands of ligand-bound protein structures in an ensemble. Due to these factors, the prospective selection of optimum ensembles is a challenging task, shown by a reassessment of published ensemble selection protocols.     

Ligand-biased ensemble receptor docking (LigBEnD): a hybrid ligand/receptor structure-based approach

Ligand docking to flexible protein molecules can be efficiently carried out through ensemble docking to multiple protein conformations, either from experimental X-ray structures or from in silico simulations. The success of ensemble docking often requires the careful selection of complementary protein conformations, through docking and scoring of known co-crystallized ligands. False positives, in which a ligand in a wrong pose achieves a better docking score than that of native pose, arise as additional protein conformations are added. In the current study, we developed a new ligand-biased ensemble receptor docking method and composite scoring function which combine the use of ligand-based atomic property field (APF) method with receptor structure-based docking. This method helps us to correctly dock 30 out of 36 ligands presented by the D3R docking challenge. For the six mis-docked ligands, the cognate receptor structures prove to be too different from the 40 available experimental Pocketome conformations used for docking and could be identified only by receptor sampling beyond experimentally explored conformational subspace.     

Utilizing experimental data for reducing ensemble size in flexible-protein docking

Efficient and sufficient incorporation of protein flexibility into docking is still a challenging task. Docking to an ensemble of protein structures has proven its utility for docking, but using a large ensemble of structures can reduce the efficiency of docking and can increase the number of false positives in virtual screening. In this paper, we describe the application of our new methodology, Limoc, to generate an ensemble of holo-like protein structures in combination with the relaxed complex scheme (RCS), to virtual screening. We describe different schemes to reduce the ensemble of protein structures to increase efficiency and enrichment quality. Utilizing experimental knowledge about actives for a target protein allows the reduction of ensemble members to a minimum of three protein structures, increasing enrichment quality and efficiency simultaneously.     

Significant enhancement of docking sensitivity using implicit ligand sampling

The efficient and accurate quantification of protein-ligand interactions using computational methods is still a challenging task. Two factors strongly contribute to the failure of docking methods to predict free energies of binding accurately: the insufficient incorporation of protein flexibility coupled to ligand binding and the neglected dynamics of the protein-ligand complex in current scoring schemes. We have developed a new methodology, named the 'ligand-model' concept, to sample protein conformations that are relevant for binding structurally diverse sets of ligands. In the ligand-model concept, molecular-dynamics (MD) simulations are performed with a virtual ligand, represented by a collection of functional groups that binds to the protein and dynamically changes its shape and properties during the simulation. The ligand model essentially represents a large ensemble of different chemical species binding to the same target protein. Representative protein structures were obtained from the MD simulation, and docking was performed into this ensemble of protein conformation. Similar binding poses were clustered, and the averaged score was utilized to rerank the poses. We demonstrate that the ligand-model approach yields significant improvements in predicting native-like binding poses and quantifying binding affinities compared to static docking and ensemble docking simulations into protein structures generated from an apo MD simulation.     

Cheminformatics meets molecular mechanics: a combined application of knowledge-based pose scoring and physical force field-based hit scoring functions improves the accuracy of structure-based virtual screening

Poor performance of scoring functions is a well-known bottleneck in structure-based virtual screening (VS), which is most frequently manifested in the scoring functions' inability to discriminate between true ligands vs known nonbinders (therefore designated as binding decoys). This deficiency leads to a large number of false positive hits resulting from VS. We have hypothesized that filtering out or penalizing docking poses recognized as non-native (i.e., pose decoys) should improve the performance of VS in terms of improved identification of true binders. Using several concepts from the field of cheminformatics, we have developed a novel approach to identifying pose decoys from an ensemble of poses generated by computational docking procedures. We demonstrate that the use of target-specific pose (scoring) filter in combination with a physical force field-based scoring function (MedusaScore) leads to significant improvement of hit rates in VS studies for 12 of the 13 benchmark sets from the clustered version of the Database of Useful Decoys (DUD). This new hybrid scoring function outperforms several conventional structure-based scoring functions, including XSCORE::HMSCORE, ChemScore, PLP, and Chemgauss3, in 6 out of 13 data sets at early stage of VS (up 1% decoys of the screening database). We compare our hybrid method with several novel VS methods that were recently reported to have good performances on the same DUD data sets. We find that the retrieved ligands using our method are chemically more diverse in comparison with two ligand-based methods (FieldScreen and FLAP::LBX). We also compare our method with FLAP::RBLB, a high-performance VS method that also utilizes both the receptor and the cognate ligand structures. Interestingly, we find that the top ligands retrieved using our method are highly complementary to those retrieved using FLAP::RBLB, hinting effective directions for best VS applications. We suggest that this integrative VS approach combining cheminformatics and molecular mechanics methodologies may be applied to a broad variety of protein targets to improve the outcome of structure-based drug discovery studies.     

Improving binding mode and binding affinity predictions of docking by ligand-based search of protein conformations: evaluation in D3R grand challenge 2015

The growing number of protein–ligand complex structures, particularly the structures of proteins co-bound with different ligands, in the Protein Data Bank helps us tackle two major challenges in molecular docking studies: the protein flexibility and the scoring function. Here, we introduced a systematic strategy by using the information embedded in the known protein–ligand complex structures to improve both binding mode and binding affinity predictions. Specifically, a ligand similarity calculation method was employed to search a receptor structure with a bound ligand sharing high similarity with the query ligand for the docking use. The strategy was applied to the two datasets (HSP90 and MAP4K4) in recent D3R Grand Challenge 2015. In addition, for the HSP90 dataset, a system-specific scoring function (ITScore2_hsp90) was generated by recalibrating our statistical potential-based scoring function (ITScore2) using the known protein–ligand complex structures and the statistical mechanics-based iterative method. For the HSP90 dataset, better performances were achieved for both binding mode and binding affinity predictions comparing with the original ITScore2 and with ensemble docking. For the MAP4K4 dataset, although there were only eight known protein–ligand complex structures, our docking strategy achieved a comparable performance with ensemble docking. Our method for receptor conformational selection and iterative method for the development of system-specific statistical potential-based scoring functions can be easily applied to other protein targets that have a number of protein–ligand complex structures available to improve predictions on binding.     

Pose prediction and virtual screening performance of GOLD scoring functions in a standardized test

The performance of all four GOLD scoring functions has been evaluated for pose prediction and virtual screening under the standardized conditions of the comparative docking and scoring experiment reported in this Edition. Excellent pose prediction and good virtual screening performance was demonstrated using unmodified protein models and default parameter settings. The best performing scoring function for both pose prediction and virtual screening was demonstrated to be the recently introduced scoring function ChemPLP. We conclude that existing docking programs already perform close to optimally in the cognate pose prediction experiments currently carried out and that more stringent pose prediction tests should be used in the future. These should employ cross-docking sets. Evaluation of virtual screening performance remains problematic and much remains to be done to improve the usefulness of publically available active and decoy sets for virtual screening. Finally we suggest that, for certain target/scoring function combinations, good enrichment may sometimes be a consequence of 2D property recognition rather than a modelling of the correct 3D interactions.     

Comparative performance of several flexible docking programs and scoring functions: enrichment studies for a diverse set of pharmaceutically relevant targets

Virtual screening by molecular docking has become a widely used approach to lead discovery in the pharmaceutical industry when a high-resolution structure of the biological target of interest is available. The performance of three widely used docking programs (Glide, GOLD, and DOCK) for virtual database screening is studied when they are applied to the same protein target and ligand set. Comparisons of the docking programs and scoring functions using a large and diverse data set of pharmaceutically interesting targets and active compounds are carried out. We focus on the problem of docking and scoring flexible compounds which are sterically capable of docking into a rigid conformation of the receptor. The Glide XP methodology is shown to consistently yield enrichments superior to the two alternative methods, while GOLD outperforms DOCK on average. The study also shows that docking into multiple receptor structures can decrease the docking error in screening a diverse set of active compounds.     

Ensemble pharmacophore meets ensemble docking: a novel screening strategy for the identification of RIPK1 inhibitors

Programmed cell death has been a fascinating area of research since it throws new challenges and questions in spite of the tremendous ongoing research in this field. Recently, necroptosis, a programmed form of necrotic cell death, has been implicated in many diseases including neurological disorders. Receptor interacting serine/threonine protein kinase 1 (RIPK1) is an important regulatory protein involved in the necroptosis and inhibition of this protein is essential to stop necroptotic process and eventually cell death. Current structure-based virtual screening methods involve a wide range of strategies and recently, considering the multiple protein structures for pharmacophore extraction has been emphasized as a way to improve the outcome. However, using the pharmacophoric information completely during docking is very important. Further, in such methods, using the appropriate protein structures for docking is desirable. If not, potential compound hits, obtained through pharmacophore-based screening, may not have correct ranks and scores after docking. Therefore, a comprehensive integration of different ensemble methods is essential, which may provide better virtual screening results. In this study, dual ensemble screening, a novel computational strategy was used to identify diverse and potent inhibitors against RIPK1. All the pharmacophore features present in the binding site were captured using both the apo and holo protein structures and an ensemble pharmacophore was built by combining these features. This ensemble pharmacophore was employed in pharmacophore-based screening of ZINC database. The compound hits, thus obtained, were subjected to ensemble docking. The leads acquired through docking were further validated through feature evaluation and molecular dynamics simulation.     

Bootstrap-based consensus scoring method for protein-ligand docking

To improve the performance of a single scoring function used in a protein-ligand docking program, we developed a bootstrap-based consensus scoring (BBCS) method, which is based on ensemble learning. BBCS combines multiple scorings, each of which has the same function form but different energy-parameter sets. These multiple energy-parameter sets are generated in two steps: (1) generation of training sets by a bootstrap method and (2) optimization of energy-parameter set by a Z-score approach, which is based on energy landscape theory as used in protein folding, against each training set. In this study, we applied BBCS to the FlexX scoring function. Using given 50 complexes, we generated 100 training sets and obtained 100 optimized energy-parameter sets. These parameter sets were tested against 48 complexes different from the training sets. BBCS was shown to be an improvement over single scoring when using a parameter set optimized by the same Z-score approach. Comparing BBCS with the original FlexX scoring function, we found that (1) the success rate of recognizing the crystal structure at the top relative to decoys increased from 33.3% to 52.1% and that (2) the rank of the crystal structure improved for 54.2% of the complexes and worsened for none. We also found that BBCS performed better than conventional consensus scoring (CS).     

Robust optimization of scoring functions for a target class

Target-specific optimization of scoring functions for protein–ligand docking is an effective method for significantly improving the discrimination of active and inactive molecules in virtual screening applications. Its applicability, however, is limited due to the narrow focus on, e.g., single protein structures. Using an ensemble of protein kinase structures, the publically available directory of useful decoys ligand dataset, and a novel multi-factorial optimization procedure, it is shown here that scoring functions can be tuned to multiple targets of a target class simultaneously. This leads to an improved robustness of the resulting scoring function parameters. Extensive validation experiments clearly demonstrate that (1) virtual screening performance for kinases improves significantly; (2) variations in database content affect this kind of machine-learning strategy to a lesser extent than binary QSAR models, and (3) the reweighting of interaction types is of particular importance for improved screening performance. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Machine learning optimization of cross docking accuracy

Performance of small molecule automated docking programs has conceptually been divided into docking -, scoring -, ranking - and screening power, which focuses on the crystal pose prediction, affinity prediction, ligand ranking and database screening capabilities of the docking program, respectively. Benchmarks show that different docking programs can excel in individual benchmarks which suggests that the scoring function employed by the programs can be optimized for a particular task. Here the scoring function of Smina is re-optimized towards enhancing the docking power using a supervised machine learning approach and a manually curated database of ligands and cross docking receptor pairs. The optimization method does not need associated binding data for the receptor-ligand examples used in the data set and works with small train sets. The re-optimization of the weights for the scoring function results in a similar docking performance with regard to docking power towards a cross docking test set. A ligand decoy based benchmark indicates a better discrimination between poses with high and low RMSD. The reported parameters for Smina are compatible with Autodock Vina and represent ready-to-use alternative parameters for researchers who aim at pose prediction rather than affinity prediction.     

Performance of multiple docking and refinement methods in the pose prediction D3R prospective Grand Challenge 2016

We describe the performance of multiple pose prediction methods for the D3R 2016 Grand Challenge. The pose prediction challenge includes 36 ligands, which represent 4 chemotypes and some miscellaneous structures against the FXR ligand binding domain. In this study we use a mix of fully automated methods as well as human-guided methods with considerations of both the challenge data and publicly available data. The methods include ensemble docking, colony entropy pose prediction, target selection by molecular similarity, molecular dynamics guided pose refinement, and pose selection by visual inspection. We evaluated the success of our predictions by method, chemotype, and relevance of publicly available data. For the overall data set, ensemble docking, visual inspection, and molecular dynamics guided pose prediction performed the best with overall mean RMSDs of 2.4, 2.2, and 2.2 Å respectively. For several individual challenge molecules, the best performing method is evaluated in light of that particular ligand. We also describe the protein, ligand, and public information data preparations that are typical of our binding mode prediction workflow.     

Scoring functions for protein-ligand docking

Virtual screening by molecular docking has become established as a method for drug lead discovery and optimization. All docking algorithms make use of a scoring function in combination with a method of search. Two theoretical aspects of scoring function performance dominate operational performance. The first is the degree to which a scoring function has a global extremum within the ligand pose landscape at the proper location. The second is the degree to which the magnitude of the function at the extremum is accurate. Presuming adequate search strategies, a scoring function's location performance will dominate behavior with respect to docking accuracy: the degree to which a predicted pose of a ligand matches experimental observation. A scoring function's magnitude performance will dominate behavior with respect to screening utility: enrichment of true ligands over non-ligands. Magnitude estimation also controls pure scoring accuracy: the degree to which bona fide ligands of a particular protein may be correctly ranked. Approaches to the development of scoring functions have varied widely, with a number of functions yielding similarly high levels of performance relating to the location issue. However, even among functions performing equally well on location, widely varying performance is observed on the question of magnitude. In many cases, performance is good enough to yield high enrichments of true ligands versus non-ligands in screening across a wide variety of protein types. Generally, performance is not good enough to correctly rank among true ligands. Strategies for improvement are discussed.     

Development and evaluation of a generic evolutionary method for protein-ligand docking

We have developed a generic evolutionary method with an empirical scoring function for the protein-ligand docking, which is a problem of paramount importance in structure-based drug design. This approach, referred to as the GEMDOCK (Generic Evolutionary Method for molecular DOCKing), combines both continuous and discrete search mechanisms. We tested our approach on seven protein-ligand complexes, and the docked lowest energy structures have root-mean-square derivations ranging from 0.32 to 0.99 A with respect to the corresponding crystal ligand structures. In addition, we evaluated GEMDOCK on crossdocking experiments, in which some complexes with an identical protein used for docking all crystallized ligands of these complexes. GEMDOCK yielded 98% docked structures with RMSD below 2.0 A when the ligands were docked into foreign protein structures. We have reported the validation and analysis of our approach on various search spaces and scoring functions. Experimental results show that our approach is robust, and the empirical scoring function is simple and fast to recognize compounds. We found that if GEMDOCK used the RMSD scoring function, then the prediction accuracy was 100% and the docked structures had RMSD below 0.1 A for each test system. These results suggest that GEMDOCK is a useful tool, and may systematically improve the forms and parameters of a scoring function, which is one of major bottlenecks for molecular recognition.     

A machine learning-based method to improve docking scoring functions and its application to drug repurposing

Docking scoring functions are notoriously weak predictors of binding affinity. They typically assign a common set of weights to the individual energy terms that contribute to the overall energy score; however, these weights should be gene family dependent. In addition, they incorrectly assume that individual interactions contribute toward the total binding affinity in an additive manner. In reality, noncovalent interactions often depend on one another in a nonlinear manner. In this paper, we show how the use of support vector machines (SVMs), trained by associating sets of individual energy terms retrieved from molecular docking with the known binding affinity of each compound from high-throughput screening experiments, can be used to improve the correlation between known binding affinities and those predicted by the docking program eHiTS. We construct two prediction models: a regression model trained using IC(50) values from BindingDB, and a classification model trained using active and decoy compounds from the Directory of Useful Decoys (DUD). Moreover, to address the issue of overrepresentation of negative data in high-throughput screening data sets, we have designed a multiple-planar SVM training procedure for the classification model. The increased performance that both SVMs give when compared with the original eHiTS scoring function highlights the potential for using nonlinear methods when deriving overall energy scores from their individual components. We apply the above methodology to train a new scoring function for direct inhibitors of Mycobacterium tuberculosis (M.tb) InhA. By combining ligand binding site comparison with the new scoring function, we propose that phosphodiesterase inhibitors can potentially be repurposed to target M.tb InhA. Our methodology may be applied to other gene families for which target structures and activity data are available, as demonstrated in the work presented here.     

Cross-docking of inhibitors into CDK2 structures. 2

In the preceding paper (Duca, J. S.; Madison, V. S.; Voigt, J. H. J. Chem. Inf. Model. 2008, 48, 659-668), the accuracy of docking and affinity predictions of the Gold and Glide programs were investigated using single protein conformations spanning 150 CDK2/inhibitor crystallographic complexes. High docking accuracy was observed with both methods; furthermore, Glide showed modest log(IC50)/score correlations. In this part of the study, the effect of combining docking results from multiple protein conformations in a consensus fashion was probed. This approach enhanced docking accuracy only for Glide, which was attributed to the nature of its scoring function. For log(IC50)/score correlations, particular emphasis was placed on considering only scores from correctly docked poses. Using multiple instead of single protein structures showed an improvement in the correlations. Validation sets and scrambling experiments were used to examine the statistical significance and predictivity of these correlations. Rather than actual improvements in scoring accuracy, docking to multiple protein conformations produced overfitting artifacts.     

Surflex-Dock: Docking benchmarks and real-world application

Benchmarks for molecular docking have historically focused on re-docking the cognate ligand of a well-determined protein-ligand complex to measure geometric pose prediction accuracy, and measurement of virtual screening performance has been focused on increasingly large and diverse sets of target protein structures, cognate ligands, and various types of decoy sets. Here, pose prediction is reported on the Astex Diverse set of 85 protein ligand complexes, and virtual screening performance is reported on the DUD set of 40 protein targets. In both cases, prepared structures of targets and ligands were provided by symposium organizers. The re-prepared data sets yielded results not significantly different than previous reports of Surflex-Dock on the two benchmarks. Minor changes to protein coordinates resulting from complex pre-optimization had large effects on observed performance, highlighting the limitations of cognate ligand re-docking for pose prediction assessment. Docking protocols developed for cross-docking, which address protein flexibility and produce discrete families of predicted poses, produced substantially better performance for pose prediction. Performance on virtual screening performance was shown to benefit by employing and combining multiple screening methods: docking, 2D molecular similarity, and 3D molecular similarity. In addition, use of multiple protein conformations significantly improved screening enrichment.