Using physics-based pose predictions and free energy perturbation calculations to predict binding poses and relative binding affinities for FXR ligands in the D3R Grand Challenge 2

Computer-aided drug design has become an integral part of drug discovery and development in the pharmaceutical and biotechnology industry, and is nowadays extensively used in the lead identification and lead optimization phases. The drug design data resource (D3R) organizes challenges against blinded experimental data to prospectively test computational methodologies as an opportunity for improved methods and algorithms to emerge. We participated in Grand Challenge 2 to predict the crystallographic poses of 36 Farnesoid X Receptor (FXR)-bound ligands and the relative binding affinities for two designated subsets of 18 and 15 FXR-bound ligands. Here, we present our methodology for pose and affinity predictions and its evaluation after the release of the experimental data. For predicting the crystallographic poses, we used docking and physics-based pose prediction methods guided by the binding poses of native ligands. For FXR ligands with known chemotypes in the PDB, we accurately predicted their binding modes, while for those with unknown chemotypes the predictions were more challenging. Our group ranked #1st (based on the median RMSD) out of 46 groups, which submitted complete entries for the binding pose prediction challenge. For the relative binding affinity prediction challenge, we performed free energy perturbation (FEP) calculations coupled with molecular dynamics (MD) simulations. FEP/MD calculations displayed a high success rate in identifying compounds with better or worse binding affinity than the reference (parent) compound. Our studies suggest that when ligands with chemical precedent are available in the literature, binding pose predictions using docking and physics-based methods are reliable; however, predictions are challenging for ligands with completely unknown chemotypes. We also show that FEP/MD calculations hold predictive value and can nowadays be used in a high throughput mode in a lead optimization project provided that crystal structures of sufficiently high quality are available.     

AutoShim: empirically corrected scoring functions for quantitative docking with a crystal structure and IC50 training data

It has been notoriously difficult to develop general all-purpose scoring functions for high-throughput docking that correlate with measured binding affinity. As a practical alternative, AutoShim uses the program Magnet to add point-pharmacophore like "shims" to the binding site of each protein target. The pharmacophore shims are weighted by partial least-squares (PLS) regression, adjusting the all-purpose scoring function to reproduce IC 50 data, much as the shims in an NMR magnet are weighted to optimize the field for a better spectrum. This dramatically improves the affinity predictions on 25% of the compounds held out at random. An iterative procedure chooses the best pose during the process of shim parametrization. This method reproducibly converges to a consistent solution, regardless of starting pose, in just 2-4 iterations, so these robust models do not overtrain. Sets of complex multifeature shims, generated by a recursive partitioning method, give the best activity predictions, but these are difficult to interpret when designing new compounds. Sets of simpler single-point pharmacophores still predict affinity reasonably well and clearly indicate the molecular interactions producing effective binding. The pharmacophore requirements are very reproducible, irrespective of the compound sets used for parametrization, lending confidence to the predictions and interpretations. The automated procedure does require a training set of experimental compounds but otherwise adds little extra time over conventional docking.     

Workflows and performances in the ranking prediction of 2016 D3R Grand Challenge 2: lessons learned from a collaborative effort

The 2016 D3R Grand Challenge 2 includes both pose and affinity or ranking predictions. This article is focused exclusively on affinity predictions submitted to the D3R challenge from a collaborative effort of the modeling and informatics group. Our submissions include ranking of 102 ligands covering 4 different chemotypes against the FXR ligand binding domain structure, and the relative binding affinity predictions of the two designated free energy subsets of 15 and 18 compounds. Using all the complex structures prepared in the same way allowed us to cover many types of workflows and compare their performances effectively. We evaluated typical workflows used in our daily structure-based design modeling support, which include docking scores, force field-based scores, QM/MM, MMGBSA, MD-MMGBSA, and MacroModel interaction energy estimations. The best performing methods for the two free energy subsets are discussed. Our results suggest that affinity ranking still remains very challenging; that the knowledge of more structural information does not necessarily yield more accurate predictions; and that visual inspection and human intervention are considerably important for ranking. Knowledge of the mode of action and protein flexibility along with visualization tools that depict polar and hydrophobic maps are very useful for visual inspection. QM/MM-based workflows were found to be powerful in affinity ranking and are encouraged to be applied more often. The standardized input and output enable systematic analysis and support methodology development and improvement for high level blinded predictions.     

Docking of small molecules to farnesoid X receptors using AutoDock Vina with the Convex-PL potential: lessons learned from D3R Grand Challenge 2

The 2016 D3R Grand Challenge 2 provided an opportunity to test multiple protein–ligand docking protocols on a set of ligands bound to farnesoid X receptor that has many available experimental structures. We participated in the Stage 1 of the Challenge devoted to the docking pose predictions, with the mean RMSD value of our submission poses of 2.9 Å. Here we present a thorough analysis of our docking predictions made with AutoDock Vina and the Convex-PL rescoring potential by reproducing our submission protocol and running a series of additional molecular docking experiments. We conclude that a correct receptor structure, or more precisely, the structure of the binding pocket, plays the crucial role in the success of our docking studies. We have also noticed the important role of a local ligand geometry, which seems to be not well discussed in literature. We succeed to improve our results up to the mean RMSD value of 2.15–2.33 Å  dependent on the models of the ligands, if docking these to all available homologous receptors. Overall, for docking of ligands of diverse chemical series we suggest to perform docking of each of the ligands to a set of multiple receptors that are homologous to the target.     

Prediction of small molecule binding property of protein domains with Bayesian classifiers based on Markov chains

Accurate computational methods that can help to predict biological function of a protein from its sequence are of great interest to research biologists and pharmaceutical companies. One approach to assume the function of proteins is to predict the interactions between proteins and other molecules. In this work, we propose a machine learning method that uses a primary sequence of a domain to predict its propensity for interaction with small molecules. By curating the Pfam database with respect to the small molecule binding ability of its component domains, we have constructed a dataset of small molecule binding and non-binding domains. This dataset was then used as training set to learn a Bayesian classifier, which should distinguish members of each class. The domain sequences of both classes are modelled with Markov chains. In a Jack-knife test, our classification procedure achieved the predictive accuracies of 77.2% and 66.7% for binding and non-binding classes respectively. We demonstrate the applicability of our classifier by using it to identify previously unknown small molecule binding domains. Our predictions are available as supplementary material and can provide very useful information to drug discovery specialists. Given the ubiquitous and essential role small molecules play in biological processes, our method is important for identifying pharmaceutically relevant components of complete proteomes. The software is available from the author upon request.     

Virtual fragment screening: exploration of MM-PBSA re-scoring

An NMR fragment screening dataset with known binders and decoys was used to evaluate the ability of docking and re-scoring methods to identify fragment binders. Re-scoring docked poses using the Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) implicit solvent model identifies additional active fragments relative to either docking or random fragment screening alone. Early enrichment, which is clearly most important in practice for selecting relatively small sets of compounds for experimental testing, is improved by MM-PBSA re-scoring. In addition, the value in MM-PBSA re-scoring of docked poses for virtual screening may be in lessening the effect of the variation in the protein complex structure used.     

Docking rigid macrocycles using Convex-PL,AutoDock Vina,and RDKit in the D3R Grand Challenge 4

The D3R Grand Challenge 4 provided a brilliant opportunity to test macrocyclic docking protocols on a diverse high-quality experimental data. We participated in both pose and affinity prediction exercises. Overall, we aimed to use an automated structure-based docking pipeline built around a set of tools developed in our team. This exercise again demonstrated a crucial importance of the correct local ligand geometry for the overall success of docking. Starting from the second part of the pose prediction stage, we developed a stable pipeline for sampling macrocycle conformers. This resulted in the subangstrom average precision of our pose predictions. In the affinity prediction exercise we obtained average results. However, we could improve these when using docking poses submitted by the best predictors. Our docking tools including the Convex-PL scoring function are available at https://team.inria.fr/nano-d/software/.     

A structure-guided approach for protein pocket modeling and affinity prediction

Binding affinity prediction is frequently addressed using computational models constructed solely with molecular structure and activity data. We present a hybrid structure-guided strategy that combines molecular similarity, docking, and multiple-instance learning such that information from protein structures can be used to inform models of structure–activity relationships. The Surflex-QMOD approach has been shown to produce accurate predictions of binding affinity by constructing an interpretable physical model of a binding site with no experimental binding site structural information. We introduce a method to integrate protein structure information into the model induction process in order to construct more robust physical models. The structure-guided models accurately predict binding affinities over a broad range of compounds while producing more accurate representations of the protein pockets and ligand binding modes. Structure-guidance for the QMOD method yielded significant performance improvements, both for affinity and pose prediction, especially in cases where predictions were made on ligands very different from those used for model induction.     

Lessons learned in induced fit docking and metadynamics in the Drug Design Data Resource Grand Challenge 2

Two of the major ongoing challenges in computational drug discovery are predicting the binding pose and affinity of a compound to a protein. The Drug Design Data Resource Grand Challenge 2 was developed to address these problems and to drive development of new methods. The challenge provided the 2D structures of compounds for which the organizers help blinded data in the form of 35 X-ray crystal structures and 102 binding affinity measurements and challenged participants to predict the binding pose and affinity of the compounds. We tested a number of pose prediction methods as part of the challenge; we found that docking methods that incorporate protein flexibility (Induced Fit Docking) outperformed methods that treated the protein as rigid. We also found that using binding pose metadynamics, a molecular dynamics based method, to score docked poses provided the best predictions of our methods with an average RMSD of 2.01 Å. We tested both structure-based (e.g. docking) and ligand-based methods (e.g. QSAR) in the affinity prediction portion of the competition. We found that our structure-based methods based on docking with Smina (Spearman ρ?=?0.614), performed slightly better than our ligand-based methods (ρ?=?0.543), and had equivalent performance with the other top methods in the competition. Despite the overall good performance of our methods in comparison to other participants in the challenge, there exists significant room for improvement especially in cases such as these where protein flexibility plays such a large role.     

Quantitative surface field analysis: learning causal models to predict ligand binding affinity and pose

We introduce the QuanSA method for inducing physically meaningful field-based models of ligand binding pockets based on structure-activity data alone. The method is closely related to the QMOD approach, substituting a learned scoring field for a pocket constructed of molecular fragments. The problem of mutual ligand alignment is addressed in a general way, and optimal model parameters and ligand poses are identified through multiple-instance machine learning. We provide algorithmic details along with performance results on sixteen structure-activity data sets covering many pharmaceutically relevant targets. In particular, we show how models initially induced from small data sets can extrapolatively identify potent new ligands with novel underlying scaffolds with very high specificity. Further, we show that combining predictions from QuanSA models with those from physics-based simulation approaches is synergistic. QuanSA predictions yield binding affinities, explicit estimates of ligand strain, associated ligand pose families, and estimates of structural novelty and confidence. The method is applicable for fine-grained lead optimization as well as potent new lead identification.     

D3R grand challenge 2015: Evaluation of protein–ligand pose and affinity predictions

The Drug Design Data Resource (D3R) ran Grand Challenge 2015 between September 2015 and February 2016. Two targets served as the framework to test community docking and scoring methods: (1) HSP90, donated by AbbVie and the Community Structure Activity Resource (CSAR), and (2) MAP4K4, donated by Genentech. The challenges for both target datasets were conducted in two stages, with the first stage testing pose predictions and the capacity to rank compounds by affinity with minimal structural data; and the second stage testing methods for ranking compounds with knowledge of at least a subset of the ligand–protein poses. An additional sub-challenge provided small groups of chemically similar HSP90 compounds amenable to alchemical calculations of relative binding free energy. Unlike previous blinded Challenges, we did not provide cognate receptors or receptors prepared with hydrogens and likewise did not require a specified crystal structure to be used for pose or affinity prediction in Stage 1. Given the freedom to select from over 200 crystal structures of HSP90 in the PDB, participants employed workflows that tested not only core docking and scoring technologies, but also methods for addressing water-mediated ligand–protein interactions, binding pocket flexibility, and the optimal selection of protein structures for use in docking calculations. Nearly 40 participating groups submitted over 350 prediction sets for Grand Challenge 2015. This overview describes the datasets and the organization of the challenge components, summarizes the results across all submitted predictions, and considers broad conclusions that may be drawn from this collaborative community endeavor.     

Simultaneous prediction of binding free energy and specificity for PDZ domain–peptide interactions

Interactions between protein domains and linear peptides underlie many biological processes. Among these interactions, the recognition of C-terminal peptides by PDZ domains is one of the most ubiquitous. In this work, we present a mathematical model for PDZ domain–peptide interactions capable of predicting both affinity and specificity of binding based on X-ray crystal structures and comparative modeling with Rosetta. We developed our mathematical model using a large phage display dataset describing binding specificity for a wild type PDZ domain and 91 single mutants, as well as binding affinity data for a wild type PDZ domain binding to 28 different peptides. Structural refinement was carried out through several Rosetta protocols, the most accurate of which included flexible peptide docking and several iterations of side chain repacking and backbone minimization. Our findings emphasize the importance of backbone flexibility and the energetic contributions of side chain-side chain hydrogen bonds in accurately predicting interactions. We also determined that predicting PDZ domain–peptide interactions became increasingly challenging as the length of the peptide increased in the N-terminal direction. In the training dataset, predicted binding energies correlated with those derived through calorimetry and specificity switches introduced through single mutations at interface positions were recapitulated. In independent tests, our best performing protocol was capable of predicting dissociation constants well within one order of magnitude of the experimental values and specificity profiles at the level of accuracy of previous studies. To our knowledge, this approach represents the first integrated protocol for predicting both affinity and specificity for PDZ domain–peptide interactions.     

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.     

Extrapolative prediction using physically-based QSAR

Surflex-QMOD integrates chemical structure and activity data to produce physically-realistic models for binding affinity prediction. Here, we apply QMOD to a 3D-QSAR benchmark dataset and show broad applicability to a diverse set of targets. Testing new ligands within the QMOD model employs automated flexible molecular alignment, with the model itself defining the optimal pose for each ligand. QMOD performance was compared to that of four approaches that depended on manual alignments (CoMFA, two variations of CoMSIA, and CMF). QMOD showed comparable performance to the other methods on a challenging, but structurally limited, test set. The QMOD models were also applied to test a large and structurally diverse dataset of ligands from ChEMBL, nearly all of which were synthesized years after those used for model construction. Extrapolation across diverse chemical structures was possible because the method addresses the ligand pose problem and provides structural and geometric means to quantitatively identify ligands within a model’s applicability domain. Predictions for such ligands for the four tested targets were highly statistically significant based on rank correlation. Those molecules predicted to be highly active (\(\hbox {pK}_i \ge 7.5\)) had a mean experimental \(\hbox {pK}_i\) of 7.5, with potent and structurally novel ligands being identified by QMOD for each target.     

Estimation of relative binding free energy based on a free energy variational principle for the FKBP-ligand system

Predicting an accurate binding free energy between a target protein and a ligand can be one of the most important steps in a drug discovery process. Often, many molecules must be screened to find probable high potency ones. Thus, a computational technique with low cost is highly desirable for the estimation of binding free energies of many molecules. Several techniques have thus far been developed for estimating binding free energies. Some techniques provide accurate predictions of binding free energies but high large computational cost. Other methods give good predictions but require tuning of some parameters to predict them with high accuracy. In this study, we propose a method to predict relative binding free energies with accuracy comparable to the results of prior methods but with lower computational cost and with no parameter needing to be carefully tuned. Our technique is based on the free energy variational principle. FK506 binding protein (FKBP) with 18 ligands is taken as a test system. Our results are compared to those from other widely used techniques. Our method provides a correlation coefficient (r ² ) of 0.80 between experimental and calculated relative binding free energies and yields an average absolute error of 0.70 kcal/mol compared to experimental values. These results are comparable to or better than results from other techniques. We also discuss the possibility to improve our method further.     

CrystalDock: a novel approach to fragment-based drug design

We present a novel algorithm called CrystalDock that analyzes a molecular pocket of interest and identifies potential binding fragments. The program first identifies groups of pocket-lining receptor residues (i.e., microenvironments) and then searches for geometrically similar microenvironments present in publically available databases of ligand-bound experimental structures. Germane fragments from the crystallographic or NMR ligands are subsequently placed within the novel binding pocket. These positioned fragments can be linked together to produce ligands that are likely to be potent; alternatively, they can be joined to an inhibitor with a known or suspected binding pose to potentially improve binding affinity. To demonstrate the utility of the algorithm, CrystalDock is used to analyze the principal binding pockets of influenza neuraminidase and Trypanosoma brucei RNA editing ligase 1, validated drug targets in the fight against pandemic influenza and African sleeping sickness, respectively. In both cases, CrystalDock suggests modifications to known inhibitors that may improve binding affinity.     

Ligand Identification Scoring Algorithm (LISA)

A central problem in de novo drug design is determining the binding affinity of a ligand with a receptor. A new scoring algorithm is presented that estimates the binding affinity of a protein-ligand complex given a three-dimensional structure. The method, LISA (Ligand Identification Scoring Algorithm), uses an empirical scoring function to describe the binding free energy. Interaction terms have been designed to account for van der Waals (VDW) contacts, hydrogen bonding, desolvation effects, and metal chelation to model the dissociation equilibrium constants using a linear model. Atom types have been introduced to differentiate the parameters for VDW, H-bonding interactions, and metal chelation between different atom pairs. A training set of 492 protein-ligand complexes was selected for the fitting process. Different test sets have been examined to evaluate its ability to predict experimentally measured binding affinities. By comparing with other well-known scoring functions, the results show that LISA has advantages over many existing scoring functions in simulating protein-ligand binding affinity, especially metalloprotein-ligand binding affinity. Artificial Neural Network (ANN) was also used in order to demonstrate that the energy terms in LISA are well designed and do not require extra cross terms.     

Docking and scoring with ICM: the benchmarking results and strategies for improvement

Flexible docking and scoring using the internal coordinate mechanics software (ICM) was benchmarked for ligand binding mode prediction against the 85 co-crystal structures in the modified Astex data set. The ICM virtual ligand screening was tested against the 40 DUD target benchmarks and 11-target WOMBAT sets. The self-docking accuracy was evaluated for the top 1 and top 3 scoring poses at each ligand binding site with near native conformations below 2?? RMSD found in 91 and 95% of the predictions, respectively. The virtual ligand screening using single rigid pocket conformations provided the median area under the ROC curves equal to 69.4 with 22.0% true positives recovered at 2% false positive rate. Significant improvements up to ROC AUC?=?82.2 and ROC((2%))?=?45.2 were achieved following our best practices for flexible pocket refinement and out-of-pocket binding rescore. The virtual screening can be further improved by considering multiple conformations of the target.