Oliver Korb

Empirical Scoring Functions for Advanced Protein−Ligand Docking with PLANTS

In this paper we present two empirical scoring functions, PLANTS(CHEMPLP) and PLANTS(PLP), designed for our docking algorithm PLANTS (Protein-Ligand ANT System), which is based on ant colony optimization (ACO). They are related, regarding their functional form, to parts of already published scoring functions and force fields. The parametrization procedure described here was able to identify several parameter settings showing an excellent performance for the task of pose prediction on two test sets comprising 298 complexes in total. Up to 87% of the complexes of the Astex diverse set and 77% of the CCDC/Astex clean listnc (noncovalently bound complexes of the clean list) could be reproduced with root-mean-square deviations of less than 2 A with respect to the experimentally determined structures. A comparison with the state-of-the-art docking tool GOLD clearly shows that this is, especially for the druglike Astex diverse set, an improvement in pose prediction performance. Additionally, optimized parameter settings for the search algorithm were identified, which can be used to balance pose prediction reliability and search speed.

An ant colony optimization approach to flexible protein-ligand docking

Abstract The prediction of the complex structure of a small ligand with a protein, the so-called protein–ligand docking problem, is a central part of the rational drug design process. For this purpose, we introduce the docking algorithm PLANTS (Protein–Ligand ANT System), which is based on ant colony optimization, one of the most successful swarm intelligence techniques. We study the effectiveness of PLANTS for several parameter settings and present a direct comparison of PLANTS’s performance to a state-of-the-art program called GOLD, which is based on a genetic algorithm and frequently used in the pharmaceutical industry for this task. Last but not least, we also show that PLANTS can make effective use of protein flexibility giving example results on cross-docking and virtual screening experiments for protein kinase A.

PLANTS: application of ant colony optimization to structure-based drug design

A central part of the rational drug development process is the prediction of the complex structure of a small ligand with a protein, the so-called protein-ligand docking problem, used in virtual screening of large databases and lead optimization. In the work presented here, we introduce a new docking algorithm called PLANTS (Protein-Ligand ANTSystem), which is based on ant colony optimization. An artificial ant colony is employed to find a minimum energy conformation of the ligand in the protein’s binding site. We present the effectiveness of PLANTS for several parameter settings as well as a direct comparison to a state-of-the-art program called GOLD, which is based on a genetic algorithm. Last but not least, results for a virtual screening on the protein target factor Xa are presented.

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.

Docking Performance of Fragments and Druglike Compounds

This paper addresses two questions of key interest to researchers working with protein-ligand docking methods: (i) Why is there such a large variation in docking performance between different test sets reported in the literature? (ii) Are fragments more difficult to dock than druglike compounds? To answer these, we construct a test set of in-house X-ray structures of protein-ligand complexes from drug discovery projects, half of which contain fragment ligands, the other half druglike ligands. We find that a key factor affecting docking performance is ligand efficiency (LE). High LE compounds are significantly easier to dock than low LE compounds, which we believe could explain the differences observed between test sets reported in the literature. There is no significant difference in docking performance between fragments and druglike compounds, but the reasons why dockings fail appear to be different.

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.

Design, synthesis, and biological evaluation of an allosteric inhibitor of HSET that targets cancer cells with supernumerary centrosomes.

Summary Centrosomes associate with spindle poles; thus, the presence of two centrosomes promotes bipolar spindle assembly in normal cells. Cancer cells often contain supernumerary centrosomes, and to avoid multipolar mitosis and cell death, these are clustered into two poles by the microtubule motor protein HSET. We report the discovery of an allosteric inhibitor of HSET, CW069, which we designed using a methodology on an interface of chemistry and biology. Using this approach, we explored millions of compounds in silico and utilized convergent syntheses. Only compound CW069 showed marked activity against HSET in vitro. The inhibitor induced multipolar mitoses only in cells containing supernumerary centrosomes. CW069 therefore constitutes a valuable tool for probing HSET function and, by reducing the growth of cells containing supernumerary centrosomes, paves the way for new cancer therapeutics.

Accelerating Molecular Docking Calculations Using Graphics Processing Units

The generation of molecular conformations and the evaluation of interaction potentials are common tasks in molecular modeling applications, particularly in protein-ligand or protein-protein docking programs. In this work, we present a GPU-accelerated approach capable of speeding up these tasks considerably. For the evaluation of interaction potentials in the context of rigid protein-protein docking, the GPU-accelerated approach reached speedup factors of up to over 50 compared to an optimized CPU-based implementation. Treating the ligand and donor groups in the protein binding site as flexible, speedup factors of up to 16 can be observed in the evaluation of protein-ligand interaction potentials. Additionally, we introduce a parallel version of our protein-ligand docking algorithm PLANTS that can take advantage of this GPU-accelerated scoring function evaluation. We compared the GPU-accelerated parallel version to the same algorithm running on the CPU and also to the highly optimized sequential CPU-based version. In terms of dependence of the ligand size and the number of rotatable bonds, speedup factors of up to 10 and 7, respectively, can be observed. Finally, a fitness landscape analysis in the context of rigid protein-protein docking was performed. Using a systematic grid-based search methodology, the GPU-accelerated version outperformed the CPU-based version with speedup factors of up to 60.

pharmACOphore: multiple flexible ligand alignment based on ant colony optimization.

Molecular alignment of biologically active compounds is a key technique in ligand-based drug design. Such alignments can be used for similarity-based identification of new compounds by using known active template structures as seeds. Furthermore, alignments allow for the identification of potential pharmacophoric features in a set of chemically divers, biologically active molecules. Here, pharmACOphore is presented, a new approach for pairwise as well as multiple flexible alignments of ligands based on ant colony optimization (ACO) [1]. Translational, rotational and torsional degrees of freedom of all ligand structures are encoded on a pheromone vector (Figure ​(Figure1).1). The artificial ant colony uses this representation to mark favourable values for each degree of freedom by depositing a pheromone trail onto the corresponding pheromone vector. The amount of pheromone deposited directly depends on the solution quality, assessed by the scoring function. Figure 1 The scoring function was parameterized by reproducing reference alignments taken from four different proteins. The performance of the new alignment algorithm will be demonstrated by pairwise alignments obtained for examples from the FlexS data set [2] and by an additional example for multiple flexible alignment.

Are predefined decoy sets of ligand poses able to quantify scoring function accuracy

Due to the large number of different docking programs and scoring functions available, researchers are faced with the problem of selecting the most suitable one when starting a structure-based drug discovery project. To guide the decision process, several studies comparing different docking and scoring approaches have been published. In the context of comparing scoring function performance, it is common practice to use a predefined, computer-generated set of ligand poses (decoys) and to reevaluate their score using the set of scoring functions to be compared. But are predefined decoy sets able to unambiguously evaluate and rank different scoring functions with respect to pose prediction performance? This question arose when the pose prediction performance of our piecewise linear potential derived scoring functions (Korb et al. in J Chem Inf Model 49:84–96, 2009) was assessed on a standard decoy set (Cheng et al. in J Chem Inf Model 49:1079–1093, 2009). While they showed excellent pose identification performance when they were used for rescoring of the predefined decoy conformations, a pronounced degradation in performance could be observed when they were directly applied in docking calculations using the same test set. This implies that on a discrete set of ligand poses only the rescoring performance can be evaluated. For comparing the pose prediction performance in a more rigorous manner, the search space of each scoring function has to be sampled extensively as done in the docking calculations performed here. We were able to identify relative strengths and weaknesses of three scoring functions (ChemPLP, GoldScore, and Astex Statistical Potential) by analyzing the performance for subsets of the complexes grouped by different properties of the active site. However, reasons for the overall poor performance of all three functions on this test set compared to other test sets of similar size could not be identified.