Analysis of data fusion methods in virtual screening: similarity and group fusion

In a recent companion paper we have related the operation of simple data fusion rules used in virtual screening to a multiple integral formalism. In this paper we extend these ideas to the analysis of data fusion methods applied to real data. We examine several cases of similarity fusion using different coefficients and different representations and consider the reasons for positive or negative results in terms of the similarity distributions. Results are obtained using the SUM-, MAX- MIN-, and CombMNZ-fusion rules. We also develop a customized fusion rule, which provides an estimate of the optimal possible result for fusing multiple searches of a specific database; this shows that similarity fusion can, in principle, achieve retrieval enhancements even if this is not achieved in practice with current fusion rules. The methods are extended to analyze the comparatively successful results of group fusion with multiple actives, and we provide a rationale for the observed superiority of the MAX-rule over the SUM-rule in this context.     

New methods for ligand-based virtual screening: use of data fusion and machine learning to enhance the effectiveness of similarity searching

Similarity searching using a single bioactive reference structure is a well-established technique for accessing chemical structure databases. This paper describes two extensions of the basic approach. First, we discuss the use of group fusion to combine the results of similarity searches when multiple reference structures are available. We demonstrate that this technique is notably more effective than conventional similarity searching in scaffold-hopping searches for structurally diverse sets of active molecules; conversely, the technique will do little to improve the search performance if the actives are structurally homogeneous. Second, we make the assumption that the nearest neighbors resulting from a similarity search, using a single bioactive reference structure, are also active and use this assumption to implement approximate forms of group fusion, substructural analysis, and binary kernel discrimination. This approach, called turbo similarity searching, is notably more effective than conventional similarity searching.     

Virtual screening data fusion using both structure- and ligand-based methods

Virtual screening is widely applied in drug discovery, and significant effort has been put into improving current methods. In this study, we have evaluated the performance of compound ranking in virtual screening using five different data fusion algorithms on a total of 16 data sets. The data were generated by docking, pharmacophore search, shape similarity, and electrostatic similarity, spanning both structure- and ligand-based methods. The algorithms used for data fusion were sum rank, rank vote, sum score, Pareto ranking, and parallel selection. None of the fusion methods require any prior knowledge or input other than the results from the single methods and, thus, are readily applicable. The results show that compound ranking using data fusion improves the performance and consistency of virtual screening compared to the single methods alone. The best performing data fusion algorithm was parallel selection, but both rank voting and Pareto ranking also have good performance.     

Analysis and use of fragment-occurrence data in similarity-based virtual screening

Current systems for similarity-based virtual screening use similarity measures in which all the fragments in a fingerprint contribute equally to the calculation of structural similarity. This paper discusses the weighting of fragments on the basis of their frequencies of occurrence in molecules. Extensive experiments with sets of active molecules from the MDL Drug Data Report and the World of Molecular Bioactivity databases, using fingerprints encoding Tripos holograms, Pipeline Pilot ECFC_4 circular substructures and Sunset Molecular keys, demonstrate clearly that frequency-based screening is generally more effective than conventional, unweighted screening. The results suggest that standardising the raw occurrence frequencies by taking the square root of the frequencies will maximise the effectiveness of virtual screening. An upper-bound analysis shows the complex interactions that can take place between representations, weighting schemes and similarity coefficients when similarity measures are computed, and provides a rationalisation of the relative performance of the various weighting schemes.     

VSDMIP: virtual screening data management on an integrated platform

A novel software (VSDMIP) for the virtual screening (VS) of chemical libraries integrated within a MySQL relational database is presented. Two main features make VSDMIP clearly distinguishable from other existing computational tools: (i) its database, which stores not only ligand information but also the results from every step in the VS process, and (ii) its modular and pluggable architecture, which allows customization of the VS stages (such as the programs used for conformer generation or docking), through the definition of a detailed workflow employing user-configurable XML files. VSDMIP, therefore, facilitates the storage and retrieval of VS results, easily adapts to the specific requirements of each method and tool used in the experiments, and allows the comparison of different VS methodologies. To validate the usefulness of VSDMIP as an automated tool for carrying out VS several experiments were run on six protein targets (acetylcholinesterase, cyclin-dependent kinase 2, coagulation factor Xa, estrogen receptor alpha, p38 MAP kinase, and neuraminidase) using nine binary (actives/inactive) test sets. The performance of several VS configurations was evaluated by means of enrichment factors and receiver operating characteristic plots. ángel R. Ortiz deceased on May 5, 2008.     

Comparison of ranking methods for virtual screening in lead-discovery programs

This paper discusses the use of several rank-based virtual screening methods for prioritizing compounds in lead-discovery programs, given a training set for which both structural and bioactivity data are available. Structures from the NCI AIDS data set and from the Syngenta corporate database were represented by two types of fragment bit-string and by sets of high-level molecular features. These representations were processed using binary kernel discrimination, similarity searching, substructural analysis, support vector machine, and trend vector analysis, with the effectiveness of the methods being judged by the extent to which active test set molecules were clustered toward the top of the resultant rankings. The binary kernel discrimination approach yielded consistently superior rankings and would appear to have considerable potential for chemical screening applications.     

Conditional probability: a new fusion method for merging disparate virtual screening results

This paper introduces a new consensus scoring approach for merging the results of different virtual screening methods based on conditional probabilities. The technique is experimentally evaluated using several ligand-based virtual screening methods and compared to two variations of the established Sum-rank fusion method where it performs as well or better than the Sum-rank methods. Our experiments confirm that consensus scoring increases the number of active compounds retrieved with respect to the best individual methods on average.     

Ranking targets in structure-based virtual screening of three-dimensional protein libraries: methods and problems

Structure-based virtual screening is a promising tool to identify putative targets for a specific ligand. Instead of docking multiple ligands into a single protein cavity, a single ligand is docked in a collection of binding sites. In inverse screening, hits are in fact targets which have been prioritized within the pool of best ranked proteins. The target rate depends on specificity and promiscuity in protein-ligand interactions and, to a considerable extent, on the effectiveness of the scoring function, which still is the Achilles' heel of molecular docking. In the present retrospective study, virtual screening of the sc-PDB target library by GOLD docking was carried out for four compounds (biotin, 4-hydroxy-tamoxifen, 6-hydroxy-1,6-dihydropurine ribonucleoside, and methotrexate) of known sc-PDB targets and, several ranking protocols based on GOLD fitness score and topological molecular interaction fingerprint (IFP) comparison were evaluated. For the four investigated ligands, the fusion of GOLD fitness and two IFP scores allowed the recovery of most targets, including the rare proteins which are not readily suitable for statistical analysis, while significantly filtering out most false positive entries. The current survey suggests that selecting a small number of targets (<20) for experimental evaluation is achievable with a pure structure-based approach.     

A theoretical interpretation of 13C screening data in unsaturated molecules

Carbon-13 screening constants are calculated within the INDO/S level of approximation to Pople's model. Satisfactory agreement is obtained in most cases between the calculated and observed screening results. An analysis of the contributions of the π → σ*, σ → π* and σ → σ* transitions to the paramagnetic term shows that a linear relationship between ¹³C chemical shifts and the lowest energy transition is not present. The average excitation energies are found to vary appreciably among the molecules studied.     

Comparison of topological, shape, and docking methods in virtual screening

Virtual screening benchmarking studies were carried out on 11 targets to evaluate the performance of three commonly used approaches: 2D ligand similarity (Daylight, TOPOSIM), 3D ligand similarity (SQW, ROCS), and protein structure-based docking (FLOG, FRED, Glide). Active and decoy compound sets were assembled from both the MDDR and the Merck compound databases. Averaged over multiple targets, ligand-based methods outperformed docking algorithms. This was true for 3D ligand-based methods only when chemical typing was included. Using mean enrichment factor as a performance metric, Glide appears to be the best docking method among the three with FRED a close second. Results for all virtual screening methods are database dependent and can vary greatly for particular targets.     

Toward the discovery of functional transthyretin amyloid inhibitors: application of virtual screening methods

Inhibition of amyloid fibril formation by stabilization of the native form of the protein transthyretin (TTR) is a viable approach for the treatment of familial amyloid polyneuropathy that has been gaining momentum in the field of amyloid research. The TTR stabilizer molecules discovered to date have shown efficacy at inhibiting fibrilization in vitro but display impairing issues of solubility, affinity for TTR in the blood plasma and/or adverse effects. In this study we present a benchmark of four protein- and ligand-based virtual screening (VS) methods for identifying novel TTR stabilizers: (i) two-dimensional (2D) similarity searches with chemical hashed, pharmacophore, and UNITY fingerprints, (ii) 3D searches based on shape, chemical, and electrostatic similarity, (iii) LigMatch, a new ligand-based method which uses multiple templates and combines 3D geometric hashing with a 2D preselection process, and (iv) molecular docking to consensus X-ray crystal structures of TTR. We illustrate the potential of the best-performing VS protocols to retrieve promising new leads by ranking a tailored library of 2.3 million commercially available compounds. Our predictions show that the top-scoring molecules possess distinctive features from the known TTR binders, holding better solubility, fraction of halogen atoms, and binding affinity profiles. To the best of our knowledge, this is the first attempt to rationalize the utilization of a large battery of in silico screening techniques toward the identification of a new generation of TTR amyloid inhibitors.     

Comprehensive comparison of ligand-based virtual screening tools against the DUD data set reveals limitations of current 3D methods

In recent years, many virtual screening (VS) tools have been developed that employ different molecular representations and have different speed and accuracy characteristics. In this paper, we compare ten popular ligand-based VS tools using the publicly available Directory of Useful Decoys (DUD) data set comprising over 100?000 compounds distributed across 40 protein targets. The DUD was developed initially to evaluate docking algorithms, but our results from an operational correlation analysis show that it is also well suited for comparing ligand-based VS tools. Although it is conventional wisdom that 3D molecular shape is an important determinant of biological activity, our results based on permutational significance tests of several commonly used VS metrics show that the 2D fingerprint-based methods generally give better VS performance than the 3D shape-based approaches for surprisingly many of the DUD targets. To help understand this finding, we have analyzed the nature of the scoring functions used and the composition of the DUD data set itself. We propose that to improve the VS performance of current 3D methods, it will be necessary to devise screening queries that can represent multiple possible conformations and which can exploit knowledge of known actives that span multiple scaffold families.     

Current trends in virtual high throughput screening using ligand-based and structure-based methods

High throughput in silico methods have offered the tantalizing potential to drastically accelerate the drug discovery process. Yet despite significant efforts expended by academia, national labs and industry over the years, many of these methods have not lived up to their initial promise of reducing the time and costs associated with the drug discovery enterprise, a process that can typically take over a decade and cost hundreds of millions of dollars from conception to final approval and marketing of a drug. Nevertheless structure-based modeling has become a mainstay of computational biology and medicinal chemistry, helping to leverage our knowledge of the biological target and the chemistry of protein-ligand interactions. While ligand-based methods utilize the chemistry of molecules that are known to bind to the biological target, structure-based drug design methods rely on knowledge of the three-dimensional structure of the target, as obtained through crystallographic, spectroscopic or bioinformatics techniques. Here we review recent developments in the methodology and applications of structure-based and ligand-based methods and target-based chemogenomics in Virtual High Throughput Screening (VHTS), highlighting some case studies of recent applications, as well as current research in further development of these methods. The limitations of these approaches will also be discussed, to give the reader an indication of what might be expected in years to come.     

Optimization of high throughput virtual screening by combining shape-matching and docking methods

Receptor flexibility is a critical issue in structure-based virtual screening methods. Although a multiple-receptor conformation docking is an efficient way to account for receptor flexibility, it is still too slow for large molecular libraries. It was reported that a fast ligand-centric, shape-based virtual screening was more consistent for hit enrichment than a typical single-receptor conformation docking. Thus, we designed a "distributed docking" method that improves virtual high throughput screening by combining a shape-matching method with a multiple-receptor conformation docking. Database compounds are classified in advance based on shape similarities to one of the crystal ligands complexed with the target protein. This classification enables us to pick the appropriate receptor conformation for a single-receptor conformation docking of a given compound, thereby avoiding time-consuming multiple docking. In particular, this approach utilizes cross-docking scores of known ligands to all available receptor structures in order to optimize the algorithm. The present virtual screening method was tested for reidentification of known PPARgamma and p38 MAP kinase active compounds. We demonstrate that this method improves the enrichment while maintaining the computation speed of a typical single-receptor conformation docking.     

Comparison of fingerprint-based methods for virtual screening using multiple bioactive reference structures

Fingerprint-based similarity searching is widely used for virtual screening when only a single bioactive reference structure is available. This paper reviews three distinct ways of carrying out such searches when multiple bioactive reference structures are available: merging the individual fingerprints into a single combined fingerprint; applying data fusion to the similarity rankings resulting from individual similarity searches; and approximations to substructural analysis. Extended searches on the MDL Drug Data Report database suggest that fusing similarity scores is the most effective general approach, with the best individual results coming from the binary kernel discrimination technique.     

FieldScreen: virtual screening using molecular fields. Application to the DUD data set

FieldScreen, a ligand-based Virtual Screening (VS) method, is described. Its use of 3D molecular fields makes it particularly suitable for scaffold hopping, and we have rigorously validated it for this purpose using a clustered version of the Directory of Useful Decoys (DUD). Using thirteen pharmaceutically relevant targets, we demonstrate that FieldScreen produces superior early chemotype enrichments, compared to DOCK. Additionally, hits retrieved by FieldScreen are consistently lower in molecular weight than those retrieved by docking. Where no X-ray protein structures are available, FieldScreen searches are more robust than docking into homology models or apo structures.