Comparing performance of computational tools for combinatorial library design

Virtual Screening Meets Deep Learning

BACKGROUND

Automated compound testing is currently the de facto standard method for drug screening, but it has not brought the great increase in the number of new drugs that was expected. Computer- aided compounds search, known as Virtual Screening, has shown the benefits to this field as a complement or even alternative to the robotic drug discovery. There are different methods and approaches to address this problem and most of them are often included in one of the main screening strategies. Machine learning, however, has established itself as a virtual screening methodology in its own right and it may grow in popularity with the new trends on artificial intelligence.

OBJECTIVE

This paper will attempt to provide a comprehensive and structured review that collects the most important proposals made so far in this area of research. Particular attention is given to some recent developments carried out in the machine learning field: the deep learning approach, which is pointed out as a future key player in the virtual screening landscape.

Screening Library Design

Thanks to technological advances and a greater understanding of the biological and chemical natures of targets and related diseases, high-throughput screening (HTS) has been allowed to be faster, cheaper, and more accessible. Yet, despite these increased technologies and understandings, the frequency of novel and drugs are being approved each year has not being increasing over the years. 2017 was considered a "bumper" year with a total of 46 approved drugs, over double that of the previous year. However, it is thought that part of the problem that HTS has not lived up to expectations is because of the contents of current chemical libraries. Therefore, new methods to design screening libraries are of great interest.

IspE inhibitors identified by a combination of in silico and in vitro high-throughput screening

CDP-ME kinase (IspE) contributes to the non-mevalonate or deoxy-xylulose phosphate (DOXP) pathway for isoprenoid precursor biosynthesis found in many species of bacteria and apicomplexan parasites. IspE has been shown to be essential by genetic methods and since it is absent from humans it constitutes a promising target for antimicrobial drug development. Using in silico screening directed against the substrate binding site and in vitro high-throughput screening directed against both, the substrate and co-factor binding sites, non-substrate-like IspE inhibitors have been discovered and structure-activity relationships were derived. The best inhibitors in each series have high ligand efficiencies and favourable physico-chemical properties rendering them promising starting points for drug discovery. Putative binding modes of the ligands were suggested which are consistent with established structure-activity relationships. The applied screening methods were complementary in discovering hit compounds, and a comparison of both approaches highlights their strengths and weaknesses. It is noteworthy that compounds identified by virtual screening methods provided the controls for the biochemical screens.

Ligand design by a combinatorial approach based on modeling and experiment: application to HLA-DR4

Combinatorial synthesis and large scale screening methods are being used increasingly in drug discovery, particularly for finding novel lead compounds. Although these "random" methods sample larger areas of chemical space than traditional synthetic approaches, only a relatively small percentage of all possible compounds are practically accessible. It is therefore helpful to select regions of chemical space that have greater likelihood of yielding useful leads. When three-dimensional structural data are available for the target molecule this can be achieved by applying structure-based computational design methods to focus the combinatorial library. This is advantageous over the standard usage of computational methods to design a small number of specific novel ligands, because here computation is employed as part of the combinatorial design process and so is required only to determine a propensity for binding of certain chemical moieties in regions of the target molecule. This paper describes the application of the Multiple Copy Simultaneous Search (MCSS) method, an active site mapping and de novo structure-based design tool, to design a focused combinatorial library for the class II MHC protein HLA-DR4. Methods for the synthesizing and screening the computationally designed library are presented; evidence is provided to show that binding was achieved. Although the structure of the protein-ligand complex could not be determined, experimental results including cross-exclusion of a known HLA-DR4 peptide ligand (HA) by a compound from the library. Computational model building suggest that at least one of the ligands designed and identified by the methods described binds in a mode similar to that of native peptides.

Virtual ligand screening: strategies, perspectives and limitations

In contrast to high-throughput screening, in virtual ligand screening (VS), compounds are selected using computer programs to predict their binding to a target receptor. A key prerequisite is knowledge about the spatial and energetic criteria responsible for protein–ligand binding. The concepts and prerequisites to perform VS are summarized here, and explanations are sought for the enduring limitations of the technology. Target selection, analysis and preparation are discussed, as well as considerations about the compilation of candidate ligand libraries. The tools and strategies of a VS campaign, and the accuracy of scoring and ranking of the results, are also considered.

Probing molecular docking in a charged model binding site

A model binding site was used to investigate charge-charge interactions in molecular docking. This simple site, a small (180A(3)) engineered cavity in cyctochrome c peroxidase (CCP), is negatively charged and completely buried from solvent, allowing us to explore the balance between electrostatic energy and ligand desolvation energy in a system where many of the common approximations in docking do not apply. A database with about 5300 molecules was docked into this cavity. Retrospective testing with known ligands and decoys showed that overall the balance between electrostatic interaction and desolvation energy was captured. More interesting were prospective docking scre"ens that looked for novel ligands, especially those that might reveal problems with the docking and energy methods. Based on screens of the 5300 compound database, both high-scoring and low-scoring molecules were acquired and tested for binding. Out of 16 new, high-scoring compounds tested, 15 were observed to bind. All of these were small heterocyclic cations. Binding constants were measured for a few of these, they ranged between 20microM and 60microM. Crystal structures were determined for ten of these ligands in complex with the protein. The observed ligand geometry corresponded closely to that predicted by docking. Several low-scoring alkyl amino cations were also tested and found to bind. The low docking score of these molecules owed to the relatively high charge density of the charged amino group and the corresponding high desolvation penalty. When the complex structures of those ligands were determined, a bound water molecule was observed interacting with the amino group and a backbone carbonyl group of the cavity. This water molecule mitigates the desolvation penalty and improves the interaction energy relative to that of the "naked" site used in the docking screen. Finally, six low-scoring neutral molecules were also tested, with a view to looking for false negative predictions. Whereas most of these did not bind, two did (phenol and 3-fluorocatechol). Crystal structures for these two ligands in complex with the cavity site suggest reasons for their binding. That these neutral molecules do, in fact bind, contradicts previous results in this site and, along with the alkyl amines, provides instructive false negatives that help identify weaknesses in our scoring functions. Several improvements of these are considered.

Here be dragons: docking and screening in an uncharted region of chemical space

To compare virtual and high-throughput screening in an unbiased way, 50,000 compounds were docked into the 3-dimensional structure of dihydrofolate reductase prospectively, and the results were compared to a subsequent experimental screening of the same library. Undertaking these calculations demanded careful database curation and control calculations with annotated inhibitors. These ultimately led to a ranked list of more likely and less likely inhibitors and to the prediction that relatively few inhibitors would be found in the empirical screen. The latter prediction turned out to be correct, with arguably no validated inhibitors found experimentally. Subsequent retesting of high-scoring docked molecules may have found 2 true inhibitors, although this remains uncertain due to experimental ambiguities. The implications of this study for screening campaigns are considered.

The Discovery of Kv1.5 Blockers as a Case Study for the Application of Virtual Screening Approaches

Different virtual screening techniques are available as alternatives to high throughput screening. These different techniques have been rarely used together on the same target. We had the opportunity to do so in order to discover novel blockers of the voltage-dependent potassium channel Kv1.5, a potential target for the treatment of atrial fibrillation. Our corporate database was searched, using a protein-based pharmacophore, derived from a homology model, as query. As a result, 244 molecules were screened in vitro, 19 of them (7.8%) were found to be active. Five of them, belonging to five different chemical classes, exhibited IC50 values under 10 microM. The performance of this structure-based virtual screening protocol has been compared with those of similarity and ligand-based pharmacophore searches. The analysis of the results supports the conventional wisdom of using as many virtual screening techniques as possible in order to maximize the chance of finding as many chemotypes as possible.

Evaluation of library ranking efficacy in virtual screening

We present the results of a comprehensive study in which we explored how the docking procedure affects the performance of a virtual screening approach. We used four docking engines and applied 10 scoring functions to the top-ranked docking solutions of seeded databases against six target proteins. The scores of the experimental poses were placed within the total set to assess whether the scoring function required an accurate pose to provide the appropriate rank for the seeded compounds. This method allows a direct comparison of library ranking efficacy. Our results indicate that the LigandFit/Ligscore1 and LigandFit/GOLD docking/scoring combinations, and to a lesser degree FlexX/FlexX, Glide/Ligscore1, DOCK/PMF (Tripos implementation), LigandFit1/Ligscore2 and LigandFit/PMF (Tripos implementation) were able to retrieve the highest number of actives at a 10% fraction of the database when all targets were looked upon collectively. We also show that the scoring functions rank the observed binding modes higher than the inaccurate poses provided that the experimental poses are available. This finding stresses the discriminatory ability of the scoring algorithms, when better poses are available, and suggests that the number of false positives can be lowered with conformers closer to bioactive ones.

Integrating cheminformatic analysis in combinatorial chemistry

Cheminformatic analysis of drug-related compound databases has enabled the identification of the physicochemical properties that have the greatest influence on determining the drug-like characteristics of a compound. This enables definition of the parameters and profiles used in constructing a high-quality combinatorial library. Awareness of the multi-objective nature of combinatorial library construction has also given rise to techniques designed to enhance the likelihood of including the best compounds in a given library.