Diversity selection algorithms

An Open-Source Implementation of the Scaffold Identification and Naming System (SCINS) and Example Applications

Organizing and partitioning sets of chemical structures is of considerable practical significance, e.g., in compound library analysis and the postprocessing of screening hit lists. Approaches such as unsupervised clustering are computationally demanding and dataset-dependent; on the other hand, rule-based methods, such as those based on Murcko scaffolds, have linear time complexity but are often too fine-grained, leading to a large number of singletons or sparsely populated classes. An alternative rule-based method that seeks to achieve an optimal balance when grouping compounds into sets is the 'Scaffold Identification and Naming System' (SCINS). To facilitate public use of this previously published method, here, we provide an open-source Python implementation of SCINS, dependent only on RDKit. We show that SCINS can be useful in identifying sparsely and densely populated regions in chemical space in large databases, here exemplified with Enamine REAL Diverse and ChEMBL. We find that Enamine REAL Diverse covers a much smaller SCINS space relative to ChEMBL, whereas the opposite is true when Murcko and generic Murcko scaffolds are considered. Additionally, we show that SCINS can result in chemically intuitive grouping of medium-sized sets of bioactive compounds, which can be useful in compound selection from virtual screening campaigns as well as postprocessing of experimental hit lists. Hence, in this work, we provide both an open-source implementation of SCINS and its characterization with relevant use cases.

Selecting molecules with diverse structures and properties by maximizing submodular functions of descriptors learned with graph neural networks

Selecting diverse molecules from unexplored areas of chemical space is one of the most important tasks for discovering novel molecules and reactions. This paper proposes a new approach for selecting a subset of diverse molecules from a given molecular list by using two existing techniques studied in machine learning and mathematical optimization: graph neural networks (GNNs) for learning vector representation of molecules and a diverse-selection framework called submodular function maximization. Our method, called SubMo-GNN, first trains a GNN with property prediction tasks, and then the trained GNN transforms molecular graphs into molecular vectors, which capture both properties and structures of molecules. Finally, to obtain a subset of diverse molecules, we define a submodular function, which quantifies the diversity of molecular vectors, and find a subset of molecular vectors with a large submodular function value. This can be done efficiently by using the greedy algorithm, and the diversity of selected molecules measured by the submodular function value is mathematically guaranteed to be at least 63% of that of an optimal selection. We also introduce a new evaluation criterion to measure the diversity of selected molecules based on molecular properties. Computational experiments confirm that our SubMo-GNN successfully selects diverse molecules from the QM9 dataset regarding the property-based criterion, while performing comparably to existing methods regarding standard structure-based criteria. We also demonstrate that SubMo-GNN with a GNN trained on the QM9 dataset can select diverse molecules even from other MoleculeNet datasets whose domains are different from the QM9 dataset. The proposed method enables researchers to obtain diverse sets of molecules for discovering new molecules and novel chemical reactions, and the proposed diversity criterion is useful for discussing the diversity of molecular libraries from a new property-based perspective.

Multi-objective optimization methods in novel drug design

Introduction: In multi-objective drug design, optimization gains importance, being upgraded to a discipline that attracts its own research. Current strategies are broadly classified into single - objective optimization (SOO) and multi-objective optimization (MOO).Areas covered: Starting with SOO and the ways used to incorporate multiple criteria into it, the present review focuses on MOO techniques, their comparison, advantages, and restrictions. Pareto analysis and the concept of dominance stand in the core of MOO. The Pareto front, Pareto ranking, and limitations of Pareto-based methods, due to high dimensions and data uncertainty, are outlined. Desirability functions and the weighted sum approaches are described as stand-alone techniques to transform the MOO problem to SOO or in combination with pareto analysis and evolutionary algorithms. Representative applications in different drug research areas are also discussed.Expert opinion: Despite their limitations, the use of combined MOO techniques, as well as being complementary to SOO or in conjunction with artificial intelligence, contributes dramatically to efficient drug design, assisting decisions and increasing success probabilities. For multi-target drug design, optimization is supported by network approaches, while applicability of MOO to other fields like drug technology or biological complexity opens new perspectives in the interrelated fields of medicinal chemistry and molecular biology.

The Symbiotic Relationship Between Drug Discovery and Organic Chemistry

All pharmaceutical products contain organic molecules; the source may be a natural product or a fully synthetic molecule, or a combination of both. Thus, it follows that organic chemistry underpins both existing and upcoming pharmaceutical products. The reverse relationship has also affected organic synthesis, changing its landscape towards increasingly complex targets. This Review article sets out to give a concise appraisal of this symbiotic relationship between organic chemistry and drug discovery, along with a discussion of the design concepts and highlighting key milestones along the journey. In particular, criteria for a high-quality compound library design enabling efficient virtual navigation of chemical space, as well as rise and fall of concepts for its synthetic exploration (such as combinatorial chemistry; diversity-, biology-, lead-, or fragment-oriented syntheses; and DNA-encoded libraries) are critically surveyed.

How Size Matters: Diversity for Fragment Library Design

Fragment-based drug discovery (FBDD) has become a major strategy to derive novel lead candidates for various therapeutic targets, as it promises efficient exploration of chemical space by employing fragment-sized (MW < 300) compounds. One of the first challenges in implementing a FBDD approach is the design of a fragment library, and more specifically, the choice of its size and individual members. A diverse set of fragments is required to maximize the chances of discovering novel hit compounds. However, the exact diversity of a certain collection of fragments remains underdefined, which hinders direct comparisons among different selections of fragments. Based on structural fingerprints, we herein introduced quantitative metrics for the structural diversity of fragment libraries. Structures of commercially available fragments were retrieved from the ZINC database, from which libraries with sizes ranging from 100 to 100,000 compounds were selected. The selected libraries were evaluated and compared quantitatively, resulting in interesting size-diversity relationships. Our results demonstrated that while library size does matter for its diversity, there exists an optimal size for structural diversity. It is also suggested that such quantitative measures can guide the design of diverse fragment libraries under different circumstances.

GRID-independent molecular descriptor analysis and molecular docking studies to mimic the binding hypothesis of γ-aminobutyric acid transporter 1 (GAT1) inhibitors

Background

The γ-aminobutyric acid (GABA) transporter GAT1 is involved in GABA transport across the biological membrane in and out of the synaptic cleft. The efficiency of this Na⁺ coupled GABA transport is regulated by an electrochemical gradient, which is directed inward under normal conditions. However, in certain pathophysiological situations, including strong depolarization or an imbalance in ion homeostasis, the GABA influx into the cytoplasm is increased by re-uptake transport mechanism. This mechanism may lead to extra removal of extracellular GABA which results in numerous neurological disorders such as epilepsy. Thus, small molecule inhibitors of GABA re-uptake may enhance GABA activity at the synaptic clefts.

Methods

In the present study, various GRID-independent molecular descriptor (GRIND) models have been developed to shed light on the 3D structural features of human GAT1 (hGAT1) inhibitors using nipecotic acid and N-diarylalkenyl piperidine analogs. Further, a binding hypothesis has been developed for the selected GAT1 antagonists by molecular docking inside the binding cavity of hGAT1 homology model.

Results

Our results indicate that two hydrogen bond acceptors, one hydrogen bond donor and one hydrophobic region at certain distances from each other play an important role in achieving high inhibitory potency against hGAT1. Our docking results elucidate the importance of the COOH group in hGAT1 antagonists by considering substitution of the COOH group with an isoxazol ring in compound 37, which subsequently leads to a three order of magnitude decrease in biological activity of 37 (IC₅₀ = 38 µM) as compared to compound 1 (IC₅₀ = 0.040 µM).

Discussion

Our docking results are strengthened by the structure activity relationship of the data series as well as by GRIND models, thus providing a significant structural basis for understanding the binding of antagonists, which may be useful for guiding the design of hGAT1 inhibitors.

A 2D-QSAR and Grid-Independent Molecular Descriptor (GRIND) Analysis of Quinoline-Type Inhibitors of Akt2: Exploration of the Binding Mode in the Pleckstrin Homology (PH) Domain

Protein kinase B-β (PKBβ/Akt2) is a serine/threonine-specific protein kinase that has emerged as one of the most important regulators of cell growth, differentiation, and division. Upregulation of Akt2 in various human carcinomas, including ovarian, breast, and pancreatic, is a well-known tumorigenesis phenomenon. Early on, the concept of the simultaneous administration of anticancer drugs with inhibitors of Akt2 was advocated to overcome cell proliferation in the chemotherapeutic treatment of cancer. However, clinical studies have not lived up to the high expectations, and several phase II and phase III clinical studies have been terminated prematurely because of severe side effects related to the non-selective isomeric inhibition of Akt2. The notion that the sequence identity of pleckstrin homology (PH) domains within Akt-isoforms is less than 30% might indicate the possibility of the development of selective antagonists against the Akt2 PH domain. Therefore, in this study, various in silico tools were utilized to explore the hypothesis that quinoline-type inhibitors bind in the Akt2 PH domain. A Grid-Independent Molecular Descriptor (GRIND) analysis indicated that two hydrogen bond acceptors, two hydrogen bond donors and one hydrophobic feature at a certain distance from each other were important for the selective inhibition of Akt2. Our docking results delineated the importance of Lys30 as an anchor point for mapping the distances of important amino acid residues in the binding pocket, including Lys14, Glu17, Arg25, Asn53, Asn54 and Arg86. The binding regions identified complement the GRIND-based pharmacophoric features.

Plate-based diversity subset screening generation 2: an improved paradigm for high-throughput screening of large compound files

High-throughput screening (HTS) is an effective method for lead and probe discovery that is widely used in industry and academia to identify novel chemical matter and to initiate the drug discovery process. However, HTS can be time consuming and costly and the use of subsets as an efficient alternative to screening entire compound collections has been investigated. Subsets may be selected on the basis of chemical diversity, molecular properties, biological activity diversity or biological target focus. Previously, we described a novel form of subset screening: plate-based diversity subset (PBDS) screening, in which the screening subset is constructed by plate selection (rather than individual compound cherry-picking), using algorithms that select for compound quality and chemical diversity on a plate basis. In this paper, we describe a second-generation approach to the construction of an updated subset: PBDS2, using both plate and individual compound selection, that has an improved coverage of the chemical space of the screening file, whilst only selecting the same number of plates for screening. We describe the validation of PBDS2 and its successful use in hit and lead discovery. PBDS2 screening became the default mode of singleton (one compound per well) HTS for lead discovery in Pfizer.

A reliable computational workflow for the selection of optimal screening libraries

Background

The experimental screening of compound collections is a common starting point in many drug discovery projects. Successes of such screening campaigns critically depend on the quality of the screened library. Many libraries are currently available from different vendors yet the selection of the optimal screening library for a specific project is challenging. We have devised a novel workflow for the rational selection of project-specific screening libraries.

Results

The workflow accepts as input a set of virtual candidate libraries and applies the following steps to each library: (1) data curation; (2) assessment of ADME/T profile; (3) assessment of the number of promiscuous binders/frequent HTS hitters; (4) assessment of internal diversity; (5) assessment of similarity to known active compound(s) (optional); (6) assessment of similarity to in-house or otherwise accessible compound collections (optional). For ADME/T profiling, Lipinski’s and Veber’s rule-based filters were implemented and a new blood brain barrier permeation model was developed and validated (85 and 74 % success rate for training set and test set, respectively). Diversity and similarity descriptors which demonstrated best performances in terms of their ability to select either diverse or focused sets of compounds from three databases (Drug Bank, CMC and CHEMBL) were identified and used for diversity and similarity assessments. The workflow was used to analyze nine common screening libraries available from six vendors. The results of this analysis are reported for each library providing an assessment of its quality. Furthermore, a consensus approach was developed to combine the results of these analyses into a single score for selecting the optimal library under different scenarios.

Conclusions

We have devised and tested a new workflow for the rational selection of screening libraries under different scenarios. The current workflow was implemented using the Pipeline Pilot software yet due to the usage of generic components, it can be easily adapted and reproduced by computational groups interested in rational selection of screening libraries. Furthermore, the workflow could be readily modified to include additional components. This workflow has been routinely used in our laboratory for the selection of libraries in multiple projects and consistently selects libraries which are well balanced across multiple parameters.Graphical abstract

Similarity-driven discovery of zeolite materials for adsorption-based separations

Crystalline porous materials can be exploited in many applications. Discovery of materials with optimum adsorption properties typically involves expensive brute-force characterization of large sets of materials. An alternative approach based on similarity searching that enables discovery of materials with optimum adsorption for CO(2) and other molecules at a fraction of the cost of brute-force characterization is demonstrated.