Comparing Molecular Patterns Using the Example of SMARTS: Applications and Filter Collection Analysis

Tackling assay interference associated with small molecules

Biochemical and cell-based assays are essential to discovering and optimizing efficacious and safe drugs, agrochemicals and cosmetics. However, false assay readouts stemming from colloidal aggregation, chemical reactivity, chelation, light signal attenuation and emission, membrane disruption, and other interference mechanisms remain a considerable challenge in screening synthetic compounds and natural products. To address assay interference, a range of powerful experimental approaches are available and in silico methods are now gaining traction. This Review begins with an overview of the scope and limitations of experimental approaches for tackling assay interference. It then focuses on theoretical methods, discusses strategies for their integration with experimental approaches, and provides recommendations for best practices. The Review closes with a summary of the critical facts and an outlook on potential future developments.

Enhanced Calculation of Property Distributions in Chemical Fragment Spaces

Chemical fragment spaces exceed traditional virtual compound libraries by orders of magnitude, making them ideal search spaces for drug design projects. However, due to their immense size, they are not compatible with traditional analysis and search algorithms that rely on the enumeration of molecules. In this paper, we present SpaceProp2, an evolution of the SpaceProp algorithm, which enables the calculation of exact property distributions for chemical fragment spaces without enumerating them. We extend the original algorithm by the capabilities to compute distributions for the TPSA, the number of rotatable bonds, and the occurrence of user-defined molecular structures in the form of SMARTS patterns. Furthermore, SpaceProp2 produces example molecules for every property bin, enabling a detailed interpretation of the distributions. We demonstrate SpaceProp2 on six established make-on-demand chemical fragment spaces as well as BICLAIM, the in-house fragment space of Boehringer Ingelheim. The possibility to search multiple SMARTS patterns simultaneously as well as the produced example molecules offers previously impossible insights into the composition of these vast combinatorial molecule collections, making it an ideal tool for the analysis and design of chemical fragment spaces.

IMPPAT 2.0: An Enhanced and Expanded Phytochemical Atlas of Indian Medicinal Plants

Compilation, curation, digitization, and exploration of the phytochemical space of Indian medicinal plants can expedite ongoing efforts toward natural product and traditional knowledge based drug discovery. To this end, we present IMPPAT 2.0, an enhanced and expanded database compiling manually curated information on 4010 Indian medicinal plants, 17,967 phytochemicals, and 1095 therapeutic uses. Notably, IMPPAT 2.0 compiles associations at the level of plant parts and provides a FAIR-compliant nonredundant in silico stereo-aware library of 17,967 phytochemicals from Indian medicinal plants. The phytochemical library has been annotated with several useful properties to enable easier exploration of the chemical space. We have also filtered a subset of 1335 drug-like phytochemicals of which majority have no similarity to existing approved drugs. Using cheminformatics, we have characterized the molecular complexity and molecular scaffold based structural diversity of the phytochemical space of Indian medicinal plants and performed a comparative analysis with other chemical libraries. Altogether, IMPPAT 2.0 is a manually curated extensive phytochemical atlas of Indian medicinal plants that is accessible at https://cb.imsc.res.in/imppat/.

A Current Encyclopedia of Bioinformatics Tools, Data Formats and Resources for Mass Spectrometry Lipidomics

Mass spectrometry is a widely used technology to identify and quantify biomolecules such as lipids, metabolites and proteins necessary for biomedical research. In this study, we catalogued freely available software tools, libraries, databases, repositories and resources that support lipidomics data analysis and determined the scope of currently used analytical technologies. Because of the tremendous importance of data interoperability, we assessed the support of standardized data formats in mass spectrometric (MS)-based lipidomics workflows. We included tools in our comparison that support targeted as well as untargeted analysis using direct infusion/shotgun (DI-MS), liquid chromatography−mass spectrometry, ion mobility or MS imaging approaches on MS¹ and potentially higher MS levels. As a result, we determined that the Human Proteome Organization-Proteomics Standards Initiative standard data formats, mzML and mzTab-M, are already supported by a substantial number of recent software tools. We further discuss how mzTab-M can serve as a bridge between data acquisition and lipid bioinformatics tools for interpretation, capturing their output and transmitting rich annotated data for downstream processing. However, we identified several challenges of currently available tools and standards. Potential areas for improvement were: adaptation of common nomenclature and standardized reporting to enable high throughput lipidomics and improve its data handling. Finally, we suggest specific areas where tools and repositories need to improve to become FAIRer.

The Torsion Library: Semiautomated Improvement of Torsion Rules with SMARTScompare

The Torsion Library is a collection of torsion motifs associated with angle distributions, derived from crystallographic databases. It is used in strain assessment, conformer generation, and geometry optimization. A hierarchical structure of expert curated SMARTS defines the chemical environments of rotatable bonds and associates these with preferred angles. SMARTS can be very complex and full of implications, which make them difficult to maintain manually. Recent developments in automatically comparing SMARTS patterns can be applied to the Torsion Library to ensure its correctness. We specifically discuss the implementation and the limits of such a procedure in the context of torsion motifs and show several examples of how the Torsion Library benefits from this. All automated changes are validated manually and then shown to have an effect on the angle distributions by correcting matching behavior. The corrected Torsion Library itself is available including both PDB as well as CSD histograms in the Supporting Information and can be used to evaluate rotatable bonds at https://torsions.zbh.uni-hamburg.de.

Influence of Template Size, Canonicalization, and Exclusivity for Retrosynthesis and Reaction Prediction Applications

Heuristic and machine learning models for rank-ordering reaction templates comprise an important basis for computer-aided organic synthesis regarding both product prediction and retrosynthetic pathway planning. Their viability relies heavily on the quality and characteristics of the underlying template database. With the advent of automated reaction and template extraction software and consequently the creation of template databases too large for manual curation, a data-driven approach to assess and improve the quality of template sets is needed. We therefore systematically studied the influence of template generality, canonicalization, and exclusivity on the performance of different template ranking models. We find that duplicate and nonexclusive templates, i.e., templates which describe the same chemical transformation on identical or overlapping sets of molecules, decrease both the accuracy of the ranking algorithm and the applicability of the respective top-ranked templates significantly. To remedy the negative effects of nonexclusivity, we developed a general and computationally efficient framework to deduplicate and hierarchically correct templates. As a result, performance improved considerably for both heuristic and machine learning template ranking models, as well as multistep retrosynthetic planning models. The canonicalization and correction code is made freely available.

Maximum Common Substructure Searching in Combinatorial Make-on-Demand Compound Spaces

Commercial make-on-demand compound spaces have become increasingly popular within the past few years. Since these libraries are too large for enumeration, they are usually accessed using combinatorial fragment space technologies like FTrees-FS and SpaceLight. Although both search types are of high practical impact, they lack the ability to search for precise structural features on the atomic level. To address this important use case, we developed SpaceMACS enabling efficient and precise maximum common induced substructure (MCIS) similarity and substructure searches within chemical fragment spaces. SpaceMACS enumerates a user-defined number of compounds in a multistep procedure. First, substructures of the query are extracted and matched to all fragments of the space. Then partial results are combined to actual compounds of the space. In this way, SpaceMACS identifies common substructures even if they cross fragment borders. We applied SpaceMACS on three commercial fragment spaces searching for the 150 000 most similar analogs to a glucosyltransferase binder from literature. We were able to find almost all building blocks used for the synthesis of the 90 listed analogs and a plethora of additional results. SpaceMACS is the missing link to enable rational drug discovery on make-on-demand combinatorial catalogs. No matter whether initial compound suggestions come from a de novo design, an AI-based compound generation, or a medicinal chemist's drawing board, the method gives access to the structurally closest chemically available analogs in seconds to at most minutes.

Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems

Machine learning models are poised to make a transformative impact on chemical sciences by dramatically accelerating computational algorithms and amplifying insights available from computational chemistry methods. However, achieving this requires a confluence and coaction of expertise in computer science and physical sciences. This Review is written for new and experienced researchers working at the intersection of both fields. We first provide concise tutorials of computational chemistry and machine learning methods, showing how insights involving both can be achieved. We follow with a critical review of noteworthy applications that demonstrate how computational chemistry and machine learning can be used together to provide insightful (and useful) predictions in molecular and materials modeling, retrosyntheses, catalysis, and drug design.

Disconnected Maximum Common Substructures under Constraints

The maximum common substructure (MCS) problem is an important, well-studied problem in cheminformatics. It is applied in several application scenarios like molecule superimposition and scaffold detection or as a similarity measure in virtual screening and clustering. In many cases, the connected MCS is preferred since it is faster to calculate and a highly fragmented MCS is not very meaningful from a chemical point of view. Nevertheless, a disconnected MCS (dMCS) can be very instructive if it consists of reasonably sized molecular parts connected by variable groups. We present a new algorithm named RIMACS, which is able to calculate the dMCS under constraints. We can control the maximum number of connected components and their minimal size using a modified local substructure mapping approach. A formal proof of correctness is provided as well as extended runtime evaluations on chemical data. The evaluation of RIMACS shows that a small number of connected components helps us to improve MCS similarity in a meaningful way while keeping the runtime requirements in a reasonable range.

Evolution of Novartis' Small Molecule Screening Deck Design

This article summarizes the evolution of the screening deck at the Novartis Institutes for BioMedical Research (NIBR). Historically, the screening deck was an assembly of all available compounds. In 2015, we designed a first deck to facilitate access to diverse subsets with optimized properties. We allocated the compounds as plated subsets on a 2D grid with property based ranking in one dimension and increasing structural redundancy in the other. The learnings from the 2015 screening deck were applied to the design of a next generation in 2019. We found that using traditional leadlikeness criteria (mainly MW, clogP) reduces the hit rates of attractive chemical starting points in subset screening. Consequently, the 2019 deck relies on solubility and permeability to select preferred compounds. The 2019 design also uses NIBR's experimental assay data and inferred biological activity profiles in addition to structural diversity to define redundancy across the compound sets.