Visualization Of Chemical Space Research Articles

Navigating pharmacophore space to identify activity discontinuities: a case study with BCR-ABL.

The exploration of chemical space is a fundamental aspect of chemoinformatics, particularly when one explores a large compound data set to relate chemical structures with molecular properties. In this study, we extend our previous work on chemical space visualization at the pharmacophoric level. Instead of using conventional binary classification of affinity (active vs inactive), we introduce a refined approach that categorizes compounds into four distinct classes based on their activity levels: super active, very active, active, and inactive. This classification enriches the color scheme applied to pharmacophore space, where the color representation of a pharmacophore hypothesis is driven by the associated compounds. Using the BCR-ABL tyrosine kinase as a case study, we identified intriguing regions corresponding to pharmacophore activity discontinuities, providing valuable insights for structure-activity relationships analysis.

Exploring the chemical space of BRAF Inhibitors: A cheminformatic and Machine learning analysis

BRAF kinase inhibitors have shown promise in treating melanoma patients with BRAF-V600 mutations. However, their clinical benefits are limited due to the development of acquired resistance. To address this issue, we conducted a systematic cheminformatic analysis and machine learning modeling to study the chemical space, scaffolds, structure–activity relationship, and landscape of human BRAF inhibitors. The final dataset comprised 3,952 molecules. Physicochemical property visualization for chemical space visualization has demonstrated that molecules from Group 1 (potent/active class) generally have slightly higher MW, RB, NumHAcceptors, and TPSA than molecules from Group 2 (intermediate/inactive class). The principal component analysis (PCA) shows that Group 1 data spreads widely and overlaps with the other group, which occupies only the left area on the plot. This suggests that molecules’ chemical structure or scaffolds are more diverse for Group 1 (Potent and Active) compounds than for Group 2 (Intermediate and Inactive) compounds. Murcko scaffold analysis has shown a greater scaffold diversity in the Active, Intermediate, and Inactive classes than in the Potent class. However, in the four bioactivity-defined classes for our BRAF dataset, molecules are uniformly distributed, with a small number of highly populated scaffolds and a large number of singletons. Furthermore, scaffold visualization has identified 12 representative Murcko scaffolds. Scaffolds 1 S1C1, S2C1, S2C2, S3C2, and S4C2 are highly favorable due to their high scaffold enrichment factor values. Based on scaffold analysis, the study investigated and summarized the local structure–activity relationships (SARs). In addition, the global SAR landscape was explored through quantitative structure–activity relationship (QSAR) modeling and structure–activity landscape visualization. Out of a total of 14 candidate models for BRAF inhibitors, a QSAR classification model that includes all 3952 molecules has been identified as the best model. The model was built using the PubChem fingerprint and extra trees algorithm and achieved an accuracy of 0.920 for the training set, 0.699 for the 10-fold cross-validation set, and 0.733 for the test set. Through a detailed analysis of the structure–activity landscapes, a total of sixteen significant consensus activity cliff (AC) generators were identified (ChEMBL molecule IDs: 4795335, 1822247, 3665863, 5083036, 5086749, 4061139, 1822242, 3665859, 1822250, 3641129, 3641043, 514688,3641178, 3697934, 3641046, and 368991). These generators offer valuable information on the SAR (Structure-Activity Relationship) for medicinal chemistry. The findings from this study provide new insights and guidelines for hit identification and lead optimization in developing novel BRAF inhibitors.

An efficient framework to design near-IR monomers for polymer solar cells with the help of machine learning, virtual screening and chemical space visualization

Organic solar cells (OSCs) have emerged as attractive alternatives for capturing solar energy. In the research work, the information from literature was collected to create a database of 700 organic semiconductors and machine learning (ML) was performed. Numerous organic molecules were designed. The chosen monomers were then subjected to cluster analysis. The silhoutte data including t-SNE data plot predict similar and dissimilar data points. In order to determine the structural diversity and similarity of the chosen monomers, chemical similarity analysis using RDkit is employed to visualize the chemical space. The hierarchical relationship and graphical representation between several groups of data was depicted using cluster graph and heat map. Finally, the synthesis accessibility of designed monomers was analyzed using Synthetic Bayesian classifier (SYBA). Through the integration of machine learning, cluster analysis, and synthetic accessibility, the methodology offers a productive and organized framework for designing polymers for OSCs.

Generation of Chemical Space of Compounds for Prostate Cancer Treatment: Biological Activity Prediction, Clustering, and Visualization of Chemical Space.

Designing molecules for pharmaceutical purposes has been a significant focus for several decades. The pursuit of novel drugs is an arduous and financially demanding undertaking. Nevertheless, the integration of computer-assisted frameworks presents a swift avenue for designing and screening drug-like compounds. Within the context of this research, we introduce a comprehensive approach for the design and screening of compounds tailored to the treatment of prostate cancer. To forecast the biological activity of these compounds, we employed machine learning (ML) models. Additionally, an automated process involving the deconstruction and reconstruction of molecular building blocks leads to the generation of novel compounds. Subsequently, the ML models were utilized to predict the biological activity of the designed compounds, and the t-SNE method was employed to visualize the chemical space covered by the novel compounds. A meticulous selection process identified the most promising compounds, and their potential for synthesis was assessed, offering valuable guidance to experimental chemists in their investigative endeavors. Furthermore, fingerprint and heatmap analysis were conducted to evaluate the chemical similarity among the selected compounds. This multifaceted approach, encompassing predictive modeling, compound generation, visualization, and similarity assessment, underscores our commitment to refining the process of identifying potential candidates for further exploration in prostate cancer treatment.

Sampling and Mapping Chemical Space with Extended Similarity Indices.

Visualization of the chemical space is useful in many aspects of chemistry, including compound library design, diversity analysis, and exploring structure-property relationships, to name a few. Examples of notable research areas where the visualization of chemical space has strong applications are drug discovery and natural product research. However, the sheer volume of even comparatively small sub-sections of chemical space implies that we need to use approximations at the time of navigating through chemical space. ChemMaps is a visualization methodology that approximates the distribution of compounds in large datasets based on the selection of satellite compounds that yield a similar mapping of the whole dataset when principal component analysis on a similarity matrix is performed. Here, we show how the recently proposed extended similarity indices can help find regions that are relevant to sample satellites and reduce the amount of high-dimensional data needed to describe a library's chemical space.

Cheminformatic Analysis and Machine Learning Modeling to Investigate Androgen Receptor Antagonists to Combat Prostate Cancer.

Prostate cancer (PCa) is a major leading cause of mortality of cancer among males. There have been numerous studies to develop antagonists against androgen receptor (AR), a crucial therapeutic target for PCa. This study is a systematic cheminformatic analysis and machine learning modeling to study the chemical space, scaffolds, structure-activity relationship, and landscape of human AR antagonists. There are 1678 molecules as final data sets. Chemical space visualization by physicochemical property visualization has demonstrated that molecules from the potent/active class generally have a mildly smaller molecular weight (MW), octanol-water partition coefficient (log P), number of hydrogen-bond acceptors (nHA), number of rotatable bonds (nRot), and topological polar surface area (TPSA) than molecules from intermediate/inactive class. The chemical space visualization in the principal component analysis (PCA) plot shows significant overlapping distributions between potent/active class molecules and intermediate/inactive class molecules; potent/active class molecules are intensively distributed, while intermediate/inactive class molecules are widely and sparsely distributed. Murcko scaffold analysis has shown low scaffold diversity in general, and scaffold diversity of potent/active class molecules is even lower than intermediate/inactive class molecules, indicating the necessity for developing molecules with novel scaffolds. Furthermore, scaffold visualization has identified 16 representative Murcko scaffolds. Among them, scaffolds 1, 2, 3, 4, 7, 8, 10, 11, 15, and 16 are highly favorable scaffolds due to their high scaffold enrichment factor values. Based on scaffold analysis, their local structure-activity relationships (SARs) were investigated and summarized. In addition, the global SAR landscape was explored by quantitative structure-activity relationship (QSAR) modelings and structure-activity landscape visualization. A QSAR classification model incorporating all of the 1678 molecules stands out as the best model from a total of 12 candidate models for AR antagonists (built on PubChem fingerprint, extra trees algorithm, accuracy for training set: 0.935, 10-fold cross-validation set: 0.735 and test set: 0.756). Deeper insights into the structure-activity landscape highlighted a total of seven significant activity cliff (AC) generators (ChEMBL molecule IDs: 160257, 418198, 4082265, 348918, 390728, 4080698, and 6530), which provide valuable SAR information for medicinal chemistry. The findings in this study provide new insights and guidelines for hit identification and lead optimization for the development of novel AR antagonists.

Scaffold and Structural Diversity of the Secondary Metabolite Space of Medicinal Fungi.

Medicinal fungi, including mushrooms, have well-documented therapeutic uses. In this study, we perform a cheminformatics-based investigation of the scaffold and structural diversity of the secondary metabolite space of medicinal fungi and, moreover, perform a detailed comparison with approved drugs, other natural product libraries, and semi-synthetic libraries. We find that the secondary metabolite space of medicinal fungi has similar or higher scaffold diversity in comparison to other natural product libraries analyzed here. Notably, 94% of the scaffolds in the secondary metabolite space of medicinal fungi are not present in the approved drugs. Further, we find that the secondary metabolites, on the one hand, are structurally far from the approved drugs, while, on the other hand, they are close in terms of molecular properties to the approved drugs. Lastly, chemical space visualization using dimensionality reduction methods showed that the secondary metabolite space has minimal overlap with the approved drug space. In a nutshell, our results underscore that the secondary metabolite space of medicinal fungi is a valuable resource for identifying potential lead molecules for natural product-based drug discovery.

Molecular Similarity for Drug Discovery, Target Prediction and Chemical Space Visualization

Similar drug molecules often have similar properties and activities. Therefore, quantifying molecular similarity is central to drug discovery and optimization. Here I review computational methods using molecular similarity measures developed in my group within the interdisciplinary network NCCR TransCure investigating the physiology, structural biology and pharmacology of ion channels and membrane transporters. We designed a 3D molecular shape and pharmacophore comparison algorithm to optimize weak and unselective inhibitors by scaffold hopping and discovered potent and selective inhibitors of the ion channels TRPV6 and TRPM4, of endocannabinoid membrane transport, and of the divalent metal transporters DMT1 and ZIP8. We predicted off-target effects by combining molecular similarity searches from different molecular fingerprints against target annotated compounds from the ChEMBL database. Finally, we created interactive chemical space maps reflecting molecular similarities to facilitate the selection of screening compounds and the analysis of screening results. These different tools are available online at https://gdb.unibe.ch/tools/.

Optimization of LC-MS2 Data Acquisition Parameters for Molecular Networking Applied to Marine Natural Products.

Since the introduction of the online open-source GNPS, molecular networking has quickly become a widely applied tool in the field of natural products chemistry, with applications from dereplication, genome mining, metabolomics, and visualization of chemical space. Studies have shown that data dependent acquisition (DDA) parameters affect molecular network topology but are limited in the number of parameters studied. With an aim to optimize LC-MS2 parameters for integrating GNPS-based molecular networking into our marine natural products workflow, a design of experiment (DOE) was used to screen the significance of the effect that eleven parameters have on both Classical Molecular Networking workflow (CLMN) and the new Feature-Based Molecular Networking workflow (FBMN). Our results indicate that four parameters (concentration, run duration, collision energy and number of precursors per cycle) are the most significant data acquisition parameters affecting the network topology. While concentration and the LC duration were found to be the two most important factors to optimize for CLMN, the number of precursors per cycle and collision energy were also very important factors to optimize for FBMN.

Extended continuous similarity indices: theory and application for QSAR descriptor selection.

Extended (or n-ary) similarity indices have been recently proposed to extend the comparative analysis of binary strings. Going beyond the traditional notion of pairwise comparisons, these novel indices allow comparing any number of objects at the same time. This results in a remarkable efficiency gain with respect to other approaches, since now we can compare N molecules in O(N) instead of the common quadratic O(N2) timescale. This favorable scaling has motivated the application of these indices to diversity selection, clustering, phylogenetic analysis, chemical space visualization, and post-processing of molecular dynamics simulations. However, the current formulation of the n-ary indices is limited to vectors with binary or categorical inputs. Here, we present the further generalization of this formalism so it can be applied to numerical data, i.e. to vectors with continuous components. We discuss several ways to achieve this extension and present their analytical properties. As a practical example, we apply this formalism to the problem of feature selection in QSAR and prove that the extended continuous similarity indices provide a convenient way to discern between several sets of descriptors.

Towards the De Novo Design of HIV-1 Protease Inhibitors Based on Natural Products.

Acquired immunodeficiency syndrome (AIDS) caused by the human immunodeficiency virus (HIV) continues to be a public health problem. In 2020, 680,000 people died from HIV-related causes, and 1.5 million people were infected. Antiretrovirals are a way to control HIV infection but not to cure AIDS. As such, effective treatment must be developed to control AIDS. Developing a drug is not an easy task, and there is an enormous amount of work and economic resources invested. For this reason, it is highly convenient to employ computer-aided drug design methods, which can help generate and identify novel molecules. Using the de novo design, novel molecules can be developed using fragments as building blocks. In this work, we develop a virtual focused compound library of HIV-1 viral protease inhibitors from natural product fragments. Natural products are characterized by a large diversity of functional groups, many sp3 atoms, and chiral centers. Pseudo-natural products are a combination of natural products fragments that keep the desired structural characteristics from different natural products. An interactive version of chemical space visualization of virtual compounds focused on HIV-1 viral protease inhibitors from natural product fragments is freely available in the supplementary material.

GenUI: interactive and extensible open source software platform for de novo molecular generation and cheminformatics

Many contemporary cheminformatics methods, including computer-aided de novo drug design, hold promise to significantly accelerate and reduce the cost of drug discovery. Thanks to this attractive outlook, the field has thrived and in the past few years has seen an especially significant growth, mainly due to the emergence of novel methods based on deep neural networks. This growth is also apparent in the development of novel de novo drug design methods with many new generative algorithms now available. However, widespread adoption of new generative techniques in the fields like medicinal chemistry or chemical biology is still lagging behind the most recent developments. Upon taking a closer look, this fact is not surprising since in order to successfully integrate the most recent de novo drug design methods in existing processes and pipelines, a close collaboration between diverse groups of experimental and theoretical scientists needs to be established. Therefore, to accelerate the adoption of both modern and traditional de novo molecular generators, we developed Generator User Interface (GenUI), a software platform that makes it possible to integrate molecular generators within a feature-rich graphical user interface that is easy to use by experts of diverse backgrounds. GenUI is implemented as a web service and its interfaces offer access to cheminformatics tools for data preprocessing, model building, molecule generation, and interactive chemical space visualization. Moreover, the platform is easy to extend with customizable frontend React.js components and backend Python extensions. GenUI is open source and a recently developed de novo molecular generator, DrugEx, was integrated as a proof of principle. In this work, we present the architecture and implementation details of GenUI and discuss how it can facilitate collaboration in the disparate communities interested in de novo molecular generation and computer-aided drug discovery.

Chemical space exploration guided by deep neural networks.

A parametric t-SNE approach based on deep feed-forward neural networks was applied to the chemical space visualization problem. It is able to retain more information than certain dimensionality reduction techniques used for this purpose (principal component analysis (PCA), multidimensional scaling (MDS)). The applicability of this method to some chemical space navigation tasks (activity cliffs and activity landscapes identification) is discussed. We created a simple web tool to illustrate our work (http://space.syntelly.com).

Exploring DrugBank in Virtual Reality Chemical Space.

The recent general availability of low-cost virtual reality headsets and accompanying three-dimensional (3D) engine support presents an opportunity to bring the concept of chemical space into virtual environments. While virtual reality applications represent a category of widespread tools in other fields, their use in the visualization and exploration of abstract data such as chemical spaces has been experimental. In our previous work, we established the concept of interactive two-dimensional (2D) maps of chemical spaces followed by interactive web-based 3D visualizations, culminating in the interactive web-based 3D visualization of extremely large chemical spaces. Virtual reality chemical spaces are a natural extension of these concepts. As 2D and 3D embeddings and projections of high-dimensional chemical fingerprint spaces have been shown to be valuable tools in chemical space visualization and exploration, existing pipelines of data mining and preparation can be extended to be used in virtual reality applications. Here we present an application based on the Unity engine and the Virtual Reality Toolkit, allowing for the interactive exploration of chemical space populated by DrugBank compounds in virtual reality. The source code of the application as well as the most recent build are available on GitHub ( https://github.com/reymond-group/virtual-reality-chemical-space ).

Platform for Unified Molecular Analysis: PUMA.

We introduce a free platform for chemoinformatic-based diversity analysis and visualization of chemical space of user supplied data sets. Platform for Unified Molecular Analysis (PUMA) integrates metrics used to characterize compound databases including visualization of chemical space, scaffold content, and analysis of chemical diversity. The user's input is a file with SMILES, database names, and compound IDs. PUMA computes molecular properties of pharmaceutical relevance, Murcko scaffolds, and diversity analysis. The user can interactively navigate through the graphs and export image files and the raw data of the diversity calculations. The platform links two public online resources: Consensus Diversity Plots for the assessment of global diversity and Activity Landscape Plotter to analyze structure-activity relationships. Herein, we describe the functionalities of PUMA and exemplify its use through the analysis of compound databases of general interest. PUMA is freely accessible at the authors web-site https://www.difacquim.com/d-tools/ .

Generative Topographic Mapping Approach to Modeling and Chemical Space Visualization of Human Intestinal Transporters

The generative topographic mapping (GTM) approach has been used both to build predictive models linking chemical structure of molecules and their ability to bind some membrane transport proteins (transporters) and to visualize a chemical space of transporters’ binders on two-dimensional maps. For this purpose, experimental data on 2958 molecules active against up to 11 transporters have been used. It has been shown that GTM-based classification (active/inactive) models display reasonable predictive performance, comparable with that of such popular machine-learning methods as Random Forest, SVM, or k-NN. Moreover, GTM offers its own models applicability domain definition which may significantly improve the models performance. GTM maps themselves represent an interesting tool of the chemical space analysis of the transporters’ ligands. Thus, with the help of class landscapes, they identify distinct zones populated by active or inactive molecules with respect to a given transporter. As demonstrated in this paper, the superposition of class landscapes describing different activities delineates the areas mostly populated by the compounds of desired pharmacological profile.

Final report: Testing a procedure for the identification of emerging chemical risks in the food chain. External Scientific Report. OC/EFSA/SCER/2014/03

Final report: Testing a procedure for the identification of emerging chemical risks in the food chain. External Scientific Report. OC/EFSA/SCER/2014/03

Following Ramachandran: exit vector plots (EVP) as a tool to navigate chemical space covered by 3D bifunctional scaffolds. The case of cycloalkanes

An approach to analysis and visualization of chemical space covered by disubstituted scaffolds, which is based on exit vector plots (EVP), is used for analysis of cycloalkane. Four clearly defined regions (α, β, γ and δ) are found in their EVP.

Visualization of chemical space for medicinal chemists.

One of the most common tasks that cheminformatics experts in pharmaceutical industry are facing practically daily is analysis and visualization of large collections of molecules. Typical areas, where this is needed are analysis and enhancement of company compound archive, analysis of high-throughput screening data, design of combinatorial libraries, chemogenomics analyses and many others. But also researchers in academia are facing similar challenges when analyzing large public molecular databases that become available recently or even structures generated in silico. This presentation will provide overview of various methods used to analyze and visualize chemical space with particular focus on needs of medicinal chemists. When displaying results, for chemists it is of great importance that the molecules are represented by their actual structures, or at least by their scaffolds and not only by points as it is common in other scientific fields. This particular requirement makes chemistry visualizations challenging because of necessity to squeeze a lot of information on rather limited computer screen real estate. In the presentation various chemistry visualization techniques will be discussed, starting from classical display of molecules as tables and grids, through visualization based on analysis of scaffold, up to advanced cheminformatics visualizations techniques recently developed at Novartis, such as a method for natural ordering or scaffolds or Molecule Cloud diagrams.

A large scale classification of molecular fingerprints for the chemical space representation and SAR analysis

Fingerprint-based structure representation has a broad range of applications including, but not limited to, diversity analysis, compound classification, chemical space visualization [1], activity landscape modelling and similarity searching. It has been shown that depending on the particular fingerprints used, the outcome of similarity searching [2] or activity landscapes [3] can be very different. Combining structure representations is a common practice to increase the performance of similarity searching [4]. Also, combining representations for activity landscape modelling has been proposed to generate robust descriptive SAR models [5]. However, the selection of fingerprints to be combined is not an easy task. As part of our efforts to select fingerprint representations to generate consensus representations of chemical space and activity landscapes [5,6] herein we discuss the results of a systematic comparison of more than 10 2D and 3D fingerprint representations in terms of performance in diversity analysis (as opposed to similarity searching). We employed more than 20 data sets from different sources relevant to drug discovery. In this work the widely used Tanimoto coefficient was employed. The approach presented here can be easily extended to other similarity measures, additional fingerprints and molecular databases. We also discuss the typical mean/median similarity values of selected fingerprints across databases from different sources.