SubMo-GNN: Property- and Structure-Aware Diverse Molecular Selection with Representation Learning and Mathematical Diverse-Selection Framework

Selecting diverse molecules from unexplored areas of chemical space is one of the most important tasks for discovering novel molecules and reactions. This paper develops a new method for selecting a diverse subset of molecules from a given molecular list by utilizing two 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 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 diverse subset of 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 method successfully selects diverse molecules from the QM9 dataset regarding the property-based criterion, while performing comparably to existing methods regarding a standard structure-based criterion. 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.

Predicting Accurate Lead Structures for Screening Molecular Libraries: A Quantum Crystallographic Approach

Optimization of lead structures is crucial for drug discovery. However, the accuracy of such a prediction using the traditional molecular docking approach remains a major concern. Our study demonstrates that the employment of quantum crystallographic approach-counterpoise corrected kernel energy method (KEM-CP) can improve the accuracy by and large. We select human aldose reductase at 0.66 Å, cyclin dependent kinase 2 at 2.0 Å and estrogen receptor β at 2.7 Å resolutions with active site environment ranging from highly hydrophilic to moderate to highly hydrophobic and several of their known ligands. Overall, the use of KEM-CP alongside the GoldScore resulted superior prediction than the GoldScore alone. Unlike GoldScore, the KEM-CP approach is neither environment-specific nor structural resolution dependent, which highlights its versatility. Further, the ranking of the ligands based on the KEM-CP results correlated well with that of the experimental IC50 values. This computationally inexpensive yet simple approach is expected to ease the process of virtual screening of potent ligands, and it would advance the drug discovery research.

Transmol: Repurposing Language Model for Molecular Generation

Recent advances in convolutional neural networks have inspired the application of deep learning to other disciplines. Even though image processing and natural language processing have turned out to be the most successful, there are many other areas that have benefited, like computational chemistry in general and drug design in particular. From 2018 the scientific community has seen a surge of methodologies related to the generation of diverse molecular libraries using machine learning. However, no algorithm used an attention mechanisms for <i>de novo</i> molecular generation. Here we employ a variant of transformers, a recent NLP architecture, for this purpose. We have achieved a statistically significant increase in some of the core metrics of the MOSES benchmark. Furthermore, a novel way of generating libraries fusing two molecules as seeds has been described.

Transmol: Repurposing Language Model for Molecular Generation

Recent advances in convolutional neural networks have inspired the application of deep learning to other disciplines. Even though image processing and natural language processing have turned out to be the most successful, there are many other areas that have benefited, like computational chemistry in general and drug design in particular. From 2018 the scientific community has seen a surge of methodologies related to the generation of diverse molecular libraries using machine learning. However, no algorithm used an attention mechanisms for <i>de novo</i> molecular generation. Here we employ a variant of transformers, a recent NLP architecture, for this purpose. We have achieved a statistically significant increase in some of the core metrics of the MOSES benchmark. Furthermore, a novel way of generating libraries fusing two molecules as seeds has been described.

Multiple C–H Activation Enabled Modular Construction of Densely Functionalized Sulfur-Contained Arenes

<div> <p>Modular construction of multiple functionalized arenes from abundant feedstocks, stands as an unremitting pursue goal in synthetic chemistry, which would accelerate the discovery of new drugs an advanced materials. Herein, by using multiple C-H activation strategy, through judicious choice of versatile imidate ester as the key directing group, expedi-ent delivery of molecular libraries of densely functionalized sulfur-contained arenes. Further synthetic application was demonstrated by multiple C-H modification of fused arenes and pharmaceuticals such as Ibuprofen, and concise con-struction of biologically active molecules, Madam dihydrochloride and Bipenamol, was also achieved.</p> </div> <br>

Multiple C–H Activation Enabled Modular Construction of Densely Functionalized Sulfur-Contained Arenes

<div> <p>Modular construction of multiple functionalized arenes from abundant feedstocks, stands as an unremitting pursue goal in synthetic chemistry, which would accelerate the discovery of new drugs an advanced materials. Herein, by using multiple C-H activation strategy, through judicious choice of versatile imidate ester as the key directing group, expedi-ent delivery of molecular libraries of densely functionalized sulfur-contained arenes. Further synthetic application was demonstrated by multiple C-H modification of fused arenes and pharmaceuticals such as Ibuprofen, and concise con-struction of biologically active molecules, Madam dihydrochloride and Bipenamol, was also achieved.</p> </div> <br>

Natural Products as a Paradigm for the Treatment of Coxsackievirus - induced Myocarditis

Coxsackievirus B3 (CVB3), a member of the Picornaviridae family, is considered to be one of the most important infectious agents to cause virus-induced myocarditis. Despite improvements in studying viral pathology, structure and molecular biology, as well as diagnosis of this disease, there is still no virus-specific drug in clinical use. Structural and nonstructural proteins produced during the coxsackievirus life cycle have been identified as potential targets for blocking viral replication at the step of attachment, entry, uncoating, RNA and protein synthesis by synthetic or natural compounds. Moreover, WIN (for Winthrop) compounds and application of nucleic-acid based strategies were shown to target viral capsid, entry and viral proteases, but have not reached to the clinical trials as a successful antiviral agent. There is an urgent need for diverse molecular libraries for phenotype-selective and high-throughput screening.

Bacterial production and direct functional screening of expanded molecular libraries for discovering inhibitors of protein aggregation

Protein misfolding and aggregation are associated with a many human disorders, including Alzheimer’s and Parkinson’s diseases. Toward increasing the effectiveness of early-stage drug discovery for these conditions, we report a bacterial platform that enables the biosynthesis of molecular libraries with expanded diversities and their direct functional screening for discovering protein aggregation inhibitors. We illustrate this approach by performing, what is to our knowledge, the largest functional screen of small-size molecular entities described to date. We generated a combinatorial library of ~200 million drug-like, cyclic peptides and rapidly screened it for aggregation inhibitors against the amyloid-β peptide (Aβ42), linked to Alzheimer’s disease. Through this procedure, we identified more than 400 macrocyclic compounds that efficiently reduce Aβ42 aggregation and toxicity in vitro and in vivo. Finally, we applied a combination of deep sequencing and mutagenesis analyses to demonstrate how this system can rapidly determine structure-activity relationships and define consensus motifs required for bioactivity.