Computational Discovery Research Articles

Transcriptomic Response of the Diazotrophic Bacteria Gluconacetobacter diazotrophicus Strain PAL5 to Iron Limitation and Characterization of the fur Regulatory Network.

Gluconacetobacter diazotrophicus has been the focus of several studies aiming to understand the mechanisms behind this endophytic diazotrophic bacterium. The present study is the first global analysis of the early transcriptional response of exponentially growing G. diazotrophicus to iron, an essential cofactor for many enzymes involved in various metabolic pathways. RNA-seq, targeted gene mutagenesis and computational motif discovery tools were used to define the G. diazotrophicus fur regulon. The data analysis showed that genes encoding functions related to iron homeostasis were significantly upregulated in response to iron limitations. Certain genes involved in secondary metabolism were overexpressed under iron-limited conditions. In contrast, it was observed that the expression of genes involved in Fe-S cluster biosynthesis, flagellar biosynthesis and type IV secretion systems were downregulated in an iron-depleted culture medium. Our results support a model that controls transcription in G. diazotrophicus by fur function. The G. diazotrophicus fur protein was able to complement an E. coli fur mutant. These results provide new insights into the effects of iron on the metabolism of G. diazotrophicus, as well as demonstrate the essentiality of this micronutrient for the main characteristics of plant growth promotion by G. diazotrophicus.

Antimony nitride discovered by theoretical structure prediction, rapid thermal annealing, and in situ X-ray diffraction

Nitride materials have promising properties for optoelectronics, but synthesis is challenging because of their metastable character. Therefore, even some simple binary nitrides have not been reported, and their properties remain unknown. Here, we report a metastable Sb–N phase. Unconstrained first-principle crystal-structure predictions suggest that a semiconducting layered SbN phase could be synthesized under activated nitrogen conditions. Experimentally, we approach the synthesis by rapid thermal annealing of an amorphous thin-film precursor. In situ X-ray diffraction monitoring reveals a transient phase appearing at 500°C and lasting for 17 s. With molecular dynamics simulations, we conclude that the phase observed is related to the predicted ground-state structure but that it contains stacking variations. This research highlights the synergy between computational materials discovery and in situ structural characterization during synthesis and processing, enabling the identification and synthesizability of new materials at the frontier of thermodynamically metastable regions.

Rapid discovery of stable materials by coordinate-free coarse graining.

A fundamental challenge in materials science pertains to elucidating the relationship between stoichiometry, stability, structure, and property. Recent advances have shown that machine learning can be used to learn such relationships, allowing the stability and functional properties of materials to be accurately predicted. However, most of these approaches use atomic coordinates as input and are thus bottlenecked by crystal structure identification when investigating previously unidentified materials. Our approach solves this bottleneck by coarse-graining the infinite search space of atomic coordinates into a combinatorially enumerable search space. The key idea is to use Wyckoff representations, coordinate-free sets of symmetry-related positions in a crystal, as the input to a machine learning model. Our model demonstrates exceptionally high precision in finding unknown theoretically stable materials, identifying 1569 materials that lie below the known convex hull of previously calculated materials from just 5675 ab initio calculations. Our approach opens up fundamental advances in computational materials discovery.

Emerging Promise of Computational Techniques in Anti-Cancer Research: At a Glance

Research on the immune system and cancer has led to the development of new medicines that enable the former to attack cancer cells. Drugs that specifically target and destroy cancer cells are on the horizon; there are also drugs that use specific signals to stop cancer cells multiplying. Machine learning algorithms can significantly support and increase the rate of research on complicated diseases to help find new remedies. One area of medical study that could greatly benefit from machine learning algorithms is the exploration of cancer genomes and the discovery of the best treatment protocols for different subtypes of the disease. However, developing a new drug is time-consuming, complicated, dangerous, and costly. Traditional drug production can take up to 15 years, costing over USD 1 billion. Therefore, computer-aided drug design (CADD) has emerged as a powerful and promising technology to develop quicker, cheaper, and more efficient designs. Many new technologies and methods have been introduced to enhance drug development productivity and analytical methodologies, and they have become a crucial part of many drug discovery programs; many scanning programs, for example, use ligand screening and structural virtual screening techniques from hit detection to optimization. In this review, we examined various types of computational methods focusing on anticancer drugs. Machine-based learning in basic and translational cancer research that could reach new levels of personalized medicine marked by speedy and advanced data analysis is still beyond reach. Ending cancer as we know it means ensuring that every patient has access to safe and effective therapies. Recent developments in computational drug discovery technologies have had a large and remarkable impact on the design of anticancer drugs and have also yielded useful insights into the field of cancer therapy. With an emphasis on anticancer medications, we covered the various components of computer-aided drug development in this paper. Transcriptomics, toxicogenomics, functional genomics, and biological networks are only a few examples of the bioinformatics techniques used to forecast anticancer medications and treatment combinations based on multi-omics data. We believe that a general review of the databases that are now available and the computational techniques used today will be beneficial for the creation of new cancer treatment approaches.

Computational Drug Discovery in Patients with Endometriosis-Induced Infertility via Text Mining and Biomedical Databases

Background: Endometriosis (EMT) is the most common benign gynecological disease among women of reproductive age, causing infertility and seriously affects women’s physical and mental health. However, the current treatment was not always effective. This study was designed to use publicly available data to identify drugs targeting the relevant gene with EMT-induced-infertility using computational tools. Methods: EMT and infertility genes were determined by text mining, and the GeneCodis program was used to analyzed gene ontology of the intersection of the two gene sets. A string database was used to analyze the protein-protein interaction network. The Drug-Gene Interaction database is queried for the rich gene set belonging to the identified pathways to find drug candidates that can be used in EMT-induced infertility. Results: Our analysis identified 550 genes common to both the EMT and infertility by text mining. Gene enrichment analysis and protein-protein interaction analysis found 39 genes potentially targetable by a total of 49 drugs that could be formulated for application, which have not been used in EMT-induced infertility. Conclusions: The findings from the present analysis can facilitate the Identification of existing drugs that have the potential of topical administration to improve EMT-induced infertility and present tremendous opportunities to study novel targets pharmacology using in silico text mining and pathway analysis tools. However, all the results were based on online bioinformatics databases, and as such require validation experiments. And some of the drugs highlighted as possibly relevant may be toxic and as such safely data is required before any experiments are undertaken in humans.

Machine-designed biotherapeutics: opportunities, feasibility and advantages of deep learning in computational antibody discovery.

Antibodies are versatile molecular binders with an established and growing role as therapeutics. Computational approaches to developing and designing these molecules are being increasingly used to complement traditional lab-based processes. Nowadays, in silico methods fill multiple elements of the discovery stage, such as characterizing antibody–antigen interactions and identifying developability liabilities. Recently, computational methods tackling such problems have begun to follow machine learning paradigms, in many cases deep learning specifically. This paradigm shift offers improvements in established areas such as structure or binding prediction and opens up new possibilities such as language-based modeling of antibody repertoires or machine-learning-based generation of novel sequences. In this review, we critically examine the recent developments in (deep) machine learning approaches to therapeutic antibody design with implications for fully computational antibody design.

Toward Automated Computational Discovery of Battery Materials

AbstractNew rechargeable batteries with high energy density and low cost have been intensively explored, but their commercialization still faces multiple challenges involving battery materials and interfaces. Some difficulties faced by battery materials are that a single material often needs to possess multiple functions, and also needs to be cheap, easy to prepare, safe, and environmentally friendly. Recent developments in workflow managers (WMs) along with continuously increasing computing power have enabled the automated computational workflow method. Using this method, the WM can execute the predesigned research workflow to study tens of thousands of materials and screen out materials that meet the multiple requirements. In this perspective, a critical overview of the automated computational workflows is presented, focusing on the high‐throughput study of battery materials. First, an introduction to the automated computational workflow as well as commonly used WMs will be given. Next, the latest works and methods to build such automated workflows are presented. Finally, an outlook on the existing challenges and future directions to drive computational and experimental developments in this nascent field is provided.

Synergistic combination of ritonavir and cisplatin as an efficacious therapy in human cervical cancer cells: a computational drug discovery and in vitro insight

HIV-protease inhibitor Ritonavir (RTV) is a clinical-stage drug. We exhibit here the synergistic effect of RTV coupled with cisplatin as potential combination therapy for treatment of cervical cancer. Knowledge about the interaction of RTV with the high-expression signatures in cancer is limited. Therefore, we utilized computational techniques to understand and assess the drug-binding affinity and drug-target interaction of RTV with these altered protein signatures. Computational studies revealed the potential interaction ability of RTV along with few other HIV protease inhibitors against these altered cancer targets. All targets exhibited good affinity towards RTV and the highest affinity was exhibited by CYP450 3A4, PDGFR and ALK. RTV established stable interaction with PDGFR and molecular dynamics simulation confirms their frequent interaction for 300 ns. Control docking of PDGFR with standard PDGFR inhibitor exhibited lower binding affinity when compared with RTV-PDGFR complex. In search of drugs as a part of combination therapy to reduce side effects of Cisplatin, this paper further evaluated the effect of combination of RTV and Cisplatin in cervical cancer cells. We propose several combination models that combines anti-viral drug RTV and standard chemotherapeutic agent, Cisplatin to be synergistic with CI value ranging from of 0.01 to 1.14. These observations suggest that anti-viral compound (RTV) could act synergistically with Cisplatin for cervical cancer therapy. However, further studies are warranted to investigate the combinatorial mode of action of RTV and Cisplatin on different molecular pathways to have a translational outcome in cervical cancer. Communicated by Ramaswamy H. Sarma

Automation of AMOEBA polarizable force field for small molecules: Poltype 2.

A next-generation protocol (Poltype 2) has been developed which automatically generates AMOEBA polarizable force field parameters for small molecules. Both features and computational efficiency have been drastically improved. Notable advances include improved database transferability using SMILES, robust torsion fitting, non-aromatic ring torsion parameterization, coupled torsion-torsion parameterization, Van der Waals parameter refinement using ab initio dimer data and an intelligent fragmentation scheme that produces parameters with dramatically reduced ab initio computational cost. Additional improvements include better local frame assignment for atomic multipoles, automated formal charge assignment, Zwitterion detection, smart memory resource defaults, parallelized fragment job submission, incorporation of Psi4 quantum package, ab initio error handling, ionization state enumeration, hydration free energy prediction and binding free energy prediction. For validation, we have applied Poltype 2 to ~1000 FDA approved drug molecules from DrugBank. The ab initio molecular dipole moments and electrostatic potential values were compared with Poltype 2 derived AMOEBA counterparts. Parameters were further substantiated by calculating hydration free energy (HFE) on 40 small organic molecules and were compared with experimental data, resulting in an RMSE error of 0.59 kcal/mol. The torsion database has expanded to include 3543 fragments derived from FDA approved drugs. Poltype 2 provides a convenient utility for applications including binding free energy prediction for computational drug discovery. Further improvement will focus on automated parameter refinement by experimental liquid properties, expansion of the Van der Waals parameter database and automated parametrization of modified bio-fragments such as amino and nucleic acids.

Investigation of novel compounds via insilico approaches of EGFR inhibitors as anticancer agents

Before the experimental studies of a compound to be synthesized from in vitro to in vivo, it is possible to save both time and money with in silico approaches only with Computer Aided Drug Design (CADD) methods. In other words, compounds that can be new drug candidates can be suggested by drug design using computational drug discovery strategies. In this study, all molecules in the ChEMBL Database were virtual screened based on drugs with inhibitory properties on the Epidermal Growth Factor Receptor (EGFR), one of the receptor tyrosine kinases, which is effective in cancer cells. During this High-Throughput Screening (HTS), the number of compounds was minimized according to the parameters of the reference drugs, physicochemical properties such as logP, M.W., HBA, HDB, RosLip. As a result of in silico approaches and molecular docking analysis, ten compounds with the highest docking scores were determined and a model compound that could be a new drug candidate was proposed.

GeomGCL: Geometric Graph Contrastive Learning for Molecular Property Prediction

Recently many efforts have been devoted to applying graph neural networks (GNNs) to molecular property prediction which is a fundamental task for computational drug and material discovery. One of major obstacles to hinder the successful prediction of molecular property by GNNs is the scarcity of labeled data. Though graph contrastive learning (GCL) methods have achieved extraordinary performance with insufficient labeled data, most focused on designing data augmentation schemes for general graphs. However, the fundamental property of a molecule could be altered with the augmentation method (like random perturbation) on molecular graphs. Whereas, the critical geometric information of molecules remains rarely explored under the current GNN and GCL architectures. To this end, we propose a novel graph contrastive learning method utilizing the geometry of the molecule across 2D and 3D views, which is named GeomGCL. Specifically, we first devise a dual-view geometric message passing network (GeomMPNN) to adaptively leverage the rich information of both 2D and 3D graphs of a molecule. The incorporation of geometric properties at different levels can greatly facilitate the molecular representation learning. Then a novel geometric graph contrastive scheme is designed to make both geometric views collaboratively supervise each other to improve the generalization ability of GeomMPNN. We evaluate GeomGCL on various downstream property prediction tasks via a finetune process. Experimental results on seven real-life molecular datasets demonstrate the effectiveness of our proposed GeomGCL against state-of-the-art baselines.

First-Principles Computational Design and Discovery of Solid-Oxide Proton Conductors

First-Principles Computational Design and Discovery of Solid-Oxide Proton Conductors

Abstract 1177: Introduction of the GA4GH Variation Representation Specification (VRS) and supporting tools for discovery and exchange of clinical genomic and cytogenomic knowledge in cancers

Abstract Precision oncology is the practice of interpreting the clinical significance of observed molecular changes in patient neoplasms, potentially impacting medical decision making and care. This process is labor-intensive and (among other challenges) involves accurately translating between variation representation conventions from one resource to the next. For example, differences in representations of Copy Number Variation (CNV) from genomic regions, cytogenomic bands, or gene features create challenges in knowledge matching due to lack of standards covering all of these modalities of observed variation.The Global Alliance for Genomics and Health (GA4GH; ga4gh.org) is an international collaborative of genomic data sharing initiatives (Driver Projects) developing genomic data sharing standards within a human rights framework. GA4GH recently published the Variation Representation Specification (VRS; pronounced “verse”), a standard for the computational representation of biomolecular variation. VRS is a terminology, schema, and associated conventions for creating uniquely identifiable and federatable representations of molecular variation. VRS has formal data classes well-suited to differentiating between variation on a single molecule (e.g. tandem duplications) from variation measured at a systemic level (e.g. genome-wide copy number variation). In addition to molecular sequence variation, VRS also supports variation on cytogenetic coordinate systems and genes, making it well-suited to representing variation associated with cancer biomarkers.We demonstrate the use of VRS to model reported gene-associated CNVs from the AACR Project GENIE cohort, to aid in the computational discovery of evidence from clinico-genomic knowledgebases with genomic or cytogenomic CNV representations. We highlight the use case of knowledge matching to the Atlas of Genetics and Cytogenetics in Oncology and Haematology (“the Atlas”; atlasgeneticsoncology.org), a cytogenetics resource historically driven by user website navigation. Using VRS search tools we developed for the Variant Interpretation for Cancer Consortium (VICC; cancervariants.org) GA4GH Driver Project, we found that 64% of GENIE samples with reported CNVs matched clinically relevant knowledge in the Atlas. This work was enabled by programmatic search tools leveraging standard VRS object structures, demonstrating how VRS enables collection of real-world evidence across more resources without manual interpretation or custom normalization methods. We conclude with a survey of open-source tools supporting this analysis as well as search of other clinico-genomic knowledgebases with VRS, including CIViC (civicdb.org), BRCA Exchange (brcaexchange.org), and the Molecular Oncology Almanac (moalmanac.org). Citation Format: Matthew Cannon, Kori Kuzma, James Stevenson, Jiachen Liu, Colin O'Sullivan, Bimal P. Chaudhari, Matthew Brush, Robert R. Freimuth, Tristan Nelson, Michael Baudis, Obi L. Griffith, Malachi Griffith, Lawrence Babb, Melissa S. Cline, Xuelu Liu, Brian Walsh, Alex H. Wagner. Introduction of the GA4GH Variation Representation Specification (VRS) and supporting tools for discovery and exchange of clinical genomic and cytogenomic knowledge in cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1177.

Computational discovery of small drug-like compounds as potential inhibitors of PD-1/PD-L1 interactions

The programmed cell death ligand protein 1 (PD-L1) is a strong immunosuppressive molecule that inactivates tumor-specific T cells by binding to the programmed cell death- 1 protein (PD-1). Cancer immunotherapy based on the monoclonal antibodies targeting the PD-1/PD-L1 pathway has demonstrated therapeutic responses without precedent over a wide range of cancers. However, the antibody-based immunotherapies have several limitations such as high production cost or the induction of severe immune-related adverse effects. Small-molecule inhibitors of the PD-1/PD-L1 pathway are a promising alternative or complementary therapeutic to antibodies. Currently, the field of developing anti-PD-1/PD-L1 small-molecule inhibitors is intensively explored. In the present study a pharmacophore model was generated based on previously developed compounds and their atomistic structures with the PD-L1 dimer. Structure-based affinity-based virtual screening of small-molecule inhibitors of the PD-1/PD-L1 pathway according to the pharmacophore model followed by a screening in terms of drug-likeness resulted in ten hit compounds of high affinity towards the PD-L1 dimer and the satisfaction to all of the drug-likeness rules. Molecular dynamics (MD) simulations showed that nine of ten compounds formed stable complexes with the PD-L1 dimer as evidenced by the analysis of MD trajectories. Molecular mechanics Poisson- Boltzmann surface area (MM-PBSA) calculation revealed very low binding energies (<-46 kcal/mol) for the interactions of these ligands with the PD-L1 dimer, suggesting that identified compounds may serve as good scaffolds for the design of novel agents of antitumor immunotherapy able to target the PD-1/PD-L1 interaction Communicated by Ramaswamy H. Sarma

Designing of disruptor molecules to restrain the protein-protein interaction network of VANG1/SCRIB/NOS1AP using fragment-based drug discovery techniques.

Governing protein-protein interaction networks are the cynosure of cell signaling and oncogenic networks. Multifarious processes when aligned with one another can result in a dysregulated output which can result in cancer progression. In the current research, one such network of proteins comprising VANG1/SCRIB/NOS1AP, which is responsible for cell migration, is targeted. The proteins are modeled using in-silico approaches, and the interaction is visualized utilizing protein-protein docking. Designing drugs for the convoluted protein network can serve as a challenging task that can be overcome by fragment-based drug designing, a recent game-changer in the computational drug discovery strategy for protein interaction networks. The model is exposed to the extraction of hotspots, also known as the restrained regions for small molecular hits. The hotspot regions are subjected to a library of generated fragments, which are then recombined and rejoined to develop small molecular disruptors of the macromolecular assemblage. Rapid screening methods using pharmacokinetic tools and 2D interaction studies resulted in four molecules that could serve the purpose of a disruptor. The final validation is executed by long-range simulations of 100ns and exploring the stability of the complex using several parameters leading to the emergence of two novel molecules VNS003 and VNS005 that could be used as the disruptors of the protein assembly VANG1/SCRIB/NOS1AP. Also, the molecules were explored as single protein targets approbated via molecular docking and 100ns molecular dynamics simulation. This concluded VNS003 as the most suitable inhibitor module capable of acting as a disruptor of a macromolecular assembly as well as acting on individual protein chains, thus leading to the primary hindrance in the formation of the protein interaction complex.

BMC128: A rationally designed live bacterial consortium for the potentiation of immune checkpoint therapy in solid tumors.

TPS2685 Background: Immune-Checkpoint inhibitors (ICI) have transformed the treatment of solid tumors, specifically in cutaneous melanoma, clear cell renal cell carcinoma (ccRCC) and non-small cell lung cancer (NSCLC). Despite these advances, a significant proportion of patients do not respond to ICI and only a fraction of responding patients will experience durable responses. Recent reports have found that patients showing poor response to ICI are characterized by a reduced gut microbiome diversity, suggesting that the gut microbiome might affect immune activation by ICI. BMC128 is a live bacterial consortium of 4 bacterial strains predicted to induce anti-tumor immune function when given in conjunction with ICI. These strains were identified by analyzing data obtained from NSCLC and ccRCC patients presenting varied responses to ICI, using a proprietary computational discovery platform enabling high-resolution functional microbiome analysis. In pre-clinical studies, treatment with BMC128 potentiated the efficacy of ICI in breast cancer and melanoma mouse models, reducing tumor volume, increasing the number of responders, and demonstrating an increase in infiltrating immunocytes: CD4, CD8 and NK cells. Methods: In this phase I, first-in-human trial, patients with ccRCC, cutaneous melanoma and NSCLC-adenocarcinoma (EGFR/ALK wild-type), who previously progressed on PD1/PDL-1 inhibitors will be treated with BMC128 8*108 CFU BID in combination with Nivolumab at a fixed-dose of 480mg q4weeks in a 3+3 design, followed by a 9 patient expansion phase. No dose escalation is planned. The primary objective is to investigate the safety and tolerability of BMC128 in combination with Nivolumab. Enrollment is intended to start in Feb, 2022. Clinical trial information: 202122207. [Table: see text]

Computational compensatory mutation discovery approach: Predicting a PARP1 variant rescue mutation

The prediction of protein mutations that affect function may be exploited for multiple uses. In the context of disease variants, the prediction of compensatory mutations that reestablish functional phenotypes could aid in the development of genetic therapies. In this work, we present an integrated approach that combines coevolutionary analysis and molecular dynamics (MD) simulations to discover functional compensatory mutations. This approach is employed to investigate possible rescue mutations of a poly(ADP-ribose) polymerase 1 (PARP1) variant, PARP1 V762A, associated with lung cancer and follicular lymphoma. MD simulations show PARP1 V762A exhibits noticeable changes in structural and dynamical behavior compared with wild-type (WT) PARP1. Our integrated approach predicts A755E as a possible compensatory mutation based on coevolutionary information, and molecular simulations indicate that the PARP1 A755E/V762A double mutant exhibits similar structural and dynamical behavior to WT PARP1. Our methodology can be broadly applied to a large number of systems where single-nucleotide polymorphisms have been identified as connected to disease and can shed light on the biophysical effects of such changes as well as provide a way to discover potential mutants that could restore WT-like functionality. This can, in turn, be further utilized in the design of molecular therapeutics that aim to mimic such compensatory effect.

Towards universal neural network potential for material discovery applicable to arbitrary combination of 45 elements

Computational material discovery is under intense study owing to its ability to explore the vast space of chemical systems. Neural network potentials (NNPs) have been shown to be particularly effective in conducting atomistic simulations for such purposes. However, existing NNPs are generally designed for narrow target materials, making them unsuitable for broader applications in material discovery. Here we report a development of universal NNP called PreFerred Potential (PFP), which is able to handle any combination of 45 elements. Particular emphasis is placed on the datasets, which include a diverse set of virtual structures used to attain the universality. We demonstrated the applicability of PFP in selected domains: lithium diffusion in LiFeSO4F, molecular adsorption in metal-organic frameworks, an order–disorder transition of Cu-Au alloys, and material discovery for a Fischer–Tropsch catalyst. They showcase the power of PFP, and this technology provides a highly useful tool for material discovery.

Femtosecond X-ray Laser Reveals Intact Sea-Island Structures of Metastable Solid-State Electrolytes for Batteries.

Experimental characterization of the nanostructure of metastable functional materials has attracted significant attention with recent advances in computational materials discovery. However, since metastable glass-ceramics are easily damaged by irradiation, damage-free nanoimaging has not been realized thus far. Herein, we propose novel high-contrast coherent diffractive imaging that quantitatively analyzes the intact internal nanostructure of metastable glass-ceramics using femtosecond X-ray pulses. The immersion of sample particles in a solvent helps enhance the reconstructed image contrast and allows us to distinguish an ∼7% electron density difference between an amorphous form and crystals. Furthermore, morphological operations with a band-pass filter quantitatively elucidate the depth information. The evaluated volume ratio of the amorphous to crystalline phases is ∼2.5:1 for the measured metastable (Li2S)70-(P2S5)30 glass-ceramic particle. Sulfide glass-ceramics are used as electrolytes for all-solid-state batteries, which are indispensable for reducing the carbon footprint. Our results will facilitate structural studies on fragile metastable materials with important scientific and industrial implications.

High-Throughput Computational Discovery and Intelligent Design of Two-Dimensional Functional Materials for Various Applications

High-Throughput Computational Discovery and Intelligent Design of Two-Dimensional Functional Materials for Various Applications