scholarly journals Transferable Neural Network Potential Energy Surfaces for Closed-Shell Organic Molecules: Extension to Ions

2021 ◽  
Author(s):  
Leif Jacobson ◽  
James Stevenson ◽  
Farhad Ramezanghorbani ◽  
Delaram Ghoreishi ◽  
Karl Leswing ◽  
...  
Keyword(s):  
Neural Network ◽  
Potential Energy Surfaces ◽  
Closed Shell ◽  
Great Promise ◽  
Discovery Research ◽  
Research Activities ◽  
Drug Discovery Research ◽  
Empirical Level ◽  
Torsion Energy ◽  
Conformational Energies

Transferable high dimensional neural network potentials (HDNNP) have shown great promise as an avenue to increase the accuracy and domain of applicability of existing atomistic force fields for organic systems relevant to life science. We have previously reported such a potential (Schrödinger-ANI) that has broad coverage of druglike molecules. We extend that work here to cover ionic and zwitterionic druglike molecules expected to be relevant to drug discovery research activities. We report a novel HDNNP architechture, which we call QRNN, that predicts atomic charges and uses these charges as descriptors in an energy model which delivers conformational energies within chemical accuracy when measured against the reference theory it is trained to. Further, we find that delta learning based on a semi-empirical level of theory approximately halves the errors. We test the models on torsion energy profiles, relative conformational energies, geometric parameters and relative tautomer errors.

scholarly journals Transferable Neural Network Potential Energy Surfaces for Closed-Shell Organic Molecules: Extension to Ions

2021 ◽  
Author(s):  
Leif Jacobson ◽  
James Stevenson ◽  
Farhad Ramezanghorbani ◽  
Delaram Ghoreishi ◽  
Karl Leswing ◽  
...  
Keyword(s):  
Neural Network ◽  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Organic Molecules ◽  
Closed Shell ◽  
Energy Surfaces

scholarly journals Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network

2021 ◽  
Vol 13 (1) ◽  
Author(s):  
Jiarui Chen ◽  
Yain-Whar Si ◽  
Chon-Wai Un ◽  
Shirley W. I. Siu
Keyword(s):  
Neural Network ◽  
Supervised Learning ◽  
Superior Performance ◽  
Chemical Toxicity ◽  
Animal Testing ◽  
Toxicity Prediction ◽  
Discovery Research ◽  
Drug Discovery Research

AbstractAs safety is one of the most important properties of drugs, chemical toxicology prediction has received increasing attentions in the drug discovery research. Traditionally, researchers rely on in vitro and in vivo experiments to test the toxicity of chemical compounds. However, not only are these experiments time consuming and costly, but experiments that involve animal testing are increasingly subject to ethical concerns. While traditional machine learning (ML) methods have been used in the field with some success, the limited availability of annotated toxicity data is the major hurdle for further improving model performance. Inspired by the success of semi-supervised learning (SSL) algorithms, we propose a Graph Convolution Neural Network (GCN) to predict chemical toxicity and trained the network by the Mean Teacher (MT) SSL algorithm. Using the Tox21 data, our optimal SSL-GCN models for predicting the twelve toxicological endpoints achieve an average ROC-AUC score of 0.757 in the test set, which is a 6% improvement over GCN models trained by supervised learning and conventional ML methods. Our SSL-GCN models also exhibit superior performance when compared to models constructed using the built-in DeepChem ML methods. This study demonstrates that SSL can increase the prediction power of models by learning from unannotated data. The optimal unannotated to annotated data ratio ranges between 1:1 and 4:1. This study demonstrates the success of SSL in chemical toxicity prediction; the same technique is expected to be beneficial to other chemical property prediction tasks by utilizing existing large chemical databases. Our optimal model SSL-GCN is hosted on an online server accessible through: https://app.cbbio.online/ssl-gcn/home.

scholarly journals Chemical Toxicity Prediction Based on Semi-supervised Learning and Graph Convolutional Neural Network

2021 ◽  
Author(s):  
Jiarui Chen ◽  
Yain-Whar Si ◽  
Chon-Wai Un ◽  
Shirley W. I. Siu
Keyword(s):  
Neural Network ◽  
Supervised Learning ◽  
Superior Performance ◽  
Chemical Toxicity ◽  
Animal Testing ◽  
Toxicity Prediction ◽  
Discovery Research ◽  
Drug Discovery Research

Abstract As safety is one of the most important properties of drugs, chemical toxicology prediction has received increasing attentions in the drug discovery research. Traditionally, researchers rely on in vitro and in vivo experiments to test the toxicity of chemical compounds. However, not only are these experiments time consuming and costly, but experiments that involve animal testing are increasingly subject to ethical concerns. While traditional machine learning (ML) methods have been used in the field with some success, the limited availability of annotated toxicity data is the major hurdle for further improving model performance. Inspired by the success of semi-supervised learning (SSL) algorithms, we propose a Graph Convolution Neural Network (GCN) to predict chemical toxicity and trained the network by the Mean Teacher (MT) SSL algorithm. Using the Tox21 data, our optimal SSL-GCN models for predicting the twelve toxicological endpoints achieve an average ROC-AUC score of 0.757 in the test set, which is a 6% improvement over GCN models trained by supervised learning and conventional ML methods. Our SSL-GCN models also exhibit superior performance when compared to models constructed using the built-in DeepChem ML methods. This study demonstrates that SSL can increase the prediction power of models by learning from unannotated data. The optimal unannotated to annotated data ratio ranges between 1:1 and 4:1. This study demonstrates the success of SSL in chemical toxicity prediction; the same technique is expected to be beneficial to other chemical property prediction tasks by utilizing existing large chemical databases.

Epigenetic Target Prediction with Accurate Machine Learning Models

2021 ◽  
Author(s):  
Norberto Sánchez-Cruz ◽  
Jose L. Medina-Franco
Keyword(s):  
Machine Learning ◽  
Small Molecules ◽  
Predictive Models ◽  
Large Scale ◽  
Target Prediction ◽  
Quantitative Measure ◽  
Learning Models ◽  
Discovery Research ◽  
Drug Discovery Research ◽  
Machine Learning Models

<p>Epigenetic targets are a significant focus for drug discovery research, as demonstrated by the eight approved epigenetic drugs for treatment of cancer and the increasing availability of chemogenomic data related to epigenetics. This data represents a large amount of structure-activity relationships that has not been exploited thus far for the development of predictive models to support medicinal chemistry efforts. Herein, we report the first large-scale study of 26318 compounds with a quantitative measure of biological activity for 55 protein targets with epigenetic activity. Through a systematic comparison of machine learning models trained on molecular fingerprints of different design, we built predictive models with high accuracy for the epigenetic target profiling of small molecules. The models were thoroughly validated showing mean precisions up to 0.952 for the epigenetic target prediction task. Our results indicate that the herein reported models have considerable potential to identify small molecules with epigenetic activity. Therefore, our results were implemented as freely accessible and easy-to-use web application.</p>

scholarly journals iPSC technology-based regenerative medicine for kidney diseases

2021 ◽  
Author(s):  
Kenji Osafune
Keyword(s):  
Regenerative Medicine ◽  
Cell Therapy ◽  
Kidney Development ◽  
Disease Modeling ◽  
Kidney Diseases ◽  
Collecting Duct ◽  
Current Status ◽  
Renal Tubules ◽  
Discovery Research ◽  
Drug Discovery Research

AbstractWith few curative treatments for kidney diseases, increasing attention has been paid to regenerative medicine as a new therapeutic option. Recent progress in kidney regeneration using human-induced pluripotent stem cells (hiPSCs) is noteworthy. Based on the knowledge of kidney development, the directed differentiation of hiPSCs into two embryonic kidney progenitors, nephron progenitor cells (NPCs) and ureteric bud (UB), has been established, enabling the generation of nephron and collecting duct organoids. Furthermore, human kidney tissues can be generated from these hiPSC-derived progenitors, in which NPC-derived glomeruli and renal tubules and UB-derived collecting ducts are interconnected. The induced kidney tissues are further vascularized when transplanted into immunodeficient mice. In addition to the kidney reconstruction for use in transplantation, it has been demonstrated that cell therapy using hiPSC-derived NPCs ameliorates acute kidney injury (AKI) in mice. Disease modeling and drug discovery research using disease-specific hiPSCs has also been vigorously conducted for intractable kidney disorders, such as autosomal dominant polycystic kidney disease (ADPKD). In an attempt to address the complications associated with kidney diseases, hiPSC-derived erythropoietin (EPO)-producing cells were successfully generated to discover drugs and develop cell therapy for renal anemia. This review summarizes the current status and future perspectives of developmental biology of kidney and iPSC technology-based regenerative medicine for kidney diseases.

scholarly journals To Stretch the Boundary of Secondary Metabolite Production in Plant Cell-Based Bioprocessing: Anthocyanin as a Case Study

2004 ◽  
Vol 2004 (5) ◽  
pp. 264-271 ◽  
Author(s):  
Wei Zhang ◽  
Chris Franco ◽  
Chris Curtin ◽  
Simon Conn
Keyword(s):  
Secondary Metabolites ◽  
Plant Cell ◽  
Metabolic Pathways ◽  
Anthocyanin Biosynthesis ◽  
Rational Approach ◽  
Great Promise ◽  
Primary Metabolism ◽  
Global Correlation ◽  
Empirical Level

Plant cells and tissue cultures hold great promise for controlled production of a myriad of useful secondary metabolites on demand. The current yield and productivity cannot fulfill the commercial goal of a plant cell-based bioprocess for the production of most secondary metabolites. In order to stretch the boundary, recent advances, new directions and opportunities in plant cell-based bioprocessing, have been critically examined for the 10 years from 1992 to 2002. A review of the literature indicated that most of the R&D work was devoted predominantly to studies at an empirical level. A rational approach to molecular plant cell bioprocessing based on the fundamental understanding of metabolic pathways and their regulations is urgently required to stimulate further advances; however, the strategies and technical framework are still being developed. It is the aim of this review to take a step forward in framing workable strategies and technologies for molecular plant cell-based bioprocessing. Using anthocyanin biosynthesis as a case study, an integrated postgenomic approach has been proposed. This combines the functional analysis of metabolic pathways for biosynthesis of a particular metabolite from profiling of gene expression and protein expression to metabolic profiling. A global correlation not only can thus be established at the three molecular levels, but also places emphasis on the interactions between primary metabolism and secondary metabolism; between competing and/or complimentary pathways; and between biosynthetic and post-biosynthetic events.

scholarly journals Representation of coupled adiabatic potential energy surfaces using neural network based quasi-diabatic Hamiltonians: 1,2 2A′ states of LiFH

2019 ◽  
Vol 21 (26) ◽  
pp. 14205-14213 ◽  
Author(s):  
Yafu Guan ◽  
Dong H. Zhang ◽  
Hua Guo ◽  
David R. Yarkony
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Adiabatic Potential ◽  
General Algorithm ◽  
Energy Surfaces

A general algorithm for determining diabatic representations from adiabatic energies, energy gradients and derivative couplings using neural networks is introduced.

scholarly journals Development of a Selection Method for Discovering Irreversible (Covalent) Binders from a DNA-Encoded Library

2018 ◽  
Vol 24 (2) ◽  
pp. 169-174 ◽  
Author(s):  
Zhengrong Zhu ◽  
LaShadric C. Grady ◽  
Yun Ding ◽  
Kenneth E. Lind ◽  
Christopher P. Davie ◽  
...  
Keyword(s):  
Magnetic Beads ◽  
Chemical Probes ◽  
Lead Discovery ◽  
3C Protease ◽  
Discovery Research ◽  
Drug Discovery Research ◽  
Affinity Resin ◽  
Covalent Ligands ◽  
Developing Method ◽  
Unique Mechanism

DNA-encoded libraries (DELs) have been broadly applied to identify chemical probes for target validation and lead discovery. To date, the main application of the DEL platform has been the identification of reversible ligands using multiple rounds of affinity selection. Irreversible (covalent) inhibition offers a unique mechanism of action for drug discovery research. In this study, we report a developing method of identifying irreversible (covalent) ligands from DELs. The new method was validated by using 3C protease (3CP) and on-DNA irreversible tool compounds (rupintrivir derivatives) spiked into a library at the same concentration as individual members of that library. After affinity selections against 3CP, the irreversible tool compounds were specifically enriched compared with the library members. In addition, we compared two immobilization methods and concluded that microscale columns packed with the appropriate affinity resin gave higher tool compound recovery than magnetic beads.

scholarly journals Modern Approaches in the Discovery and Development of Plant-Based Natural Products and Their Analogues as Potential Therapeutic Agents

Molecules ◽  
2022 ◽  
Vol 27 (2) ◽  
pp. 349
Author(s):  
Asim Najmi ◽  
Sadique A. Javed ◽  
Mohammed Al Bratty ◽  
Hassan A. Alhazmi
Keyword(s):  
Natural Products ◽  
Drug Discovery ◽  
Natural Product ◽  
Therapeutic Agents ◽  
Throughput Capacity ◽  
Scientific Interest ◽  
Comprehensive Overview ◽  
Natural Sources ◽  
Discovery Research ◽  
Drug Discovery Research

Natural products represents an important source of new lead compounds in drug discovery research. Several drugs currently used as therapeutic agents have been developed from natural sources; plant sources are specifically important. In the past few decades, pharmaceutical companies demonstrated insignificant attention towards natural product drug discovery, mainly due to its intrinsic complexity. Recently, technological advancements greatly helped to address the challenges and resulted in the revived scientific interest in drug discovery from natural sources. This review provides a comprehensive overview of various approaches used in the selection, authentication, extraction/isolation, biological screening, and analogue development through the application of modern drug-development principles of plant-based natural products. Main focus is given to the bioactivity-guided fractionation approach along with associated challenges and major advancements. A brief outline of historical development in natural product drug discovery and a snapshot of the prominent natural drugs developed in the last few decades are also presented. The researcher’s opinions indicated that an integrated interdisciplinary approach utilizing technological advances is necessary for the successful development of natural products. These involve the application of efficient selection method, well-designed extraction/isolation procedure, advanced structure elucidation techniques, and bioassays with a high-throughput capacity to establish druggability and patentability of phyto-compounds. A number of modern approaches including molecular modeling, virtual screening, natural product library, and database mining are being used for improving natural product drug discovery research. Renewed scientific interest and recent research trends in natural product drug discovery clearly indicated that natural products will play important role in the future development of new therapeutic drugs and it is also anticipated that efficient application of new approaches will further improve the drug discovery campaign.

Sign in / Sign up

Export Citation Format

Share Document