Abstract
It is more pressing than ever to reduce the time and costs for the development of lead compounds in the pharmaceutical industry. The co-occurrence of advances in high-throughput screening and the rise of deep learning (DL) have enabled the development of large-scale multimodal predictive models for virtual drug screening. Recently, deep generative models have emerged as a powerful tool to explore the chemical space and raise hopes to expedite the drug discovery process. Following this progress in chemocentric approaches for generative chemistry, the next challenge is to build multimodal conditional generative models that leverage disparate knowledge sources when mapping biochemical properties to target structures.
Here, we call the community to bridge drug discovery more closely with systems biology when designing deep generative models. Complementing the plethora of reviews on the role of DL in chemoinformatics, we specifically focus on the interface of predictive and generative modelling for drug discovery. Through a systematic publication keyword search on PubMed and a selection of preprint servers (arXiv, biorXiv, chemRxiv, and medRxiv), we quantify trends in the field and find that molecular graphs and VAEs have become the most widely adopted molecular representations and architectures in generative models, respectively. We discuss progress on DL for toxicity, drug-target affinity, and drug sensitivity prediction and specifically focus on conditional molecular generative models that encompass multimodal prediction models. Moreover, we outline future prospects in the field and identify challenges such as the integration of deep learning systems into experimental workflows in a closed-loop manner or the adoption of federated machine learning techniques to overcome data sharing barriers. Other challenges include, but are not limited to interpretability in generative models, more sophisticated metrics for the evaluation of molecular generative models, and, following up on that, community-accepted benchmarks for both multimodal drug property prediction and property-driven molecular design.
Keywords: Deep learning, drug discovery, generative models, review, QSAR modelling, Machine learning, drug design.
Current Medicinal Chemistry
Title:Trends in Deep Learning for Property-driven Drug Design
Volume: 28 Issue: 38
Author(s): Jannis Born*Matteo Manica
Affiliation:
- IBM Research Europe, Zurich,Switzerland
Keywords: Deep learning, drug discovery, generative models, review, QSAR modelling, Machine learning, drug design.
Abstract:
It is more pressing than ever to reduce the time and costs for the development of lead compounds in the pharmaceutical industry. The co-occurrence of advances in high-throughput screening and the rise of deep learning (DL) have enabled the development of large-scale multimodal predictive models for virtual drug screening. Recently, deep generative models have emerged as a powerful tool to explore the chemical space and raise hopes to expedite the drug discovery process. Following this progress in chemocentric approaches for generative chemistry, the next challenge is to build multimodal conditional generative models that leverage disparate knowledge sources when mapping biochemical properties to target structures.
Here, we call the community to bridge drug discovery more closely with systems biology when designing deep generative models. Complementing the plethora of reviews on the role of DL in chemoinformatics, we specifically focus on the interface of predictive and generative modelling for drug discovery. Through a systematic publication keyword search on PubMed and a selection of preprint servers (arXiv, biorXiv, chemRxiv, and medRxiv), we quantify trends in the field and find that molecular graphs and VAEs have become the most widely adopted molecular representations and architectures in generative models, respectively. We discuss progress on DL for toxicity, drug-target affinity, and drug sensitivity prediction and specifically focus on conditional molecular generative models that encompass multimodal prediction models. Moreover, we outline future prospects in the field and identify challenges such as the integration of deep learning systems into experimental workflows in a closed-loop manner or the adoption of federated machine learning techniques to overcome data sharing barriers. Other challenges include, but are not limited to interpretability in generative models, more sophisticated metrics for the evaluation of molecular generative models, and, following up on that, community-accepted benchmarks for both multimodal drug property prediction and property-driven molecular design.
Export Options
About this article
Cite this article as:
Born Jannis *, Manica Matteo , Trends in Deep Learning for Property-driven Drug Design, Current Medicinal Chemistry 2021; 28 (38) . https://dx.doi.org/10.2174/0929867328666210729115728
DOI https://dx.doi.org/10.2174/0929867328666210729115728 |
Print ISSN 0929-8673 |
Publisher Name Bentham Science Publisher |
Online ISSN 1875-533X |
Call for Papers in Thematic Issues
Advances in Medicinal Chemistry: From Cancer to Chronic Diseases.
The broad spectrum of the issue will provide a comprehensive overview of emerging trends, novel therapeutic interventions, and translational insights that impact modern medicine. The primary focus will be diseases of global concern, including cancer, chronic pain, metabolic disorders, and autoimmune conditions, providing a broad overview of the advancements in ...read more
Approaches to the treatment of chronic inflammation
Chronic inflammation is a hallmark of numerous diseases, significantly impacting global health. Although chronic inflammation is a hot topic, not much has been written about approaches to its treatment. This thematic issue aims to showcase the latest advancements in chronic inflammation treatment and foster discussion on future directions in this ...read more
Cellular and Molecular Mechanisms of Non-Infectious Inflammatory Diseases: Focus on Clinical Implications
The Special Issue covers the results of the studies on cellular and molecular mechanisms of non-infectious inflammatory diseases, in particular, autoimmune rheumatic diseases, atherosclerotic cardiovascular disease and other age-related disorders such as type II diabetes, cancer, neurodegenerative disorders, etc. Review and research articles as well as methodology papers that summarize ...read more
Chalcogen-modified nucleic acid analogues
Chalcogen-modified nucleosides, nucleotides and oligonucleotides have been of great interest to scientific research for many years. The replacement of oxygen in the nucleobase, sugar or phosphate backbone by chalcogen atoms (sulfur, selenium, tellurium) gives these biomolecules unique properties resulting from their altered physical and chemical properties. The continuing interest in ...read more
![](/images/wayfinder.jpg)
- Author Guidelines
- Graphical Abstracts
- Fabricating and Stating False Information
- Research Misconduct
- Post Publication Discussions and Corrections
- Publishing Ethics and Rectitude
- Increase Visibility of Your Article
- Archiving Policies
- Peer Review Workflow
- Order Your Article Before Print
- Promote Your Article
- Manuscript Transfer Facility
- Editorial Policies
- Allegations from Whistleblowers
- Announcements
Related Articles
-
Macrocyclic Histone Deacetylase Inhibitors
Current Topics in Medicinal Chemistry An Updated Review on <i>Betacoronavirus</i> Viral Entry Inhibitors: Learning from Past Discoveries to Advance COVID-19 Drug Discovery
Current Topics in Medicinal Chemistry Reversing Aberrant Methylation Patterns in Cancer
Current Medicinal Chemistry The Role of Microbiota and Intestinal Permeability in the Pathophysiology of Autoimmune and Neuroimmune Processes with an Emphasis on Inflammatory Bowel Disease Type 1 Diabetes and Chronic Fatigue Syndrome
Current Pharmaceutical Design Regulators of Chemokine Receptor Activity as Promising Anticancer Therapeutics
Current Cancer Drug Targets Fibromyalgia and Bipolar Disorder: Emerging Epidemiological Associations and Shared Pathophysiology
Current Molecular Medicine Exploring the Potential of Neuroproteomics in Alzheimer's Disease
Current Topics in Medicinal Chemistry Hyperuricaemia and Non-Alcoholic Fatty Liver Disease (NAFLD): A Relationship with Implications for Vascular Risk?
Current Vascular Pharmacology Potential of Gene Therapy for the Treatment of Pituitary Tumors
Current Gene Therapy Meet Our Editorial Board Member:
Current Drug Targets Sealing the Broken Barrier in IBD: Intestinal Permeability, Epithelial Cells and Junctions
Current Drug Targets Metabolic Obesity: The Paradox Between Visceral and Subcutaneous Fat
Current Diabetes Reviews Candidate Genes Implicated in Type 1 Diabetes Susceptibility
Current Diabetes Reviews Endothelin-1 and Endothelin Converting Enzyme-1 in Human Atherosclerosis - Novel Targets for Pharmacotherapy in Atherosclerosis
Current Vascular Pharmacology A Review on Aptamers Selection and Application in Heart DiseasesDiagnosis
Current Topics in Medicinal Chemistry Effects of Hypoxia and Ischemia on MicroRNAs in the Brain
Current Medicinal Chemistry Energy Expenditure of Hunter-Gatherers: When Statistics Turns to be Unreliable
Endocrine, Metabolic & Immune Disorders - Drug Targets The Acquired Deficiency of C1-Inhibitor: Lymphoproliferation and Angioedema
Current Molecular Medicine Receptor-Based Design of Cytokine Therapeutics
Current Pharmaceutical Biotechnology Lessons from Nature: Sources and Strategies for Developing AMPK Activators for Cancer Chemotherapeutics
Anti-Cancer Agents in Medicinal Chemistry