Curated Papers > Highly Cited Authoritative Literature on "AI Drug Discovery"

Deep learning in drug discovery: an integrative review and future challenges.

Literature Information

DOI	10.1007/s10462-022-10306-1
PMID	36415536
Journal	Artificial intelligence review
Impact Factor	13.9
JCR Quartile	Q1
Publication Year	2023
Times Cited	70
Keywords	Artificial intelligence, Deep learning, Digital twining, Drug discovery, Drug dosing optimization
Literature Type	Journal Article
ISSN	0269-2821
Pages	5975-6037
Issue	56(7)
Authors	Heba Askr, Enas Elgeldawi, Heba Aboul Ella, Yaseen A M M Elshaier, Mamdouh M Gomaa, Aboul Ella Hassanien

TL;DR

This systematic literature review examines the application of deep learning (DL) technologies in drug discovery, highlighting their role in areas such as drug-target interactions and side effect predictions, based on an analysis of over 300 articles published from 2000 to 2022. The findings underscore the potential of AI to enhance drug development efficiency while also identifying future challenges and research directions, including the integration of explainable AI and digital twinning in the field.

Search for more papers on MaltSci.com

Artificial intelligence · Deep learning · Digital twining · Drug discovery · Drug dosing optimization

Abstract

Recently, using artificial intelligence (AI) in drug discovery has received much attention since it significantly shortens the time and cost of developing new drugs. Deep learning (DL)-based approaches are increasingly being used in all stages of drug development as DL technology advances, and drug-related data grows. Therefore, this paper presents a systematic Literature review (SLR) that integrates the recent DL technologies and applications in drug discovery Including, drug-target interactions (DTIs), drug-drug similarity interactions (DDIs), drug sensitivity and responsiveness, and drug-side effect predictions. We present a review of more than 300 articles between 2000 and 2022. The benchmark data sets, the databases, and the evaluation measures are also presented. In addition, this paper provides an overview of how explainable AI (XAI) supports drug discovery problems. The drug dosing optimization and success stories are discussed as well. Finally, digital twining (DT) and open issues are suggested as future research challenges for drug discovery problems. Challenges to be addressed, future research directions are identified, and an extensive bibliography is also included.

MaltSci.com AI Research Service

Intelligent ReadingAnswer any question about the paper and explain complex charts and formulas

Locate StatementsFind traces of a specific claim within the paper

Add to KBasePerform data extraction, report drafting, and advanced knowledge mining

Primary Questions Addressed

1. What are the specific challenges faced when integrating deep learning into drug discovery processes?
2. How do different deep learning architectures compare in their effectiveness for predicting drug-target interactions?
3. What role does explainable AI play in enhancing the interpretability of deep learning models in drug discovery?
4. In what ways can digital twinning be applied to optimize drug dosing and improve clinical outcomes?
5. How can the insights from drug-drug similarity interactions inform the development of combination therapies?

Key Findings

Research Background and Purpose

The integration of artificial intelligence (AI) in drug discovery has garnered significant attention due to its potential to reduce the time and costs associated with developing new drugs. Deep learning (DL) techniques are increasingly employed across various stages of drug development, driven by advancements in technology and the growing volume of drug-related data. This paper presents a systematic literature review (SLR) that consolidates recent DL technologies and their applications in drug discovery, focusing on drug-target interactions (DTIs), drug-drug interactions (DDIs), drug sensitivity, and side effect predictions. The review also discusses the use of explainable AI (XAI) and digital twinning (DT) as future research directions.

Main Methods/Materials/Experimental Design

The review is based on an extensive analysis of over 300 articles published between 2000 and 2022. The authors organized the findings into six main building blocks:

1. Deep Learning Algorithms: Overview of various DL models used in drug discovery.
2. Drug Discovery Problem Categories: Classification of problems such as DTIs, DDIs, drug sensitivity, and side effects.
3. Benchmark Datasets and Databases: Summary of datasets used in drug discovery.
4. Evaluation Metrics: Description of metrics for assessing the performance of models.
5. Drug Dosing Optimization: Discussion on optimizing drug dosages using DL.
6. Success Stories: Examples of successful applications of DL in drug discovery.

The methodological framework can be visualized as follows:

Key Results and Findings

• DL Algorithms: Various DL techniques, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and graph neural networks (GNNs), are identified as effective for predicting drug interactions and responses.
• Applications: The review highlights the applicability of DL in predicting DTIs, DDIs, drug sensitivity, and side effects, with specific models outperforming traditional methods.
• Benchmarking: A range of datasets, including the Tox21 and Davis datasets, are used for training and validating DL models.
• Success Cases: Notable examples include the use of AI by AstraZeneca and other companies in expediting drug development processes, particularly during the COVID-19 pandemic.

Main Conclusions/Significance/Innovation

The integration of DL into drug discovery represents a transformative approach that significantly enhances the efficiency and effectiveness of the drug development process. The review emphasizes the potential of XAI to improve the interpretability of DL models, fostering trust and adoption in clinical settings. Furthermore, the concept of digital twinning is introduced as a novel approach to simulate and optimize drug interactions and responses in real-time.

Research Limitations and Future Directions

Despite the advancements, the review acknowledges several limitations:

• Data Scarcity: The availability of high-quality, annotated datasets remains a challenge.
• Model Interpretability: Many DL models still function as "black boxes," necessitating further research into XAI.
• Standardization: The lack of standardized evaluation metrics and datasets complicates comparative studies.

Future research should focus on enhancing model interpretability, expanding datasets, and exploring the full potential of digital twinning in drug discovery.

Summary Table of Key Findings

Section	Key Points
Deep Learning Algorithms	Various models like CNNs, RNNs, and GNNs are effective in predicting drug interactions.
Applications	Successful predictions of DTIs, DDIs, drug sensitivity, and side effects using DL.
Benchmarking	Utilization of datasets such as Tox21 and Davis for model training and validation.
Success Cases	Companies like AstraZeneca effectively employed AI for rapid drug development during COVID-19.
Future Directions	Emphasis on XAI, digital twinning, and improving dataset availability and model interpretability.

This structured summary encapsulates the critical aspects of the review, highlighting the transformative role of deep learning in drug discovery while identifying avenues for future research.

Literatures Citing This Work

1. An extensive survey on the use of supervised machine learning techniques in the past two decades for prediction of drug side effects. - Pranab Das;Dilwar Hussain Mazumder - Artificial intelligence review (2023)
2. A Systematic Review of Deep Learning Methodologies Used in the Drug Discovery Process with Emphasis on In Vivo Validation. - Nikoletta-Maria Koutroumpa;Konstantinos D Papavasileiou;Anastasios G Papadiamantis;Georgia Melagraki;Antreas Afantitis - International journal of molecular sciences (2023)
3. A deep learning predictive model for public health concerns and hesitancy toward the COVID-19 vaccines. - Heba Mamdouh Farghaly;Mamdouh M Gomaa;Enas Elgeldawi;Heba Askr;Yaseen A M M Elshaier;Hassan Aboul Ella;Ashraf Darwish;Aboul Ella Hassanien - Scientific reports (2023)
4. Making sense of chemical space network shows signs of criticality. - Nicola Amoroso;Nicola Gambacorta;Fabrizio Mastrolorito;Maria Vittoria Togo;Daniela Trisciuzzi;Alfonso Monaco;Ester Pantaleo;Cosimo Damiano Altomare;Fulvio Ciriaco;Orazio Nicolotti - Scientific reports (2023)
5. Artificial intelligence-driven new drug discovery targeting serine/threonine kinase 33 for cancer treatment. - Na Ly Tran;Hyerim Kim;Cheol-Hee Shin;Eun Ko;Seung Ja Oh - Cancer cell international (2023)
6. Advancing Biomedicine with Graph Representation Learning: Recent Progress, Challenges, and Future Directions. - Fang Li;Yi Nian;Zenan Sun;Cui Tao - Yearbook of medical informatics (2023)
7. Deep learning algorithms applied to computational chemistry. - Abimael Guzman-Pando;Graciela Ramirez-Alonso;Carlos Arzate-Quintana;Javier Camarillo-Cisneros - Molecular diversity (2024)
8. Artificial Intelligence in Drug Discovery: A Bibliometric Analysis and Literature Review. - Baoyu He;Jingjing Guo;Henry H Y Tong;Wai Ming To - Mini reviews in medicinal chemistry (2024)
9. Advancing bioinformatics with large language models: components, applications and perspectives. - Jiajia Liu;Mengyuan Yang;Yankai Yu;Haixia Xu;Tiangang Wang;Kang Li;Xiaobo Zhou - ArXiv (2025)
10. Techniques and Strategies in Drug Design and Discovery. - George Mihai Nitulescu - International journal of molecular sciences (2024)

... (60 more literatures)

---

© 2025 MaltSci - We reshape scientific research with AI technology