Drug Development and Discovery Considering Artificial Intelligence: A Through Analysis

Abstract

Artificial Intelligence (AI) is increasingly reshaping drug discovery and development by offering new computational capabilities that significantly enhance efficiency, accuracy, and innovation. This comprehensive review discusses the evolving role of AI across various stages of pharmaceutical R&D—from early target identification and validation to lead optimization, preclinical assessment, and clinical trials. With the growing complexity and costs associated with traditional drug development pipelines, AI presents powerful alternatives through machine learning (ML), deep learning (DL), and natural language processing (NLP) tools that enable rapid data analysis, compound generation, and predictive modeling. In target discovery, AI algorithms analyze vast omics datasets to identify novel biological targets, while virtual screening models streamline high-throughput screening of chemical libraries with improved hit rates. Lead optimization benefits from AI’s ability to predict ADMET (absorption, distribution, metabolism, excretion, and toxicity) profiles, thus reducing the failure rate in later stages. In clinical research, AI assists in patient stratification, real-time monitoring, and biomarker identification, accelerating trial timelines and enhancing patient safety. This review evaluates prominent AI platforms such as Deep Chem, Atom Net, Schrödinger’s suite, and AlphaFold, along with case studies from industry leaders like Pfizer, Novartis, and Insilco Medicine. Challenges such as data quality, model interpretability, algorithmic bias, and regulatory concerns are critically analyzed. The paper concludes by identifying future research opportunities and emphasizing the need for collaborative frameworks between AI developers, biologists, and regulatory bodies. As AI continues to evolve, its integration into the drug discovery lifecycle holds the promise of significantly transforming pharmaceutical innovation, enabling more targeted therapies and advancing the vision of precision medicine.

Keywords

Introduction

Table 1: Key AI Techniques and Their Applications in Drug Discovery

2. Overview of the Drug Development Pipeline

Table 2: AI Applications and Emerging Software in Drug Development Stages

Table 3: Comprehensive Overview of AI in Target Identification & Validation

3.1 Machine Learning for Multi-Omics Integration Machine learning approaches like random forests (RF), ensemble learning models, and support vector machines (SVM) are frequently employed to classify disease-relevant genes or proteins from massive multi-omics datasets. These methods can extract features, cluster disease subtypes, and prioritize potential drug targets based on their functional roles and network topologies [1,2]. For example, tools like DeepTarget leverage neural networks to combine gene expression, mutation frequency, and pathway involvement to predict viable targets in cancer therapy (Zeng et al., 2020) [3]. Similarly, PANDAomics by Insilico Medicine utilizes AI to rank targets based on disease association, biological relevance, druggability, and novelty [4].

3.2 Protein-Protein Interaction (PPI) Networks and AI Deep learning has shown promise in modeling protein-protein interactions, a crucial aspect of identifying nodes central to disease progression. Platforms such as STRING, BioGRID, and HINT offer curated PPI databases that, when coupled with graph neural networks (GNNs), can reveal hidden relationships within the proteome. These insights enable the identification of key regulatory proteins and interaction hubs [5,6].

3.3 Predictive Modeling and Literature Mining Models for natural language processing (NLP), among them Bidirectional Encoder Representations from Transformers (BERT) and SciBERT, are applied to mine biomedical literature and databases like PubMed, identifying emerging targets, associated pathways, and biomarkers [7]. AI-driven tools like IBM Watson Discovery can analyze thousands of scientific papers to extract meaningful patterns and hypotheses [8].

3.4 Structural and Functional Annotation with AI AI also contributes to functional annotation and structural prediction of potential targets. AlphaFold, developed by DeepMind, revolutionized protein structure prediction with over 90% accuracy, enabling the visualization of binding sites and aiding in structure-based drug design. [9]

Case Example: A study by Aliper et al. (2016) demonstrated the use of DL models trained on transcriptomic profiles to distinguish between cancerous and non-cancerous cells and identify differential gene expression patterns that could serve as target leads for specific cancer types. These models significantly outperformed conventional clustering techniques [10]. Overall, AI empowers researchers to overcome data complexity and variability in biological systems, enhancing the efficiency and accuracy of target identification and validation. It paves the way for personalized therapeutic strategies by identifying targets that are context-specific and more likely to succeed in downstream development stages.

4. AI in Virtual Screening and Drug Design Virtual screening (VS) and drug design are critical components of the early drug discovery process. These techniques aim to identify potential lead compounds with high affinity for a target protein by screening large chemical libraries. Traditional approaches include ligand-based and structure-based virtual screening using molecular docking and pharmacophore modeling. However, these methods often suffer from limited accuracy, high false-positive rates, and significant computational burden. Artificial Intelligence (AI) addresses these challenges by enabling more precise predictions of molecular interactions, binding affinities, and drug-likeness properties. With advancements in deep learning, generative modeling, and reinforcement learning, AI has transformed both virtual screening and de novo drug design.

4.1 AI-Driven Virtual Screening Deep learning models such as convolutional neural networks (CNNs), graph neural networks (GNNs), and recurrent neural networks (RNNs) are employed to predict the bioactivity of compounds against a given target. Tools like AtomNet utilize 3D CNNs for structure-based virtual screening, enabling accurate identification of active compounds [11]. DeepChem and Chemprop are other open-source platforms that provide ML-based frameworks for property prediction, binding affinity estimation, and molecular classification [12,13]. Graph-based deep learning models excel at representing molecular structures and interactions. GNNs consider atom and bond features as nodes and edges, respectively, and can predict activity, toxicity, and solubility with high accuracy. These models significantly outperform traditional quantitative structure-activity relationship (QSAR) methods.

4.2 AI in De Novo Drug Design AI models like Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Reinforcement Learning (RL) are increasingly used to generate novel chemical structures with desired properties. The generative model learns the chemical space and generates synthetically feasible molecules optimized for drug-likeness, ADMET properties, and target binding. One prominent example is Insilico Medicine, which designed and synthesized potent DDR1 kinase inhibitors using a generative pipeline combining RL and GANs in less than 18 months—a process that traditionally takes 4–6 years [14]. Similarly, BenevolentAI and Exscientia use AI for automated compound generation and optimization, achieving high hit-to-lead ratios.

4.3 AI for Docking and Binding Affinity Prediction AI-powered docking algorithms use DL models to predict ligand-protein binding poses and scoring functions. Tools like DeepDock, OnionNet, and KDEEP have been shown to outperform classical scoring functions in blind docking challenges [15,16]. These models learn spatial features from protein-ligand complexes and generalize across diverse targets.

4.4 Case Studies and Applications

• COVID-19 Drug Discovery: AI-based drug screening was rapidly employed to identify potential inhibitors of SARS-CoV-2 proteins. BenevolentAI identified baricitinib as a repurposing candidate for COVID-19 treatment, later validated in clinical settings [17].
• AI-Based Fragment Screening: Schrödinger’s Glide and DeepDock tools were used in campaigns for oncology and CNS diseases, integrating fragment-based drug discovery (FBDD) with DL models to improve early-stage hit generation.

AI's role in virtual screening and molecular design is revolutionizing lead identification and optimization. It offers unprecedented scalability, adaptability, and predictive accuracy, reducing time-to-hit and increasing the probability of downstream success.

5. AI in Preclinical Testing and Safety Assessment Preclinical testing is a crucial phase in the drug development process, involving in vitro (cell culture) and in vivo (animal) studies to evaluate the safety, toxicity, pharmacokinetics (PK), and pharmacodynamics (PD) of drug candidates. This phase is vital for understanding how a drug behaves in a biological system and determining whether it is safe enough to progress to human clinical trials. However, traditional preclinical models are costly, time-consuming, ethically controversial, and not always predictive of human outcomes. Artificial Intelligence (AI) is increasingly being applied to improve the predictive accuracy of preclinical assessments and reduce dependence on animal testing. AI enables the analysis of large-scale toxicogenomic, pharmacogenomic, and bioassay data to forecast potential adverse events, optimize compound dosing, and model drug metabolism.

5.1 In Silico Toxicity Prediction Machine learning models such as random forests, support vector machines (SVM), and deep neural networks (DNN) are widely used to predict toxicity endpoints including hepatotoxicity, cardiotoxicity, genotoxicity, and nephrotoxicity. Tools such as DeepTox, ProTox-II, and ADMETlab apply AI to analyze chemical structures and predict toxicological outcomes before any laboratory testing [18,19].

5.2 Predictive Pharmacokinetics and Metabolism AI models can simulate Absorption, Distribution, Metabolism, Excretion, and Toxicity (ADMET) profiles of drug candidates. For instance, the pkCSM tool uses graph-based signatures to forecast oral bioavailability, blood-brain barrier penetration, and cytochrome P450 interactions [20]. Deep learning approaches like those used by ADMET Predictor (Simulations Plus) further enhance prediction accuracy and guide compound optimization.

5.3 Systems Biology and Organs-on-Chips Systems biology modeling, powered by AI, integrates omics and physiological data to simulate tissue-level responses to drugs. When combined with microfluidic technologies (organ-on-chip), AI can model complex biological interactions to predict human-relevant outcomes with greater reliability. For example, AI-enhanced liver-on-chip models can detect hepatotoxicity more accurately than standard animal models [21].

5.4 Drug-Drug Interaction (DDI) Prediction AI also plays a vital role in predicting potential drug-drug interactions, a common reason for post-marketing drug withdrawals. Deep learning models trained on electronic health records (EHRs), pharmacovigilance databases, and molecular data can anticipate adverse interactions early in the pipeline. DeepDDI is one such model using deep neural networks to identify clinically relevant DDIs [22].

5.5 Case Studies

• Merck has integrated AI models with high-content screening data to predict neurotoxicity and prioritize safe leads earlier in the process.
• Novartis and Atomwise collaborate to use AI in predicting mitochondrial toxicity, one of the leading causes of late-stage failure.

Figure 3: AI in Preclinical Safety Workflow (A flowchart depicting data ingestion → feature extraction → toxicity/PK/PD modeling → output visualization and risk scoring.)

By leveraging AI in preclinical testing, pharmaceutical companies can identify safety liabilities earlier, minimize reliance on animal models, and accelerate regulatory submissions with higher confidence.

6. AI in Clinical Trials and Patient Stratification Clinical trials are essential for evaluating the safety and efficacy of new drugs in humans, but they are also among the most expensive and time-consuming phases of drug development. Traditional trial designs often suffer from high failure rates, low recruitment efficiency, and lack of personalized treatment strategies. Artificial Intelligence (AI) offers novel solutions to these challenges by enhancing patient recruitment, optimizing trial design, stratifying patients, and enabling real-time monitoring.

6.1 Patient Recruitment and Matching Recruiting eligible participants is a major bottleneck in clinical trials. AI-driven natural language processing (NLP) systems can analyze electronic health records (EHRs), clinical notes, and diagnostic reports to match patients with trial inclusion and exclusion criteria. For instance, IBM Watson for Clinical Trial Matching has shown a 70% reduction in trial screening time by automatically identifying qualified participants from clinical databases [23].

6.2 Trial Design Optimization Machine learning models are used to simulate various trial scenarios, allowing researchers to predict potential outcomes and adapt protocols accordingly. Bayesian adaptive trial designs, supported by AI, can adjust randomization probabilities based on interim results, improving trial efficiency and ethical considerations. AI also helps in selecting endpoints, optimizing dosing regimens, and predicting dropout rates [24].

6.3 Patient Stratification and Precision Medicine AI can segment patients into subgroups based on biomarkers, genetic profiles, lifestyle data, and disease phenotypes. Unsupervised learning algorithms like clustering and t-SNE are used to identify hidden patient subpopulations that may respond differently to treatments. This stratification supports precision medicine, ensuring the right drug is given to the right patient at the right time [25].

6.4 Real-Time Monitoring and Remote Data Collection With the rise of wearable devices and digital health platforms, AI facilitates continuous remote monitoring of trial participants. Deep learning algorithms can analyze data from heart rate monitors, glucose sensors, and sleep trackers to detect early signs of adverse events or non-compliance. These tools enhance patient safety and reduce the need for in-person visits [26].

6.5 Case Study: AI in Oncology Trials Pfizer partnered with Concerto HealthAI to use real-world evidence and AI to optimize oncology trials. The collaboration helped improve trial feasibility, identify responsive patient groups, and reduce protocol amendments, which are typically time-consuming and costly.

By integrating AI into clinical trials, the pharmaceutical industry is moving toward more dynamic, data-driven, and patient-centric research models. These innovations can shorten trial durations, improve outcome predictability, and increase regulatory acceptance of trial results.

7. Challenges, Limitations, and Ethical Considerations While the integration of Artificial Intelligence (AI) into drug discovery and development offers remarkable potential, several challenges and ethical concerns must be addressed to ensure responsible and effective implementation.

7.1 Data Quality and Standardization AI models rely heavily on high-quality, well-annotated datasets. Inconsistent data formatting, missing values, and heterogeneous sources (e.g., clinical, genomic, imaging) can reduce model accuracy and reproducibility. Moreover, lack of standardization in data collection protocols across institutions hinders the integration and generalization of AI models [27].

7.2 Interpretability and Transparency Most deep learning models function as “black boxes,” making it difficult to interpret how specific decisions are made. This lack of transparency can hinder the trust of regulatory bodies, clinicians, and patients. Developing explainable AI (XAI) models is crucial to ensure traceability, accountability, and informed decision-making in healthcare applications [28].

7.3 Regulatory and Validation Frameworks AI-based tools used in drug discovery must adhere to stringent validation standards before regulatory acceptance. However, there is a lack of clear regulatory guidelines tailored specifically for AI systems. The U.S. FDA and European Medicines Agency (EMA) are actively developing frameworks, but harmonization across global agencies remains a challenge [29].

7.4 Bias and Generalizability AI models can inadvertently reflect biases present in training datasets, leading to unequal performance across different patient populations. This is especially concerning in precision medicine, where biased models can result in suboptimal or unsafe treatment recommendations for underrepresented groups [30].

7.5 Ethical and Privacy Concerns AI applications in healthcare often require access to sensitive patient data. Ensuring patient privacy, data ownership, and compliance with regulations such as the General Data Protection Regulation (GDPR) and Health Insurance Portability and Accountability Act (HIPAA) is essential. Additionally, ethical dilemmas arise when AI systems are involved in life-altering decisions without sufficient human oversight [31].

7.6 Talent and Infrastructure Gaps The implementation of AI technologies in drug development requires a skilled workforce proficient in data science, biology, and regulatory science. Many pharmaceutical companies face challenges in building interdisciplinary teams and developing the necessary computational infrastructure to support AI workflows [32].

7.7 Cost and Resource Allocation Although AI promises long-term savings, its initial implementation can be costly. Investments are needed in data infrastructure, high-performance computing, model training, and integration into existing R&D pipelines. For smaller firms and academic institutions, such costs may be prohibitive without collaborative partnerships. Despite these challenges, continued innovation, policy development, and interdisciplinary collaboration can help address limitations and foster responsible AI use in drug discovery.

8. Future Directions and Conclusion

The application of Artificial Intelligence (AI) in drug discovery and development is rapidly evolving, offering transformative potential across the entire pharmaceutical value chain. Looking ahead, several promising directions are expected to redefine the landscape of biomedical research and therapeutic innovation. The future of AI in drug discovery and development holds tremendous promise. Integration of multi-omics data-combining genomic, proteomic, metabolomic, and other biological information-will enable more comprehensive modeling of disease mechanisms and drug effects. This holistic approach will allow for more precise target identification and personalized treatment strategies. Advances in quantum computing may further accelerate AI applications in drug discovery by enabling more complex molecular simulations and property predictions. Federated learning approaches could facilitate collaborative research while preserving data privacy, allowing organizations to collectively train AI models without sharing sensitive information. The combination of AI with laboratory automation represents another promising direction. AI-guided robotic systems can design, execute, and analyze experiments with minimal human intervention, creating a closed-loop discovery process that iteratively improves based on experimental results. As AI technologies continue to evolve, their integration into the drug development pipeline will likely become more seamless and comprehensive. This evolution will require ongoing collaboration between AI researchers, drug developers, clinicians, and regulatory authorities to ensure that AI-driven approaches deliver safe, effective, and accessible therapeutic innovations.

8.1 Integration with Multi-Modal Data The future of AI in drug discovery lies in the integration of diverse data modalities, including genomics, proteomics, transcriptomics, metabolomics, imaging, and electronic health records (EHRs). Multi-modal AI models will enable a more holistic understanding of disease mechanisms and drug responses, allowing for precise, patient-specific therapeutic interventions. Tools like Google's DeepMind and Meta AI are advancing this capability with deep learning architectures capable of processing heterogeneous biomedical data at scale.

8.2 Federated and Privacy-Preserving Learning Federated learning models are gaining attention for their ability to train AI algorithms across decentralized datasets without sharing sensitive patient data. This approach helps overcome data privacy concerns and facilitates cross-institutional collaborations. As privacy regulations become more stringent, such models will play a critical role in clinical AI applications.

8.3 AI-Augmented Drug Repurposing AI is expected to significantly enhance drug repurposing strategies by identifying new therapeutic uses for existing drugs. This is particularly valuable in pandemic scenarios and rare diseases, where time and resources are limited. Platforms like BenevolentAI and Healx are already pioneering this field with AI models trained on biomedical knowledge graphs and real-world evidence.

8.4 Digital Twins and Personalized Drug Testing The concept of digital twins—virtual replicas of individual patients—combined with AI could revolutionize personalized medicine. These models simulate disease progression and drug response in silico, enabling personalized treatment regimens, dosing strategies, and adverse event prediction before administration.

8.5 Regulatory Harmonization and Ethical AI Future success of AI in drug development will depend on the establishment of globally harmonized regulatory frameworks that address data integrity, model validation, bias mitigation, and ethical considerations. Collaboration between regulators, industry stakeholders, and academic institutions will be essential to achieve trust and widespread adoption.

8.6 Enhanced Human-AI Collaboration Rather than replacing scientists, AI will increasingly serve as a decision-support partner, augmenting human capabilities. The synergy between domain experts and AI tools will accelerate hypothesis generation, streamline experimentation, and improve R&D productivity.

CONCLUSION-AI has emerged as a powerful force in transforming drug discovery and development, from target identification to clinical trials. By enabling faster, cheaper, and more accurate research processes, AI holds the promise of accelerating the delivery of safer and more effective therapies to patients. Despite existing challenges related to data quality, model transparency, and regulatory readiness, the future is optimistic. As technologies mature and collaborative ecosystems evolve, AI will become an indispensable pillar of precision medicine and pharmaceutical innovation. Artificial intelligence has emerged as a transformative force in drug discovery and development, offering innovative solutions to longstanding challenges in pharmaceutical research. By accelerating target identification, enabling virtual screening of vast compound libraries, predicting molecular properties, optimizing clinical trials, and advancing personalized medicine, AI technologies are reshaping the entire drug development landscape. While challenges related to data quality, regulatory considerations, and ethical concerns persist, the rapid pace of innovation in AI methodologies suggests that many of these limitations will be addressed in the coming years. The integration of diverse AI techniques-from machine learning and deep learning to generative AI and large language models-provides a rich toolkit for addressing complex problems across the drug development pipeline. The successful implementation of AI in pharmaceutical research requires interdisciplinary collaboration, combining expertise in computer science, biology, chemistry, medicine, and regulatory affairs^]. By fostering such collaborative approaches and continuing to advance AI technologies, the field stands poised to dramatically accelerate the discovery and development of novel therapies, potentially transforming patient care and reducing the global burden of disease. As we look to the future, the synergy between human expertise and artificial intelligence will likely define a new paradigm for drug discovery and development-one characterized by greater efficiency, reduced costs, higher success rates, and more personalized therapeutic approaches. This evolution promises to benefit not only the pharmaceutical industry but also healthcare systems and, most importantly, patients awaiting new treatments for challenging medical conditions.

REFERENCES

1. Subramanian I, Verma S, Kumar S, Jere A, Anamika K. Multi-omics Data Integration, Interpretation, and Its Application. Bioinform Biol Insights. 2020;14:1177932219899051.
2. Angermueller C, Pärnamaa T, Parts L, Stegle O. Deep learning for computational biology. Mol Syst Biol. 2016;12(7):878.
3. Zeng X, Song X, Ma T, Pan X, Zhou Y, Hou Y, Zhang Z. DeepTarget: End-to-end learning framework for predicting protein targets of small molecules. J Cheminform. 2020;12(1):49.
4. Zhavoronkov A, Vanhaelen Q, Oprea TI. Will Artificial Intelligence for Drug Discovery Impact Clinical Pharmacology? Clin Pharmacol Ther. 2020;107(4):780–785.
5. Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, et al. STRING v11: protein–protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607–D613.
6. Stark C, Breitkreutz BJ, Reguly T, Boucher L, Breitkreutz A, Tyers M. BioGRID: a general repository for interaction datasets. Nucleic Acids Res. 2006;34(Database issue):D535–D539.
7. Beltagy I, Lo K, Cohan A. SciBERT: A Pretrained Language Model for Scientific Text. arXiv preprint arXiv:1903.10676. 2019.
8. Ferrucci D. Introduction to “This is Watson”. IBM J Res Dev. 2012;56(3.4):1:1–1:15.
9. Jumper J, Evans R, Pritzel A, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596(7873):583–589.
10. Aliper A, Plis S, Artemov A, Ulloa A, Mamoshina P, Zhavoronkov A. Deep learning applications for predicting pharmacological properties of drugs and drug repurposing using transcriptomic data. Mol Pharm. 2016;13(7):2524–2530.
11. Wallach I, Dzamba M, Heifets A. AtomNet: A deep convolutional neural network for bioactivity prediction in structure-based drug discovery. arXiv preprint arXiv:1510.02855. 2015.
12. Ramsundar B, Eastman P, Walters P, Pande V. Deep Learning for the Life Sciences. O'Reilly Media; 2019.
13. Yang K, Swanson K, Jin W, Coley CW, Eiden P, Gao H, et al. Analyzing Learned Molecular Representations for Property Prediction. J Chem Inf Model. 2019;59(8):3370–3388.
14. Zhavoronkov A, Ivanenkov YA, Aliper A, et al. Deep learning enables rapid identification of potent DDR1 kinase inhibitors. Nat Biotechnol. 2019;37(9):1038–1040.
15. Jiménez J, Skalic M, Martinez-Rosell G, De Fabritiis G. KDEEP: Protein–ligand absolute binding affinity prediction via 3D-convolutional neural networks. J Chem Inf Model. 2018;58(2):287–296.
16. Liu Z, Li Y, Han L, Li J, Liu J, Zhao Z, et al. PDB-wide collection of binding data: current status of the PDBbind database. Bioinformatics. 2015;31(3):405–412.
17. Richardson P, Griffin I, Tucker C, Smith D, Oechsle O, Phelan A, et al. Baricitinib as potential treatment for 2019-nCoV acute respiratory disease. Lancet. 2020;395(10223):e30–e31.
18. Mayr A, Klambauer G, Unterthiner T, Hochreiter S. DeepTox: Toxicity prediction using deep learning. Front Environ Sci. 2016;3:80.
19. Banerjee P, Eckert AO, Schrey AK, Preissner R. ProTox-II: A webserver for the prediction of toxicity of chemicals. Nucleic Acids Res. 2018;46(W1):W257–W263.
20. Pires DE, Blundell TL, Ascher DB. pkCSM: Predicting small-molecule pharmacokinetic and toxicity properties using graph-based signatures. J Med Chem. 2015;58(9):4066–4072.
21. Zhang B, Korolj A, Lai BFL, Radisic M. Advances in organ-on-a-chip engineering. Nat Rev Mater. 2018;3(8):257–278.
22. Ryu JY, Kim HU, Lee SY. Deep learning improves prediction of drug–drug and drug–food interactions. Proc Natl Acad Sci USA. 2018;115(18):E4304–E4311.
23. Ross C, Swetlitz I. IBM’s Watson supercomputer recommended ‘unsafe and incorrect’ cancer treatments, internal documents show. STAT News. 2018.
24. Bhatt DL, Mehta C. Adaptive Designs for Clinical Trials. N Engl J Med. 2016;375(1):65–74.
25. Shah P, Kendall F, Khozin S, Goosen R, Hu J, Laramie J, et al. Artificial intelligence and machine learning in clinical development: a translational perspective. NPJ Digit Med. 2019;2:69.
26. Montalbano A, Toumazou C. Wearable sensors for real-time clinical trial monitoring: current trends and future opportunities. ACS Sens. 2020;5(7):2244–2260.
27. Beam AL, Kohane IS. Big data and machine learning in health care. JAMA. 2018;319(13):1317–1318.
28. Holzinger A, Biemann C, Pattichis CS, Kell DB. What do we need to build explainable AI systems for the medical domain? Rev Life Sci. 2017;36:6–10.
29. U.S. Food & Drug Administration. Artificial Intelligence and Machine Learning in Software as a Medical Device. 2021.
30. Gianfrancesco MA, Tamang S, Yazdany J, Schmajuk G. Potential biases in machine learning algorithms using electronic health record data. JAMA Intern Med. 2018;178(11):1544–1547.
31. Price WN, Cohen IG. Privacy in the age of medical big data. Nat Med. 2019;25(1):37–43.
32. Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44–56.
33. IBM Watson Health, Mayo Clinic. AI to accelerate patient matching. [Online]. Available: https://www.ibm.com/watson-health/
34. FDA. Artificial Intelligence/Machine Learning-Based Software as a Medical Device (SaMD) Action Plan. 2021.
35. EMA. Big Data Task Force. European Medicines Agency, 2020.
36. Insilico Medicine. AI-designed drug enters clinical trials. [Press Release]. 2021.
37. Artificial intelligence in drug discovery and development. PubMed Central (PMC). Published October 21, 2020. Available from: https://pmc.ncbi.nlm.nih.gov/articles/PMC7577280/
38. Artificial Intelligence in Pharmaceutical Technology and Drug Development. PubMed Central (PMC). Published July 10, 2023. Available from: https://pmc.ncbi.nlm.nih.gov/articles/PMC10385763/
39. Early evidence and emerging trends: how AI is shaping drug discovery and clinical development. Drug Target Review. Published March 26, 2025. Available from: https://www.drugtargetreview.com/article/158593/early-evidence-and-emerging-trends-how-ai-is-shaping-drug-discovery-and-clinical-development/
40. Artificial intelligence in drug discovery and development. In: Artificial Intelligence in Drug Discovery and Development. ScienceDirect. Available from: https://www.sciencedirect.com/science/article/abs/pii/B9780443223082000184
41. Artificial intelligence for drug discovery and development in neurodegenerative diseases. ScienceDirect. Available from: https://www.sciencedirect.com/science/article/pii/S0959440X24000034
42. Artificial intelligence in drug discovery and development. PubMed. Available from: https://pubmed.ncbi.nlm.nih.gov/33099022/
43. Artificial intelligence in drug development: reshaping the therapeutic landscape. Therapeutic Advances in Drug Safety. Published February 13, 2025. Available from: https://journals.sagepub.com/doi/full/10.1177/20420986251321704
44. Nelson JL, Smith BL, Jared JD, Younger JG. Prospective trial of real-time electronic surveillance to expedite early care of severe sepsis. Ann Emerg Med. 2011;57:500–504.
45. Tan C, Chen H, Xia C. The prediction of cardiovascular disease based on trace element contents in hair and a classifier of boosting decision stumps. Biol Trace Elem Res. 2008;129:9–19.
46. Chen H, Tan C. Prediction of Type-2 diabetes based on several element levels in blood and chemometrics. Biol Trace Elem Res. 2011;147:67–74.
47. Hooper MH, Weavind L, Wheeler AP, et al. Randomized trial of automated, electronic monitoring to facilitate early detection of sepsis in the intensive care unit. Crit Care Med. 2012;40:2096–2101.
48. Giannini HM, Ginestra JC, Chivers C, et al. A machine learning algorithm to predict severe sepsis and septic shock: development, implementation, and impact on clinical practice. Crit Care Med. 2019.
49. Naushad SM, Janaki Ramaiah M, Pavithrakumari M, et al. Artificial neural network-based exploration of gene-nutrient interactions in folate and xeno-biotic metabolic pathways that modulate susceptibility to breast cancer. Gene. 2016;580:159–168.
50. Vamathevan J, Clark D, Czodrowski P, et al. Applications of machine learning in drug discovery and development. Nat Rev Drug Discov. 2019;18(6):463–477.
51. Gawehn E, Hiss JA, Schneider G. Deep learning in drug discovery. Mol Inform. 2016;35(1):3–14.
52. Gómez-Bombarelli R, Wei JN, Duvenaud D, et al. Automatic chemical design using a data-driven continuous representation of molecules. ACS Cent Sci. 2018;4(2):268–276.
53. Schneider G. Automating drug discovery. Nat Rev Drug Discov. 2018;17(2):97–113.
54. Miotto R, Li L, Kidd BA, Dudley JT. Deep Patient: an unsupervised representation to predict the future of patients from the electronic health records. Sci Rep. 2016;6:26094.
55. Bate A, Lindquist M, Edwards IR, et al. A Bayesian neural network method for adverse drug reaction signal generation. Eur J Clin Pharmacol. 1998;54(4):315–321.
56. Izmailova ES, Wagner JA, Perakslis ED. Wearable devices in clinical trials: hype and hypothesis. Clin Pharmacol Ther. 2018;104(1):42–52.
57. Wang Y, Zeng D, Wang L. Social media analytics for drug safety signal detection. ACM Trans Manag Inf Syst. 2018;9(4):1–19.
58. Harpaz R, Callahan A, Tamang S, et al. Text mining for adverse drug events: the promise, challenges, and state of the art. Drug Saf. 2014;37(10):777–790.
59. Botsis T, Nguyen MD, Woo EJ, et al. Text mining for the biopharmaceutical domain: a systematic review of recent advances. Drug Saf. 2011;34(10):809–823.
60. Luo H, Wang J, Li M, et al. Drug repositioning based on comprehensive similarity measures and bi-random walk algorithm. Bioinformatics. 2016;32(17):2664–2671.
61. Carvalho C, et al. Functional genomics and CRISPR screens in drug discovery: Synergy with AI. Trends Pharmacol Sci. 2021;42(10):765–781.
62. Lee J, et al. BioBERT: A pre-trained biomedical language representation model for biomedical text mining. Bioinformatics. 2020;36(4):1234–1240.
63. Zhou Y, et al. Network-based drug repurposing for novel coronavirus 2019-nCoV/SARS-CoV-2. Cell Discov. 2020;6:14.
64. Chen H, Engkvist O, Wang Y, Olivecrona M, Blaschke T. The rise of deep learning in drug discovery. Drug Discov Today. 2018;23(6):1241–1250.
65. Ekins S, Puhl AC, Zorn KM, et al. Exploiting machine learning for end-to-end drug discovery and development. Nat Mater. 2019;18(5):435–441.
66. Stokes JM, Yang K, Swanson K, et al. A deep learning approach to antibiotic discovery. Cell. 2020;180(4):688–702.e13.
67. Vamathevan J, Clark D, Czodrowski P, et al. Machine learning in drug discovery: progress and prospects. Drug Discov Today. 2019;24(3):589–601.
68. Altae-Tran H, Ramsundar B, Pappu AS, Pande V. Low data drug discovery with one-shot learning. ACS Cent Sci. 2017;3(4):283–293.
69. Walters WP, Barzilay R. Applications of deep learning in molecule generation and molecular property prediction. Acc Chem Res. 2021;54(2):263–270.
70. Zhavoronkov A. Artificial intelligence for drug discovery, biomarker development, and generation of novel chemistry. Mol Pharm. 2018;15(10):4311–4313.
71. Ragoza M, Hochuli J, Idrobo E, Sunseri J, Koes DR. Protein–ligand scoring with convolutional neural networks. J Chem Inf Model. 2017;57(4):942–957.
72. Kuenemann M, Schmidl S, Korotcov A, et al. Evaluation of graph neural networks for predicting drug–target interactions. J Chem Inf Model. 2021;61(12):5858–5868.
73. Chen L, Cruz A, Ramsey S, et al. The rise of generative models in drug design. Drug Discov Today. 2020;25(6):1034–1041.
74. Liu Q, Wu H, Wang X, et al. Application of artificial intelligence in the diagnosis and treatment of COVID-19: A systematic review. Int J Med Sci. 2021;18(3):678–689.
75. Ma J, Sheridan RP, Liaw A, Dahl GE, Svetnik V. Deep neural nets as a method for quantitative structure–activity relationships. J Chem Inf Model. 2015;55(2):263–274.
76. Elton DC, Boukouvalas Z, Fuge MD, Chung PW. Deep learning for molecular design—a review of the state of the art. Mol Syst Des Eng. 2019;4(4):828–849.
77. Gilmer J, Schoenholz SS, Riley PF, Vinyals O, Dahl GE. Neural message passing for quantum chemistry. Proceedings of the 34th International Conference on Machine Learning. 2017:1263–1272.
78. Schwaller P, Gaudin T, Lányi D, Bekas C, Laino T. “Found in translation”: predicting outcomes of complex organic chemistry reactions using neural sequence-to-sequence models. Chem Sci. 2018;9(28):6091–6098.