Pharma Focus Asia

AI Cures

Repurposing Old Drugs for New Threats

Anuradha Singh, Department of Chemistry Sadanlal Sanwaldas Khanna Girl's Degree College (A Constituent College of the University of Allahabad)

The growing threat of emerging infectious diseases and antimicrobial resistance demands innovative solutions in drug discovery. Artificial intelligence (AI) is revolutionising this field by accelerating drug repurposing, a strategy that identifies new uses for existing medications. AI's ability to analyse vast datasets, uncover hidden patterns, and predict drug-target interactions is transforming how we combat infectious threats. This approach offers faster, safer, and more cost-effective treatments by leveraging existing drug knowledge. AI's multifaceted role in drug repurposing spans target identification, virtual screening, drug optimisation, and clinical trial design. Real-world examples demonstrate AI's potential to address evolving pathogenic challenges and deliver innovative therapies for infectious diseases.

Beyond Serendipity

Ever wondered why the specter of infectious diseases like COVID-19 still haunts us, despite decades of medical advancements? The answer lies in the escalating threat of antimicrobial resistance and the limitations of traditional drug discovery, a process often as serendipitous as finding a needle in a haystack. But what if we could repurpose existing drugs, breathing new life into them with the help of artificial intelligence (AI)?

Repurposing Existing Medications: A Double-Edged Sword in the Fight against Disease

Drug repurposing, the identification of new therapeutic uses for existing drugs, offers a compelling strategy to accelerate drug development and address unmet medical needs. By leveraging prior research and established safety profiles, this approach can potentially streamline timelines and reduce costs compared to de novo drug discovery. The success story of sildenafil (Viagra), originally developed for angina but repurposed for erectile dysfunction, exemplifies the transformative potential of drug repurposing. Hydroxychloroquine anantimalarial drug has been explored for various conditions, including COVID-19, although its efficacy remains controversial and requires further rigorous investigation.

Advantages and Opportunities

  • Accelerated Development: Repurposed drugs often bypass early-stage clinical trials, saving valuable time and resources.
  • Reduced Costs: Leveraging existing research significantly lowers the financial burden of drug development.
  • Novel Treatment Options: Repurposing can unlock unexpected therapeutic applications for existing drugs, expanding treatment options for various diseases.

Challenges and Considerations

  • Limited Specificity: Some repurposed drugs may lack optimal target specificity, potentially leading to adverse effects or suboptimal efficacy. Thalidomide, initially used for morning sickness but later linked to birth defects, tragically illustrates this risk.
  • Unintended Consequences: Repurposing can uncover unforeseen side effects or unintended therapeutic applications. Off-target effects, unintended interactions with molecules other than the intended target, can lead to unwanted side effects. Minoxidil, originally an antihypertensive, surprisingly promoted hair growth, highlighting the importance of thorough investigation.
  • Intellectual Property and Regulatory Hurdles: Existing patents and the need for additional clinical trials can pose barriers to the successful repurposing of some drugs.

AI: A Catalyst for Drug Repurposing

AI is transforming drug repurposing, offering a faster, safer, and more cost-effective approach to drug discovery. By analysing vast datasets, AI algorithms swiftly identify promising drug candidates for new therapeutic applications, bypassing the lengthy and expensive traditional development process.

Predict and mitigate off-target effects

A critical advantage of AI in drug repurposing lies in its ability to predict and mitigate off-target effects, a significant concern when applying existing drugs to new conditions. AI achieves this by scrutinising drug structures, predicting interactions with proteins, and comparing drug profiles to anticipate potential adverse reactions. This early identification of potential risks enables researchers to develop safer and more effective clinical trials.

Baricitinib, initially developed for rheumatoid arthritis, serves as a prime example of AI's efficacy in predicting and mitigating side effects. AI analysis revealed its potential as a COVID-19 treatment due to its anti-inflammatory and antiviral properties, paving the way for its successful repurposing. Conversely, AI analysis of Lopinavir/Ritonavir, an HIV drug explored for SARS-CoV-2 treatment, predicted potential liver toxicity, highlighting the need for more targeted inhibitors.

Expanding the Candidate Pool

AI's impact extends beyond risk assessment. It significantly expands the pool of potential drug candidates by identifying unexpected connections between existing drugs and novel targets. Traditional approaches focused on drugs with mechanisms of action directly related to the target disease, limiting the scope of exploration. AI, however, delves deeper by analysing vast datasets on existing drugs, including their structures, interactions with human proteins, and potential side effects. By comparing this information with data on the target disease, AI uncovers previously unknown therapeutic potential, even in drugs originally intended for unrelated conditions. A compelling example is Alendronate, an osteoporosis drug. AI analysis revealed its ability to inhibit an enzyme crucial for cancer cell growth, opening up new avenues for its repurposing as an anticancer therapy.

Unlocks Hidden Cures in Expired Patents

Drug patents provide valuable information on a drug's structure, mechanism, and original use. Once expired, the drug becomes generic and accessible for repurposing. AI excels at mining this patent data to identify drugs with potential new uses, accelerating development of affordable therapies for emerging pathogens. Teicoplanin, an antibiotic, serves as a prime example. AI analysis unveiled its potential antiviral properties against Zika, demonstrating AI's power in repurposing drugs to combat viral threats.

AI-Augmented Drug Repurposing Trials

AI is transforming clinical trial design for repurposed drugs through data-driven optimisation. By analysing extensive patient datasets, AI algorithms precisely identify ideal patient cohorts, ensuring targeted and conclusive results. Furthermore, AI optimises trial parameters, such as dosage and duration, streamlining data collection and maximising resource efficiency. This approach accelerates the evaluation and potential approval of repurposed drugs, expediting the delivery of safe and effective therapies to patients.

AI-powered Drug Repurposing methods

AI is revolutionising drug repurposing, propelling the identification of existing medications for novel therapeutic applications. By harnessing sophisticated algorithms and vast datasets, AI-driven approaches (Table 1) accelerate drug discovery, offering a promising avenue to address unmet medical needs and combat emerging diseases.

Key AI-Powered Drug Repurposing Methods:

  • Machine learning (ML) and deep learning (DL) algorithmsin Virtual Screening:AI algorithms rapidly screen extensive molecular libraries, identifying potential drug candidates with enhanced efficiency.ML models analyse known interactions to predict the efficacy of existing drugs against new targets, while DL algorithms, particularly deep neural networks, excel at identifying subtle patterns in complex datasets, such as protein structures and molecular interactions.

Table 1 AI-Driven Strategies for Drug Repurposing: Software and Algorithms




Machine Learning (ML) and Deep Learning (DL) in Virtual Screening


Predicts potential drug-target interactions using deep learning on protein and molecule representations.



Uses 3D convolutional neural networks to predict protein-ligand binding affinity for virtual screening.


Generative Tensorial Reinforcement Learning



Employs generative models to design new molecules with desired properties for specific targets.


Deep Docking

Combines deep learning with molecular docking to improve accuracy and efficiency in virtual screening.



Reinforcement learning-based approach for de novo drug design and lead optimisation.

Natural Language Processing (NLP) for Knowledge Extraction


Pre-trained language model specifically for scientific text, useful for information extraction.



Variant of BERT fine-tuned on biomedical corpora, enhancing performance on biomedical NLP tasks.



Another BERT variant optimised for biomedical NLP, offering state-of-the-art results on various tasks.



Extracts clinical information from electronic health records (EHRs) for drug repurposing studies.

Network Analysis and Systems Biology


Open-source software for visualising and analysing complex biological networks.



Models and optimises signaling pathways for drug target identification.



Infers regulatory networks from gene expression data for drug repurposing.


KeyPathwayMiner (KPM)

Identifies key pathways involved in disease for potential drug targets.



Comprehensive tool for network analysis and visualisation in systems biology.

Predictive Modeling of Drug Safety and Efficacy


Predicts drug-drug interactions to assess potential side effects.


Tox21 Challenge Winners

Ensemble of machine learning models for predicting toxicity of chemicals.


PBPK models (e.g., GastroPlus)

Simulate drug absorption, distribution, metabolism, and excretion to predict drug behavior in humans.


QSAR models

Quantitative structure-activity relationship models predict biological activity based on chemical structure.


Survival analysis models

Predict patient survival outcomes based on various factors, including drug response.










































  • Natural Language Processing (NLP) for Knowledge Extraction: NLP algorithms mine the wealth of information in biomedical literature and databases, automating literature reviews, summarising key findings, and identifying potential drug repurposing opportunities. By extracting relevant entities and relationships from text, NLP uncovers hidden connections between existing drugs and novel antiviral targets.
  • Network Analysis and Systems Biology: AI-powered network analysis tools map intricate biological networks, identifying key nodes and pathways that can be targeted by existing drugs. Systems biology approaches integrate diverse data types, such as genomic, proteomic, and metabolomic data, to build comprehensive models of disease processes, which can then be used to identify drug repurposing opportunities.
  • Predictive Modeling of Drug Safety and Efficacy: AI algorithms leverage patient data and clinical trial results to predict the safety and efficacy of repurposed drugs in new indications. These models can help identify potential side effects and interactions, allowing researchers to make informed decisions about which drugs to prioritise for further investigation.

These AI-driven methodologies represent a paradigm shift in drug repurposing, enabling the rapid identification and evaluation of potential therapeutic agents, ultimately accelerating the development of innovative treatments for a wide range of diseases.

Beyond the hype: tangible results of AI in drug repurposing

The transformative power of AI in drug repurposing is already tangible, extending beyond the notable examples of remdesivir and favipiravir for COVID-19. AI algorithms have also been instrumental in identifying potential repurposing candidates for other viral infections. For instance, the anti-parasitic drug niclosamide, identified through AI-powered screening, is being investigated for its potential against COVID-19 and other coronaviruses. Additionally, AI-driven analysis has suggested the antidepressant fluvoxamine as a potential therapeutic agent for COVID-19, prompting further clinical evaluation. In the realm of bacterial infections, AI-driven drug repurposing has led to the identification of the antibiotic teixobactin, originally developed for Gram-positive bacteria, as a potential treatment for tuberculosis. These diverse examples highlight the versatility and efficacy of AI in uncovering hidden therapeutic potential within existing drugs, offering hope for accelerating drug development and addressing urgent global health challenges.

AI-driven drug repurposing has yielded promising leads across diverse diseases. For instance, the anti-inflammatory drug baricitinib, initially approved for rheumatoid arthritis, was identified through AI as a potential COVID-19 treatment, later confirmed by clinical trials. Additionally, the AI-powered platform "Drug Repurposing Hub" identified several existing drugs, including apilimod and mefloquine, as potential candidates for combating SARS-CoV-2. These real-world examples underscore AI's capacity to accelerate drug discovery and unveil therapeutic potential hidden within existing medications.

Challenges and Future Directions

While AI holds immense promise, challenges remain. The quality and availability of data are critical for the accuracy and reliability of AI models. Additionally, the "black box" nature of some AI algorithms can make it difficult to interpret their findings and understand the underlying mechanisms. Addressing these challenges requires ongoing research and collaboration between AI experts, pharmaceutical companies, and regulatory agencies.

The future of AI in drug repurposing is bright. As AI technologies continue to advance, we can expect even more efficient and targeted approaches to drug discovery. AI-powered drug repurposing has the potential to revolutionise the way we combat infectious diseases, offering hope for a healthier future.


Singh, A. (2024). Artificial intelligence for drug repurposing against infectious diseases. Artificial Intelligence Chemistry, 2(2), 100071.

Petrova, N. (2024). AI-driven Drug Repurposing for Novel Therapeutic Applications: Utilizes AI algorithms to identify existing drugs with potential therapeutic applications in new disease areas. Journal of Artificial Intelligence Research and Applications, 4(1), 182-190.

Kazma, J. M., Van Den Anker, J., & Ahmadzia, H. K. (2023). Pharmacoethics and pregnancy: Overcoming the therapeutic orphan stigma. British journal of clinical pharmacology, 89(2), 483-490.

Gupta, A. K., Talukder, M., Shemar, A., Piraccini, B. M., & Tosti, A. (2023). Low-dose oral minoxidil for alopecia: a comprehensive review. Skin Appendage Disorders, 9(6), 423-437.

Wang, L., Liu, W. Q., Broussy, S., Han, B., & Fang, H. (2024). Recent advances of anti-angiogenic inhibitors targeting VEGF/VEGFR axis. Frontiers in Pharmacology, 14, 1307860.


Anuradha Singh

Anuradha Singh is an Assistant professor at the Department of Chemistry, S Khanna Girls’ Degree College, University of Allahabad. She is well versed in synthetic chemistry, computer-aided drug design and other techniques used in the field of medicinal chemistry. Dr Singh has published more than 30 research papers/book chapters, in addition to several articles, in journals/books of national and international repute. She has been selected for the prestigious Indian National Science Academy's Visiting Scientist fellowship in 2023. She is a member of several national bodies and referee to various national/ international journals. She successfully led research projects funded by DST and UGC, demonstrating strong leadership and project management skills.

magazine-slider-imageCytiva - Supor Prime filtersMFA + MMA 20244th Annual Cleaning Validation 20242nd Annual Pharma Impurity Conclave 2024CPHI Korea 2024CHEMICAL INDONESIA 2024World Orphan Drug Congress Europe 2024INALAB 2024Thermo Fisher - Drug Discovery and the impact of mAbsAdvanced Therapies USA 2024ISPE Singapore Affiliate Conference & Exhibition 20242024 PDA Aseptic Manufacturing Excellence Conference2024 PDA Aseptic Processing of Biopharmaceuticals Conference