The development processes of new drugs are lengthy, expensive, and characterized by low success rates. Developing a drug takes 10-15 years, with an estimated cost of $1.5-2 billion. Only about 10% of drugs that reach the clinical development stage successfully pass the regulatory approval process. These low success rates and the high financial costs they entail drive the need to develop and adopt new approaches.
Integration of AI in Drug Development
Artificial Intelligence (AI) and Machine Learning (ML) can potentially reduce the time and cost of drug development and increase the percentage of drugs receiving regulatory approval for marketing. Indeed, this field has gained momentum in recent years.
AI tools have the potential to be integrated into all stages of drug development, starting from the identification and validation of target molecules, the search and optimization of active compounds, synthesis processes of active ingredients, and evaluation of the drug’s safety, toxicity, absorption, distribution, excretion, and metabolism in the body (ADMET). AI can also be applied in clinical trial stages, including identifying suitable target populations for clinical trials and discovering biomarkers to predict patient responsiveness to drugs.
AI-Based Approaches
1. Identification of New Therapeutic Targets
AI can analyze large datasets of biological information (OMICS – including gene expression, genomic, and proteomic data) and protein interaction networks to find causal relationships between diseases and genes/proteins. Additionally, AI models are used to assess the suitability of proteins as targets for therapeutic intervention. Another approach to identifying therapeutic target-disease pairs involves text processing from scientific literature using natural language processing (NLP).
Examples of companies in the field: BenevolentAI and Insilico Medicine. In Israel: CytoReason.
2. Drug Discovery
In this field, artificial intelligence (AI) plays a crucial role in the discovery, design, and optimization of molecules that can function as drugs for specific therapeutic targets. AI enhances the drug discovery process by broadening the chemical space available for exploration and efficiently identifying promising candidates. This involves virtual screening of vast libraries, comprising hundreds of millions of molecules, to predict their binding affinities to therapeutic target proteins using structural and computational methods. By prioritizing potential candidate molecules for laboratory testing, AI significantly accelerates the path from discovery to clinical application.
Another strategy focuses on the de novo design and assembly of novel molecules using chemical building blocks tailored for specific therapeutic targets. This process utilizes machine learning algorithms trained on extensive datasets of protein-small molecule interaction pairs and empirical laboratory results. Such approaches not only expand the available chemical space but also facilitate the discovery of molecules that can effectively bind to therapeutic targets previously deemed undruggable. By leveraging advanced computational techniques, researchers can innovate new drug candidates that push the boundaries of current medicinal chemistry.
AI models can also be used for modeling the relationship between a molecule’s chemical structure and its biological activity (Quantitative Structure-Activity Relationship, QSAR), thereby enabling the optimization of drug candidate molecules according to desired properties, such as improving binding strength, enhancing selectivity, stability, solubility, and pharmacokinetic profiles.
In addition to AI platforms designed for the discovery of small-molecule drugs, there are unique AI platforms for the discovery and design of non-small-molecule drugs, including peptides, antibodies, and nucleic acids.
Examples of companies in the field: Atomwise, Exscientia, Schrodinger.
3. Predicting Efficacy/Safety in the Preclinical Development Stage
In this field, a key approach combines laboratory experiments with the analysis of their results through AI and machine learning tools. These approaches often involve human cell-based technologies, such as organoids, which are three-dimensional tissue structures derived mainly from stem cells or patient biopsies. These organoids mimic, both morphologically and functionally, a human organ or diseased tissue (e.g., cancerous tumors). Diverse metrics collected from the organoids, both with and without drug exposure, are analyzed using AI tools, yielding insights into the drug’s efficacy, mechanism of action, and/or toxicity.
Examples of companies worldwide: Emulate, InSphero, Mimetas.
Examples of companies in Israel: CuResponse, Quris-AI, Tissue Dynamics.
4. Artificial Intelligence in the Clinical Development Stage
AI tools can be employed to identify relevant biomarkers for the disease by analyzing data from electronic medical records alongside various biological datasets – including historical clinical trial data and information collected during ongoing trials. These advanced analytics enable the prediction of treatment efficacy and potential toxic effects, thereby aiding in the selection of the most suitable patient population for clinical trials. This targeted approach enhances the likelihood of successful outcomes and improves patient safety.
AI can also improve parameters related to the logistical management of clinical development, including trials and studies involving humans. AI enhancement in areas such as selecting medical centers, accelerating patient recruitment, and reducing dropout rates can increase the likelihood of successful and efficient completion of clinical trials.
Examples of companies in the field: IQVIA ,OWKIN ,CytoReason ,Saama, trials.ai
Market Trends
The excitement surrounding AI’s potential in drug development extends well beyond the scientific community; investors are also demonstrating substantial interest in this area. According to Morgan Stanley, even “modest improvements in early-stage drug development success rates enabled by the use of artificial intelligence and machine learning could lead to an additional 50 novel therapies over a 10-year period, which could translate to a more than $50 billion opportunity.”
Third-party investments in companies focused on AI-supported drug development have doubled over the past five years. Total assets reached $2.4 billion in 2020 and exceeded $5.2 billion by 2021.
These figures do not include the the investments of pharmaceutical companies in their internal capabilities or investments by tech giants, which are also expanding their AI activities into biology and drug development. For example, in 2021, Alphabet (Google) launched Isomorphic Labs, building on breakthroughs in AI achieved at its research lab, DeepMind.
Collaborations Between Pharmaceutical Companies in Recent Years
In the past two years, numerous collaborations have been established between pharmaceutical companies and researchers or organizations specializing in artificial intelligence. Below are a few examples:
A collaboration between AMGEN (pharma) and the research institute Mila (Canada) focused on artificial intelligence. A joint lab was established where researchers from both organizations work side by side to advance innovative drug development using AI.
SANOFI (pharma) expanded its collaboration with Exscientia (AI-driven drug discovery). This partnership includes leveraging Exscientia’s AI platform to address complex challenges in developing new drugs.
Novo Nordisk (pharma) signed a partnership agreement with Microsoft, integrating AI models developed by Microsoft into Novo Nordisk’s research and development processes.
The growing number of collaboration deals aimed at integrating AI into drug development processes underscores pharmaceutical companies’ increasing interest in AI capabilities.
Notable Players in the Global Market
An analysis of investments and funding in the field reveals approximately 50 players who have raised hundreds of millions of dollars to support their ongoing development programs (including the Israeli company CytoReason).
These companies specialize in different areas of AI expertise, such as Advanced AI tools for specific use cases, Advanced AI systems with multiple models, and End-to-end AI.
Accordingly, they integrate at various stages of the drug development process, including clinical pipelines (phases 1-2), validated R&D use cases, and preclinical pipelines.
Market Size
In terms of revenue, the global market size for AI-driven drug discovery was $0.6 billion in 2022 and is expected to reach $4 billion by 2027, reflecting an average annual growth rate of 45.7%
Global Overview of Activities and Trends in the Field
An analysis of the number of companies established between 2010 and 2021 reveals a steady increase in the formation of companies until 2018. However, a decline began in 2019, primarily attributed to the COVID-19 pandemic and its market implications.
Despite the decrease in newly founded companies, investments in the field grew tenfold between 2017 and 2021, with a CAGR of 71%. There was also a significant increase in investments in mature companies through public offerings or venture capital funding.
Key Milestones in AI-Based Drug Development
Clinical Trials of AI-Developed Drugs
Clinical trials of drugs developed using AI are still in their early stages but are becoming increasingly common, with some already reporting positive results in Phase I.
In an opinion piece published in September 2023, clinical trials at various stages conducted by 10 different companies were highlighted. These trials involve new drugs developed with the assistance of AI, targeting a wide range of diseases, including various types of cancer, diabetes, heart and kidney diseases, infectious diseases, and more.
Protein Structure Prediction
In July 2021, DeepMind‘s AI system, AlphaFold, predicted the structure of 330,000 proteins, including all 20,000 proteins in the human genome. Since then, the AlphaFold database has expanded to include over 200 million proteins, covering nearly all cataloged proteins known to science.
This computational tool has immensely contributed to drug development by enabling the modeling of the structure of selected therapeutic targets whose structures were previously unknown. This allows for the design of drugs targeting these proteins based on structural compatibility.
Recognition by Regulatory Authorities
In May 2023, the FDA published an article addressing current and potential applications of AI in drug discovery, preclinical and clinical trials, post-marketing safety studies, and drug marketing. The agency explained that the publication’s purpose was to gather information from the industry and other stakeholders regarding the opportunities and challenges in drug development using AI.
The FDA’s publication emphasized that AI and machine learning are no longer futuristic concepts but have become integral to how we live and work today.
Artificial intelligence (AI) and machine learning (ML) are no longer futuristic concepts; they are now part of how we live and work
Patrizia Cavazzoni, director of the Center for Drug Evaluation and Research (CDER).
Activity in Israel
AION Labs – An Innovation Lab Supported by the Israel Innovation Authority
AION Labs is a first-of-its-kind innovation lab, formed through a collaboration of four pharmaceutical companies – AstraZeneca, Merck, Pfizer, and Teva – together with the Israeli biotech fund Amiti Ventures, Amazon Web Services (AWS), and BioMed X. Supported by the Israel Innovation Authority, the lab aims to develop and adopt groundbreaking AI technologies to revolutionize the drug discovery process.
Located in Rehovot, the lab is designed to establish 4-6 companies annually. The selection process for companies within the lab follows a structured model developed by the BioMedX research institute, comprising the following stages:
- Identifying challenges in drug development where AI has the potential to provide solutions, conducted by the partnering pharma companies and the lab team
- Publishing a call for proposals for one challenge each quarter
- Running a structured “Boot Camp” process with entrepreneurs who submit proposals
- Entrepreneurs present their projects to an investment committee, which selects the entrepreneur/project to address the challenge
Since its formation in June 2021, the lab has issued six challenges addressing various stages of new drug development. It has established six new companies offering solutions to these challenges: DenovAI, Omec.AI, CoBind, TenAces, CombinAble.AI, and Promise Bio.
Trends in Israel
A total of 27 companies operating in Israel were identified, with 19 (70%) receiving support from the Israel Innovation Authority. The chart below illustrates that most of these companies (20) were founded between 2016 and 2023, with 2021 being the peak year, during which five companies were established. Notably, 2021 also marked the launch of the innovation lab AION Labs; however, the six companies founded within this lab are not included in this count.
The situation in Israel differs significantly from the global landscape. In 2023, approximately 800 companies were active in this field worldwide, with 53% of them located in the United States. Globally, investments in the sector grew tenfold between 2017 and 2021. In contrast, Israel has a relatively small number of companies in this domain, with only a few new ventures emerging each year, and the anticipated growth in this market has yet to materialize locally. This disparity highlights the unique challenges and opportunities facing the Israeli landscape in comparison to the global market.
Summary
Although no drugs developed with AI have received market approval yet, numerous candidates are currently in various stages of clinical trials, with several anticipated to be launched in the coming years. Many experts are optimistic that AI has the potential to transform the pharmaceutical industry and revolutionize the drug discovery process. This innovative approach could lead to faster, more efficient development of new treatments, ultimately improving patient outcomes and reshaping the landscape of medicine.
Achieving success in this field requires expertise in software and AI to create and train practical algorithms, combined with deep knowledge in biology and pharmaceuticals to understand complex biological mechanisms. This integration aligns closely with the bio-convergence field, which is being promoted as part of a long-term national plan in Israel.
With its strong high-tech foundation, Israel has the potential to propel this promising field forward by allocating resources and talent to drug development applications.
* All information provided in this article is correct as of the date of writing and according to the data available to the author. The Innovation Authority or anyone on its behalf is not responsible for the accuracy, truthfulness, and precision of the data, in whole or in part. The article is published for the public’s benefit, and no commercial use should be made of it, including for its sale, distribution, or presentation.