The impact of AI and ML on the future of drug discovery  

Over the past several years, Dr Jo Varshney, Founder of VeriSIM Life, has front-led numerous collaborative engagements with a host of biopharma companies, government agencies, academic/medical institutes and industry innovators. Here, she discusses with DDW’s Megan Thomas the impact of AI and ML on the drug discovery process.  

MT: How does AI and ML accelerate the drug discovery process?

JV: A key bottleneck in the research and development of a drug is the lengthy experimentation that is required to gain the confidence to run clinical trials. That experimentation is riddled with both failures and successes – largely based on animal studies – which do not translate to humans. AI and ML speeds up the process significantly by using pattern matching and superior computational methods to guide experimentation so that scientists can avoid redundancies and mistakes.   

 MT: What does VeriSIM Life do, and how does this influence drug discovery?

JV: Our mission is to solve the translational challenges that every biotech and pharma company faces, and get more novel therapies to patients, faster. We believe deep technology coupled with rigorous science can help reduce inaccuracy and waste in the drug discovery process. VeriSIM Life has developed a hybrid biosimulation and AI-based platform that can assess the likely success of drug candidates, similar to a credit score.  

MT: How does VeriSIM Life help to advance biosimulation models?

JV: We use a first-in-class hybrid AI-driven bio-simulation platform which combines machine learning, mechanistic modeling, secure data storage and massive computational power. By providing insights to prioritise the most informative experiments much earlier in the drug development process, this platform allows us to translate, scale, and accelerate drug development insights and de-risk R&D decisions by predicting the clinical benefits before human trials. 

MT: Has the company enabled any recent breakthroughs with its technology?

JV: We’ve helped clients save millions of dollars in drug development and significantly decrease their time for clinical application. A few examples include $3M in savings for conventional drug testing for a top 25 global pharma company by predicting translatable candidates in just two months; $2M in conventional drug testing for a mid-sized biotech company by predicting optimal compound combination and dosing and the elimination of 2.5 years of in-house testing for a top five global pharma company by identifying biochemical descriptors in complex compounds.  

MT: Where does VeriSIM source the data used for its AI and ML technology?

JV: We’re proud of having developed a data lake of the largest proprietary database in the pharma space consisting of public and private datasets including Pubchem, toxcast, publications, clinical trials. It contains structure-related data for more than 3 million compounds, over 5000 unique animal and human validation datasets, and physiological parameters of 196 different subject populations (different species, gender, strain, sub-strain) and counting. Additionally, we have built a library of extensive synthetic data to create massive training sets and extensive validation and statistical inferences. Recently, we announced a partnership with Clarivate, a leader in drug intelligence data, to jointly provide clients with additional insights early in their drug investigation. 

MT: Where are the biggest opportunities for biosimulation models?

JV: We believe that traditional biosimulation and computational models have reached a plateau of capability and suffer from significant limitations. With few exceptions, clinical and experimental data relating to translatability is extremely limited. When run on small datasets, traditional model predictions are more variable, making them error prone, and are frequently overfitted, making results difficult to interpret. Integrating artificial intelligence with biosimulation can better predict clinical translational outcomes, deriving highly accurate predictions of drug toxicity, bioavailability and efficacy based only on structural molecular data.  

MT: What are the challenges to overcome when it comes to AI and ML technology?

Explainability and visibility into the “black box” of AI and ML technology is a significant challenge. In drug development, this is especially important for creating reproducibility and experimental confidence. AI systems should be able to describe how a single concrete prediction or classification was made in a language that drug knowledge expert audiences can understand. They should also embed methods to audit the determination of results or behaviours that occurred during modeling.   

MT: Are there ethical considerations that need attention in this field?

There are continued concerns with things like data quality, regulatory hurdles, and algorithm bias but at VeriSIM Life, we’re addressing these by establishing several benchmarks to confirm our continuous system validation simulations, as well as designing explainability into our systems. This allows us to reduce bias and ensure that results are reliable. Ultimately, the use of AI does not replace good science, nor the rigorous process of ensuring drug safety. We believe it should complement these processes.  

MT: What is your perception of the regulatory ecosystem when it comes to advancing AI and ML technology?

In the United States, the FDA has long acknowledged that model-informed drug development (MIDD) approaches can be used to support decisions on whether, when, and how to conduct certain pharmacology studies, and to support dosing recommendations in product labeling, among other applications. The use of AI and ML techniques has also been recognised as increasingly playing a key role in drug discovery and development. Data and analysis derived from these approaches have already supported hundreds of new drug applications, as well as clinical trial design and specific regulatory approval decisions. We believe regulators will continue refining guidance for the responsible integration of AI and ML in the drug development process, working in close collaboration with the industry.   


Jo Varshney, VeriSIMLifeDr. Jo Varshney is the Founder and CEO of VeriSIM Life, and the inventor of VeriSIM Life’s BIOiSIM core technology. A virtual drug development engine that harnesses the power of traditional statistical modeling along with artificial intelligence and machine learning, BIOiSIM enables pharma and biotech professionals to bring more lifesaving and cost-efficient therapies to patients faster than ever before.

Related Articles

Join FREE today and become a member
of Drug Discovery World

Membership includes:

  • Full access to the website including free and gated premium content in news, articles, business, regulatory, cancer research, intelligence and more.
  • Unlimited App access: current and archived digital issues of DDW magazine with search functionality, special in App only content and links to the latest industry news and information.
  • Weekly e-newsletter, a round-up of the most interesting and pertinent industry news and developments.
  • Whitepapers, eBooks and information from trusted third parties.
Join For Free