Making the AI Breakthrough in Life Sciences R&D

Use Cases & Projects, Scaling AI, Featured Kelci Miclaus

In the quest to combat diseases that afflict humanity, the biopharmaceutical industry has long faced significant challenges, not the least of which is the arduous process of drug discovery and the time-consuming and costly methodologies of traditional scientific experimentation. However, a new era is emerging as AI promises to be a transformative force that could revolutionize all aspects of medicine, from diagnosis to treatment.

The just-released Season 2 of the Dataiku web series AI&Us opens with Episode 1, “Making the Breakthrough,” which sheds light on not just the groundbreaking impact, but also practical and realistic effects of AI in drug discovery, featuring a myriad of experts from the field who share their insights into the unprecedented possibilities that AI presents.

Episode 1 features John Apathy (Digital Transformation Leader at XponentL Data), Rory Kelleher (Global Head of Business Development for Healthcare and Life Sciences at NVIDIA), Dr. Christopher E. Mason (Professor Genomics, Physiology, and Biophysics), Dr. Shahar Keinan (CEO and co-founder at Polaris Quantum Biotech), Kyle Tretina (Alliance Manager at Insilico Medicine), Harini Gopalakrishnan (Field CTO, Life Sciences at Snowflake), and Lurong Pan (CEO and founder at Ainnocence), along with industry experts.

This article shares a high-level summary of the episode and some of our favorite sound bites, as well as the episode video for convenient viewing!

→ Watch All 4 New Episodes of AI&Us Now

Can AI Truly Revolutionize Drug Discovery? 

The episode opens with a staggering statistic: Over 15,000 categorized diseases afflict our world, that is a daunting reality. The mission is to accelerate drug discovery and, with AI leading the charge, the landscape of medicine is undergoing a profound transformation. AI has not merely impacted but has begun to disrupt nearly every facet of medicine. From speeding up the discovery of new medicines to expediting their development, the potential of AI is unparalleled. 

The traditional timeline of drug discovery and development, which often spans five to seven years just to bring a drug candidate to clinical testing, is starting to reflect a new pattern of speed and efficiency. Thanks to emergent technologies, coupled with AI, we are witnessing a paradigm shift in how we understand diseases and develop new drugs. Laborious manual experimentation processes are becoming regularly augmented by AI, scientists now wield the power of thousands of virtual workers, exponentially increasing productivity and efficiency.

When I have an AI tool at my fingertips, I effectively have hundreds or thousands of workers, if you will, that can go do work on my behalf. It’s not that computational biologists or researchers or radiologists are going to be replaced, it’s that they are going to become far more productive, both within their own organization and for society at large.

-Rory Kelleher, Global Head of Business Development for Healthcare and Life Sciences, NVIDIA

The conversation delves deeper into the realm of molecular design, where AI's capabilities shine brightest. AlphaFold, an AI system developed by DeepMind, exemplifies this breakthrough, accurately predicting the functional form of millions of proteins and unlocking new avenues and providing new data for AI-driven drug design.

For a molecule to become a drug, it needs to have many properties, starting from the biological activity, either stopping a disease or slowing a disease. How easy it is to synthesize, toxicity, and side effects. You’re looking for many properties for that single molecule. Looking at a very large number of molecules and finding the ones that have all the properties you want, very fast, this is what we do and we use quantum computers to do that.

-Dr. Shahar Keinan, CEO and co-founder, Polaris Quantum Biotech 

Accelerating Hypothesis & Experimentation 

The impact of AI extends beyond prediction; it fundamentally transforms the drug discovery process. By analyzing vast datasets at unprecedented speeds, AI accelerates the scientific cycle of hypothesis and experiment, propelling us towards faster and more efficient drug development.

Yet, amid the excitement surrounding AI's potential lies a crucial caveat: the need for validation. While algorithms continue to improve, real-world validation remains essential. As the episode emphasizes, AI is not a panacea; it requires rigorous testing and refinement to ensure its efficacy.

The first three molecules invented by AI have all now failed in further downstream testing. So, this is still a very complex set of processes that all have to come together to invent a new medicine.” 

-John Apathy, Digital Transformation Leader, XponentL Data

AI Cannot Work in Isolation

The episode concludes with a poignant reminder: While AI holds immense promise, it cannot work in isolation. Success hinges on collaboration between data scientists and bench scientists, bridging the gap between computational predictions and real-world outcomes.

When it comes to small molecule design, the situation is a lot more optimistic in the short-term because we can validate in the lab pretty much right away. Our algorithms only take minutes to hours to imagine all the molecules that you need for your experiments. The computational result can be validated in the wet lab with a very high success rate.

-Kyle Tretina, Alliance Manager, Insilico Medicine 

In essence, the journey towards revolutionizing drug discovery through AI is not without its challenges. However, as we harness the power of AI to predict, triage, and prioritize experiments, we inch closer to a future where medicines are more accessible, more available, and brought to market faster than ever before.

Be sure to check out the full episode here for yourself:


You May Also Like

Democratizing Access to AI: SLB and Deloitte

Read More

Secure and Scalable Enterprise AI: TitanML & the Dataiku LLM Mesh

Read More

Solving the Ocean Plastic Pollution Problem With Data

Read More

Revolutionizing Renault: AI's Impact on Supply Chain Efficiency

Read More