In recent years, artificial intelligence (AI) has made significant strides in the field of medical research and drug discovery. The increasing engagement of AI is particularly prominent when addressing difficult diseases that previously lacked effective treatment options. This article serves as the fourth installment in a six-part series exploring how AI is reshaping the landscape of medical research and treatment methodologies.
One notable figure in this domain is Dr. Alex Zhavoronkov, co-founder and CEO of Insilico Medicine, who has demonstrated the efficacy of AI in drug development. Through a virtual discussion, Dr. Zhavoronkov showcased a small, green, diamond-shaped pill which targets idiopathic pulmonary fibrosis (IPF), a rare lung disease characterized by the progressive scarring of lung tissue without an identifiable cause. Although this drug is still undergoing clinical trials, initial results indicate a substantial improvement in patient outcomes.
As Dr. Zhavoronkov eloquently notes, the field is witnessing an influx of drugs, thanks in part to AI’s vital role in their discovery. “While we can’t claim to have the first AI-discovered and designed molecule approved for use, we might be ahead on this journey,” he stated. This aligns with a broader trend where myriad biotech companies are leveraging AI to replace traditional tasks historically performed by medicinal chemists.
Among the new major players in this arena is Isomorphic Labs, a UK-based drug discovery venture backed by Alphabet, Google’s parent company. Headed by CEO Demis Hassabis, who notably shared the Nobel Prize in Chemistry for contributions related to AI models in drug design, this initiative is indicative of the industry’s shift towards technology-driven research methodologies.
Chris Meier from the Boston Consulting Group (BCG) attests to the transformative potential of AI in this space. He highlights that the conventional process of bringing a new drug to market can take upwards of 10 to 15 years and cost over $2 billion, all while facing a staggering 90% rate of failure in clinical trials. By implementing AI, the hope is to reduce both time and financial constraints while increasing success rates, leading to quicker therapeutic advancements.
According to Charlotte Deane, a professor of structural bioinformatics at Oxford University, we are at the dawn of realizing just how beneficial AI can be in drug discovery. However, she notes that this change is likely to complement, rather than replace, pharmaceutical scientists. The genuine savings are expected to arise from minimizing the number of failed drugs, necessitating a collaborative approach between human experts and AI systems.
Recent analyses have shown that at least 75 AI-discovered molecules have entered clinical trials, marking a significant milestone for the industry. However, clarity over what constitutes an “AI-discovered drug” remains elusive, as human expertise is still deeply entwined with the AI processes.
Within drug discovery, AI is predominantly employed at two critical stages: firstly, in identifying potential therapeutic targets at the molecular level, and secondly, in designing the molecules that will interact with those targets. This is achieved through generative AI technologies—similar to those behind platforms like ChatGPT—which can envision new molecular designs, replacing the need for physical experimentation.
Insilico Medicine, established in 2014 and funded with over $425 million, utilizes AI across both steps of its drug discovery process, including predicting clinical trial success rates. Currently, the firm has six molecules in clinical trials, including one for IPF that is poised for its next trial phase. Moreover, four additional molecules have been approved for trials, while nearly 30 others are emerging as promising candidates.
Despite these advancements, challenges persist, particularly regarding data availability for AI learning. Recursion Pharmaceuticals is actively addressing this issue through automated experiments designed to generate extensive datasets that reveal complex interrelationships among various molecules in the human body. Their developments have led to promising results, including a molecule targeting a novel gene associated with both lymphoma and solid tumors, highlighting the potential of AI to revolutionize cancer treatment.
As Recursion co-founder Chris Gibson states, the ultimate test of this technology lies in demonstrating that AI-discovered molecules not only succeed in clinical trials but also surpass the success rates of traditional drug development approaches. When that pivotal moment arrives, it will solidify AI’s role as an invaluable tool in modern medicine.