Model Medicines Unveils Groundbreaking AI Drug Discovery Advances in New Chemistry and New Biology with Two New Preprint Papers

LA JOLLA, CA — Model Medicines, a leading human health company specializing in generative AI-driven drug discovery, today announced the publication of two significant preprint papers on bioRxiv detailing major advances in antiviral and oncology drug discovery. 

The preprint, titled "Discovery of RdRp Thumb-1 as a novel broad-spectrum antiviral family of targets and MDL-001 as a potent broad-spectrum inhibitor thereof - Part I: A Bioinformatics and Deep Learning Approach", showcases how Model Medicines researchers leveraged their proprietary GALILEO AI platform to identify the RdRp Thumb-1 site as a novel target for broad-spectrum antivirals against a range of RNA viruses. Using bioinformatics and deep learning, they also found MDL-001 to be a promising orally bioavailable broad-spectrum antiviral that blocks the RdRp Thumb-1 site. Importantly, the research demonstrates how their AI approach bypassed pitfalls of conventional in silico screening methods to zero in on MDL-001, which has now demonstrated potent in vitro and in vivo antiviral activity.

Figure 1 Both the structure and sequence of the 𝜦1-loop interaction with the Thumb-1 pocket show remarkable conservation within the families Coronaviridae, Flaviviridae, and Picornaviridae.

In the second preprint, "ChemPrint: An AI-Driven Framework for Enhanced Drug Discovery", Model Medicines details their foundational GALILEO AI platform and its core ChemPrint deep learning model for molecular property prediction. By employing adaptive molecular embeddings and rigorous model training and validation methods, ChemPrint substantially outperformed conventional AI approaches. In a prospective study, ChemPrint achieved a remarkable 45.5% in vitro hit rate in identifying novel inhibitors of AXL and BRD4, two highly challenging, previously “undruggable” oncology targets. Notably, the discovered hits were chemically distinct from known inhibitors, demonstrating ChemPrint's ability to explore novel chemical spaces for drug discovery effectively.

Figure 2: In vitro validated hits for both AXL and BRD4 displaying their maximum Tanimoto similarity score to respective training datasets used to discover them.

"These two preprint papers, combined with our recent success in discovering new pan-antiviral chemical entities with an unprecedented 66.7% hit rate, represent a quantum leap in our ability to leverage AI for transformative drug discoveries," said Dr. Daniel Haders, CEO and Founder of Model Medicines. "By efficiently navigating deep chemical space and accurately predicting the most promising compounds, we can significantly accelerate the development of novel, best-in-class therapeutics for diseases of high, unmet medical need, from broad-spectrum antivirals to cancer treatments. We are excited to advance MDL-001 as a potential broad-spectrum oral antiviral and to leverage our platform for further groundbreaking discoveries."

Davey Smith, MD, MAS, FACP, FIDSA, Senior Clinical Advisor to Model Medicines and a renowned infectious disease specialist at UC San Diego Health, commented: "Model Medicines' discovery of MDL-001 as a potent broad-spectrum antiviral inhibiting the novel RdRp Thumb-1 target is a game-changer. As we've seen with the devastating consequences of the COVID-19 pandemic, the world desperately needs rapidly deployable, broad-spectrum antivirals that can effectively combat novel pathogens from the moment an outbreak occurs. Model Medicines' AI-driven approach is poised to revolutionize how we develop and deploy antiviral therapeutics, potentially saving millions of lives and transforming our capacity to confront future pandemics head-on."