How generative AI is transforming early Drug Discovery: Insights from Novartis’ Jeremy Jenkins at Hariri Institute, Boston

At the Hariri Institute for Computing at Boston University in November 2024, keynote speaker Jeremy Jenkins offered an invigorating glimpse into how generative AI is transforming the earliest, and often most challenging, phases of drug discovery. And we're so excited that he mentioned Causaly!

Titled “Generative AI Applications in Early Drug Discovery,” Jenkins’ talk underscored the shift from conventional discriminative machine learning models to generative approaches that can synthesize novel biological and chemical insights. While traditional models predict outcomes from labeled data, generative models can sample from probability distributions to suggest innovative hypotheses, such as new drug targets or molecular structures.

Jenkins illustrated this power with examples spanning genomics, protein structure, and molecular design. He described how AI can now interpolate between molecular structures in latent space to invent entirely new compounds, and how population-scale genomic data can fuel novel discoveries through genome-wide association studies (GWAS). He also highlighted how models trained on gene expression data can infer cell types and regulatory networks without explicit supervision.

Causaly was one of the AI platforms he highlighted:

“We have an interest in this particular commercial product, Causaly.”

Causaly integrates advanced AI models with a high-precision biomedical knowledge graph, empowering life sciences teams to de-risk decision-making and scale R&D productivity with confidence. There are many different ways scientists use Causaly, an AI platform for scientists, in their research.  The use case this talk focused on is using Causaly Discover, for research during drug discovery: