NVIDIA Advances Healthcare AI with Protein Design

NVIDIA is expanding its healthcare AI initiatives through new partnerships focused on protein design and personalized cancer treatment. The company has unveiled an open protein design model, RFD4-Proteina, developed with the Institute for Protein Design at the University of Washington School of Medicine.

This model, part of NVIDIA’s BioNeMo platform, combines programmable protein design with search capabilities. It enables scientists to generate and refine protein designs using molecular inputs, constraints, and scoring objectives. Potential applications include creating proteins to interrupt disease processes or enhance diagnostic tests.

Protein Design Model Demonstrates Success in Testing

In testing, RFD4-Proteina achieved a higher computational design success rate compared to evaluated baselines across protein binding, small-molecule binding, enzyme scaffolding, and DNA binding. The model is now accessible through NVIDIA’s partners, including Amazon Bio Discovery, Anthropic, Benchling, Databricks, Google Cloud, Hugging Face, Microsoft Azure, Nebius, OpenAI, and Tamarind Bio.

NVIDIA has also partnered with Siemens Healthineers to develop adaptive radiotherapy, utilizing AI-assisted treatment planning, patient-specific simulation, and real-time computing. This collaboration aims to streamline workflows and broaden access to personalized cancer care.

Adaptive Radiotherapy: Tailoring Treatment to Patients

The adaptive radiotherapy initiative utilizes NVIDIA’s Nemotron open foundation models to assist oncologists in assessing and improving personalized treatment. The companies will explore integrating these capabilities with CUDA-accelerated radiation dose calculations and the Holoscan platform to combine live sensor data with AI and simulation.

“Adaptive radiotherapy is built on a single principle: treatment should adapt to the patient, not the other way around,” said Arthur Kaindl, head of Varian, the Siemens Healthineers cancer-care business. “Our collaboration with NVIDIA builds on that progress, exploring how AI and real-time computing could provide clinicians with greater speed and insight to make informed decisions, adapt treatment with greater confidence, and deliver personalized care to more patients.”

These partnerships expand on NVIDIA’s recent healthcare AI efforts. In January, NVIDIA and Eli Lilly and Co. announced a $1 billion investment in a San Francisco-based AI co-innovation lab to accelerate drug development using the BioNeMo platform.

In March, Roche extended its partnership with NVIDIA to deploy a hybrid-cloud AI factory, aiming to improve drug discovery and clinical trial efficiency. In July, NVIDIA and Bristol Myers Squibb announced plans to build a powerful AI factory for the life sciences industry, utilizing NVIDIA’s DGX SuperPOD and DGX Vera Rubin NVL72 systems.

NVIDIA’s David Niewolny, head of healthcare and medical, emphasized the impact of AI in drug discovery, noting the convergence of compute power and innovation in this field. Last year, NVIDIA became the world’s first company to reach a $5 trillion valuation, with its stock currently trading at around $216.51 per share.

NVIDIA is broadening its healthcare AI initiatives through collaborations with leading pharmaceutical companies.