NVIDIA Introduces Large Language Models and Generative AI Service to Propel Life Sciences R&D
NVIDIA introduced an extended range of generative AI cloud services called BioNeMo™ Cloud, tailored for refining AI foundation models and expediting the development of new drugs, proteins, and therapeutic solutions. This innovation, under NVIDIA’s AI Foundations, facilitates AI model training and inference, significantly hastening the typically time-consuming and expensive phases of drug discovery. The service enables researchers to customize generative AI applications using their exclusive data and conduct AI model inference through web browsers or new cloud APIs. Kimberly Powell, NVIDIA’s VP of healthcare, highlighted the potential of generative AI in revolutionizing the life sciences and pharmaceutical sectors, stating that BioNeMo Cloud Service functions as an AI drug discovery lab.
Amgen Among Early Users
Amgen, a leading biotech company, is utilizing BioNeMo to enhance its research and development endeavors in biologics discovery. The service allows them to pretrain large language models on their proprietary data, facilitating the exploration and development of therapeutic proteins for advanced medicine.
Generative AI Supercharges Drug Discovery Pipeline
BioNeMo Cloud Service offers pretrained AI models that support various stages of drug development and has been embraced by companies like Evozyne and Insilico Medicine for data-driven drug design. These generative AI models expedite the identification and design of potential drug molecules, predict protein structures, and analyze molecular interactions, thereby significantly impacting the drug discovery pipeline.
New Generative AI Models Available With BioNeMo Service Early Access
BioNeMo now has six new optimized, open-source models, in addition to its previously announced MegaMolBART generative chemistry model, ESM1nv protein language model and OpenFold protein structure prediction model. They include:
The BioNeMo Service provides easy access to these generative AI models through a browser-based interface for interactive inference and protein structure visualization. Pairing BioNeMo with NVIDIA’s DGX Cloud supercomputing resources enables researchers to customize models using NVIDIA’s managed software services.
Pharma Companies, Startups Tap BioNeMo to Optimize AI Workflows
Several pharmaceutical companies and startups, including Amgen, Evozyne, and Insilico Medicine, have experienced notable benefits using BioNeMo. Amgen, for instance, reduced the time required to train custom models for molecule screening from three months to a few weeks using DGX Cloud. Evozyne collaborated with NVIDIA to create a Protein Transformer Variational AutoEncoder, enhancing synthetic variant designs compared to natural enzymes. Meanwhile, Insilico Medicine accelerated early drug discovery, reducing time and costs significantly, and is on track to advance a preclinical drug candidate to phase 2 clinical trials faster and at a lower expense than traditional methods.
*Credit goes to NVIDIA Corporation for their insightful article titled “NVIDIA Unveils Large Language Models and Generative AI Service to Advance Life Sciences R&D.” We’re delighted to share it with our readers.