Bionemo Cloud

[Image courtesy of Nvidia]

During Nvidia’s (Nasdaq:NVDA) GTC event, the company introduced BioNeMo Cloud, a new component to their AI Foundations suite. This service, designed to streamline life sciences research, drug discovery and protein engineering, provides researchers with access to pretrained AI models, allowing for customization with proprietary data. Offered as a cloud service, BioNeMo Cloud enables accelerated drug discovery workflows, with multinational biotech company Amgen (Nasdaq:AMGN) and several startups already utilizing the platform.

BioNeMo Cloud features pretrained AI models for molecular biology, chemistry and molecular dynamics. The platform allows researchers to fine-tune models with proprietary data, executing AI model inference via web browsers or new cloud APIs. This accelerates drug discovery, molecular identification and protein 3D structure prediction.

Featured models in BioNeMo Cloud include AlphaFold2, DiffDock, ESMFold, ESM2, MoFlow and ProtGPT-2. These models encompass various aspects of molecular biology, chemistry and protein engineering, offering unique capabilities for researchers in the field.

Early BioNeMo Cloud adopters include mix of Big Pharma and startups

Several companies, such as Evozyne and Insilico Medicine, have adopted BioNeMo Cloud to enhance data-driven drug design. Generative AI models, trained on large-scale datasets, can quickly identify potential drug molecules and predict protein 3D structures and molecular interactions.

Generative AI has grown more popular in recent months thanks to the widespread success of ChatGPT, an advanced language model that had more than 100 million users two months after the service launched on November 30, 2022.

Early access users of BioNeMo Cloud include AstraZeneca (Nasdaq:AZN), Insilico Medicine, Evozyne, Inophor and Alchemab Therapeutics. These companies cover various areas within the life sciences and pharmaceutical industries.

BioNeMo Cloud offers six new optimized, open-source models in addition to its previously announced models. Accessible through a browser-based interface, these models enable interactive inference and protein structure visualization. By integrating with Nvidia DGX Cloud, researchers can customize their models using a fully managed software service.

Huang said the BioNeMo service demonstrated the potential of generative AI in transforming the life science and pharmaceutical industries, leading to faster and more efficient research and development for new treatments and therapies.

Nvidia also announced a collaboration with Japanese conglomerate Mitsui to launch Tokyo-1, an initiative aimed at accelerating drug discovery. This project will provide Japanese pharmaceutical companies and startups access to Nvidia DGX AI supercomputers and technologies, with training and technical support available for researchers.

“Generative AI is a new computing platform, like PC, internet, mobile and cloud,” said Huang in a keynote address. “As first movers create new applications and found new companies, they will capitalize on generative AI’s ability to automate and co-create the world.”