Scientists have explored new insights into proteins, DNA and other building blocks of life. The NVIDIA BioNeMo framework, announced at NVIDIA GTC 2022, is designed to accelerate their research.
NVIDIA BioNeMo is the foundation for learning and implementing large biomolecular language models
The Large Language Model (LLM) will support chemistry, protein, DNA and RNA data formats to help scientists better understand disease and find treatments for patients.
This will be part of NVIDIA’s Clara Discovery applications, a collection of AI frameworks and models for drug discovery. As artificial intelligence learns to understand human languages through MF, it also learns the languages of chemistry and biology. This system will facilitate the training of massive neural networks based on biomolecular data and help researchers discover new patterns and understanding of biological sequences. With these fresh discoveries, researchers can tap into biological functions, including human health. “As we see more and more widespread adoption of large language models in the protein space,
The ability to effectively train LMSs and rapidly modulate model architecture is becoming more and more
important,” said Istvan Redl, lead machine learning specialist at Pepton, a biotech startup in NVIDIA’s
Creation Program. “We believe that these two aspects of engineering – scalability and rapid
experiments are exactly what BioNeMo can provide.”
The framework will be available on the NVIDIA GPU Cloud, a hub for GPU-optimized software. A wave of biotech and pharma experts such as AstraZeneca, OpenFold, Peptone and Evozyne, including researchers from the Broad Institute at MIT and Harvard, are adopting NVIDIA BioNeMo to support drug discovery research.