NVIDIA Clara Powers New AI-Driven Genomic Analysis on Terra Platform to Accelerate Biomedical Discovery

On September 20, 2022 (Pacific Time), NVIDIA announced a collaboration with the Broad Institute of MIT and Harvard to provide AI and acceleration tools for the Terra cloud platform, enabling rapid analysis of massive medical datasets. This initiative will benefit more than 25,000 users of the platform, including biomedical researchers from academia, startups, and large pharmaceutical companies.

This collaboration aims to connect NVIDIA’s expertise in AI and its medical computing platforms with the globally renowned researchers, scientists, and open platform of the Broad Institute, focusing on three key areas:

•NVIDIA Clara Parabricks Available on the Terra Platform: Parabricks is a GPU-accelerated software suite for secondary analysis of sequencing data, now accessible through six new Terra workflows. Users can now leverage Clara Parabricks to complete whole-genome analysis in just over an hour, compared to 24 hours required in CPU-based environments, while reducing computational costs by more than half.

•Building Large Language Models (LLMs): To delve deeper into human biology, researchers will use the AI application framework for biological LLM models released today—NVIDIA BioNeMo—to develop foundation models for DNA and RNA, the “building blocks of life.”

•Bringing More Powerful Deep Learning to the Genome Analysis Toolkit (GATK): NVIDIA is working to develop new deep learning models for the Broad Institute’s GATK, an industry-standard toolkit used by over 100,000 researchers, to help identify disease-associated genetic variants. This will support drug developers in researching new therapies.

Kimberly Powell, Vice President of Healthcare at NVIDIA, stated, “The entire healthcare ecosystem requires more advanced computational tools to enable breakthroughs in our understanding of diseases, the advancement of diagnostics, and the delivery of therapeutic solutions. By expanding our collaboration with the Broad Institute, we can harness the power of large language models to ultimately deliver joint solutions that translate researchers’ deep insights into tangible benefits for patients.”

The Broad Institute aims to enable a new generation of collaborative biomedical research by providing an open cloud platform that connects researchers with one another and links them to the datasets and tools necessary for achieving scientific breakthroughs.

Anthony Philippakis, Chief Data Officer at the Broad Institute, stated, “The life sciences sector is undergoing a data revolution, and researchers urgently need a new approach to integrating machine learning into biomedicine. Through this collaboration, we aim to further advance our mission of ‘data sharing and collaborative processes,’ thereby expanding genomic research.”

NVIDIA’s BioNeMo framework includes pre-trained large language models (LLMs) for protein and chemistry domains, simplifying training, inference, and scaling. BioNeMo is an extension of the NVIDIA NeMo Megatron framework tailored for chemistry, proteins, and DNA/RNA sequences.

Through BioNeMo, developers can effectively train and deploy biological large language models (LLMs) with billions of parameters.

Building on this collaboration, the two teams will jointly develop new models, incorporate them into the BioNeMo collection, and make them available on the Terra platform.

NVIDIA Parabricks’ GPU-accelerated workflows provide researchers with faster turnaround times and lower costs for extensive genomic data analysis. Within the Broad Institute’s GATK Best Practices pipeline for germline variant calling, Parabricks accelerates analysis on GPUs by 24-fold while halving costs.

Researchers at the Broad Institute also have access to MONAI, an open-source deep learning framework for medical imaging AI, and NVIDIA RAPIDS, a GPU-accelerated data science toolkit that accelerates data preparation. The latter can be used for single-cell genomic analysis.

VCBeat believes that NVIDIA has been continuously introducing and implementing more applications in medical AI, leveraging its unique approach to support healthcare by providing the entire medical ecosystem with more powerful computing capabilities and advanced computational tools. After years of consistent and focused development, its solutions have become deeply integrated into every stage of medical workflows, making significant contributions to breakthroughs in disease diagnosis and treatment.

Learn MoreIntegration of Clara Parabricks with Terra, and submit a registration applicationNVIDIA BioNeMo LLM Service Early Access。

Watch the GTC keynote address by Jensen Huang, founder and CEO of NVIDIA, to learn more about the collaboration between NVIDIA and Bodhi Institute.