Healthcare innovation gets a boost from NVIDIA AI

• The compound annual growth rate of healthcare data will be 36% by 2025
• NVIDIA partners with the Broad Institute of MIT to accelerate genome analysis workflows

Any advancements, no matter how simple or complicated, that enhance patient experiences and health outcomes are considered healthcare innovation. However, many business executives and healthcare experts are turning to new technology in response to the challenges they are currently encountering, including strict regulations, privacy concerns, and rising expense.

Artificial intelligence (AI) that can identify early diseases faster than medical professionals and new technologies that use virtual reality to speed up recovery in rehab are just a few of the developments that are now revolutionizing medicine.

Healthcare innovation is important to driving transformation in Asia. More than a billion people are impacted by digital health now, and projections indicate that by 2025, the value of the sector in Asia might reach $100 billion, up from $37 billion in 2020.

​

While healthcare is being impacted by technology innovations, they eventually generate a lot of data. A report shows that the compound annual growth rate of healthcare data will be 36% by 2025.

To handle this ever-growing healthcare data mass, NVIDIA recently announced a partnership with the Broad Institute of MIT and Harvard to equip the Terra cloud platform and its over 25,000 users with AI and acceleration tools. These users include biomedical researchers in academia, startups, and large pharma companies.

The partnership that will bring innovation to healthcare

With a focus on three key areas, the collaboration aims to link the Broad Institute’s renowned academics, scientists, and open platforms with NVIDIA’s AI expertize and healthcare computing platforms.

• Six new Terra workflows now support Parabricks, a GPU-accelerated software suite for secondary processing of sequencing data. With Clara Parabricks, users can now analyze an entire genome in less than an hour as opposed to 24 hours in a CPU-based environment, and they can cut the compute cost in half.
• Using NVIDIA BioNeMo, an AI application framework released for large language models (LLMs) in biology, researchers will create fundamental models for DNA and RNA, the components of life, to more thoroughly comprehend human biology.
• NVIDIA is contributing a new deep learning model directly to the Broad Institute’s Genome Analysis Toolkit (GATK), the industry standard used by more than 100,000 researchers, which helps identify genetic variants that are associated with diseases. This will help drug discovery researchers develop new therapies.

“There’s a need across the healthcare ecosystem for better computational tools to enable breakthroughs in the way we understand disease, develop diagnostics and deliver treatments,” said Kimberly Powell, vice president of healthcare at NVIDIA.

Powell noted that by extending its partnership with the Broad Institute, the company can leverage the strength of large language models to eventually deliver shared solutions and close the gap between research findings and tangible benefits for patients.

By offering an open cloud platform that connects researchers as well as the datasets and tools they require to make scientific breakthroughs, the Broad Institute hopes to allow the next generation of collaborative biomedical research.

“Life sciences are in the midst of a data revolution, and researchers are in critical need of a new approach to bring machine learning into biomedicine,” said Anthony Philippakis, chief data officer of the Broad Institute. “In this collaboration, we aim to expand our mission of data sharing and collaborative processes to scale genomics research.”

How NVIDIA’s technologies study diseases

The BioNeMo framework from NVIDIA provides pretrained LLMs for proteins and chemistry that make scalability, inference, and training easier. The NVIDIA NeMo Megatron framework’s extension, BioNeMo, is domain-specific for chemistry, proteins, and DNA/RNA sequences.

By using billions of parameters, developers may efficiently train and use biology LLMs thanks to BioNeMo. Building on this work, teams from both organizations will produce new models to be added to the BioNeMo collection and made accessible through the Terra platform.

Researchers at the Broad Institute will also have access to MONAI, an open-source deep learning framework for medical imaging AI, and NVIDIA RAPIDS, a GPU-accelerated data science toolbox for quicker data preparation and genomic single-cell analysis.

---

---

---

---