NYU Langone Health Accelerates Genomic Analysis With NVIDIA Parabricks

Overview

NYU Langone Health's deciphEHR program is pioneering genomic medicine by sequencing patient genomes and linking them to electronic health records (EHRs). The program aims to sequence 100,000 genomes by the end of its pilot phase, using low-coverage sequencing for cost-effectiveness and scalability. However, computational bottlenecks in alignment and variant calling slowed progress.

The Challenge

• Alignment took 27 minutes per genome on CPUs.
• Variant calling required 7 hours per genome.
• Scalability was hindered by high contention for shared HPC resources.

The Solution: NVIDIA Parabricks

NYU Langone implemented NVIDIA Parabricks, a GPU-accelerated genomics software suite, to streamline secondary analysis. The team tested performance across multiple NVIDIA GPUs, including:

• NVIDIA A100
• NVIDIA L40
• NVIDIA V100

Results

• Alignment runtime reduced by 5x (27 minutes → 5 minutes).
• Variant calling runtime reduced by 10x (7 hours → 40 minutes).
• Force-calling mode was added to Parabricks' HaplotypeCaller, maintaining consistency with CPU-based results.

Key Benefits

• Simplified workflows by reducing container dependencies.
• Enabled flexible deployment alongside CPU resources.
• Accelerated processing to meet sequencing goals faster.

"Variant calling used to take over seven hours per genome. After integrating NVIDIA Parabricks and leveraging GPUs, we reduced that to just 40 minutes—a dramatic improvement for our workflow."
—Jonathan McCafferty, Senior Research Scientist, deciphEHR

Future Outlook

With faster processing, NYU Langone can focus on advancing genomic research and clinical applications, bringing precision medicine closer to reality.

Learn more about NVIDIA solutions for genomics.