White Paper: Medical

Accelerating Next-Generation Sequencing Secondary Analysis with Dell EMC Isilon and NVIDIA Parabricks


Next-Generation Sequencing (NGS) is a combination of laboratory instrumentation technologies and analysis methods to identify patterns in DNA, the code of life, at dramatically increased resolution and quality. The cost of acquiring NGS data continues to decline exponentially, while the volume of NGS data is doubling every year. The latest NGS instrumentation produces five times more data than the previous generation of instrumentation. As the capacity to sequence DNA continues to increase, organizations like the Global Alliance for Genomics and Health (GA4GH) estimate that over 60 million patients will have their DNA sequenced in a healthcare context by 2025.

However, Secondary Analysis, the conversion of raw NGS data into a usable DNA sequence and compared to a reference, may require a significant amount of time (CPU-hours) to complete. Depending on available computing and storage resources, software, and analysis methodology, secondary analysis time can range from minutes to days. Ideally, there are enough computing and storage resources such that the output of secondary analysis keeps pace with the rate of raw NGS data generation. The goal is to avoid a secondary analysis backlog and ensure processed data are sent further to downstream analysis and interpretation as fast as possible.

This white paper describes a modular, scale-out solution architecture composed of NVIDIA Parabricks application software, NVIDIA V100 Tesla GPUs, and Dell EMC Isilon network-attached storage (NAS) capable of analyzing the daily output of an Illumina NovaSeq 6000 system, or approximately 24 40X whole human genomes sequences (WGS) per day. This solution architecture can scale-out to process over 1,000 WGS per week. This paper also highlights variables to consider when building out a technical computing environment designed to accelerate the secondary analysis of NGS data.

Don't have an account? Sign up here.