Oxford Nanopore Collaborates With NVIDIA To Integrate DGX Station A100 With An Ultra-High-Throughput Sequencing System, PromethION

0
1613
Source: https://nanoporetech.com/about-us/news/oxford-nanopore-and-nvidia-collaborate-partner-dgx-ai-compute-system-ultra-high

Oxford Nanopore Technologies and NVIDIA collaborate to integrate the NVIDIA DGX Station A100 into Oxford Nanopore’s ultra-high-throughput sequencing system, PromethION. The collaboration aims to deliver the world’s most robust sequencer that supports real-time analyses at any scale and can analyze any length fragment of DNA/RNA. 

The ever-increasing availability of nanopore sequencing data is supported by the use of accelerated computing and AI to efficiently sequence DNA or RNA, at scale, to a variety of high-throughput users. Researchers are increasingly using Oxford Nanopore’s technology to analyze thousands of genomes for understanding genetic diversity and discover new variants. Sequencing is also used to generate rapid insights into healthcare settings, food safety, or environmental analysis.

Recently launched NVIDIA DGX Station A100 is a data-center-grade, GPU-powered, multi-user workgroup appliance to tackle the various complex AI workloads. It plugs directly into an office or laboratory outlet and is tranquil due to its refrigerant-based cooling system. It offers 320GB of GPU memory by having four NVIDIA A100 80GB GPUs, fully connected via NVIDIA NVLink.

The modern analytical ability for ultra-high-throughput sequencing

NVIDIA’s 2.5 petaFLOP AI compute system grants unprecedented compute density, performance, and flexibility in a benchtop format. Oxford Nanopore’s PromethION P48 sequencing device can generate as much as 10 Terabases of DNA information per 72-hour run. This breakthrough represents a 25% increase in data output. Regular advances in flow cell chemistry have made this increase. These developments have been exhibited in customer data, with increasing yields reported over a scope of sequencing applications. 

PromethION is supplied in a P24 and a P48 format and is widely being deployed into high-throughput projects, where the rich sequencing data provided by Oxford Nanopore can be delivered at very high throughput. 

The technology helps academic societies, core facilities, and service providers realize the value of sequencing every length of fragments, from short fragments to over 100,000 bases long, and characterizing base modifications, coupled with high-efficiency single nucleotide. Oxford Nanopore sequencing systems are using NVIDIA GPUs are already running real-time sequencing analysis at various scales. 

The solutions for high accuracy sequence data

Using robust AI systems substantially improves the accuracy of Oxford Nanopore’s sequencing data. The revised analysis algorithms can result in higher accuracy of the very same sequence data. Bonito CRF, Oxford Nanopore’s new machine-learning driven analysis algorithm, has reported >98% single read base-calling accuracy. Base-calling is the method of identifying the sequence of bases on an individual molecule of DNA. The revised Bonito is trained with a more extensive and diverse data set and builds on earlier work to achieve enhanced performance.

With the latest technology release, variant calling performance is also improving. Structural variation (SV) accuracy has reached the gold standard of 96% with 30X rather than 60X coverage using the latest tools. Oxford Nanopore has now achieved SNV at 99.92% compared to traditional SBS accuracy. 

Besides, Oxford Nanopore also generated a modal single-read accuracy of 99.1% (99%=Q20) using new chemistry with Bonito, presented on internal validation sets, with a large proportion of the raw reads above Q20. Oxford Nanopore and NVIDIA are using the latest AI advancements to make biological analysis available to anyone.

Source: https://nanoporetech.com/about-us/news/oxford-nanopore-and-nvidia-collaborate-partner-dgx-ai-compute-system-ultra-high

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.