Nvidia Gives Health Researchers Access to Supercomputer in UK

Cambridge-1, the UK’s fastest supercomputer, is officially operational. But forget about ultra-precise weather forecasts or dinosaur extinction surveys – this supercomputer is exclusively intended to advance healthcare and boost the life science industry.

Launched by Nvidia, which has invested a total of $ 100 million in the design and construction of the device, Cambridge-1 is expected to build research capacity in areas such as digital biology, genomics and drug discovery. .

The system was built in just 20 weeks, despite restrictions caused by the Covid-19 pandemic, which prevented Nvidia engineers from crossing the Atlantic to help their UK-based colleagues during construction. Instead, the company used a number of technologies, ranging from sophisticated computer modeling to remote-controlled mobile robots, to ensure Cambridge-1 could be delivered and launched on time.

41e place in the Top500

Healthcare data is growing exponentially – a trend that has only accelerated during the Covid-19 pandemic. Used wisely, this data can be invaluable: it can help inform decisions in the laboratory when developing a new drug, to improving patient care through personalized therapy.

But according to Nvidia, until now British researchers have not had access to sufficient computing power to process the huge amounts of data they have and learn useful lessons that could unlock these opportunities.

With Cambridge-1 now live, that could change. The supercomputer delivers 400 petaflops of AI performance and eight petaflops of Linpack performance, and is powered by 80 Nvidia DGX A100 systems, which are designed to build and run large-scale AI projects – in this case, large-scale AI projects. machine learning applications dedicated specifically to advancing health care research.

Only 20 DGX A100s provide the equivalent of hundreds of processors. Nvidia presented this system as the fastest in the country and Cambridge-1 which also occupies the 41e place on the Top500 list, which lists the 500 most powerful supercomputers in the world. “Cambridge-1 will enable leading researchers in business and education to do their lifetimes’ work on the UK’s most powerful supercomputer, and uncover clues to diseases and illnesses. processing at a scale and at a speed hitherto impossible in the UK, ”says Jensen Huang, Founder and CEO of Nvidia.

Five organizations were granted early access

A recent report by business consultancy Frontier Economics shows that the supercomputer has the potential to create up to £ 600million (around € 701million) of value in the UK over the next 10 years.

Five organizations have had early access to the system since it was announced last year. One of these is Oxford-based start-up Oxford Nanopore Technologies, which leveraged the computing power of Cambridge-1 to perform genomic sequencing, allowing scientists to understand how the virus Covid-19 is evolving to determine the design of vaccines and drug treatments. According to Nvidia, the supercomputer can help researchers change the algorithm used for genomic sequencing in hours, rather than days, to achieve the highest levels of precision and speed. In the field, this allows scientists to react more quickly to mutations in viruses.

Cambridge-1 is also expected to advance healthcare beyond the immediate crisis. Another early partnership between Nvidia and AstraZeneca, for example, focused on using Cambridge-1 to run next-generation algorithms, which can accelerate drug discovery by predicting reactions or optimizing molecules.

Called MegaMolBART, this algorithm is an improvement of the MolBART algorithm, which AstraZeneca is already testing. Instead of using large datasets that have been laboriously hand-labeled, MolBART is a self-supervised program, meaning it can browse a database of compounds and start learning the relationships between different chemical structures. . Ultimately, the algorithm can give researchers ideas for molecules that do not exist in the databases but that could be potential drug candidates.

Advances at the clinical level

At the clinical level, Nvidia also has a long-standing relationship with King’s College London and the Guy’s and St Thomas’ NHS Foundation Trust, where scientists are already using Cambridge-1 to improve their diagnosis of various diseases. After training AI models on tens of thousands of MRI brain scanners, the researchers used the supercomputer to generate synthetic brain images that, according to Nvidia, even radiologists were unable to differentiate from genuine images.

This synthetic data allows scientists to gain a more nuanced understanding of what illnesses like dementia, stroke, brain cancer or multiple sclerosis look like because they have access to never-before-seen brain images showing different characteristics, such as age or type of disease. “The power of artificial intelligence in healthcare will accelerate diagnosis for patients, improve services like breast cancer screening, and support the way we assess risks and classify them. patients in order of priority, according to their clinical needs, ”says Ian Abbs, Managing Director of Guy’s and St Thomas’ NHS Foundation Trust.

Nvidia’s fifth Cambridge-1 partner to date, pharmaceutical giant GSK, has focused on using Cambridge-1 to analyze genetic data, with the aim of predicting the success of new drugs. According to the company, the supercomputer-based approach could help design drugs that are twice as likely to be clinically successful and become authorized therapies.

With Cambridge-1 now officially live, Nvidia expects more research projects to be launched to make the most of the health data that is currently accumulating, with the aim of improving patient outcomes. . For example, the company predicts that the supercomputer could play a leading role in modifying existing Covid-19 vaccines to combat the new variants. It could also advance scientists’ understanding of the long-term effects of the disease, by predicting what those effects might be and what proportion of the population might be at risk.

Source: .com

Back to top button