{"id":3455,"date":"2024-05-13T09:49:05","date_gmt":"2024-05-13T09:49:05","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2024\/05\/13\/drug-discovery-stat-nvidia-recursion-speed-pharma-rd-with-ai-supercomputer\/"},"modified":"2024-05-13T09:49:05","modified_gmt":"2024-05-13T09:49:05","slug":"drug-discovery-stat-nvidia-recursion-speed-pharma-rd-with-ai-supercomputer","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2024\/05\/13\/drug-discovery-stat-nvidia-recursion-speed-pharma-rd-with-ai-supercomputer\/","title":{"rendered":"Drug Discovery, STAT! NVIDIA, Recursion Speed Pharma R&D\u00a0With AI Supercomputer"},"content":{"rendered":"
\n\t\t<\/span> <\/span> <\/span><\/span><\/p>\n

Described as the largest system in the pharmaceutical industry, BioHive-2 at the Salt Lake City headquarters of Recursion debuts today at No. 35, up more than 100 spots from its predecessor on the latest TOP500 list of the world\u2019s fastest supercomputers.<\/p>\n

The advance represents the company\u2019s most recent effort to accelerate drug discovery with NVIDIA technologies.<\/p>\n

\u201cJust as with large language models, we see AI models in the biology domain improve performance substantially as we scale our training with more data and compute horsepower, which ultimately leads to greater impacts on patients\u2019 lives,\u201d said Recursion\u2019s CTO, Ben Mabey, who\u2019s been applying machine learning to healthcare for more than a decade.<\/p>\n

BioHive-2 packs 504 NVIDIA H100 Tensor Core GPUs<\/a> linked on an NVIDIA Quantum-2<\/a> InfiniBand network to deliver 2 exaflops of AI performance. The resulting NVIDIA DGX SuperPOD<\/a> is nearly 5x faster than Recursion\u2019s first-generation system, BioHive-1.<\/p>\n

Performance Powers Through Complexity<\/b><\/h2>\n

That performance is key to rapid progress because \u201cbiology is insanely complex,\u201d Mabey said.<\/p>\n

Finding a new drug candidate can take scientists years performing millions of wet-lab experiments.<\/p>\n

That work is vital; Recursion\u2019s scientists run more than 2 million such experiments a week. But going forward, they\u2019ll use AI models on BioHive-2 to direct their platform to the most promising biology areas to run their experiments.<\/p>\n

\u201cWith AI in the loop today, we can get 80% of the value with 40% of the wet lab work, and that ratio will improve going forward,\u201d he said.<\/p>\n

Biological Data Propels Healthcare AI<\/b><\/h2>\n

Recursion is collaborating with biopharma companies such as Bayer AG, Roche and Genentech. Over time, it also amassed a more than 50-petabyte database of biological, chemical and patient data, helping it build powerful AI models that are accelerating drug discovery.<\/p>\n

\u201cWe believe it\u2019s one of the largest biological datasets on Earth \u2014 it was built with AI training in mind, intentionally spanning biology and chemistry,\u201d said Mabey, who joined the company more than seven years ago in part due to its commitment to building such a dataset.<\/p>\n

Creating an AI Phenomenon<\/b><\/h2>\n

Processing that data on BioHive-1, Recursion developed a family of foundation models<\/a> called Phenom. They turn a series of microscopic cellular images into meaningful representations for understanding the underlying biology.<\/p>\n

A member of that family, Phenom-Beta<\/a>, is now available as a cloud API and the first third-party model on NVIDIA BioNeMo<\/a>, a generative AI<\/a> platform for drug discovery.<\/p>\n

Over several months of research and iteration, BioHive-1 trained Phenom-1<\/a> using more than 3.5 billion cellular images. Recursion\u2019s expanded system enables training even more powerful models with larger datasets in less time.<\/p>\n

The company also used NVIDIA DGX Cloud<\/a>, hosted by Oracle Cloud Infrastructure, to provide additional supercomputing resources to power their work.<\/p>\n

\"Animation<\/a>
Much like how LLMs are trained to generate missing words in a sentence, Phenom models are trained by asking them to generate the masked out pixels in images of cells.<\/figcaption><\/figure>\n

The Phenom-1 model serves Recursion and its partners in several ways, including finding and optimizing molecules to treat a variety of diseases and cancers. Earlier models helped Recursion predict drug candidates for COVID-19 nine out of 10 times.<\/p>\n

The company announced its collaboration with NVIDIA<\/a> in July. Less than 30 days later, the combination of BioHive-1 and DGX Cloud screened and analyzed<\/a> a massive chemical library to predict protein targets for approximately 36 billion chemical compounds.<\/p>\n

In January, the company demonstrated<\/a> LOWE, an AI workflow engine with a natural-language interface to help make its tools more accessible to scientists. And in April it described<\/a> a billion-parameter AI model it built to provide a new way to predict the properties of key molecules of interest in healthcare.<\/p>\n

Recursion uses NVIDIA software to optimize its systems.<\/p>\n

\u201cWe love CUDA and NVIDIA AI Enterprise<\/a>, and we\u2019re looking to see if NVIDIA NIM<\/a> can help us distribute our models more easily, both internally and to partners,\u201d he said.<\/p>\n

A Shared Vision for Healthcare<\/b><\/h2>\n

The efforts are part of a broad vision that Jensen Huang, NVIDIA founder and CEO, described in a fireside chat<\/a> with Recursion\u2019s chairman as moving toward simulating biology.<\/p>\n

\u201cYou can now recognize and learn the language of almost anything with structure, and you can translate it to anything with structure \u2026 This is the generative AI revolution,\u201d Huang said.<\/p>\n

\u201cWe share a similar view,\u201d said Mabey.<\/p>\n

\u201cWe are in the early stages of a very interesting time where just as computers accelerated chip design, AI can speed up drug design. Biology is much more complex, so it will take years to play out, but looking back, people will see this was a real turning point in healthcare,\u201d he added.<\/p>\n

Learn about <\/i>NVIDIA\u2019s AI platform for healthcare and life sciences<\/i><\/a> and subscribe to <\/i>NVIDIA healthcare news<\/i><\/a>.<\/i><\/p>\n

Pictured at top: BioHive-2 with a few members of the Recursion team (from left) Paige Despain, John Durkin, Joshua Fryer, Jesse Dean, Ganesh Jagannathan, Chris Gibson, Lindsay Ellinger, Michael Secora, Alex Timofeyev, and Ben Mabey.\u00a0<\/i><\/p>\n

\t\t<\/p><\/div>\n","protected":false},"excerpt":{"rendered":"

https:\/\/blogs.nvidia.com\/blog\/drug-discovery-recursion-supercomputer\/<\/p>\n","protected":false},"author":0,"featured_media":3456,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[3],"tags":[],"_links":{"self":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/3455"}],"collection":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/comments?post=3455"}],"version-history":[{"count":0,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/3455\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media\/3456"}],"wp:attachment":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media?parent=3455"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/categories?post=3455"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/tags?post=3455"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}