Generative AI is remodeling drug analysis and growth, enabling new discoveries quicker than ever — and Amgen, one of many world’s main biotechnology firms, is tapping the expertise to energy its analysis.
Amgen will construct AI fashions skilled to investigate one of many world’s largest human datasets on an NVIDIA DGX SuperPOD, a full-stack knowledge heart platform, that will likely be put in at Amgen’s deCODE genetics’ headquarters in Reykjavik, Iceland. The system will likely be named Freyja in honor of the highly effective, life-giving Norse goddess related to the power to foretell the longer term.
Freyja will likely be used to construct a human range atlas for drug goal and disease-specific biomarker discovery, offering very important diagnostics for monitoring illness development and regression. The system may even assist develop AI-driven precision medication fashions, probably enabling individualized therapies for sufferers with critical illnesses.
Amgen plans to combine the DGX SuperPOD, which is able to function 31 NVIDIA DGX H100 nodes totaling 248 H100 Tensor Core GPUs, to coach state-of-the-art AI fashions in days relatively than months, enabling researchers to extra effectively analyze and be taught from knowledge of their seek for novel well being and therapeutics insights.
“For greater than a decade, Amgen has been making ready for this hinge second we’re seeing within the trade, powered by the union of expertise and biotechnology,” stated David M. Reese, govt vice chairman and chief expertise officer at Amgen. “We look ahead to combining the breadth and maturity of our world-class human knowledge capabilities at Amgen with NVIDIA’s applied sciences.”
The objective of deCODE founder and CEO Kári Stefánsson in beginning the corporate was to know human illness by trying on the range of the human genome. He predicted in a latest Amgen podcast that inside the subsequent 10 years, docs will routinely use genetics to discover unusual illnesses in sufferers.
“This SuperPOD has the potential to speed up our analysis by coaching fashions extra rapidly and serving to us generate questions we would not have in any other case thought to ask,” stated Stefánsson.
Placing the Tech in Biotechnology
Since its founding in 1996, deCODE has curated greater than 200 petabytes of de-identified human knowledge from practically 3 million people.
The corporate began by gathering de-identified knowledge from Icelanders, who’ve a wealthy heritage in genealogies that stretch again for hundreds of years. This population-scale knowledge from analysis volunteers offers distinctive insights into human range because it applies to illness.
deCODE has additionally helped sequence greater than half one million human genomes from volunteers within the UK Biobank.
However drawing insights from this a lot knowledge requires highly effective AI methods.
By integrating highly effective new expertise, Amgen has a possibility to speed up the invention and growth of life-changing medicines. In March 2023, NVIDIA introduced that Amgen turned one of many first firms to make use of NVIDIA BioNeMo, which researchers have used to construct generative AI fashions to speed up drug discovery and growth. Amgen researchers have additionally been accessing BioNeMo by way of NVIDIA DGX Cloud, an AI supercomputing service.
“Fashions skilled in BioNeMo can advance drug discovery on a number of fronts,” stated Marti Head, govt director of computational and knowledge sciences at Amgen. “Along with serving to develop medication which can be simpler, they’ll additionally assist keep away from negative effects like immune responses, and new biologics may be made in quantity.”
By adopting DGX SuperPOD, Amgen is poised to realize unprecedented knowledge insights with the potential to vary the tempo and scope of drug discovery.
“The fusion of superior AI, groundbreaking developments in biology and molecular engineering and huge portions of human knowledge usually are not simply reshaping how we uncover and develop new medicines — they’re redefining medication,” Reese stated.
Study NVIDIA’s AI platform for healthcare and life sciences.
