GTC 2020 Part II: Race to the Cure with AI
NVIDIA CEO Jensen Huang’s GTC 2020 keynote address generated major excitement for anyone who relies on its technology. The address was divided into four segments (each available on YouTube) and our first blog on the address centered on the metaverse and NVIDIA Omniverse. This blog focuses on AI and how it’s being employed to combat disease.
The second installment of Jensen’s GTC 20 keynote began with Eroom’s Law, which in short, means that despite advances in technology, drug discovery is becoming slower and more expensive, i.e., over a decade to develop, $2.5 billion in research and development, and costs doubling every nine years. Worst of all, 90% of efforts fail. The good news is, through AI, simulation can accelerate the search for new drugs before entering the slow process of in vivo testing (procedures done on a living organism).
It’s hard to find the protein implicated in the disease, hard to find a small molecule chemical compound to bind with to activate or deactivate the protein, and (due to the human body’s immune system), hard to get the small molecule inside the cell. Next, pharmacokinetics (PK) comes into play which studies how your body safely distributes, absorbs, metabolizes, and excretes the compound.
Data analytics screens potential compounds for “fingerprints” which may lead to a drug candidate. Deep learning generates new leads learning from known chemicals. Docking predicts the strength of the bond between the chemical and the protein. Natural language processing (NLP) studies literature to find existing chemicals that might also be effective against other diseases.
Jensen then unveiled NVIDIA Clara Discovery, a new suite of tools for scientists to discover lifesaving drugs and detailed how NVIDIA is working with Cambridge to build Cambridge 1, the UK’s most powerful AI supercomputer equipped with 80 NVIDIA DGX A100 systems. The systems will also include the DGX Superpod because AI software is computer written, organizations will need powerful AI computers to write software.
When most industry pros think of BOXX customers, they assume the majority are either employed in media & entertainment (M&E) , manufacturing & product design (MPD), or architecture, engineering and construction (AEC). That’s an accurate assessment, but also incomplete. BOXX has a sizable number of customers throughout other markets including academia, energy, government & defense, and…medical science.
When deep learning and AI first appeared on the horizon, BOXX (through our close partnership with NVIDIA) quickly adopted the latest technology, manufacturing a line of data science systems ideal for scientists and medical researchers. As NVIDIA continues to introduce new advancements used in the war against disease, BOXX will join the fight by quickly adopting those technologies and providing state-of-the-art systems for universities, research laboratories, and other medical science facilities.