Before generative AI chatbots disrupted the AI field, tech giants were working on something more concrete — protein folding. Just last year, Google DeepMind’s AlphaFold solved the 50-year-old grand challenge of protein folding. Unfortunately, it did not yet receive the attention it deserved. 

However, its potential extends beyond just medicine and drug discovery. And that is where NVIDIA comes into the picture. CEO Jensen Huang believes in the incredible potential of generative AI in protein folding. 

“There’s a significant shift from traditional software and product design to the design of biological solutions. AI can be used to address environmental issues, such as reducing plastic waste in the oceans, capturing carbon emissions, and finding new solutions to tackle these challenges,” Huang, the chief of NVIDIA, told AIM in an exclusive interview during his recent visit to Bengaluru last week. 

But why is he so keen on this?

Beyond Medicine

Protein folding research offers significant indirect contributions. By comprehending protein folding dynamics, scientists can engineer enzymes for efficient biofuel production, capture and store carbon dioxide, enhance agricultural practices for climate-resilient crops, facilitate bioremediation of pollutants, and promote sustainable protein sources. 

Additionally, protein folding insights aid in drug discovery for climate-related diseases. While not a standalone climate solution, this research provides crucial tools and knowledge to develop technologies and strategies to address climate-related challenges and underscores the value of interdisciplinary approaches in tackling global environmental issues.

NVIDIA’s Big Bets Healthcare

Back in March at the GTC conference, NVIDIA launched a set of generative AI cloud services known as BioNeMo Cloud consisting of AlphaFold2 by DeepMind, DiffDock by MIT, ESMFold and ESM2 by Meta, MoFlow by Cornell University and ProtGPT-2. This service accelerates various aspects of drug discovery, including protein and therapeutics research, genomics, chemistry, biology, and molecular dynamics, making them easily accessible through an interactive interface. Researchers can fine-tune generative AI models on proprietary data, run AI model inference through web browsers or cloud APIs, and access pretrained models for drug development. 

Since then several pharmaceutical companies like Amgen have used BioNeMo to reduce drug discovery times, while startups like Evozyne and Insilico Medicine have harnessed it to design therapeutic candidates more efficiently and cost-effectively. 

In July, Recursion, an AI-driven drug discovery company, expanded its capabilities through a partnership with NVIDIA whereby the latter would invest $50 million in Recursion and provide access to its cloud-based AI tools for drug discovery. The companies plan to develop new AI models for drug discovery on NVIDIA’s DGX Cloud, using the former’s extensive biological and chemical dataset. Recursion also intends to leverage BioNeMo for its own drug discovery projects. 

Not just BioNeMo, NVIDIA also plays a pivotal role in advancing drug discovery through its GPU-accelerated platform, NVIDIA Clara which integrates AI, data analysis, simulations, and visualisation to swiftly sift through extensive chemical libraries for potential drug candidates, modelling protein structures and dynamics to comprehend their roles and interactions with medications, crafting novel molecules with desired attributes, simulating drug effects within the human body, and visualising findings from drug discovery experiments. 

Leveraging NVIDIA Clara, healthcare professionals and organisations have accomplished significant breakthroughs like the creation of plans for two innovative proteins using BioNeMo, the successful execution of a groundbreaking surgical procedure with Holoscan, and the implementation of cutting-edge MONAI-driven solutions within radiology departments.

Deloitte utilises BioNeMo for 3D protein structure prediction, while Innophore analyses protein cavities with it. Holoscan enhances medical devices for real-time AI applications, as exemplified by a successful robot-assisted surgery. Parabricks accelerates genomic analysis, enabling faster and more accurate diagnoses, with Form Bio and PacBio as prime examples. 

Beyond software, NVIDIA also offers a suite of hardware solutions, encompassing GPUs, servers, and storage systems tailored to meet the exact computational demands of drug discovery, consequently enhancing the process’s speed and efficiency.

It is safe to say that NVIDIA is only going to get bigger and better with their protein folding missions. 
Read more: Meet the Genius behind Med-PaLM 2

The post Does Jensen Huang Love Protein Folding? appeared first on Analytics India Magazine.