Sharetreck – Tracking Genomic Investments & clinical trials for HIV.

Introduction

Nvidia, a global leader in computing technology, is making waves yet again with its latest venture into the field of generative AI, this time targeting drug discovery. Just over a year since its announcement, the company has unveiled the BioNeMo AI platform, a revolutionary toolkit designed to transform the way researchers approach drug development.

Understanding BioNeMo

At its core, BioNeMo is a suite of AI tools that enables researchers to delve deep into molecular design. Similar to how language models like ChatGPT are trained to understand language nuances, BioNeMo can decipher the complex connections within biological systems, including DNA, RNA, and proteins. Its primary aim? To empower drug discovery researchers by facilitating virtual experiments before physical trials.

Capabilities and Components

The BioNeMo Cloud service is the cornerstone of this platform, offering pretrained AI models to assist researchers in building AI pipelines for drug development. These models have already found homes in prominent drug-discovery companies such as Evozyne and Insilico Medicine, where they support data-driven drug design for new therapeutic candidates.

Among the standout features of BioNeMo are its six new optimized, open-source models:

1. AlphaFold2: A deep learning model that accelerates protein structure determination from years to minutes, developed by DeepMind. Protein structure determination is the process of determining the three-dimensional arrangement of atoms in a protein molecule. This is crucial because a protein’s structure often dictates its function, and understanding the structure can provide insights into how it interacts with other molecules in biological processes.
2. DiffDock: Predicts the 3D orientation and docking interaction of small molecules with high accuracy and computational efficiency. Protein molecules typically dock with other molecules, such as small molecules or other proteins, through a process known as molecular docking. This interaction plays a crucial role in various biological processes, including enzyme-substrate binding, signal transduction, and protein-protein interactions. The docking process involves complementary shapes and electrostatic properties between the molecules, allowing them to bind together in a specific orientation and form stable complexes. In drug discovery, understanding how a drug molecule docks with its target protein is essential for designing effective therapeutics.
3. ESMFold: Estimates the 3D structure of a protein based on a single amino acid sequence, without requiring examples of similar sequences.
4. ESM2: A protein language model used for inferring machine representations of proteins, useful for downstream tasks such as protein structure prediction and molecular docking.
5. MoFlow: Generates diverse chemical structures for potential therapeutics through molecular optimization and small molecule generation.
6. ProtGPT-2: Generates novel protein sequences to aid in designing proteins with unique structures, properties, and functions.

These models, coupled with BioNeMo’s browser-based interface, make interactive inference and protein structure visualization seamless.

Collaborations and Industry Impact

Nvidia’s ambitions extend far beyond its own platform. Through collaborations with industry giants like Amgen’s deCode, the company aims to expand the reach of its genomics foundation models powered by BioNeMo. Over 100 biopharma and AI drug discovery companies, including Amgen, Astellas Pharma, and Recursion, have embraced the platform, recognizing its potential to revolutionize drug development.

Unlocking the Future of Drug Discovery

With BioNeMo, Nvidia is not only redefining drug discovery but also democratizing access to cutting-edge AI technologies. By providing cloud-based enterprise services and foundational models like MolMIM, the company is paving the way for a new era of innovation in healthcare. As Nvidia healthcare chief Kimberly Powell aptly puts it, BioNeMo’s expanded offerings mark a pivotal step towards empowering healthcare companies to harness the full potential of their data—a mission poised to shape the future of medicine for years to come.

As per NVIDIA, more than a 100 companies are already using BioNemo. Nvidia is leveraging the work done on its Cambridge one supercomputer to enable BioNemo go beyond molecules and into DNA.

All that we spoke about today was NVIDIA in drug discovery. However, when you look at the depth of work NVIDIA has done with AI in other industries, it soon becomes apparent how much of a lead NVIDIA has on its peers. This reflects in its share price.

I am thinking that soon a stock split should be in the works in order to enable sufficient volumes and greater participation.