Transforming Biology and Drug Discovery: AlphaFold 3 by Google DeepMind and Leading HealthTech Companies to Watch

Published in

Sanjeev Arora

4 min readJun 11, 2024

Disclaimer: Opinions expressed are solely my own and do not express the views or opinions of my employer or any other institution. Any information shared on a publicly listed company or digital assets is for informational purposes only and is not a recommendation of an investment strategy or to buy or sell any security or digital asset.

For background on AlphaFold —

My previous article — Deepmind’s AlphaFold AI system solved biology’s biggest problem
Google Deepmind — https://deepmind.google/technologies/alphafold/

What is the update on AlphaFold 3?

In 2020, at CASP14, DeepMind introduced AlphaFold 2, showcasing an unprecedented level of accuracy that led the scientific community to deem the protein-folding problem resolved. The methods paper detailing AlphaFold 2 has garnered over 20,000 citations in scientific literature, ranking it among the top 500 most-cited papers across all fields.

May 2024, with AlphaFold 3, Google’s DeepMind unveiled an AI model that goes beyond proteins to a broad spectrum of biomolecules, including large biomolecules such as DNA and RNA, as well as small molecules, also known as ligands, which encompass many drugs.

In partnership with Isomorphic Labs, AlphaFold 3 can predict the structure and interactions of all of life’s molecules. This technology will potentially enable scientists to determine biomolecular structures, revolutionizing drug design and the ability to target diseases.

At the core of the model is an improved version of our Evoformer module — a deep learning architecture that underpinned AlphaFold 2’s incredible performance. AlphaFold 3’s updated diffusion-based architecture can predict not only structures involving proteins, nucleic acids, small molecules, ions, and modified residues, but also protein-ligand, protein-nucleic acid, and antibody-antigen interactions. These capabilities suggest that deep learning will create models with high accuracy across the biomolecular space.

AlphaFold 3’s prediction visualization -

Transforming and accelerating drug discovery -

AlphaFold 3 enhances drug design by predicting the behavior of molecules often used in medications, such as ligands and antibodies, which bind to proteins and influence their interactions in human health and disease.

According to Deepmind, AlphaFold 3 is 50% more accurate than the best traditional methods on the PoseBusters benchmark without needing the input of any structural information, making AlphaFold 3 the first AI system to surpass physics-based tools for biomolecular structure prediction.

Isomorphic Labs is already leveraging AlphaFold 3 to expedite and enhance drug design, aiding in the understanding of new disease targets and creating innovative strategies to tackle existing ones that were previously unattainable.

AlphaFold 3 offers a high-definition view of the biological world, enabling scientists to observe cellular systems in all their complexity, including structures, interactions, and modifications.

The true impact of AlphaFold 3 and the AlphaFold Server (a free and easy-to-use research tool) will be seen in how they empower the scientific community to accelerate discoveries and explore new research avenues in biology.

Prominent AI HealthTech Companies in Drug Discovery to Watch —

AI will help generate insights to identify drug targets, predict novel compound behavior, repurpose old drugs, forecast side effects, and determine patient-specific drug efficacy.

Selected list of companies —

Atomwise, San Francisco — Leveraging the power of AI to revolutionize small molecule drug discovery (source: raised over $170m).
BenevolentAI, London — Applying advanced AI to accelerate biopharma drug discovery (Euronext Amsterdam: BAI).
C4X Discovery, UK — Experts in discovering and developing small molecule medicines for immuno-inflammatory diseases.
Exscientia, UK — Applying AI to precision engineer medicines more rapidly and efficiently (NASDAQ: EXAI).
Genesis Therapeutics, San Francisco — AI for biotech, 3D spatial graph modeling with cutting-edge molecular simulation to generate novel drug candidates (source: raised over $250m).
Insilico Medicine, Hong Kong, and New York — Generative AI platform to rapidly and efficiently advance fully self-generated AIDD pipeline primarily composed of novel drug candidates (source: raised over $400m).
Insitro, San Francisco — Decoding human biology with machine learning (source: raised over $640m).
OrphAI Therapeutics, Connecticut —A clinical-stage biopharmaceutical company developing novel therapeutics for rare diseases (source: raised over $98m). OrphAI is also associated with Dr. Jonathan Rothberg, the scientist, engineer, and entrepreneur who is the founder of multiple life science and medical device companies including CuraGen, 454 Life Sciences, Ion Torrent, RainDance Technologies, ClariFI, Butterfly, (world’s first personal ultrasound-on-a-chip that’s FDA cleared for 13 indications), Quantum-Si, Hyperfine (an innovative portable magnetic-resonance-imaging (MRI) machine called Swoop), Detect, Liminal Sciences, Identifeye Health, Protein Evolution, and 4Catalyzer.
Owkin, Paris- Leveraging artificial intelligence to identify new treatments, optimize clinical trials, and develop AI diagnostics (source: raised over $300m).
PathAI, Boston, Massachusetts — Using AI-powered pathology to advance every phase of drug and diagnostic development (source: raised over $350m).
Recursion Pharmaceuticals, Salt Lake City, Utah — Leveraging AI and automation to map and navigate biology. Learn about Recursion OS (NASDAQ: RXRX).
Relay Therapeutics, Cambridge, Massachusetts — Advancing a pipeline of promising therapeutic candidates, with an initial focus on precision oncology and genetic disease (NASDAQ: RLAY).
Schrodinger, New York — Scientific software specializing in developing computational tools and software for drug discovery and materials science (NASDAQ: SDGR).
Tempus, Chicago, Illinois — Advancing precision medicine through the practical application of artificial intelligence in healthcare. (source: raised over (source: raised over $1.3B).

Notes -

Research Institutes, Academic Researchers, and more —

The Francis Crick Institute — an independent charity, established to be a UK flagship for discovery research in biomedicine.
University of California, San Diego — has developed the POLYpharmacology Generative Optimization Network (POLYGON), a generative AI reinforcement learning model that enables systematic discovery and generation of polypharmacological compounds.
Generate Biomedicines — a Boston-based startup, revealed a program called Chroma, which the company describes as the “DALL-E 2 of biology.”
University of Washington led by biologist David Baker — De novo design of protein structure and function with RFdiffusion