This AI Can Design the Machinery of Life With Atomic Precision

Key Points:

  • Proteins rapidly transform in structure and interact with other biomolecules to drive biological processes.
  • The new AI model, RoseTTAFold All-Atom, broadens biomolecular design by incorporating small molecules essential for protein functions.
  • By expanding AI’s scope beyond proteins to include small molecules, the model enables the creation of novel bio-components and potential new therapies.


A recent study published in Science has expanded the capabilities of artificial intelligence (AI) in modeling various biomolecules that interact with proteins. Led by Dr. David Baker at the University of Washington, the AI system, known as RoseTTAFold All-Atom, goes beyond focusing solely on proteins and now includes other biomolecules like DNA, RNA, and iron-containing small molecules crucial for certain protein functions.


By learning from the sequences and structures of these biomolecules, RoseTTAFold All-Atom can map out complex molecular assemblies at the atomic level. In collaboration with generative AI, the system successfully designed proteins that could bind to heart disease medication, regulate heme for oxygen transport in blood cells, and interact with bilin, a light-absorbing molecule found in plants and bacteria.


The researchers behind this advancement have made the AI tool available to the public, enabling scientists to create more intricate bio-components with the potential to lead to innovative therapies. By broadening the scope of AI modeling beyond proteins, the study aims to facilitate the development of sophisticated treatments and functional molecules that could revolutionize various fields, including medicine and materials science.


In the past, AI technologies like AlphaFold and RoseTTAFold have revolutionized protein structure prediction, unlocking vast possibilities in designing novel proteins with diverse functions. By integrating small molecules into AI modeling, the recent study introduces a new dimension to custom protein design, allowing for the creation of proteins that interact with essential molecules like heme and bilin.


The upgraded AI system, RoseTTAFold All-Atom, excelled in predicting interactions between proteins and non-protein molecules, showcasing its potential in drug discovery and biomolecular engineering. The ability of the AI to generate novel proteins that bind with small molecules accurately at the atomic level paves the way for groundbreaking discoveries and practical applications that could redefine the future of various scientific disciplines.



Prompt Engineering Guides



©2024 The Horizon