A recent article in hLife explores the advancements in protein science recognized by the 2024 Nobel Prize in Chemistry.
Study: Artificial intelligence is reshaping the study of proteins: Structures and beyond. Image Credit: Anggalih Prasetya/Shutterstock.com
The paper highlights the pivotal contributions of David Baker, Demis Hassabis, and John Jumper, whose work in computational protein design and AI-driven structure prediction has significantly advanced the field. By leveraging artificial intelligence (AI) and deep learning, tools like AlphaFold2 and generative models have tackled longstanding challenges in protein folding and design.
Understanding Protein Structure and Design
Proteins are the molecular machines that drive life, and their function depends entirely on how they fold into precise three-dimensional (3D) shapes—determined by their amino acid sequences. Predicting these structures from sequences has long been one of biology’s biggest challenges, while protein design takes it a step further by identifying sequences that create specific structures or functions.
For years, scientists relied on physics-based models like Baker’s Rosetta to simulate protein folding and design. While these methods made progress, they struggled with accuracy and couldn’t fully explore the vast range of possible structures. That changed with AlphaFold2, developed by Hassabis and Jumper’s team, which used deep learning to predict protein structures with near-experimental accuracy.
At the same time, generative AI models have revolutionized protein design by solving key challenges in backbone generation and sequence optimization, opening up new possibilities in biotechnology and medicine.
Advances in Protein Structure Prediction
Historically, predicting protein structures required minimizing energy functions to identify stable conformations, a process complicated by the immense number of possible structures. Early innovations like comparative modeling and residue contact prediction using multiple sequence alignments (MSA) helped narrow this search, but true accuracy remained out of reach.
AlphaFold2 changed the game at the 14th CASP (Critical Assessment of Protein Structure Prediction) competition. Its end-to-end deep learning approach directly predicted 3D structures from amino acid sequences, including MSAs, bypassing intermediate steps that had previously constrained accuracy. As a result, AlphaFold2 produced a database covering virtually all known protein structures, providing an invaluable resource for research.
Beyond solving long-standing challenges, it set the stage for future developments like AlphaFold3, which aims to model protein interactions with other molecules. These breakthroughs underscore the impact of deep learning on structural biology.
Transformations in Protein Design
While protein structure prediction has made remarkable progress, protein design has also been transformed by AI. Traditionally, computational protein design relied on energy-function-based methods, such as Baker’s RosettaDesign. This approach optimized amino acid sequences for predefined backbone structures by iteratively minimizing energy functions, achieving milestones like de novo-designed proteins, enzymes, and assemblies.
However, these methods faced limitations in computational efficiency and the generation of initial backbone structures.
Generative deep learning has addressed these challenges, offering new capabilities for protein design. Denoising diffusion probabilistic models can now generate physically viable backbone structures without relying on existing templates, significantly expanding design flexibility. Meanwhile, graph neural networks have improved sequence optimization, outperforming traditional methods.
These advances make protein design more efficient and accessible, with applications spanning drug discovery, synthetic biology, and beyond.
The Future of AI in Protein Science
AI and deep learning are redefining protein science, as demonstrated by the Nobel-recognized breakthroughs in structure prediction and design. AlphaFold2 and generative models have overcome key challenges, enabling unprecedented accuracy and efficiency.
Looking ahead, these advancements will continue to drive innovation in life sciences, from dynamic protein modeling to biotechnology applications.
Journal Reference
Liu, H., Chen, Q., & Liu, Y. (2025). Artificial intelligence is transforming the study of proteins: Structures and beyond. HLife. DOI:10.1016/j.hlife.2025.01.002 https://www.sciencedirect.com/science/article/pii/S2949928325000021?dgcid=api_sd_search-api-endpoint
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.