Intrinsically Disordered Proteins: Design & Properties
- Advances in artificial intelligence have revolutionized protein design in synthetic and structural biology.
- These proteins, known as intrinsically disordered proteins (IDPs), do not settle into a fixed shape.
- Paulson School of Engineering and Applied Sciences (SEAS) and Northwestern University have developed a new machine learning method capable of designing IDPs with tailored properties.
“`html
New Machine Learning Method Designs ‘Disordered’ Proteins for Biomedical Advances
The Challenge of Intrinsically Disordered proteins
Advances in artificial intelligence have revolutionized protein design in synthetic and structural biology. Computers can now accurately predict the 3D structure of proteins – from antibodies to blood clotting agents – based on their amino acid sequence.However, approximately 30% of proteins expressed by the human genome remain a significant challenge for even the most powerful AI tools, including the Nobel-winning AlphaFold.
These proteins, known as intrinsically disordered proteins (IDPs), do not settle into a fixed shape. Rather, they constantly shift, making them arduous to predict and design. Despite their instability, IDPs are crucial for numerous biological functions, including molecular cross-linking, sensing, and signaling.
Illustration depicting the dynamic, shifting structure of an intrinsically disordered protein compared to a well-defined, folded protein. (image credit: Harvard John A.Paulson School of Engineering and Applied Sciences)
Breakthrough at Harvard and Northwestern
researchers at the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) and Northwestern University have developed a new machine learning method capable of designing IDPs with tailored properties. This work promises to deepen our understanding of these enigmatic biomolecules and possibly unlock new insights into the origins and treatments of diseases.
The research, published in Nature Computational Science, was co-led by Ryan Krueger, a graduate student at SEAS, and Krishna Shrinivas, formerly an NSF-Simons QuantBio Fellow and now an assistant professor at Northwestern University, in collaboration with Michael Brenner, the Catalyst Professor of Applied Mathematics and Applied physics at SEAS.
Why IDPs are Difficult to Model
Shrinivas explained his interest in IDPs stems from their resistance to current AI-based protein prediction and design methods, such as Google DeepMind’s AlphaFold. Despite this challenge, IDPs play vital roles in fundamental biological processes.Mutations in these proteins have been linked to diseases like cancer and neurodegeneration.
Alpha-synuclein, a disordered protein long implicated in Parkinson’s disease and other neurodegenerative conditions, serves as a prime example. To design IDPs for synthetic or therapeutic applications, Shrinivas stated, “we needed to either come up with better AI models, or, we needed to come up with a way to actually take…