A recently developed AI model is providing new insights into protein localization including the effects of disease-associated mutations. Understanding where proteins localize or what it co-localizes with provides critical information to understand its role in healthy and diseased cells.
The model, published in Science and dubbed ProtGPS, is capable of predicting which of the 12 known compartments a protein will localize to as well as if a disease associated mutation will change that localization. Additionally, using a generative algorithm developed by the team, researchers can design novel proteins that will localize with specific compartments.
"My hope is that this is a first step towards a powerful platform that enables people studying proteins to do their research," said Whitehead Institute Member Richard Young, "and that it helps us understand how humans develop into the complex organisms that they are, how mutations disrupt those natural processes, and how to generate therapeutic hypotheses and design drugs to treat dysfunction in a cell."
Being able to develop functional proteins with the method could improve future development of novel therapies. The researchers anticipate the model will be used to further understand localization and could be expanded to include more compartments to test additional treatment hypotheses.
"Now that we know that this protein code for localization exists, and that machine learning models can make sense of that code and even create functional proteins using its logic, that opens up the door for so many potential studies and applications," Henry Kilgore, a postdoc researcher at Massachusetts Institute of Technology (MIT).