A new artificial intelligence (AI) tool developed by scientists at the Institute of Bioengineering of Catalonia and the Center for Genomic Regulation is transforming how we understand protein aggregation, a process linked to Alzheimer’s and over 50 other diseases. Named CANYA, this AI tool not only predicts when and why proteins clump together but also explains the chemical patterns that drive or prevent this harmful folding.
Protein aggregation, also known as amyloid aggregation, is a major health risk that disrupts normal cell function. The new tool, built on an extensive dataset of protein aggregation, offers insights into the molecular mechanisms that cause this clumping. The potential applications are vast, from advancing neurodegenerative disease research to improving drug production, as many therapeutic proteins can malfunction due to aggregation.
CANYA is designed to be “explainable,” offering transparency in decision-making, which is rare in many AI models. This makes CANYA more understandable and valuable for researchers. The tool was trained using a large-scale experiment that generated over 100,000 random protein fragments. These fragments were tested for aggregation in yeast cells, creating a comprehensive catalog of protein variations that cause aggregation.
Through its innovative approach, CANYA has not only identified known patterns but also uncovered new rules about protein aggregation. It showed, for instance, that certain hydrophobic regions in amino acid chains are more likely to trigger aggregation, and that charged amino acids can actually promote aggregation in certain contexts.
Looking ahead, the team aims to refine CANYA to predict the speed of aggregation, not just its likelihood. This development could play a critical role in understanding how amyloid formation contributes to neurodegenerative diseases like Alzheimer’s. Ultimately, this AI tool could accelerate biomedical research, providing researchers with deeper insights into diseases and helping design better therapeutic proteins with fewer aggregation-related issues.
Article written by Salud a Diario Team
03/05/2025
Source:
SaludADiario.es