Accurate prediction of protein beta-aggregation with generalized statistical potentials

Area of Science:

• Biochemistry
• Computational Biology
• Proteomics

Background:

• Protein beta-aggregation is crucial in diseases and physiological processes but its mechanisms remain poorly understood.
• Identifying aggregation-prone regions is challenging, with current methods often insufficient for practical applications.

Purpose of the Study:

• To develop and present AgMata, an unsupervised tool for identifying protein aggregation-prone regions from amino acid sequences.
• To validate AgMata's performance against existing state-of-the-art methods.

Main Methods:

• AgMata utilizes a generalized definition of statistical potentials incorporating biophysical information.
• The tool operates in an unsupervised manner, analyzing amino acid sequences.

Main Results:

• AgMata demonstrated superior performance compared to state-of-the-art methods on two benchmark datasets.
• Applied to human ataxin-3, AgMata identified aggregation-prone residues with similar structural environments.
• The tool successfully predicted outcomes of in vitro mutagenesis experiments, including mutations altering aggregation propensity.

Conclusions:

• AgMata offers an effective solution for identifying protein aggregation-prone regions.
• The tool's ability to predict mutation effects provides valuable insights into protein aggregation and related diseases.