BAMPDA: a Matrix Refactoring Framework with Heterogeneous Inference for Potential PTM-disease Association...

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Post-translational modifications (PTMs) are critical protein alterations influencing human health.
• Extensive biomedical data now links PTMs to various diseases, creating opportunities for predictive modeling.
• Existing methods for predicting PTM-disease associations require enhancement.

Purpose of the Study:

• To develop BAMPDA, a novel matrix refactoring framework for improved inference of PTM-disease associations.
• To integrate multi-source data, including protein and disease characteristics, for robust association prediction.
• To validate the predictive power of BAMPDA against existing methods and in real-world disease scenarios.

Main Methods:

• BAMPDA utilizes a matrix refactoring approach with heterogeneous inference.
• The framework integrates multi-source data on protein and disease characteristics.
• Two successive matrix-based algorithms are employed for association prediction, validated by 5-fold cross-validation.

Main Results:

• BAMPDA significantly outperforms five recent baseline methods in predicting PTM-disease associations.
• The framework demonstrated robust predictive capabilities when applied to four representative diseases.
• PTMs were validated as key mediators in the studied disease relationships.

Conclusions:

• BAMPDA offers a powerful and accurate computational tool for PTM-disease association inference.
• The framework's ability to integrate diverse data sources enhances its practical applicability.
• This work highlights the significant role of PTMs in disease pathogenesis and provides a method for their investigation.