Mapping the proteome-wide PTM landscape of molecular aging through explainable deep learning approaches

Post-translational modifications (PTMs) are crucial for protein function and regulation, with their dysregulation increasingly recognized as a hallmark of aging and diseases like Alzheimer’s and Parkinson’s. While mass spectrometry is a powerful tool for studying PTMs, it is inherently biased toward detectable peptides, leaving many PTM sites unexplored. To overcome this limitation, I propose developing a next-generation PTM predictor using advanced, yet explainable, deep learning techniques. This predictor will extend beyond mass spectrometry observations to generate a comprehensive map of potential PTM sites. By fine-tuning the model with aging-specific datasets from C. elegans, we aim to uncover PTM patterns associated with molecular aging. This approach will expand our understanding of PTMs, reveal novel aging-related mechanisms, and provide freely accessible tools and datasets to advance diagnostics and therapeutic target discovery in aging research.

Keywords:Post-translational modifications, Mass spectrometry, Deep learning

Disciplines:Computational biomodelling and machine learning, Proteomics, Development of bioinformatics software, tools and databases, Posttranslational modifications