PTMClust

PTMClust refines analysis of post-translational modifications (PTMs) in proteomic datasets using a machine learning algorithm that suppresses noise and clusters peptides with identical underlying modifications into distinct PTM groups.

Key Features:

• Noise Suppression: Reduces data noise to enhance the accuracy of PTM identification.
• Peptide Clustering: Groups peptides sharing the same modification into distinct PTM groups.
• Performance Superiority: Outperforms two standard clustering algorithms on simulated datasets and improves sensitivity and specificity in real-world applications.
• Novel Discovery Facilitation: Reduces false assignments and improves detection coverage, aiding detection of novel PTMs including terminus modifications.

Scientific Applications:

• Large-scale yeast MS/MS proteome profiling: Reveals numerous known and previously uncharacterized PTMs.

Methodology:

The algorithm processes outputs of blind PTM search methods using a machine learning approach to suppress noise and cluster peptides with identical underlying modifications into distinct PTM groups.