Classification of B and Y Ions in Peptide MS/MS Spectra Based on Machine Learning
peer-reviewed · Journal of Computer and Communications · 2023
| Date | 2023-03-30 |
| Type | peer-reviewed |
| Venue | Journal of Computer and Communications |
| Publisher | Scientific Research Publishing, Inc. |
| Contribution | adjacent |
| DOI | 10.4236/jcc.2023.113008 |
| Citations (OpenAlex) | 2 |
Abstract
In proteomics, b and y ions serve as the backbone ions for peptide sequencing in tandem mass spectrometry. Leveraging the existing ion recognition and separation methods, this article proposes a novel ion classification approach that combines machine learning with graph theory. By incorporating graph features, the method achieves higher accuracy and efficiency in ion type recognition, with the graph features playing a critical role in the classification process. Specifically, the method achieves a recall rate of nearly 90% for b and y ions, demonstrating its effectiveness in pre-processing de novo sequencing and improving its accuracy. The proposed method represents advancement in ion classification and has the potential to improve the accuracy and efficiency of de novo sequencing.
Methods and tools
- b/y ion classification with graph features: Combines machine learning with spectrum graph features to classify b and y ions, as a preprocessing step for de novo sequencing.