An iterative algorithm to quantify factors influencing peptide fragmentation during tandem mass spectrometry

Area of Science:

• Proteomics
• Computational Biology
• Analytical Chemistry

Background:

• Accurate theoretical spectrum prediction is crucial for protein identification using tandem mass spectrometry.
• Current database searching methods rely on simple statistical models, often leading to significant deviations from experimental spectra for certain peptides.

Purpose of the Study:

• To develop an improved prediction model for peptide fragmentation in mass spectrometry.
• To quantify factors influencing peptide fragmentation using a non-linear programming approach.

Main Methods:

• Utilized a non-linear programming model to quantify peptide fragmentation factors.
• Developed an iterative algorithm to solve the resulting optimization problem.
• Trained and tested the model on experimental mass spectrometry data.

Main Results:

• The model's predictions showed good agreement with established peptide fragmentation principles, including central cleavage tendencies and proline's N-terminal cleavage preference.
• Validation on a testing set of 941 spectra demonstrated that the method generates reasonable theoretical spectrum predictions compared to experimental data.

Conclusions:

• The proposed non-linear programming model offers a significant improvement over existing methods for theoretical spectrum prediction in peptide identification.
• This enhanced prediction capability can benefit both database searching and de novo sequencing approaches in proteomics.
• The findings contribute to more accurate and reliable protein identification through mass spectrometry analysis.