Combination of continuous wavelet transform and genetic algorithm-based Otsu for efficient mass spectrometry peak...

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Signal Processing

Background:

• Mass spectrometry (MS) data analysis is challenged by noise and baseline fluctuations, complicating spectral peak detection.
• Accurate identification of weak and overlapping peaks is crucial for reliable MS data interpretation.

Purpose of the Study:

• To develop an efficient peak detection algorithm for mass spectrometry data.
• To improve the detection of weak and overlapping spectral peaks.
• To enhance the accuracy of characteristic peak identification in MS.

Main Methods:

• Proposed a novel algorithm combining Continuous Wavelet Transform (CWT) with genetic algorithm-based threshold segmentation (WSTGA).
• Utilized Mexican Hat wavelet and identified ridges/valleys in the 2D wavelet coefficient matrix.
• Implemented an improved Otsu method with a genetic algorithm for optimal threshold segmentation.

Main Results:

• The WSTGA algorithm demonstrated superior peak detection performance compared to Multi-Scale Peak Detection (MSPD) and CWT-IS methods.
• WSTGA effectively identified more weak and overlapping peaks in MALDI-TOF spectra.
• The algorithm maintained a lower false peak detection rate, validating its efficacy.

Conclusions:

• The WSTGA algorithm offers significant advantages for characteristic peak identification in mass spectrometry.
• This method enhances the reliability and accuracy of MS data analysis, particularly for complex spectra.
• WSTGA provides a robust solution for overcoming challenges posed by noise and overlapping peaks in MS.