NECTAR: A New Algorithm for Characterizing and Correcting Noise in QToF-Mass Spectrometry Imaging Data

Area of Science:

• Analytical Chemistry
• Biomedical Imaging
• Data Science

Background:

• Mass spectrometry imaging (MSI) generates vast datasets with significant noise.
• Effective noise reduction is crucial for accurate peak identification and data interpretation.
• Existing methods may not adequately address diverse noise types in MSI QToF data.

Purpose of the Study:

• To develop and validate a novel algorithm, NECTAR (NoisE CorrecTion AlgoRithm), for characterizing and correcting noise in MSI QToF data.
• To differentiate and quantify random, chemical, and background noise.
• To improve the reliability of spectral baseline and noise level determination.

Main Methods:

• Statistical characterization of random, chemical, and background noise in MSI QToF data.
• Development of NECTAR algorithm for noise correction.
• Application of NECTAR to MALDI and DESI QToF data from mouse pancreatic tissue.
• Analysis of denoised data using Principal Component Analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE).

Main Results:

• Random noise dominates lower m/z values (~50-400 Da), while chemical noise prevails at higher m/z values (>400 Da).
• NECTAR reduced chemical noise by approximately a factor of 3.
• Noise reduction led to more reliable spectral baselines and noise level estimations.
• Post-processing analysis confirmed effective removal of uninformative peaks and selection of relevant peaks.

Conclusions:

• NECTAR is effective in characterizing and reducing multiple noise types in MSI QToF data.
• The algorithm improves the signal-to-noise ratio, facilitating more accurate data analysis.
• NECTAR enhances the selection of biologically relevant peaks from complex MSI datasets.