Interactive Visual Exploration of 3D Mass Spectrometry Imaging Data Using Hierarchical Stochastic Neighbor Embedding...

Area of Science:

• Biomedical Imaging
• Computational Biology
• Analytical Chemistry

Background:

• Technological advancements have increased interest in 3D mass spectrometry imaging (MSI).
• Analyzing large 3D MSI datasets is computationally challenging due to size and complexity.
• Identifying informative molecular patterns in 3D MSI data requires efficient analytical tools.

Purpose of the Study:

• To demonstrate the utility of hierarchical stochastic neighbor embedding (HSNE) for analyzing large 3D MSI datasets.
• To showcase HSNE's capability in identifying regions of interest and characterizing molecules within complex MSI data.
• To evaluate HSNE's robustness against common measurement artifacts in 3D MSI.

Main Methods:

• Application of hierarchical stochastic neighbor embedding (HSNE), a nonlinear dimensionality reduction technique.
• Analysis of multiple publicly available 3D MSI datasets from diverse biological systems and ionization techniques.
• Benchmarking HSNE performance on large, high-dimensionality datasets at full spectral and spatial resolution.

Main Results:

• HSNE successfully analyzed large 3D MSI datasets with manageable computational complexity.
• The technique rapidly identified regions of interest and associated molecular ions.
• HSNE demonstrated robustness against measurement artifacts, including batch effects and noise.

Conclusions:

• HSNE is a powerful and efficient tool for the analysis and visualization of large 3D MSI data.
• HSNE facilitates the discovery of biologically relevant molecular patterns in complex imaging datasets.
• The method offers a robust approach to MSI data analysis, improving reliability and interpretability.