rMSIKeyIon

rMSIKeyIon filters and prioritizes significant ions in untargeted MALDI-MS imaging data to enable robust comparison of regions of interest (ROIs) in metabolomics studies.

Key Features:

• R implementation: Provided as an R package for computational analysis of MALDI-MS imaging datasets.
• Ion Filtering Algorithm: Integrates percentage of undetected pixels, the Mann-Whitney U test, and fold-change to identify ions differentially expressed between ROIs within and across tissue sections.
• Handling Undetected Ions: Discards null or undetected peaks based on pixel-wise detection rates prior to statistical analysis to reduce bias from missing signals.
• Unsupervised Segmentation (k-means): Uses k-means clustering for unsupervised delineation of ROIs within tissue sections.
• Statistical Criteria: Combines nonparametric testing (Mann-Whitney U), fold-change assessment, and detection-rate thresholds to prioritize significant ions.

Scientific Applications:

• Untargeted metabolomics: Identification and prioritization of metabolites in untargeted MALDI-MS imaging studies.
• Comparative tissue-region analysis: Detection of ions differentially expressed between case versus control or between defined ROIs across tissue sections.
• Mouse brain and liver studies: Application to complex biological samples such as mouse brain tissue and liver sections exposed to thirdhand smoke.
• Lipid discovery: Detection of key lipid compounds, including triacylglycerols and phosphatidylcholines, implicated in experimental conditions.

Methodology:

Segmentation of tissue sections using k-means clustering to define ROIs; application of an ion filtering algorithm that uses pixel-wise detection rates (percentage of undetected pixels), the Mann-Whitney U test, and fold-change calculations; exclusion of undetected signals prior to statistical testing.