rMSIfragment: improving MALDI-MSI lipidomics through automated in-source fragment annotation

Area of Science:

• Biochemistry
• Analytical Chemistry
• Computational Biology

Background:

• Matrix-Assisted Laser Desorption Ionization Mass Spectrometry Imaging (MALDI-MSI) is crucial for spatial chemical analysis of tissues.
• Lipids play vital roles in biological processes, making their identification in MALDI-MSI highly significant.
• Current lipid identification in MALDI-MSI is challenging due to limited separation and untargeted tandem mass spectrometry.

Purpose of the Study:

• To develop an open-source R package, rMSIfragment, for confident lipid annotation in MALDI-MSI.
• To leverage in-source fragmentation data for improved lipid structural inference and identification.
• To enhance the accuracy and reliability of lipid profiling in complex biological samples.

Main Methods:

• Developed rMSIfragment, an R package utilizing known adducts and fragmentation pathways for lipid annotation.
• Implemented a novel scoring system for ranking lipid annotations, validated using HPLC-MS and Target-Decoy approaches.
• Applied the package across diverse MALDI-MSI sample types and experimental setups.

Main Results:

• The rMSIfragment package successfully annotates lipids in MALDI-MSI data by exploiting in-source fragments.
• The novel scoring system achieved an area under the curve of 0.7 in ROC analyses, indicating high annotation performance.
• Demonstrated that incorporating in-source fragments significantly reduces incorrect lipid annotations compared to workflows that ignore them.

Conclusions:

• rMSIfragment offers a robust solution for increasing confidence in lipid identification within MALDI-MSI studies.
• Considering in-source fragmentation data is essential for accurate lipid annotation and minimizing false positives.
• The package is broadly applicable to various MALDI-MSI experiments, improving the overall quality of lipidomic analyses.