AUTOMATED VISUALIZATION, EXPLORATION AND MATERIAL SEGMENTATION OF ION- MOBILITY MASS SPECTROMETRY IMAGING DATA

Význam tématu

Ion mobility–mass spectrometry imaging (IM-MSI) integrates gas-phase separation based on size, shape and charge with spatial localization of analytes, enabling differentiation of isomeric and isobaric species in complex samples. As IM-MSI datasets grow in dimensionality and volume, manual interpretation becomes impractical. Automated computational tools are therefore crucial to extract meaningful chemical and spatial information efficiently.

Cíle a přehled studie / článku

This work presents enhancements to Waters™ MSI Analyte Browser and MSI Segmentation MicroApps, incorporating unsupervised machine learning methods to automate visualization, exploration and segmentation of IM-MSI data. The approach is demonstrated on a rat brain tissue sample to showcase pixel-level object detection, cluster-based feature grouping and generation of average m/z-drift time profiles for each region.

Použitá metodika a instrumentace

Processed analyte text files exported from Waters HDI™ software or custom csv files containing (x,y) coordinates with intensity values at each peak-picked m/z and drift time pair serve as input. Data analysis proceeds in two main steps:

• Dimensionality reduction via UMAP to project high-dimensional m/z-drift time spectra into a low-dimensional embedding that preserves neighborhood structure.
• Density-based clustering with HDBSCAN to group similar pixels into segments without prior knowledge of cluster number.

Použitá instrumentace

• Waters™ MSI Analyte Browser
• Waters™ MSI Segmentation MicroApps
• Waters HDI™ software (data preprocessing)
• UMAP (dimensionality reduction algorithm)
• HDBSCAN (density-based clustering algorithm)

Hlavní výsledky a diskuse

Automated pixel segmentation of a rat brain section produced spatially coherent clusters correlated with anatomical regions. Each cluster was characterized by an average m/z-drift time spectrum, enabling identification of feature-specific ions. Subsequent clustering of individual analyte images grouped ion images sharing similar spatial distributions across the tissue. These analyses streamline detection of colocalized analytes and enhance interpretation of isomeric or isobaric species in complex tissues.

Přínosy a praktické využití metody

• Significant reduction of manual workload by automating segmentation and feature grouping.
• Objective discovery of spatially distinct chemical signatures for biomarker or drug distribution studies.
• Exportable csv outputs facilitate integration with downstream statistical or bioinformatics pipelines.
• Applicable in QA/QC, pharmaceutical tissue imaging, metabolomics and lipidomics workflows.

Budoucí trendy a možnosti využití

Further developments may include integration of supervised learning for targeted analyte classification, real-time cloud-based processing, multimodal data fusion with microscopy or spatial transcriptomics, and enhanced algorithms to capture rare or low-abundance features in large tissue cohorts.

Závěr

The updated MSI Analyte Browser and Segmentation MicroApps, leveraging UMAP and HDBSCAN, provide an efficient unsupervised framework to visualize, explore and segment ion mobility MSI data. This automation accelerates discovery of spatially resolved chemical information while maintaining flexibility for downstream analysis.

Reference

Trinkle S., Shrestha B., Chapman R. Automated Visualization, Exploration and Material Segmentation of Ion-Mobility Mass Spectrometry Imaging Data. Waters Corporation, 2023.