Back

MSI-VISUAL: New visualization methods for mass spectrometry imaging and tools for interactive mapping and exploration of m/z values

Gildenblat, J.; Pahnke, J.

2025-03-19 bioinformatics
10.1101/2025.03.18.643852 bioRxiv
Show abstract

Mass Spectrometry Imaging (MSI) provides complex molecular profiling pivotal for advanced biochemical analysis and diagnostics. We have identified an urgent need for computationally sophisticated, interactive, and automated visualization tools specifically designed for MSI analysis that employ dimensionality reduction (DR) methods to preserve global data structure. We present MSI-VISUAL, an open-source framework that enables rapid and intuitive MSI data exploration, providing panels of truthful visualizations and virtual pathology stains (VPS), with automatic and manual region selection, real-time verification, and statistical comparisons. Integrating established DR techniques, MSI-VISUAL introduces four novel visualization strategies: Saliency (SALO) and Spearman (SPEAR) Optimizations enhance global structure preservation across various distance metrics, addressing significant visualization challenges; TOP3 and Percentile Ratio (PR3D) provide memory-efficient and computationally rapid visualizations for large-scale datasets. In this work, comprehensive benchmarking and biological examples confirm that our DR visualizations surpass existing methods, facilitating the determination and quantification of subtle tissue differences at a single-cell level, revealing novel biological insights, and enabling diagnostic applications.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.