AI-driven framework to map the brain metabolome in three dimensions

Area of Science:

• Spatial biology
• Metabolomics
• Biotechnology

Background:

• High-resolution spatial imaging is crucial for understanding biological systems.
• Spatial metabolomics analyzes metabolic heterogeneity in tissues.
• Current methods primarily focus on 2D spatial data.

Purpose of the Study:

• To introduce MetaVision3D, a novel pipeline for creating 3D spatial metabolomics data.
• To enable the transformation of serial 2D MALDI mass spectrometry imaging sections into a high-resolution 3D spatial metabolome.
• To provide a comprehensive 3D model of metabolites in host tissues.

Main Methods:

• Utilizing computer vision and artificial intelligence for image workflow.
• Employing advanced algorithms for image registration, normalization, and interpolation.
• Integrating serial 2D tissue sections to generate a 3D model.

Main Results:

• Successfully generated a 3D mouse brain metabolome atlas.
• Created an interactive online database and web server for the atlas.
• Demonstrated the capability to model diverse metabolites across tissues at submesoscale.

Conclusions:

• MetaVision3D enables the extension of spatial metabolomics into three dimensions.
• The generated 3D brain atlas advances research in brain metabolism.
• This technology has broad applications in understanding complex biological systems.