Deep Visual Proteomics defines single-cell identity and heterogeneity

Area of Science:

• Proteomics
• Biotechnology
• Cancer Research

Background:

• Spatial proteomics methods struggle to link imaging with single-cell protein data.
• A gap exists in connecting cellular phenotypes to precise protein abundance at the single-cell level.

Purpose of the Study:

• Introduce Deep Visual Proteomics (DVP) to bridge imaging and single-cell spatial proteomics.
• Link cellular phenotypes to protein abundance while maintaining spatial context.
• Analyze proteomic changes in melanoma progression.

Main Methods:

• Combined AI-driven image analysis of cellular phenotypes with laser microdissection and mass spectrometry.
• Automated single-cell/nucleus isolation for proteomic profiling.
• Applied DVP to cell cultures and archived melanoma tissue.

Main Results:

• Classified distinct cell states and their proteomic profiles.
• Identified spatially resolved proteome changes during melanoma progression.
• Revealed altered mRNA splicing, reduced interferon signaling, and antigen presentation in metastatic melanoma.

Conclusions:

• DVP successfully links cellular phenotypes to spatial proteomic data at single-cell resolution.
• DVP provides insights into cancer progression pathways.
• DVP has significant implications for molecular profiling of clinical samples.