Single-cell spatial multi-omics molecular pathology enabled by SuperFocus

Area of Science:

• Computational Biology
• Pathology
• Genomics

Background:

• Histopathology and molecular pathology are currently distinct diagnostic fields.
• Integrating spatial context with molecular data at single-cell resolution is a key goal for advanced pathology.

Purpose of the Study:

• To present SuperFocus, a computational platform for integrating histopathology with single-cell spatial multi-omics.
• To enable cell-resolved molecular analyses across whole tissue sections.

Main Methods:

• SuperFocus is a modality-agnostic platform using constrained cascading imputation and quality control.
• It generates spatial multi-omics from spot-based measurements on adjacent tissue sections.
• No external reference data is required.

Main Results:

• SuperFocus significantly improves accuracy metrics (28-73%) on spatial transcriptomics benchmarks.
• It enables cell-resolved analyses in various datasets, including MALT lymphoma, hippocampus, MASH, and Parkinsonian mouse brain.
• Demonstrates integration of transcriptomic and metabolomic data.

Conclusions:

• SuperFocus provides scalable whole-slide, single-cell spatial multi-omics integrated with histopathology.
• It bridges the gap between histology and genome-scale molecular profiling for next-generation molecular pathology.