RESTORE

RESTORE normalizes staining intensity in multiplexed imaging datasets by identifying marker-specific negative control cells to infer background signal and remove inter-sample unwanted variation while preserving biological variation.

Key Features:

• Robust normalization: Mitigates unwanted inter-sample variation in staining intensity and background fluorescence.
• Automated negative-control identification: Identifies negative control cells specific to each marker within the same tissue sample.
• Background inference: Infers background signal levels from marker-specific negative controls to estimate non-biological signal.
• Per-marker intensity normalization: Normalizes intensity profiles per marker across samples using inferred background levels.
• Preservation of biological integrity: Removes technical artefacts while preserving biological variation in cell states and intercellular signals.
• Validation on real datasets: Demonstrated robustness using tissue microarray data and longitudinal biopsies.

Scientific Applications:

• Clinical quantitative integration: Enables precise quantitative integration of multiplexed imaging data for clinical studies.
• Tissue microarray analysis: Improves comparability of staining across tissue microarray experiments.
• Longitudinal biopsy comparison: Facilitates longitudinal comparison of biopsies by removing technical variation across time points.
• Single-cell and intercellular dynamics: Supports analysis of individual cell states and intercellular interactions within tissue contexts.

Methodology:

Automated identification of marker-specific negative control cells within each tissue sample; analysis of these controls to infer background signal levels; normalization of per-marker intensity profiles across samples using the inferred backgrounds.