TRACERx-PHLEX

TRACERx-PHLEX performs deep learning-based cell segmentation, automated cell-type annotation, and spatial analysis of multiplexed imaging data as three independent but interoperable modules to generate single-cell identities, cell densities, marker positivity calls, and spatial metrics.

Key Features:

• Deep learning-based cell segmentation (deep-imcyto): Performs cell segmentation on multiplexed imaging data using deep learning approaches.
• Automated cell-type annotation (TYPEx): Automatically annotates cell types from multiplexed imaging data without requiring manual assessment or pathology expertise.
• Interpretable spatial analysis (Spatial-PHLEX): Generates spatial metrics including cellular barrier scores, cell densities within tissue compartments, and marker positivity calls with summary graphs and spatial visualisations.
• Modular architecture: Comprises three independent but interoperable modules that together provide an end-to-end analytical solution.
• Containerised Nextflow pipeline: Implemented as an automated pipeline that does not require programming skills for operation.

Scientific Applications:

• Multiplexed imaging analysis: Processes highly multiplexed imaging data from technologies including imaging mass cytometry (IMC) and Co-detection by indexing (CODEX).
• Tumor microenvironment characterization: Developed and applied in the TRACERx study for deep cell phenotyping and spatial analysis of cancer tissues.
• Cross-platform validation: Validated using published CODEX, IMC and orthogonal datasets, and benchmarked against state-of-the-art approaches across different tissue types, tissue fixation conditions, image sizes and antibody panels.
• Clinical insight generation: Provides clinically relevant insights through comprehensive single-cell and spatial analysis outputs.

Methodology:

TRACERx-PHLEX was developed using imaging mass cytometry (IMC) in the TRACERx study, validated using published Co-detection by indexing (CODEX), IMC and orthogonal data, and benchmarked against state-of-the-art approaches.