A New Tool to Decrease Interobserver Variability in Biomarker Annotation in Solid Tumor Tissue for Spatial...

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Spatial transcriptomics and immunofluorescence have different resolutions, complicating data integration.
• Accurate identification of DNA damage regions is crucial for understanding cellular responses in irradiated tissues.
• Manual annotation of immunofluorescence data in spatial transcriptomics exhibits significant interobserver variability.

Purpose of the Study:

• To develop a reproducible method for integrating spatial transcriptomics and immunofluorescence data.
• To overcome interobserver variability in annotating DNA damage markers like γH2AX in spatial transcriptomic spots.
• To enable accurate comparison of gene expression between DNA-damaged and undamaged regions.

Main Methods:

• Coupling spatial transcriptomics of irradiated glioblastoma with immunofluorescence for γH2AX.
• Developing a MATLAB tool for spot-wise image analysis and annotation based on intensity thresholds and cell counts.
• Comparing gene expression profiles in γH2AX-positive and negative regions.

Main Results:

• Significant interobserver variability (Kappa = 0.345) was observed in manual annotation of γH2AX positivity.
• The developed MATLAB tool achieved reproducible annotation of spots, even in regions with high manual variability.
• The tool facilitated consistent identification of genes associated with DNA repair pathways.

Conclusions:

• A novel MATLAB tool significantly improves the reproducibility of integrating spatial transcriptomics and immunofluorescence data.
• This tool addresses the challenge of interobserver variability in manual annotation, leading to more reliable downstream analyses.
• The developed method enhances the study of DNA damage responses and associated gene expression in glioblastoma.