Standardized metrics for assessment and reproducibility of imaging-based spatial transcriptomics datasets

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Spatial transcriptomics technologies lack standardized evaluation metrics.
• In situ hybridization (ISH) platforms vary in chemistry and performance.
• Cross-site comparability of spatial omics data is a significant challenge.

Purpose of the Study:

• To develop standardized metrics and operating procedures for evaluating imaging-based spatial transcriptomics.
• To create a comprehensive dataset and open-source software for data analysis and comparison.
• To establish best practices for integrating multi-omics data in spatial studies.

Main Methods:

• Generated the Spatial Touchstone (ST) dataset across six tissue types and multiple global sites.
• Analyzed data using Xenium and CosMx platforms, assessing key performance metrics.
• Developed the open-source SpatialQM software for standardized data evaluation and cell annotation.

Main Results:

• Established standardized operating procedures (STSOPs) for spatial transcriptomics.
• The ST dataset includes 254 spatial profiles with diverse tissue types and platforms.
• SpatialQM enables reproducible evaluation of reproducibility, sensitivity, dynamic range, SNR, FDR, and cell annotation congruence.

Conclusions:

• Standardized metrics and tools are crucial for reliable spatial transcriptomics data.
• The ST dataset and SpatialQM software facilitate cross-platform and cross-site data comparison.
• Best practices are defined for integrating multi-omics data into spatial transcriptomics and proteomics.