Motivation

Uncovering the spatial biology for various tissue states requires aggregation and comparison of data from many tissue samples. The Explorer uses the metadata provided by the user to aggregate data from multiple samples, and enables quantitative comparisons between groups of samples (i.e., cohorts).

All of the previous analysis steps (segmentation, Cell QC, phenotyping, and Cell Neighborhood analysis) are ways to quantify the raw image data based on cell-level abstraction.

Methods

The Explorer draws data from the table of cells that was generated through segmentation and the additional labels assigned to each cell through Cell QC, phenotyping, and Cell Neighborhood analysis pipelines. The Explorer also takes the metadata that was entered through the Designer to assign additional non-spatially-derived sample/region-level information to the cells. This allows plots for cell-level and sample-level data to be generated.

Suggested Workflow

Generally, any trends that you find through the Explorer should always be confirmed through the Visualizer to ensure that the trends reflect real biology and not acquisition or analysis artifacts. The following are suggestions for steps to take when exploring your data through the Explorer.

1. Ensure that your metadata is relevant and allows reasonable cohorts to be formed for comparisons. Correct any issues with the metadata.
2. Generate plots for Phenotypes, and use the Volcano plot to determine what the most significant trends between sample cohorts are.
3. Generate scatter plots for individual phenotypes to get a better sense of the spread between samples within a cohort. Alternatively, seek out the data points for your specific cell phenotypes of interest.
4. Look at the Visualizer to see if the trends reflect a true biological insight, and see if there are any spatially striking reasons for the observed trends.
5. Repeat steps 2-4 for Cell Interactions and Cell Neighborhoods.

Expected Outcomes

Data and plots can be exported from the Explorer for inclusion in presentations and publications. Weak trends may indicate that more statistical power (i.e., samples) are needed, and can suggest future directions for additional studies.

You may find that the raw image data are the most compelling evidence for an insight. Specific ROIs and viewing parameters can be saved as a Note in a Story through the Visualizer, allowing you to walk others through your insights. Specific views can also be captured and downloaded through the Visualizer for inclusion in publications.