BANKSY - Spatial Domain Identification

Spatial domains are regions within a tissue that exhibit coherent gene expression patterns and often correspond to distinct anatomical or functional areas. Spatial domain identification analysis can:

• Reveal tissue architecture and micro-environments.

• Enable downstream analyses like differential expression and pathway enrichment.

• Support integration with histology for biological interpretation.

BANKSY is a method for clustering spatial transcriptomic data by augmenting the transcriptomic profile of each cell with an average of the transcriptomes of its spatial neighbors [1].

A known bug currently prevents running BANKSY on multiple spatial samples in the same node. Please use the 'Split by attribute' function before running the task on each individually.

Running BANKSY

The 'BANKSY - Spatial domain Identification' task can be run from any non-normalised node containing spatial data. It is recommended to run the task after filtering low quality cells and low expression genes. The analysis is species-agnostic.

• Select the data node

• Click Classification > BANKSY - Spatial Domain Identification

• Adjust the task parameters as needed. Note particularly the Clustering resolution and Lambda parameter, more information on these can be found in the origina publication [1]

• You can uncheck the Compute spatial biomarkers task if you prefer. If the option is checked (default) the task will compute differentially expressed genes between each of the spatial domains identified, similarly to the 'Compute biomarkers' task found in the 'Statistics' menu.

Figure 1. BANKSY task setup screen

• Configure the Advanced options if needed

Figure 2. BANKSY task advanced options

• Click Finish

Task report and visualisation

If the 'Compute spatial biomarkers' option was checked the task will produce two separate nodes:

Figure 3. BANKSY task output nodes. If the Compute spatial biomarkers option was left unchecked, the task will only produce the Spatial Domains node

Clicking twice on the Spatial biomarkers node will open a table of differentially expressed genes per cluster:

Figure 4. Table of Spatial biomarkers outputted by the BANKSY task

Additionally the task results can be visualised in the Data viewer section of your Analysis, common visualisations may be plotting the spatial data colored by the spatial_domain attribute (note this will only be available in the Spatial domains node), and a table of spatial biomarkers:

Figure 5. Common visualisations for the output of a spatial domains analysis

References

[1] Singhal, Vipul, et al. "BANKSY unifies cell typing and tissue domain segmentation for scalable spatial omics data analysis." Nature genetics 56.3 (2024): 431-441. https://www.nature.com/articles/s41588-024-01664-3