Fig. 4: SpatialTopic recovers spatial domain architecture from cell-type spatial layout in healthy lung. | Nature Communications

Fig. 4: SpatialTopic recovers spatial domain architecture from cell-type spatial layout in healthy lung.

From: Scalable topic modelling decodes spatial tissue architecture for large-scale multiplexed imaging analysis

Fig. 4

A Spatial distribution of cell type annotations, manual domain annotations, and spatial domains recovered by SpatialTopic and UTAG from the healthy lung tissue samples. B Comparing manual spatial domain to the cell type annotation, SpatialTopic, and UTAG with and without ad hoc relabel, using Adjusted Rand Index across 26 regions of interest (n = 26). The box plots show the median (center line), 25th and 75th percentiles (box bounds), and whiskers extending to the minimum and maximum values excluding outliers. UTAG results are obtained from the original publication. C Heatmap shows per-topic cell type composition for the four topics identified by SpatialTopic. Source data are provided as a Source Data file.

Back to article page