Fig. 3: Performance of spatial clustering of different methods in SMI Lung cancer. | Nature Communications

Fig. 3: Performance of spatial clustering of different methods in SMI Lung cancer.

From: Unveiling fine-scale spatial structures and amplifying gene expression signals in ultra-large ST slices with HERGAST

Fig. 3

a Spatial regions of ground truth provided by NanoString CosMx. b Performance comparison of all methods using ARI, NMI, FMI, and HC. Source data are provided as a Source Data file. c Visualization of spatial clustering results of top 3 methods in ARI performance. Numbers in title are the corresponding ARI score. Full visualization results can be seen in Supplementary Fig. 10. d d-1 shows the number of clusters of different methods from Leiden clustering with different resolution parameter in simulated data. Two dashed lines indicate the viable boundary (3 and 20). The black line shows number of methods in the viable interval. d-2 shows the performance of different methods (ARI) with different resolution parameters in simulated data. The x-axis is shared between two panels. e Same as a but in real-world data. Source data are provided as a Source Data file.

Back to article page