Fig. 5: Spatial visualization of different cell types from snRNA-seq data.

A, B H&E staining images of sections showing the distinct pathological features of ST1 and ST2 samples. C, D Spatial spots are separated into different regions with distinct pathological features through unbiased clustering in ST1 and ST2 samples. E, F Heatmaps showing the mean signature scores of 10 cell types in each divided region. G, H Signature scores of the E_C3 cluster in each spot on ST sections from ST1 and ST2 samples. I, J Scores of the pEMT signature in each spatial spot on ST sections from ST1 and ST2 samples. K, L The correlations for signature score of pEMT and E_C3 cluster in each spatial spot from ST1 and ST2 samples, respectively. The error band shows the 95% confidence interval. M, N Signature scores of the E_C4 cluster in each spot on ST sections from ST1 and ST2 samples. O–P Signature scores of the TAM_GPNMB cluster in each spot on ST sections from ST1 and ST2 samples. Q, R The correlations for the signature scores of the TAM_GPNMB and E_C4 clusters in each spatial spot from the ST1 and ST2 samples.