Fig. 7: SpotClean processed data leads to improved identification of malignant spots and increased specificity in clustering in human colorectal cancer sample human_colorectal.
From: SpotClean adjusts for spot swapping in spatial transcriptomics data
a H&E image (left) and spots annotated as tumor vs. non-tumor via a pathologist’s visual inspection (right). Red boxes highlight the spots belonging to SpotClean’s tumor cluster 1 (panel c). b Malignant spot composition as estimated via SPOTlight22 is shown for the raw (left) and SpotClean decontaminated data (right). SpotClean results in higher malignancy scores in tumor regions, and lower in normal regions. c BayesSpace25 clustering for the raw data (top) and SpotClean decontaminated data (bottom). The SpotClean decontaminated data identifies a cluster not identified in previous work (SpotClean tumor_1, red boxes). The SpotClean tumor_1 spots are distinct on the H&E image (red boxes in a) and likely contain tumor-infiltrating immune cells as evidenced by high expression in the immunoglobulin markers shown in d. Scale bars in a, b, d are 1 mm.
