Fig. 8: Application of pan-cancer single-cell landscape.
a, Schematic overview of the workflow using pan-cancer single-cell landscapes to analyze the ecotypes of bulk RNA-seq datasets and automatic annotation for scRNA-seq datasets. b, UMAPs showing cell-type annotation by original paper (left) and SELINA prediction (right). c, Bar plots comparing the accuracy of reference-pretrained SELINA-predicted cell types against original annotations in the BRCA_GSE176078 and NSCLC_GSE146100 datasets. SELINA was trained on reference datasets from TabulaTIME, BRCA_EMTAB8107 and NSCLC_GSE131907.
