Fig. 7
From: Explainable multiview framework for dissecting spatial relationships from highly multiplexed data
Application of MISTy to a spatial transcriptomics dataset. A Schematic of the MISTy pipeline used in Visium 10x slides. Each visium spot profiles the gene expression of up to 10 cells. Pathway activities were estimated with PROGENy and a MISTy model was built to predict them using two spatially contextualized views. B Changes in R2 observed in each predicted pathway after using the multiview model, reflecting the importance of the spatial context (upper panel). Contribution of each view to the prediction of the pathway activities in the meta-model. The stacked barplot represents normalized values of the fusion coefficients of the respective views for each pathway (lower panel). C Variable importances for the intraview. D Intrinsic associations of pathway activity scores of NFkB and TNFa, and p53 and MAPK. Spatial distribution of pathway activities from the first section. Circled areas exemplify niches where coordinated activities were observed. Scatterplots show the within-spot relationship between each pair of pathways. E Variable importances for the pathway paraview. F Spatial distribution of estrogen pathway activities and scaled gene expression of SCT1 (top predictor of Estrogen in the ligand paraview) and TNF