Fig. 1

Schematic overview of the procedure, from tissue collection to final visualization of the data analysis results. a Thin tissue sections are placed on the ST/Visium array. Barcoded capture-probes store spatial information which is added to the captured transcript prior to sequencing. Imaging data is obtained by microscopy of stained tissue sections. The sequencing data is used as input to demultiplexing and transcript quantification pipelines. The count data together with the image data are used as inputs to STUtility. Image processing (including masking and alignment), and all further data analysis (e.g. dimensionality reduction, factor analysis, identification of spatially correlated genes) is conducted within R. b Spatial autocorrelation. Two vectors are defined: (i) the original expression vector for each gene and each capture-spot and (ii) the Spatial lag expression vector, which for each capture-spot and gene takes the summed expression of up to six neighbors. Spatial autocorrelation is defined as the Pearson correlation between the two vectors (i) and (ii) with the rationale that genes with spatial structure will display a higher correlation to their neighbors. c The aligned images can be visualized in a turntable 3D model within R in which a combination of features can be visualized. Here, the NMF factors of the tissue are shown in the HSV color scale