Interactive clustering analysis of MELC image data (single-cell- & Mean Fluorescence Intensity-based)

Registered and standardized 2D fluorescence images generated by MELC and available as Related Identifiers in Zenodo were first segmented. Segmentation was performed in a two-step process, a signal-classification step using Ilastik 1.3.2, followed by an object-recognition step using CellProfiler 3.1.8. Thereby, nuclei and cellular object masks were generated and superimposed on the individual fluorescence images acquired for each marker, in order to extract single-cell information, i.e. mean fluorescent intensity (MFI) of each marker per segmented cell. Single-cell, MFIs of 19 immune-relevant markers for 8 human kidney samples are available here as excel tables (4 tumor and 4 patient-matched peri-tumor samples, named _singlecell_MFI_InputOrange). All data sets were analyzed in Orange 3.26.0 21 using several algorithms for dimensionality reduction, which use the list of MFI values as single-cell features. Data was transformed using the hyperbolic arcsine function with a scale argument of 0.2 previous to cluster analysis. K-means clustering of the kidney single-cell data was performed for all segmented cells from eight kidney data-sets pooled together (4 tumor and 4 patient-matched peri-tumor samples). T-distributed Stochastic Neighbor Embedding (t-SNE) of CD69+CD103+ tissue-resident (TR) cells was performed with 8 principal components, a perplexity of 30 and 1000 iterations. The work-flow used in Orange for the analysis is also available here as .ows file named MELC_Kidney_singlecell_MFIs_Orange_clustering_analysis