TissueProf: An ImageJ/Fiji Plugin for Tissue Profiling Based on Fluorescent Signals

Fluorescence immunohistochemistry to detect multiple molecules of interest (e.g., proteins and RNA) has been an essential experimental method used to analyse cell populations in tissues. There are two challenges in the image analysis of tissues due to the high density of cells and the higher background of signals that originate from extracellular spaces such as extracellular matrix. These are cell identification and analysis of marker coexpression. Although some programmes are available for the analysis of microscopy images, tools that support automated, yet flexible, image analysis are needed to reduce the workload of researchers. In this study, we have developed a user-friendly ImageJ/Fiji plugin that provides a semiautomated image analysis pipeline with a flexibility to reflect inputs from users. The plugin consists of three steps: segmentation of cells expressing each molecule, manual correction of cell segmentation if needed and molecule coexpression analysis. The output of the pipeline comprises Excel files containing the number of cells which express each molecule and/or combination of molecules and their signal intensities. It does so by automatizing the identification of region-of-interests (ROI) based on fluorescent signals and the process of counting cells expressing various combinations of these molecules in each zone the user is interested in. The automatization of localization of fluorescent signals relies on available deep learning networks and the analysis of coexpression from the ROIs is based on spatial analysis of ROIs. This plugin mitigates the workload and time-consumption of the analysis of multichannel microscopy images, which are widely used in neuroscience.