Image analysis pipelines

Download pdf version Pipeline logic Roughly recognize what is a signal coming from cardiomyocytes and use it for mask (Primary object: cells) Identify nuclei in the sarcomere signal masked image (Primary object: CM_nuclei) Merge close-by nuclei as one object (SplitOrMergeObjects: CM_nuclei_merged) Segment sarcomere signal channel into cells based on the merged nuclei (Secondary object: cells_final) Now there are 3 different objects nucleus (Primary object) cytoplasm (Tertiary object) cell (Secondary object) Measure YAP1 signal in nucleus, cytoplasm, and cell Save image for checking cell segmentation: Left: Sarcomeric signal overlayed with CM_merged_nucleicells_final` outlines Right: YAP1 signal overlayed with CM_merged_nuclei and cells_final outlines object numbers Save files with measurements, set that all measurements are exported Reproducible structure of the project source folder contains all the images output_<group-name> folders where output from different pipelines go CP-<Pipeline-ID>_<group-specification> one or more pipelines with clearly indicated what group was which pipeline used on Pipeline Quality control Go throgh the exported overlays and check for segmentation If there are some really badly segmented cells (more then 5% per image) remove them from the measurment files create a filter file: first column for image number, second column for object number to be removed in case of all to be removed write all in second column save and run the filterCP script on the files to remove the files Add QC-passed to all output folders which were checked Check the number of columns cells_final.csv (BJ last column) CM_nuclei_merged.csv (BL last column) Post processing Stack all the cells_final.csv and CM_nuclei_merged.csv csvstack -g *cells*final*.csv > cells_final.csv csvstack -g *CM*merged*.csv > CM_nuclei_merged.csv Select the columns for intensities and ratios in the CM_nuclei_merged.csv and add them to the end of cells_final.csv selectR -c 1,40,47,63 CM_nuclei_merged.csv ...