Calcium Signal Propagation speed analysis pipeline

Used in the paper "Multi-chamber cardioids unravel human heart development and cardiac defects" to analyse the speed of calcium signal propagation in cardioids.

Installation

Have the uv python package manager installed. Execute the command uv sync in the calsiprovis subdirectory. All local dependencies (present in the directories calsipro, runnel, rapids and hoernchen) are build automatically.

Running

CLI usage

• Choose a writeable path that will hold a JSON file identifying all images that belong to this experiment, referred to as $EXPERIMENT in the next steps.
• Add GCamp files via calsiprovis add-gcamp $EXPERIMENT $PATH_GCAMP_FILE, with $PATH_GCAMP_FILE being a path to a gcamp (czi) file
• Use calsiprovis add-brightfield $EXPERIMENT $PATH_BF_FILE to add brightfield (czi) files
• Use calsiprovis add-fluro $EXPERIMENT $FLURO_NAME $PATH_FLURO_FILE to add flurophore files. $FLURO_NAME is the name of the flurophor used, e.g. bfp
• Use calsiprovis panalyse --cores $CORES $EXPERIMENT to analyse all wells of the experiment on $CORES many cores in parallel

GUI Usage

Run the following command in the calsiprovis subdirectory:

uv run runnel ../rapids/dist/rapids-0.5.0-py3-none-any.whl dist/calsiprovis-0.18.0-py3-none-any.whl ../calsipro/dist/calsipro-0.16.0-py3-none-any.whl --foreground=calsiprovis.frontend:run --pyodide-dir=../runnel/src/runnel/pyodide/ --foreground-arg=$REFS --background=rapids.socket.backend:run

where $REFS is a writeable path that will hold a JSON containing references to multiple experiment files.

In the panel to add an experiment fill the top line with an experimental name, the bottom line with a path to the experiment file (corresponding to $EXPERIMENT in the CLI usage). Click the button Add Experiment to add the experiment to the $REFS file. Below the form to add an experiment is a list of buttons with all experiments. Click the button of an experiment to load it. This adds three new forms and a list of experiment read outs. The first form adds gcamp files to the experiment, fill in the path to the gcamp file and click the Add Gcamp button. The second form adds fluro files to the experiment, fill in the path to the czi file, the flurophore type (e.g. bfp) and click the Add Fluro button. The third form adds a brightfield file to the experiment, fill in the path to the czi file and click on Add Brightfield.

Below the forms are 6 buttons plus one button per added fluropore type and loads intermediate results for quality control. Before loading intermediate results, run the analysis with the analyse or panalyse CLI commands. Clicking the max button shows the maximal intensity projection of the gcamp files. Clicking the mask button displays the mask (identified organoid). Clicking the video button shows the gcamp timecourse as a video. Clicking the peaks button shows the peaks identified with the default peak finding algorithm. Clicking the peaks simple button shows the peaks identified with the simple algorithm. Clicking the dist button shows the maximal intensity histograms used to find the organoid mask. Clicking the beat button shows the beat image of the first identified beat. Clicking one of the flurophore buttons shows the image of that flurophore.

All analysis results are saved next to the $EXPERIMENT file.

Packages overview

Most of the algorithms are present in the calsipro package, with some in calsiprovis.library. The rest of calsiprovis is the interactive web application, based on rapids and runnel. runnel starts a webserver and provides all necesarry tools to start python in a browser, whereas rapids has some helper functions to have applications run with runnel. Finally hoernchen is an sqlite-based ORM.

Changelog

Version 2

Version 2 incorporates some bugfixes and an improved README. No algorithms are changed in version 2.

• The packages calsipro and calsiprovis had incorrectly defined dependencies, which made them impossible to run without modifying the code. This version corrects the dependencies so the packages can be build and run. Thanks to Abhijeet Krishna for notifying me of this problem.
• In version 1 building and packaging was done by poetry. In version 2 building and packaging is done by uv, which also handles python versions.

Version 1

Initial publication