README: Neutrophils actively swell to potentiate rapid migration

https://doi.org/10.7272/Q6NS0S5N

Primary neutrophils were isolated from the blood of volunteers and then gently injected into the microfluidic chips. Mechanical stimulation activates the cells so extreme care was taken to keep the cells quiescent and unactivated. The FxM chips were prepped such that they always had 3 lanes per chip and all lanes were imaged concurrently to minimize the impact of timing and, if possible, all lanes were included in the analysis. One lane was usually an untreated (aka WT) condition and the other two lanes were different drug conditions (see fxm_dataset_info.csv for details). If lanes were discarded, the reason is included in the fxm_dataset_info.csv in the notes section. The primary reason for not including a lane was due to incomplete bonding of the microfluidic chip or due to flow in the microfluidic chip. The chips lack a resistive element so any bubbles in the system would lead to very large flows and the cells could be blown away in the matter of seconds.

While this README lays out the high-level organization of the datasets, the best way to learn how to interact with the data is to look at the code used to generate the figures in the Code/Software section below.

Description of the data and file structure

There are two related FxM datasets used in the paper:

1. fxm_uncaging_* : This dataset is at a lower time resolution (every 12 seconds), but all lanes were imaged simultaneously. The chemoattractant was uncaged after the first 15 minutes and then the cell response was imaged for another 30 minutes.
2. fxm_highres_*: After the uncaging dataset was collected, certain lanes were then immediately imaged at a higher time resolution (every 3 seconds) to capture the motility-driven volume fluctuations. Each lane was imaged sequentially, not in parallel.

Uncaging FxM Dataset

The uncaging FxM microscopy dataset consists of 30 sets of movies organized by date, volunteer, and condition. The details of these datasets and any datasets that were discarded for any reason are in fxm_dataset_info.csv.

Each of the 30 sets include:

1. fxm_uncaging...raw_rgb.tif A RGB TIFF file of the raw movie with the nuclear (Hoechst), cytoplasmic (Calcein Red-Orange), and FxM (AlexaFluor-647 10k MW dextran) channels encoded as red, green, and blue, respectively.
2. fxm_uncaging...fxmcorr.tif A fully processed, denoised, and flatfield corrected version of the FxM channel that was used for segmentation also as a TIFF file.

Additionally, there are two CSV files* that contained the extracted information from the movies

1. fxm_uncaging_augmented.csv - contains the extracted data from the FxM movies. Key columns include particle (the ID of a cell), x and y (the position of the cell), time_s (the time in seconds since the start of the movie), abs_volume_um3(the absolute volume of the cell in um^3), area_um2 (the footprint area of the cell in um^2), isencapsulated (whether the cell is fully separated from other cells),
2. fxm_uncaging_augmented_w_footprints.csv - same as above but with an additional column called footprint_cart which is a list of the XY coordinates of pixels "belonging" to the cell. Useful for plotting cell footprints over the original movies.

*All CSVs can have missing values (i.e. two consecutive delimiters ,,) if a value is absent or could not be computed for a situation. This values are treated as missing in Julia.

FxM HighRes dataset

The highres dataset (described in fxm_highres_datasets.csv) consists of 4 sets of movies organized the same way as the uncaging dataset from above with both fxm_highres...rawrgb.tif and fxm_highres...fxmcorr.tif versions for each set.

The extracted information from these datasets is available in fxm_highres_augmented_w_localities.csv which is identical to the CSVs described above with the important addition of locality_cart, which is a list of pixels "belonging" to the local area around a cell. This local area is used to "fill in" the cells footprint to compute the counterfactual of what the signal would be if the cells wasn't there.

Coulter dataset

All in-suspension volume measurements were taken on a Beckman-Coulter Z2 machine. For each run, the cell volumes were measured three times prior to stimulation with chemoattractant and then was sampled every minute for the following 12+ minutes. The coulter_datasets.csv file contains the metadata for each run with the date, volunteer, and condition (aka sample). It also includes a SampleCode which is the file name prefix for each measurement in a run. Finally, DataPath is the path to the folder containing the raw data files in coulter_data.zip. The raw files are Beckman-Coulter Z2 files, which can be loaded into Julia using tlnagy/Coulter.jl.

Code/Software

The best way to interact with the data is to explore the code used to generate the figures in the paper. The code is available as a mini-site ( https://tamasnagy.com/Nagy_2023_SwellMigration/) which is built automatically on Github Actions to validate reproducibility and portability. To run the analysis locally, first git download the repository and then download this dataset into the data folder. You'll need a recent version of Julia installed (tested on 1.9.3) and then you can run the following code in the terminal from the root of the repository:

julia --project="." -e 'using Pkg; Pkg.instantiate()'

which will install all the packages you need, followed by

julia --project="." site/make.jl

This will run all the analysis code and build the website. You can follow this by running

julia --project=build -e 'using Pkg; Pkg.instantiate()'
julia --project=build -e 'using LiveServer; serve(dir = "site/build")'

which will fire up a mini-server and allow you to browse the site locally if you navigate to the URL that is printed out.