ALSdisMNs

Automated and unbiased classification of motor neuron phenotypes with single cell resolution in ALS tissue.

Overview

This repository contains source code to implement automated identification of MNs exhibiting aberrant phenotype from high-content microscopy data. It contains two Jupyter notebook allowing to reproduce the figures in manuscript Automated and unbiased classification of motor neuron phenotypes with single cell resolution in ALS tissue, Hageman et al. 2020 (BioRxiv). All data required to reproduce the manuscript can be download on Zenodo.

Repository content

• Data: this folder should containing all raw data acquired from CellProfiler on multichanel fluorescent images segmented using the pipeline if one wants to replicate the work. All scripts as well as the raw images required to reproduce these can be downloaded from Zenodo together with the Data folder under the accession number 3985099.
• Scripts: Jupyter Notebook containing R and Python codes to automatically identify MNs subpopulation from the single-cell measurements data outputted by CellProfiler.
• Figures: raw figures as outputted by the analysis and used for the paper.

Dependencies

It is recommended to create a virtual environment to ensure complete reproducibility of the project, which relies on R (tested on R-3.3.1) and Python (tested on Python 3.5.3).

install.packages(c('data.table', 'dplyr','knitr','chron','colortools','RColorBrewer','corrplot','geneplotter',
             'lme4','beeswarm','rlang','stargazer','viridis','mclust','ape','dendextend','wordcloud','reshape'))

Here is how to create a virtual environment called env

1. If virtualenv is not installed, python3 -m pip install --user -U virtualenv
2. cd $MY_PATH where $MY_PATH is the location of the repository
3. virtualenv -p ``which python3`` env
4. source ./env/bin/activate
5. pip3 install --upgrade -r requirements.txt

The file requirements.txt contains all the Python librariries.

Then you can create and work as you would do normally. The next time

1. cd $MY_PATH
2. source ./env/bin/activate

To work on Jupyter Notebook remotely

On your remote machine (access via SHH for example):

1. cd $MY_PATH
2. source ./env/bin/activate
3. jupyter notebook --no-browser --port=8786

On your local machine:

1. ssh -N -L localhost:8787:localhost:8886 <[email protected]>
2. Open your browser on the local machine and type in the address bar: localhost:8787