Cellar Documentation

Cellar is an interactive tool for analyzing single-cell omics data. Cellar is written in Python using Dash.

Getting started

Web Server

Our web server running a copy of Cellar is free for everyone to use and contains several test datasets. Most of these datasets were generated by the Human BioMolecular Atlas Program (HuBMAP) and can be downloaded via Cellar’s interface.

Video tutorials are also available on YouTube.

Local Installation: Docker

Cellar can also be installed locally. This has the advantage of being able to use on premises resources and avoid network latency. The easiest way to install Cellar locally is via docker.

Pull the image
```
$ docker pull euxhen/cellar
```

Start a Cellar server on port 8050

$ docker run --rm -p 8050:8050 euxhen/cellar

# You can also bind mount a local directory of `.h5ad` files to be read by Cellar
$ docker run --rm -p 8050:8050 -v /path/to/directory:/home/nonroot/cellar/data/server euxhen/cellar

Wait for a few seconds and point your web browser to localhost:8050

Local Installation: Conda

An alternative to a Docker installation is to create a local conda virtual environment with all the required dependencies. If you do not have conda installed, please follow these instructions first. Assuming conda is now available, follow the steps below.

Download Cellar’s installer script

$ wget https://raw.githubusercontent.com/euxhenh/cellar/main/installer.sh

Create a conda environment running python 3.8
```
$ conda create -n cellar python=3.8
```
This will create an environment named cellar.
Activate the environment
```
$ conda activate cellar
```
Install packages
```
$ bash installer.sh
```

Download the install_Rdeps.py script

$ wget https://raw.githubusercontent.com/euxhenh/cellar/main/install_Rdeps.py

Install R dependencies
```
$ python install_Rdeps.py
```

Download Cellar’s source code

$ git clone https://github.com/euxhenh/cellar

Run Cellar
```
$ cd cellar
$ python main.py
```
Wait for about 10 seconds and visit localhost:8050 on your web browser.

About the project

Cellar is developed and maintained by the Systems Biology Group at Carnegie Mellon University. It was developed to enable research groups to compare and integrate data from different modalities by using a simple and intuitive interface. Cellar is meant to be a one-stop solution to the single cell analysis pipeline by offering a wide selection of tools that follow best practices agreed upon in the community.

An accompanying paper is available via Nature Communications.

If you use Cellar in your work, we would appreciate citations to Cellar’s paper

@article{Hasanaj2022,
    author = {Euxhen Hasanaj and Jingtao Wang and Arjun Sarathi and Jun Ding and Ziv Bar-Joseph},
    issn = {2041-1723},
    issue = {1},
    journal = {Nature Communications 2022 13:1},
    month = {4},
    pages = {1-6},
    publisher = {Nature Publishing Group},
    title = {Interactive single-cell data analysis using Cellar},
    volume = {13},
    year = {2022},
}

License

Cellar is distributed by an MIT license.