ST Spot Detector

A web tool for automatic spot detection and positional adjustments for ST datasets.

The arrays used to generate ST datasets may contain positional variations due to printing artifacts. This web tool aims to detect correct spot positions using the images generated from the ST protocol. In order to obtain relevant experimental data, it is also possible to automatically select the spots which are located under the tissue, using a corresponding HE image. The spot positions and selections are further adjustable to one's own needs. A file is generated which contains the corrected spot coordinates of the ST data as adjusted array coordinates and pixel coordinates as well as file containing a 3x3 affine matrix to transform array coordinates to pixel coordinates which can be useful for downstream analysis.

Dependencies

The server uses Python 2.7 with the libraries OpenCV-Python and Pillow (PIL Fork) for image processing. These can easily be installed within a virtual environment using pip and requirements.txt (see Usage). It also uses another tissue recognition library.

A modern browser with HTML5 support is required for the front-end interface. The web app has been tested on the lastest version of Chrome and Firefox.

Usage

If you want to use deploy the ST Spot detector locally on your computer you can use this singularity container. Otherwise follow the deployment instructions below.

Download

Clone the repository

git clone https://github.com/SpatialTranscriptomicsResearch/st_spot_detector.git

Move into the directory
```
cd st_spot_detector
```

Server setup (all `pip` and `Python` commands assume Python 2)

Install the necessary packages (OS-dependent command) e.g. on Ubuntu/Debian
```
    sudo apt install python2.7 python-pip nodejs npm make
```

Create and activate a Python virtual environment

pip install virtualenv
virtualenv venv
source venv/bin/activate

Install the dependencies in requirements.txt.
```
cd server
pip install -r requirements.txt
```
Install the tissue recognition library (still within the Python virtual environment). Follow the instructions here.
Build the client side files The client side uses the Node.js package manager and a Makefile to build
```
cd ../client
npm install
make DEVEL=1
```
Set up uWSGI a uWSGI daemon can be installed, e.g. on Ubuntu/Debian
```
sudo apt install uwsgi uwsgi-core uwsgi-plugin-python
```
or it may be installed within the Python virtual environment
```
pip install uwsgi
```
NOTE: the application may also be run directly as a Bottle application but performance is limited:
```
cd ../server
python server.py
```

Edit the uWSGI configuration file uwsg-server-config.ini The following lines require editing

chdir = /path/to/server/file/directory/
logto = /path/to/log/file/directory/%n.log
plugin = /path/to/python/plugin/ 
virtualenv = /path/to/virtual/env

They may for example look like

chdir = /home/user/st_spot_detector/server
logto = /home/user/.st_log/%n.log
plugin = /usr/lib/uwsgi/plugins/python_plugin.so
virtualenv = /home/user/st_spot_detector/venv/

Run the server uWSGI may then either be daemonized, e.g. on Ubuntu/Debian

# the file must be copied to the uwsgi directory
sudo cp uwsgi-server-config.ini /etc/uwsgi/apps-available/
# and symlinked to the apps-enabled folder
sudo ln -s /etc/uwsgi/apps-available/uwsgi-server-config.ini /etc/uwsgi/apps-enabled/uwsgi-server-config.ini
# then the service may be started
sudo service uwsgi start

or run locally

uwsgi uwsgi-server-config.ini

Optional The server may be configured to run with nginx or Apache (not covered here). It may also be desirable to configure port-forwarding to be able to access the web tool through port 80, e.g. on Ubuntu/Debian
```
sudo iptables -t nat -A PREROUTING -p tcp --dport 80 -i eth0 -j DNAT --to 0.0.0.0:8080
```

For any queries or concerns, feel free to contact the authors at the addresses given below.

Manual

For a guide on using the ST spots detection tool, please refer to this guide.

License

MIT (see LICENSE).

Authors

See AUTHORS.

Contact

Kim Wong kim.wong@scilifelab.se Jose Fernandez jose.fernandez.navarro@scilifelab.se

elhb/st_spot_detector