DeepCell Spots

deepcell-spots is a deep learning package for analyzing spatial transcriptomics data sets. Its application, Polaris, allows you to apply pre-existing models and train new deep learning models for fluorescent spot detection. Polaris also contains a probabilistic method for combinatorial barcode assignment. It is written in Python and leverages a number of packages, including: TensorFlow, Keras, PyTorch, and DeepCell. More detailed documentation of deepcell-spots is available here.

Getting Started

Install with pip

The simplest way to install deepcell-spots is with pip:

pip install deepcell-spots

Install with Docker

The deepcell-spots Docker container is also available on DockerHub. For more details for running DeepCell Docker containers, please see the deepcell-tf README.

DeepCell Spots Applications

deepcell-spots contains an applications that greatly simplify the implementation of machine learning models for spot detection and decoding. SpotDetection contains a pre-trained model for fluorescent spot detection on images derived from assays such as RNA FISH and in-situ sequencing. This model returns a list of coordinate locations for fluorescent spots detected in the input image. SpotDecoding fit a probabilistic model for spot decoding with SVI. It returns a gene identity for each detected spot for multiplex FISH data sets. Polaris pairs these two applications with DeepCell models for nuclear and cytoplasmic segmentation. For example implementation, please see the example notebooks.

DeepCell Spots for Developers

Build and run a local docker container, similarly to the instructions for deepcell-tf. The relevant parts are copied here with modifications to work for deepcell-spots. For further instructions, see the deepcell-tf README.

Build a local docker container, specifying the deepcell version with DEEPCELL_VERSION

The deepcell-spots Docker image is built using the deepcell-tf Docker images as a base image, so a DEEPCELL_VERSION must be specified in the build command.

git clone https://github.com/vanvalenlab/deepcell-spots.git
cd deepcell-spots
docker build --build-arg DEEPCELL_VERSION=0.12.9-gpu -t $USER/deepcell-spots . 

Run the new docker image

# '"device=0"' refers to the specific GPU(s) to run DeepCell-Spots on, and is not required
docker run --gpus '"device=0"' -it \
-p 8888:8888 \
$USER/deepcell-spots

It can also be helpful to mount the local copy of the repository and the notebooks to speed up local development.

# you can now start the docker image with the code mounted for easy editing
docker run --gpus '"device=0"' -it \
    -p 8888:8888 \
    -v $PWD/deepcell-spots/deepcell_spots:/usr/local/lib/python3.8/dist-packages/deepcell_spots \
    -v $PWD/notebooks:/notebooks \
    -v /$PWD:/data \
    $USER/deepcell-spots

Copyright

Copyright © 2019-2024 The Van Valen Lab at the California Institute of Technology (Caltech), with support from the Shurl and Kay Curci Foundation, Google Research Cloud, the Paul Allen Family Foundation, & National Institutes of Health (NIH) under Grant U24CA224309-01. All rights reserved.

License

This software is licensed under a modified APACHE2. See LICENSE for full details.

Trademarks

All other trademarks referenced herein are the property of their respective owners.

Credits