The implementations of SpatialZ, as well as tutorials, can be found at https://spatialz-tutorial.readthedocs.io/en/latest/.

1 Installation Guide

This document provides detailed instructions for installing and setting up the SpatialZ project. Follow the steps below to configure your environment and install dependencies.

Step 1: Create a new virtual environment called spatialz with Python 3.9.19

conda create -n spatialz python=3.9.19 -y

Step 2: Activate the spatialz1 environment

conda activate spatialz

Step 3: Install PyTorch with CUDA 11.7 support

pip install torch==1.13.0+cu117 -f https://download.pytorch.org/whl/cu117/torch_stable.html

Step 4: Clone the SpatialZ repository and navigate to the project directory

git clone [https://github.com/senlin-lin/SpatialZ.git] /path/to/your/SpatialZ_code
cd /path/to/your/SpatialZ_code

Step 5: Install the project dependencies from requirements.txt

pip install -r requirements.txt

2 Deploying a Docker Image on a New Server

We also provide a Docker image that encapsulates our code and demo data, making it easier for users to directly download and use the provided resources. This image ensures a consistent and reproducible environment, allowing users to seamlessly run the code and explore the demo data without needing to configure dependencies or environments manually.

Step 1: Install Docker on the New Server

The following commands illustrate the basic steps to install Docker on Ubuntu system (Ubuntu system required):

sudo apt-get update
sudo apt-get install docker-ce

Step 2: Pull the Image from Docker Hub

To download the Docker image, execute the following command:

sudo docker pull linsenlin/spatialz:latest

Step 3: Launch the Docker Container on the New Server

Once the image is pulled, users can start the Docker container on the new server. The following command will run the container and map port 8888 of the server to port 8888 of the container:

sudo docker run --gpus all -p 8888:8888 linsenlin/spatialz:latest

Step 4: Access Jupyter Notebook

After launching the Docker container, users can access Jupyter Notebook by navigating to port 8888 on the server. If the server's IP address is 'server_ip', simply enter the following URL in a web browser:

http://server_ip:8888

Reference: Lin, S., Wang, Z., Cui, Y., Zou, Q., Han, C., Yan, R., … & Yuan, Z. (2024). Bridging the Dimensional Gap from Planar Spatial Transcriptomics to 3D Cell Atlases. bioRxiv, 2024-12.