Spotiphy is a Python-based pipeline designed to enhance our understanding of biological tissues by integrating sequencing-based spatial transcriptomics data, scRNA-seq data, and high-resolution histological images. Employing a probabilistic model, Bayesian inference, and advanced image processing techniques, Spotiphy primarily executes three key tasks:
- Deconvolution: Spotiphy estimates the abundance of each cell type in each capture area of spatial tissue.
- Decomposition: Spotiphy decomposes spatial transcriptomics data to the single-cell level.
- Pseudo single-cell resolution image: Spotiphy generates a pseudo single-cell resolution image to reconstruct cell neighbors.
With these outputs, Spotiphy facilitates numerous downstream analyses. For more detailed information, please refer to the associated research paper.
The following tutorial are available:
- Deconvolution and decomposition of mouse cortex with Spotiphy [document][Google Colab]
For more details, please refer to the documents.
To install Spotiphy, it is recommended to create a separate conda environment. This approach helps to manage dependencies and avoid conflicts with other packages.
conda create -n Spotiphy-env python=3.9
conda activate Spotiphy-env
Spotiphy is built based on Pytorch. Although installing Spotiphy automatically includes PyTorch, it is recommended that users manually install PyTorch (link) to allow for more flexibility, particularly for those who wish to utilize CUDA capabilities. We offer two methods for installing the Spotiphy package:
- Install from GitHub: This method allows you to install the latest version directly from the source code hosted on GitHub.
pip install git+https://github.com/jyyulab/Spotiphy.git
- Install from PyPI: This approach is for installing the Spotiphy package from the Python Package Index (PyPI), which is more streamlined for users who prefer standard package installations.
pip install spotiphy
To test the Installation, try to import Spotiphy in Python.
import spotiphy
Answers to frequently asked questions can be found here.
Should you have any further questions, feel free to start a discussion or reach out directly to the package authors:
- Ziqian Zheng - [email protected]
- Jiyuan Yang - [email protected]
Pending