PROST: quantitative identification of spatially variable genes and domain detection in spatial transcriptomics

Overview

PROST is a flexible framework to quantify gene spatial expression patterns and detect spatial tissue domains using spatially resolved transcriptomics with various resolutions. PROST consists of two independent workflows: PROST Index (PI) and PROST Neural Network (PNN).

Using PROST you can do:

• Quantitative identification of spatial patterns of gene expression changes by the proposed PROST Index (PI).

• Unsupervised identification of spatial tissue domains using a PROST Neural Network (PNN).

Installation and Turorials

If you want to run PROST, please visit our Document , which contains the installation, usage, and tutorials of PROST.

Easy Start

After installation, we suggest downloading the complete example files from zenodo (The dataset is too large to upload to github, there only 1 case (151672) of DLPFC data).
Similarly, you can download the dataset for each turorial individually via the google drive in the tutorial.

Reference and Citation

Liang, Y., Shi, G., Cai, R. et al. PROST: quantitative identification of spatially variable genes and domain detection in spatial transcriptomics. Nat Commun 15, 600 (2024). https://doi.org/10.1038/s41467-024-44835-w

Improvements

We welcome any comments about PROST, and if you find bugs or have any ideas, feel free to leave a comment FAQ. PROST doesn't fully test on macOS.