Deciphering spatial domains from spatially resolved transcriptomics through spatially regularized deep graph networks

Overview

SR-DGN utilizes graph attention networks (GATs) to adaptively aggregate gene expression in-formation from neighboring spots or cells. We introduced spatial regularization constraints in SR-DGN to ensure the consistency of neighborhood relationships between the original space and embedding space. Additionally, SR-DGN employs cross-entropy (CE) loss to learn gene expression states, mitigating the impact of noisy gene dropouts. We evaluated the performance of SR-DGN across multiple datasets. The results demonstrate that SR-DGN can accurately identify tissue structures in SRT data of varying resolutions. Overall, SR-DGN is a powerful tool for precise spatial domain identification.

Requirements

You'll need to install the following packages in order to run the codes.

• pandas~=1.3.5
• numpy~=1.21.5
• scanpy~=1.9.3
• matplotlib~=3.5.2
• scikit-learn~=1.0.2
• scipy~=1.7.3
• seaborn~=0.12.0
• torch~=1.11.0
• tqdm~=4.64.1
• R==4.2.1

Datasets

All datasets used in this paper are available for download from the following websites.

(1) Stereo-seq mouse olfactory bulb data: https://github.com/JinmiaoChenLab/SEDR_analyses.

(2) Slide-seqV2 mouse olfactory bulb data: https://singlecell.broadinstitute.org/single_cell/study/SCP815/highly-sensitive-spatial-transcriptomics-at-near-cellular-resolution-with-slide-seqv2#study-summary.

(3) Slide-seqV2 mouse hippocampus dataset: https://singlecell.broadinstitute.org/single_cell/study/SCP354/slide-seq-study.

(4) osmFISH mouse somatosensory cortex data: https://linnarssonlab.org/osmFISH/.

(5) Stereo-seq mouse embryos data: https://db.cngb.org/stomics/mosta/.

(6) MERFISH mouse frontal cortex data: https://cellxgene.cziscience.com/collections/31937775-0602-4e52-a799-b6acdd2bac2e/.

Tutorial

Reproduce the result of article.

• olfactory bulb(Stereo-seq):

  python  run_mob.py

• hippocampus:

  python  run_slide.py

• embryos:

  python run_embryo.py

• somatosensory cortex:

  python run_osm.py

• olfactory bulb(Slide-seqV2):

  python run_mob2.py

• mouse frontal cortex (MERFISH 3D):

  python run_3D.py

Contact

Feel free to submit an issue or contact us at gaorui@sdu.edu.cn for problems about the packages.