scMGCA

scMGCA is a Python package containing tools for clustering single-cell data based on a graph-embedding autoencoder that simultaneously learns cell–cell topology representation and cluster assignments.

• Overview
• System Requirements
• Installation Guide
• Usage
• Data Availability
• License

Overview

Single-cell RNA sequencing provides high-throughput gene expression information to explore cellular heterogeneity at the individual cell level. A major challenge in characterizing high-throughput gene expression data arises from challenges related to dimensionality, and the prevalence of dropout events. To address these concerns, we develop a deep graph learning method, scMGCA, for single-cell data analysis. scMGCA is based on a graph-embedding autoencoder that simultaneously learns cell-cell topology representation and cluster assignments. We show that scMGCA is accurate and effective for cell segregation and batch effect correction, outperforming other state-of-the-art models across multiple platforms. In addition, we perform genomic interpretation on the key compressed transcriptomic space of the graph-embedding autoencoder to demonstrate the underlying gene regulation mechanism. We demonstrate that in a pancreatic ductal adenocarcinoma dataset, scMGCA successfully provides annotations on the specific cell types and reveals differential gene expression levels across multiple tumor-associated and cell signalling pathways.

System Requirements

Hardware requirements

scMGCA package requires only a standard computer with enough RAM to support the in-memory operations.

Software requirements

OS Requirements

This package is supported for Linux. The package has been tested on the following systems:

• Linux: Ubuntu 18.04

Python Dependencies

scMGCA mainly depends on the Python scientific stack.

numpy
scipy
tensorflow
scikit-learn
pandas
scanpy
anndata

For specific setting, please see <a href="https://github.com/Philyzh8/scMGCA/blob/master/requirements.txt">requirement</a>.

Installation Guide:

Install from PyPi

$ conda create -n scMGCA_env python=3.6.8
$ conda activate scMGCA_env
$ pip install -r requirements.txt
$ pip install scMGCA

Usage

scMGCA is a deep graph embedding learning method for single-cell clustering, which can be used to:

• Single-cell data clustering. The example can be seen in the <a href="https://github.com/Philyzh8/scMGCA/blob/master/tutorial/demo.py">demo.py</a>.
• Correct the batch effect of data from different scRNA-seq protocols. The example can be seen in the <a href="https://github.com/Philyzh8/scMGCA/blob/master/tutorial/demo_batch.py">demo_batch.py</a>.
• Analysis of the mouse brain data with 1.3 million cells. The example can be seen in the <a href="https://github.com/Philyzh8/scMGCA/blob/master/tutorial/demo_scale.py">demo_scale.py</a>.
• Provide an automatic hyperparameter search algorithm. The example can be seen in the <a href="https://github.com/Philyzh8/scMGCA/blob/master/tutorial/demo_para.py">demo_para.py</a>.

We give users some suggestions for running in the <a href="https://github.com/Philyzh8/scMGCA/blob/master/tutorial/tutorial.md">tutorial.md</a>.

Data Availability

The real data sets we used can be download in <a href="https://doi.org/10.5281/zenodo.7475687">data</a>.

License

This project is covered under the MIT License.

Citation

@article{yu2023topological,
  title={Topological identification and interpretation for single-cell gene regulation elucidation across multiple platforms using scMGCA},
  author={Yu, Zhuohan and Su, Yanchi and Lu, Yifu and Yang, Yuning and Wang, Fuzhou and Zhang, Shixiong and Chang, Yi and Wong, Ka-Chun and Li, Xiangtao},
  journal={Nature Communications},
  volume={14},
  number={1},
  pages={400},
  year={2023},
  publisher={Nature Publishing Group UK London}
}