scConGraph is a scalable bi-layer graph model that efficiently integrates cross-time context information, enabling the comprehensive analysis of tumor cell dynamic responses from paired perturbed or time-seiries single-cell transcriptomics.

scConGraph is currently available for Linux systems, as it relies on LINE, a C++-based embedding method that depends on the GSL package for Linux. If you use the downstream analysis functions in scConGraph, which are implemented in Python, there are no system restrictions.

Tested System Configuration:

OS: Linux 3.10.0-1160.el7.x86_64
Python Version: 3.9.19
Processor: x86_64
CPU Cores: 36
Logical CPUs: 36
Total RAM (GB): 251.38

scConGraph requires python version 3.7+. Install directly via pip:

pip install scConGraph

If you prefer not to install scConGraph， you can download the script directly from the GitHub repository: scConGraph/scConGraph.py. Then, manually import the module in your Python environment:

import sys
sys.path.append('./scConGraph-main/scConGraph/')
import scConGraph as scg

Importantly, the LINE toolkit (LINUX version) must be downloaded and installed from LINE GitHub Repository (https://github.com/tangjianpku/LINE.git) before using scConGraph.

The vignette of scConGraph can be found in the project Wiki.

The R scripts used for analyzing drug responses are located in the Analysis/Codes folder. The raw and intermediate data are stored in the Analysis/Data folder. Larger datasets are saved on the cloud.
If you have any questions about the codes, please contact the author.