Skip to content
/ GGOT Public

GGOT: A Gaussian graphical optimal transport method to detecting disease tipping points

License

Notifications You must be signed in to change notification settings

huawenbo/GGOT

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GGOT: Detecting Disease Critical Transitions Using Gaussian Graphical Optimal Transport Embedded With Protein Interaction Networks.

This repository contains our proposed methods Gaussian Graphical Optimal Transport (GGOT) for detecting critical transitions and identifying trigger molecules of diseases, the corresponding pre-processed diseases datasets from TCGA and GEO databases, and pre-processed Protein-Protein Interaction (PPI) Networks of the Human and Mouse Genomes.

Code for the paper:

Wenbo Hua, Ruixia Cui, Heran Yang, Jingyao Zhang, Chang Liu, Jian Sun. "Detecting Disease Critical Transitions Using Gaussian Graphical Optimal Transport Embedded With Protein Interaction Networks"

Overview

GGOT uses Gaussian graphical model, incorporating the gene interaction network, to model the data distributions at different disease stages. Then we use population-level optimal transport to calculate the Wasserstein distance and transport maps between stages, enabling us to detect critical transitions. By analyzing the per-molecule transport map, we quantify the importance of each molecule and identify the trigger molecules. Moreover, GGOT predicts the occurrence of critical transitions in unseen samples and visualizes the disease progression process.

Overview

Datasets

Clicking on the name of the corresponding dataset will redirect you to the website to download the corresponding dataset. The dataset of XJTUSepsis needs to contact [email protected] to get access.

Dataset Description Species stages Type of disease
GSE48452 non-alcoholic fatty liver Home sapiens 4 Chronic progressive benign disease
GSE2565 lung injury Mus musculus 10 Acute progressive noncritical disease
LUAD lung adenocarcinoma Home sapiens 8 Chronic progressive malignant disease
COAD colon adenocarcinoma Home sapiens 8 Chronic progressive malignant disease
XJTUSepsis sepsis Home sapiens 8 Acute progressive critical disease
GSE154918 sepsis Home sapiens 4 Acute progressive critical disease

Prerequisites

1 Create your environment and install required packages

conda env create -f environment.yaml -n yourenv

2 Download the pre-processed Protein-Protein Interaction (PPI) Networks of the Human and Mouse Genomes into "PPI/". We construct the PPI networks using STRING database.

PPI Networks of two species:[GGOT_PPI]

3 Download the RNA-seq dataset of different diseases into "data/source/".

Raw datasets: [GGOT_datasets]

Make sure your data file tree is similar to the following.

GGOT
├──run_model.py
├──data/
│  ├──source
│  │  ├──GSE48452
│  │  │  ├──GSE48452.csv (RNA-seq expression)
│  │  │  ├──group.csv (stage information)
│  │  ├──GSE2565
│  │  │  ├──GSE48452.csv
│  │  │  ├──group.csv
│  │  ├──LUAD
│  │  │  ├──LUAD.csv
│  │  │  ├──group.csv
│  │  ├──......
│  ├──......
├──PPI
│  ├──Human
│  │  ├──ppi_database.npy
│  │  ├──map_gene_protein.csv
│  │  ├──map_gene_protein_full.csv
│  │  ├──symbol2id.csv
│  ├──Mus
│  │  ├──ppi_database.npy
│  │  ├──map_gene_protein.csv
│  │  ├──map_gene_protein_full.csv
│  │  ├──symbol2id.csv
├──......

Running GGOT for datasets with different types

  • running model for the chronic progressive non-critical disease, GSE48452
python run_model.py -d GSE48452 -s Human
  • running model for the acute progressive non-critical disease, GSE2565
python run_model.py -d GSE2565 -s Mus
  • running model for the chronic progressive critical disease, LUAD, COAD
python run_model.py -d LUAD -s Human
python run_model.py -d COAD -s Human
  • running model for the acute progressive critical disease, XJTUSepsis, GSE154918
python run_model.py -d XJTUSepsis -s Human
python run_model.py -d GSE154918 -s Human

Making the visualization

python visualization.py -d GSE2565

Results (example on simulation dataset)

Overview

If you have any problems with this repository, please send emails to [email protected] for discussions.

About

GGOT: A Gaussian graphical optimal transport method to detecting disease tipping points

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages