GitHub - open-prophetdb/biomedgps: A knowledge graph system with graph neural network for drug repurposing and disease mechanism.

BioMedGPS Platform

A knowledge graph system with graph neural network for drug repurposing, disease mechanism.

NOTE: NOT READY FOR PRODUCTION YET.

Features

Knowledge graph studio for graph query, visualization and analysis.
Graph neural network for drug discovery, disease mechanism, biomarker screening and discovering response to toxicant exposure.
Support customized knowledge graph schema and data source.
Support customized graph neural network model.
Support customized omics datasets.
Integrated large language models (such as vicuna, rwkv, chatgpt etc. more details on chat-publications) for answering questions.

Ecosystem

Name	Language	Description
BioMedGPS Data	Python	For building the knowledge graph of BioMedGPS and training the graph neural network models.
Chat Publications	Python	Ask questions and get answers from publications.
BioMedical Knowledgebases	Markdown	A collection of biomedical knowledgebases, ontologies, datasets and publications etc.
R Omics Utility	R	Utilities for omics data with R. It will be part of biomedgps system and provide visulization and analysis functions of multi-omics data.

Demo

Ask questions with chatbot

Find similar diseases with your queried disease

Predict drugs and related genes for your queried disease

Find potential paths between two nodes

For Users

If you want to use the platform, please follow the instructions below. You can also join us to develop the platform, see here.

Step 1: You will need to install PostgreSQL and Neo4j. You can also use docker to launch the database, see here. After that, you can run the following command to launch the database server.

mkdir biomedgps && cd biomedgps
wget -O docker-compose.yml https://raw.githubusercontent.com/biomedgps/biomedgps/master/docker-compose.yml

docker-compose up -d

Step 2: Download the latest release from here. You can also build the platform from source code, see here.

Step 3: Init Database, Check and Import Data

# Init database
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli initdb

# Check and import data, -t is the table name, -f is the data file path, -D is the delete flag
# Import entity data
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli importdb -f /data/entity.tsv -t entity -D

# Import relation data
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli importdb -f /data/relation.tsv -t relation -D --dataset drkg

# Generate metadata for entity
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli importdb -f /data/entity.tsv -t entity_metadata -D

# Generate metadata for relation
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli importdb -f /data/relation_types.tsv -t relation_metadata -D

# Import entity2d data
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli importdb -f /data/entity2d.tsv -t entity2d -D

# Import entity embeddings
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli importdb -f /data/entity_embeddings.tsv -t entity_embedding -D

# Import relation embeddings
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps-cli importdb -f /data/relation_embeddings.tsv -t relation_embedding -D

# Import entity data to graph database
export NEO4J_URL=neo4j://neo4j:password@localhost:7687 && biomedgps-cli importgraph -f /data/entities.tsv -t entity -b 1000

# Import relation data to graph database
export NEO4J_URL=neo4j://neo4j:password@localhost:7687 && biomedgps-cli importgraph -f /data/relations.tsv -t relation -b 1000

Step 4: Launch the platform, see more details on usage here.

export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && export NEO4J_URL=neo4j://neo4j:password@localhost:7687/test_biomedgps && biomedgps -H

Usage

Run the following command to see the usage.

$ biomedgps --help
biomedgps 0.1.0
Jingcheng Yang <yjcyxky@163.com>
BioMedGPS backend server

USAGE:
    biomedgps [FLAGS] [OPTIONS]

FLAGS:
    --debug          Activate debug mode short and long flags (--debug) will be deduced from the field's name
    -h, --help       Prints help information
    -o, --openapi    Activate openapi mode
    -u, --ui         Activate ui mode
    -V, --version    Prints version information

OPTIONS:
    -d, --database-url <database-url>    Database url, such as postgres://user:pass@host:port/dbname. You can also set
                                         it with env var: DATABASE_URL
    -H, --host <host>                    127.0.0.1 or 0.0.0.0 [default: 127.0.0.1]  [possible values: 127.0.0.1,
                                         0.0.0.0]
    -g, --neo4j-url <neo4j-url>          Graph Database url, such as neo4j://user:pass@host:port/dbname. You can also
                                         set it with env var: NEO4J_URL
    -p, --port <port>                    Which port [default: 3000]

Example

If you want to launch the server with openapi and ui mode, run the following command.

export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && biomedgps -H 0.0.0.0 -p 8888 --openapi --ui

For Linux with systemd

# You need to place the binary file in /opt/local/bin/biomedgps and all data files in /opt/local/data/biomedgps.
# Or you can change the path in the service file.
sudo cp build/biomedgps.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable biomedgps
sudo systemctl start biomedgps

Install Docker and Docker Compose

# For MacOSx
brew install docker docker-compose

# For Linux [https://docs.docker.com/engine/install/debian/]

For Developers

Build docker image

git clone https://github.com/yjcyxky/biomedgps

cd biomedgps
bash build-docker.sh

Requirements

If you want to build the platform from source code, please follow the instructions below. You can also join us to develop the platform, see here.

Server

Rust
Cargo
PostgreSQL
Neo4j

Frontend

nodejs == 16.13.1
yarn
Reactjs
Ant Design
Plotly.js
G6 - Graph Visualization Engine

Installation (Development) for Server

1. Install Rust and Cargo

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

2. Install PostgreSQL (use docker)

make test-db

3. Run the server

export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && cargo run --bin biomedgps -- -H 0.0.0.0 -p 8888 --openapi --debug

Installation (Development) for Frontend

1. Install nodejs and yarn

curl -fsSL https://deb.nodesource.com/setup_16.x | sudo -E bash -
sudo apt-get install -y nodejs
npm install -g yarn

# Check the version. Make sure the nodejs version is 16.x.x. Other versions may not work.
node -v

2. Install dependencies

cd studio
git submodule update --init --recursive

3. Change the auth0 config

Change the auth0_config in studio/src/app.tsx to your own auth0 config. Replace '' and '' with your own auth0 client id and domain.

4. Run the frontend

yarn start:local-dev

Build for Production

Change the auth0 config

Change the auth0_config in studio/package.json to your own auth0 config. Replace UMI_APP_AUTH0_CLIENT_ID=Y08FauV1dAEiocNIZt5LiOifzNgXr6Uo and UMI_APP_AUTH0_DOMAIN=biomedgps.jp.auth0.com with your own auth0 client id and domain.

Compile the `biomedgps` and `biomedgps-cli` binaries

After installing the dependencies, run the following command to build the frontend and backend. The output will be in target/release folder. You will get a binary file named biomedgps.

# For MacOSx
make build-mac

# For Linux
make build-linux

Install docker and docker-compose

# For MacOSx
brew install docker docker-compose

# For Linux [https://docs.docker.com/engine/install/debian/]

Launch the database and server

Download docker directory
Run docker-compose up -d to launch the database
```
cd docker && docker-compose up -d
```
Upload biomedgps and biomedgps-cli binaries to the server. /opt/local/bin/biomedgps and /opt/local/bin/biomedgps-cli are the default paths.

Run the following command to initialize the database

export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && /opt/local/bin/biomedgps-cli initdb

Run the following command to upload the data

Assume all data files are in /data folder. You need to ensure that the data directory is mounted to the docker container.

# Upload entity data
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && /opt/local/bin/biomedgps-cli importdb -f /data/entity -t entity -D

# Upload entity2d data
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && /opt/local/bin/biomedgps-cli importdb -f /data/entity2d.tsv -t entity2d -D

# Upload relation data
export DATABASE_URL=postgres://postgres:password@localhost:5432/test_biomedgps && /opt/local/bin/biomedgps-cli importdb -f /data/relation -t relation -D

...

Configure a nginx server to serve the frontend and backend
- Download biomedgps_nginx.conf and modify it to fit your environment
- Upload the modified biomedgps_nginx.conf to /etc/nginx/config.d/biomedgps_nginx.conf
- Restart nginx
```
systemctl restart nginx
```

Use docker to build cross-platform binaries for debian os on macosx with m2 chip

docker pull messense/rust-musl-cross:x86_64-musl
alias rust-musl-builder='docker run --rm -it -v "$(pwd)":/home/rust/src messense/rust-musl-cross:x86_64-musl'
rust-musl-builder cargo build --release

make build-studio && rust-musl-builder cargo build --release

Name		Name	Last commit message	Last commit date
Latest commit History 179 Commits
.github		.github
build		build
docker		docker
examples		examples
migrations		migrations
src		src
studio		studio
.dockerignore		.dockerignore
.gitignore		.gitignore
Cargo.lock		Cargo.lock
Cargo.toml		Cargo.toml
Dockerfile		Dockerfile
LICENSE		LICENSE
Makefile		Makefile
README.dev.md		README.dev.md
README.md		README.md
build-docker.sh		build-docker.sh

License

open-prophetdb/biomedgps

Folders and files

Latest commit

History

Repository files navigation

BioMedGPS Platform

Features

Ecosystem

Demo

Ask questions with chatbot

Find similar diseases with your queried disease

Predict drugs and related genes for your queried disease

Find potential paths between two nodes

For Users

Usage

Example

For Linux with systemd

Install Docker and Docker Compose

For Developers

Build docker image

Requirements

Server

Frontend

Installation (Development) for Server

1. Install Rust and Cargo

2. Install PostgreSQL (use docker)

3. Run the server

Installation (Development) for Frontend

1. Install nodejs and yarn

2. Install dependencies

3. Change the auth0 config

4. Run the frontend

Build for Production

Change the auth0 config

Compile the biomedgps and biomedgps-cli binaries

Install docker and docker-compose

Launch the database and server

Use docker to build cross-platform binaries for debian os on macosx with m2 chip

Join Us

Contributing

License

About

Topics

Resources

License

Stars

Watchers

Forks

Releases 1

Packages 0

Languages

Compile the `biomedgps` and `biomedgps-cli` binaries

Packages