molIEreVIS is a prototype implementation of an interactive visualization workflow designed to help domain experts explore, interpret, and validate drug repurposing hypotheses generated by analyzing multimodal evidence.
🚀 You can try molIEreVIS yourself at: https://molierevis.jku-vds-lab.at
The application welcomes you with an overview of molIEreVIS and its features.
- Ranked indication exploration based on multimodal consensus scoring
- Evidence-flow diagram that exposes how evidence influences scoring
- Direct drill-down into qualitative evidence (e.g., highlighted abstracts)
- Validation of MoA–indication plausibility using PrimeKG path queries
- Interactive MeSH ontology browser for disease selection
| Component | Technology |
|---|---|
| Frontend | React, visyn_scripts, D3.js |
| Backend | FastAPI (visyn_core) |
| Databases | PostgreSQL (pipeline scores), Neo4j (PrimeKG) |
| Package Management | Node 16 + yarn 3 |
| CI | GitHub Actions |
This project builds upon the datavisyn app_template.
git clone -b develop https://github.com/jku-vds-lab/molierevis.git
cd molierevisThe frontend is a React application built and managed via visyn_scripts. All relevant scripts can be found in the package.json.
yarn 3 is used as package manager, such that you can simply install the frontend via yarn install.
The application has many package.json scripts available, with one of them being yarn start. This will start the dev-server.
The backend is a FastAPI server managed via visyn_core. All relevant scripts can be found in the Makefile.
It is recommended to create a virtual environment to avoid cluttering the global installation directory.
python -m venv .venv # create a new virtual environment
source .venv/bin/activate # active it
make develop # install all dependenciesTo start the development server, run python molierevis, which will execute a uvicorn runner.