# BRAinS

This repo contains the code for creating the BRAinS-Graph (*Biomedical Knowledge Graph for Recommending and Analysing Health Studies*). It loads data from four sources into a Neo4j graph database:
- [ClinicalTrials.gov](https://clinicaltrials.gov/)
- [The Portal of Medical Data Models (MDM Portal)](https://medical-data-models.org/)
- [Medical Subject Headings (MeSH)](https://www.nlm.nih.gov/mesh/meshhome.html)
- [Unified Medical Language System (UMLS)](https://www.nlm.nih.gov/research/umls/index.html)

## Structure of the repository

The repository consists of three dataloaders (`study2neo4j`, `moi`, `umls2neo4j`) and a postprocessing script (`postprocessing`). 

> [!NOTE]
> The dataloader for the MDM Portal, `mdm2neo4j`, can be found in an additional repository: [mdm2neo4j repo](https://git.uni-greifswald.de/MILA_public/mdm2neo4j).
> It should be run as well to create the BRAinS-Graph.

### File Structure

```
src/
│
├── study2neo4j
│   ├── run.py                  # Main script to execute the ClinicalTrials.gov import
│   ├── ct2neo4j.py             # Functions for database connecting and data importing
│   └── README.md               # Documentation for setup and usage
│
├── moi
│   ├── moi.py                  # Main script to execute the ontology import and processing
│   ├── methods.py              # Functions for configuring, importing, and processing the graph
│   └── README.md               # Documentation for setup and usage
│
├── umls2neo4j
│   ├── umls2neo4j.py           # Main script to execute the umls import
│   ├── methods_umls2neo4j.py   # Functions for loading cui names, parsing relations, loading into neo4j
│   └── README.md               # Documentation for setup and usage
│
└── postprocessing
    └── postprocess.py          # Main script for postprocessing (creating relationships between data sources)
```

## How to

Create a configuration file, storing your details for the database connection, e.g., in your home-directory with the name `brains.conf`.
It should have the following structure:

```ini
[neo4j]
uri = bolt://localhost:7687
username = neo4j
password = myfancypassword
```

Run the dataloaders `study2neo4j`, `moi`, `umls2neo4j`, which can be found under :file_folder: `src`.
Instructions for usage are provided in the individual README.md files.
Run the dataloader [mdm2neo4j repo](https://git.uni-greifswald.de/MILA_public/mdm2neo4j).

> [!IMPORTANT]  
> In general, the individual dataloader (`study2neo4j`, `moi`, `umls2neo4j`) can be run in any order. 
> The postprocessing only works if data has been loaded into the Neo4j. It should be run last (after all dataloaders have been run).
> Make sure not to forget to run mdm2neo4j as a dataloader.

## Requirements
- make sure `python3` is installed
- install the required libraries with `pip install -r requirements.txt`
- have a running Neo4j DB (Neo4j version 5)
- create the configuration file as described in the [How To section](#how-to)

Have a running Neo4j instance with APOC and neosemantics installed.

## Licence

This program is released under [Version 3 of the GPL or any later version](https://www.gnu.org/licenses/gpl-3.0.en.html).