Codes for HBDB Relationship network

Data Requirements

Download the hbdb2.sql and put it in the same directory

Environment Setup

Create a new conda environment

conda create --name hbdb_env python=3.10

Activate the conda environment

conda activate hbdb_env

Install necessary package

pip insall mysql-connector-python==9.1.0
pip install openai==1.63.0

Modify gen_scores.py and build_net.py to match your dataset format

Replace the following line in the script with your own MySQL settings:

connection = mysql.connector.connect(host='localhost', port='3306',user='root'password='XXXXXXXX')

Ensure the host, port, user, and password match your database configuration.

• host: The hostname or IP address of your MySQL server.
• port: The port number your MySQL server listens on (default is 3306).
• user: Your MySQL username.
• password: Your MySQL password (replace XXXXXXXX with the actual password).

Set your OPENAI_API_KEY in prompt.py

os.environ["OPENAI_API_KEY"] = "Your API KEY"

Build Network Steps

1. Use gpt-4o-mini to generate scores and summaries:

python gen_scores.py

2. Process the generated data:

python extract.py

3. Build the network structure in the Cytoscape.js format:

python build_net.py

4. Run the server to diaplay network

npm install express
node server.js

5. View the network

The URL format is http://localhost:8001/index.html?file={compound_id}_{compound_name} Take acetone for example, you can open the following URL in your browser to view its network:

http://localhost:8001/index.html?file=28_acetone

Evaluation Steps

Prepare Dataset

download BC8_BioRED_Subtask1_BioCJSON/bc8_biored_task1_train.json and put it in the same directory

BioCreative VIII BioRED: https://doi.org/10.1093/database/baae069 BC8_BioRED dataset is available at https://codalab.lisn.upsaclay.fr/competitions/13377#learn_the_details-dataset

Run Evaluation

python eval.py

Notes

HBDB snapshot (hbdb2.sql) serves as a snapshot of the Human Breathomics Database (HBDB). In this snapshot, all literature information was collected and organized in a MySQL database, along with all recognized biomedical terms and sentences. The full text of some literature was retrieved using Elsevier's Text and Data Mining (TDM) service in HBDB. Redistribution is limited to 200 characters due to the terms of use of the TDM service; therefore, this file is not publicly available.

For the snapshot of HBDB without full text and manually curated dataset dervied from HBDB, please refer to the zenodo record: https://zenodo.org/records/14958797. To access the complete context (full text) in the manually curated dataset, please retrieve the original text from the original paper by mapping Reference ID to DOI/URL/PubMed ID in the HBDB snapshot.

Notice from Elsevier TDM service: Some rights reserved. This work permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited.