Skip to content

A TRAPI-compliant component that uses AnswerCoalesce for edge inference

Notifications You must be signed in to change notification settings

ranking-agent/edgar

Repository files navigation

Enrihment Driven Graph Recommender- EDGAR

Explores pathway enrichment strategies in biomedical Knowledge Graphs (KGs) as a versatile link-prediction approach, with drug repurposing exemplifying a significant application. Leveraging systems biology, network expression analysis, pathway analysis (PA), and machine learning (ML) methods, KGs aid in uncovering novel interactions among biomedical entities of interest. While these approaches excel in inferring missing edges within the KG, PA may overlook candidates with similar pathway effects. By utilizing enrichment-driven analyses on KG data from ROBOKOP, our EDGAR paper applied this method on Alzheimer's disease case study, demonstrating the efficacy of enrichment strategies in linking entities for drug discovery. Our approach is validated through literature-based evidence derived from clinical trials, showcasing the potential of enrichment-driven strategies in linking biomedical entities. s

Demonstration

A Live version is available at: https://edgar.apps.renci.org/

Local Development Steps

Clone the repo: git clone https://github.com/ranking-agent/edgar.git

cd to the dir: cd edgar

Install requirements: pip install -r requirements.txt

on the terminal, run python app.py

DEPLOYMENT

Build the Docker image: docker build -t edgar:latest .

Push the Docker image: docker push edgar:latest

About

A TRAPI-compliant component that uses AnswerCoalesce for edge inference

Resources

Stars

Watchers

Forks

Packages