DGLinker

DGLinker predicts novel candidate genes associated with human diseases by integrating known disease-gene associations with biological and phenotypic data into knowledge-graphs and applying machine learning models to prioritize gene-disease associations.

Key Features:

• Knowledge-graph generation: Constructs knowledge-graphs that integrate biomedical information and known disease-gene associations.
• Machine learning-based prediction: Applies machine learning models trained on existing genetic information to predict novel candidate disease genes.
• Integration of biological and phenotypic data: Incorporates extensive biological and phenotypic data, including data from high-throughput technologies, into analyses.
• Use of known disease-gene associations: Leverages curated known disease-gene associations as primary input for model training and prediction.
• Result exploration and interpretation tools: Provides analytical outputs to support interpretation of predicted gene-disease associations.
• Publication-ready outputs: Produces figures and outputs suitable for inclusion in scientific publications.

Scientific Applications:

• Gene prioritization: Prioritizes novel candidate genes for specific human diseases based on integrated data and machine learning scores.
• Mechanistic insight and target discovery: Supports generation of hypotheses about disease mechanisms and potential therapeutic targets through predicted associations.
• Large-scale data integration: Enables analysis and integration of large biological and phenotypic datasets to expand knowledge of disease genetics.

Methodology:

Integrates known disease-gene associations with biological and phenotypic data to build knowledge-graphs, then analyzes those graphs using machine learning models trained on existing genetic information to predict new candidate genes.