PharMeBINet

PharMeBINet integrates and analyzes heterogeneous pharmacological, medical, and biochemical networks to enable network-based exploration of drug–disease–gene relationships and adverse drug reactions.

Key Features:

• Integration of Diverse Databases: Incorporates Hetionet (which aggregates entities and relationships from 29 public resources) plus 19 additional pharmacological, medical, and biological databases including CTD, DrugBank, and ClinVar.
• Advanced Data Mapping Techniques: Employs external identifier systems and name-mapping techniques to harmonize identifiers and entity names across sources.
• Extensive Network Structure: Implemented as an open-source Neo4j database comprising 2,869,407 nodes with 66 labels, 15,883,653 relationships, and 208 edge types representing entities such as ADRs, diseases, drugs, genes, gene variations, and proteins.
• Interconnected Entity Relationships: Encodes complex relations including drug–drug interactions and drug–causes–ADR edges to represent pharmacological and biomedical associations.
• Potential for Advanced Data Analysis: Provides a comprehensive graph suitable for downstream analyses including machine learning and predictive modeling of biomedical interactions.

Scientific Applications:

• Drug Discovery and Development: Supports identification of therapeutic targets and prediction of adverse effects through analysis of drug interactions and ADR associations.
• Genomic and Proteomic Research: Facilitates studies linking gene variations and protein data to diseases and drug responses.
• Personalized Medicine: Enables correlation of genetic information with drug response patterns to inform personalized treatment strategies.
• Machine Learning Applications: Serves as a rich training and feature source for developing machine learning models to predict complex biological interactions and outcomes.

Methodology:

Built by integrating Hetionet and 19 additional databases, mapping entities using external identifier systems and name-mapping techniques, and implementing the merged graph in a Neo4j database.