primekg

PrimeKG Precision Medicine Knowledge Graph Skill — Multiscale Biomedical Data Query Platform

Skill Overview

The PrimeKG precision medicine knowledge graph skill integrates 20+ authoritative biomedical databases, providing multiscale biological data querying across genes, drugs, diseases, and phenotypes. It supports drug target discovery, disease association analysis, and drug repositioning studies.

Use Cases

Drug Discovery and Target Identification: By querying disease–gene associations, drug–target relationships, and protein–protein interaction networks, it quickly identifies potential therapeutic targets and mechanisms, accelerating the new drug R&D process.

Drug Repositioning Research: Based on existing drugs’ target and indication data, it uncovers new indication potential. By analyzing drug–disease path mining, it reveals opportunities for repositioning.

Multiscale Biological Data Integration: Connects molecular-level gene targets with clinical-level disease phenotypes, supporting end-to-end analysis from basic research to clinical applications. It is especially suitable for network pharmacology and systems biology research.

Core Capabilities

Intelligent Node Search and Association Query: Supports precise searches for entities such as genes, proteins, drugs, diseases, and phenotypes. With a large-scale network of 100K+ nodes and 4M+ relationship edges, it quickly retrieves direct neighbors and clinical-evidence associations.

Comprehensive Disease Context Analysis: Provides high-level summaries of disease associations, including related gene, drug, and phenotype information, helping researchers quickly understand disease biological mechanisms and potential intervention strategies.

Network Exploration Across Multiple Relationship Types: Covers 29 relationship types (e.g., drug–target, disease–gene, protein–protein interactions, GWAS evidence, etc.). It supports filtering by relationship type so users can focus on specific evidence types for in-depth analysis.

Frequently Asked Questions

What types of data does the PrimeKG knowledge graph include?

PrimeKG integrates multiple biomedical entities, such as genes, proteins, drugs, diseases, and phenotypes, containing 100,000+ nodes and 4 million+ relationship edges. Data sources include 20+ authoritative databases (e.g., drug databases, genomics databases, clinical databases) and high-quality scientific literature. It covers 29 different relationship types, including drug–target associations, disease genetic associations, protein–protein interactions, phenotype–disease relationships, GWAS evidence, and more.

How do I query drug targets using PrimeKG?

Use the search_nodes function to search for the disease or gene of interest and obtain its unique identifier. Then use get_neighbors to retrieve all associated nodes for that entity. For drug target discovery, you can filter the drug_protein relationship type to obtain drug–target association data. A more efficient approach is to directly use the get_disease_context function to retrieve, in one step, the disease-associated genes, drugs, and phenotype information—quickly building target hypotheses.

How is PrimeKG different from other biomedical databases?

PrimeKG’s core advantage lies in its multiscale integration capability: it connects molecular-level gene target data with clinical-level disease phenotype data within a unified network. Compared with single-purpose functional databases, PrimeKG supports cross-level querying across 29 relationship types, making it especially suitable for research scenarios that need to consider gene, drug, and disease relationships across multiple dimensions. In terms of scale, the 100K+ nodes and 4M+ relationship edges provide broader network coverage, while the Python interface makes data access more convenient.