Overview

The Knowledge Graph Viewer helps medical professionals explore relationships among drugs, genes, diseases, and laboratory indicators. It supports rapid hypothesis generation, mechanism exploration, and drug repurposing insights.

How to Use

1. In the sidebar, select a center node type and a specific center node. Optionally filter by MOA (All/Positive/Negative). Click “Submit”.
2. Open “Main page” to view the Knowledge Graph and the Downloadable Table.
3. Use the legend to interpret colors and shapes. Click lab group nodes to expand their children.
4. Export the filtered table with “Download Filtered CSV”.

Typical Use Cases

• Drug Prioritization and Repurposing
  • Start from a target gene or disease, discover positively/negatively modulating drugs, assess maturity via drug_max_phase.
• Mechanism Exploration
  • Start from a drug, inspect linked targets and disease associations to support mechanistic hypotheses.
• Biomarker Investigation
  • Expand lab group nodes to view related indicators and their relationships.

Key Features

• Center Selection and Filtering
  • Choose a center node type (disease/drug/gene/lab) and a specific center node (search-enabled).
  • Filter relationships by mechanism of action (MOA): Positive or Negative.
• Graph Visualization
  • Interactive network graph of entities related to the selected center node.
  • Node shapes reflect entity types; edge colors reflect MOA.
  • Click group nodes (only for lab groups) to expand their children.
• Table View and Download
  • Scrollable table summarizing filtered edges with rich context for target nodes.
  • Export the filtered result to CSV.

Visual Encoding

• Node Shapes
  • Drug: dot
  • Gene: box
  • Disease/Lab: triangle
  • Center Node: star
• Node Colors
  • Drug nodes use a five-level red gradient mapped to drug_max_phase: -1 (lightest), ½/3 (darker), 4 (darkest) to quickly assess maturity.
  • Non-drug nodes use distinct colors by type.
• Edge Colors
  • Positive modulation: magenta
  • Negative modulation: deep blue

Data and Interpretation

• Nodes (nodes.csv)
  • id, term (name/description), type, variable/value (additional info： drug_max_phase/safety_signal), group (used for lab nodes).
• Edges (edges_d.csv, edges_p.csv)
  • from, to, moa (Positive/Negative modulation), beta_hit.
• Important Note
  • Graphs represent data-driven associations, not causality. Validate with literature and experiments.