What it is: A prototype, designed around the concept that medical datasets can be graphed in 3D space to see new connections or novel patterns.
Large medical datasets are difficult to synthesize or understand as a whole, sometimes there's so much data that it's not possible to see nuanced connections or patterns. The tool is built around the promise to visualize and see medical datasets in a novel way.
When a dataset is graphed, the nodes organize themselves based on their connections with other nodes and relatedness, similarity or likeness. I believe that this organization can reveal novel connection and patterns, leading to investigations that were not thought of before.
Tools like the Cognir MedGraph do exist, though, I think the novelty lies in the fact that the Cognir MedGraph can be rendered in a single browser tab.
Taking the current prototype into consideration:
This is essentially a web-based biomedical knowledge graph visualizer focused on Alzheimer's disease. It displays disease mechanisms, symptoms, genes, drugs, and research connections as interactive 3D nodes that you can click, drag, and explore. You can add or edit nodes, create relationships between them, adjust evidence levels, and export/import the graph as JSON.
The UI is clean with a dark theme, a sidebar for browsing nodes, a 3D canvas using Three.js for visualization, and a details panel. It comes pre-loaded with an example Alzheimer's dataset of about 20 nodes and 24 connections showing how different biological factors relate to the disease.