Data-Centric Cross-Species Target Discovery with Polly KG
Key Highlights
A Boston-based biotech focused on immune and metabolic diseases was hampered by siloed non-model datasets; our scalable ETL pipeline and Base-KG unified them into one AI-ready foundation from day one.
With code-free natural-language querying on Base-KG, scientists explored cross-species gene–disease–drug relationships in plain English—no custom integrations or scripting required.
Co-building the customer’s Polly KG took the graph to production scale—~31M nodes and ~60M relationships in six months—supporting pathway, PPI, and functional analyses in a single workspace.
A transparent cross-species target-ranking framework (phase, traceability, novelty/safety, disease relevance, drug uniqueness) made every priority auditable and grounded in harmonized multi-omics evidence.
In six months, the program prioritized five high-value targets and surfaced a PDE4 inhibitor repurposing signal; the case study details the scoring schema and NLQ prompts used to get there.
.svg)
.svg)
%201%20(1).svg)
.svg)
%201.svg){kind=icon}
.svg)
%20(1).svg)
