Inside Biology Space: talking to 100+ tools in plain English

Biology Space is where a discovery project begins. Before you design a molecule, you need a target worth hitting — and the confidence that it's druggable. This is the workspace that answers "what should we go after, and why?" in plain English.

Ask, don't script

Every interaction is a conversation. Type "What targets are associated with KRAS-mutant colorectal cancer?" and the agent queries OpenTargets, NCBI, and the literature, then hands back a ranked, cited answer. No pipelines to wire, no formats to convert.

What it can do

• Target identification — Surface disease-associated genes and proteins from genomic and literature evidence.
• Structure prediction — Fold a sequence with ESMFold or AlphaFold when no experimental structure exists.
• Pathway analysis — Trace interactions and pathways to understand mechanism and off-target risk.
• Knowledge lookup — Pull records from UniProt, RCSB PDB, ChEMBL, and GEO — grounded, not guessed.

The agent doesn't just fetch a structure — it decides whether it needs to predict one, which model to use, and how to check the result before showing you.

Grounded by design

Every answer carries its evidence. Claims trace back to a database record or a live computation, visible in the run trace. When the agent folds a protein, you see the model, the inputs, and the confidence — not a black-box assertion.

Handing off to Chemistry

A validated target isn't the finish line — it's a brief. Because Biology and Chemistry share the same platform and session memory, the target you land on here flows straight into Chemistry Space without losing context. That continuity is the whole point.

Want the full tool list and query patterns? See the Biology Space docs.