Skip to content
Discovery Stage 04 · Known Actives

Start from what already binds

Measured potency from prior art beats guessing from a blank scaffold. NovoMCP pulls known actives with reported IC50 and Ki straight from ChEMBL, so lead optimization begins on chemistry that has already shown activity against the target.

“Pull known actives for EGFR with measured IC50 under 100 nM.”

2.4M
ChEMBL bioactives
IC50/Ki
Measured potency
Prior art
Starting scaffolds
Inline
No export step
The funnel

How it works

01

Name the target

Ask for actives against a target. The engine queries ChEMBL bioactivity for compounds with measured potency — IC50, Ki, and related endpoints — rather than predicted activity.

02

Filter by what matters

Narrow by potency threshold, assay type, or confidence. The result is a real set of molecules with known activity, not a generated guess.

03

Feed straight into optimization

Confirmed actives become the starting scaffolds for lead optimization — each one already enriched with ADMET and compliance, with no export or re-import between tools.

Proof

search_chembl retrieves measured bioactivity (IC50/Ki) across 2.4M ChEMBL compounds.

Every returned active is a known compound in the 122M-molecule layer — so ADMET, drug-likeness, and FAVES compliance are already computed and returned inline.

Starting from measured actives anchors lead optimization in real structure-activity, not a cold start.

Use this when you need to

Anchor a campaign in compounds with measured activity

Find starting scaffolds instead of generating from scratch

Compare a novel idea against the known active set

Pull potency data without leaving the conversation

Research preview

Begin on chemistry that has already worked.

2.4M measured actives. Real IC50/Ki. Lead optimization that starts from activity, not a guess.