Ranked lists are easy. Knowing why one compound beat another is not.

Structure-based screening you can interrogate

See the toolkit

Pose, pocket, and model comparison in one place. For teams who cannot treat a score as the answer.

The screen, explained

See the app

Bioforager structure-based screen with ranked ligands and a pose view — **Screen results**
A ranked batch, the pose, and the contacts beside it.

Atom-level explanation of a ligand pose, with residue energies — **Atom explainer**
Which residues help, which clash, and by how much.

Who it's for

When two models disagree

Bioforager is not an end-to-end discovery suite. It is the step where a shortlist has to be defended.

Computational chemists

Run the batch and compare scoring engines on the same ligands
See which poses agree, and which pocket makes a model fail
Hand over a shortlist with the disagreement already marked

Medicinal chemists

Read anchor residues and interaction patterns, not only a rank
See which chemotype the pocket actually rewards
Decide what to make before a week of manual pose review

Circe reads the complex. Genesys is one of the scorers. Both sit inside Bioforager.

Three passes

i. Read

Circe pulls interaction rules from the structure: hydrogen bonds, aromatic pockets, anchor residues, and chemistry that repeats across a cluster.

ii. Score

The same batch runs through more than one engine. Genesys is our scorer. Established tools such as AutoDock Vina, DiffDock, and GNINA sit beside it.

iii. Compare

Poses are cross-checked at residue and interaction level. Agreement is a signal. Disagreement is the part you inspect.

In a run

What the screen actually shows

Interaction map of common and unique contacts across ligands — **Shared contacts**
What the batch agrees on, and what only one ligand uses.

Chemotypes and anchor residues from a completed screening batch — **Chemotypes**
Scaffolds, anchor residues, and how much of the batch shares a motif.

Homolog and interactor map for a query protein — **Target context**
Homologs and interactors around the protein you are screening against.

Ligand bound between two protein surfaces

The point

Why Bioforager?

Early screening is less stuck on compute than on interpretation. When two models split on a compound, the choice is to trust the score or spend days opening the pose by hand. Bioforager is built for the second job, done before that week starts.