Mine enzymes for your target function from a proteome.
A five-stage funnel narrows a whole proteome down to a handful of confident, docked enzyme candidates — each stage annotated in this entry.
Identify the carbohydrate-active enzyme family each sequence belongs to.
Align against known target enzymes to lock in homologous candidates.
Flag secreted enzymes and remove near-duplicate sequences.
Predict binding pockets and score substrate binding.
Cross-verify the structure and scores of the highest-confidence candidates — including your own AlphaFold3 models.
Screen a whole proteome FASTA to surface candidate enzymes for one target catalytic function.
Classify sequences into carbohydrate-active enzyme families before any downstream filtering.
Flag secreted candidates, then predict confident structures (ipTM / pLDDT) for the survivors.
Score substrate binding in the predicted pocket, then cross-check it with a second scoring function.
Provide a proteome, optionally a reference enzyme and a substrate — then run the funnel.
Each stage narrows the candidate pool. Run the funnel to watch it advance stage by stage.
Top 20 candidates, sorted and filtered below. Expand a card for structure, sequence, and docking details.
Each saved candidate becomes a reproducible entry with an Entry ID you can cite in your methods. Handoff to codon optimization and expression cassette design will prefill the selected sequence.
Workspace integration arrives in a later task — this page is currently a front-end preview with sample data.