You're scrolling for your protein. It was second from the top — but your colleagues uploaded four new ones since then. The dashboard feels cluttered.
You're looking into the binding-site predictions. Which molecules are from the literature, which ones are novel predictions? Can you trust how this protein grabs the peptide you've been screening? Which construct will actually express better in cell-free?
We kept hearing the same four problems on calls. All four shipped over the last couple of weeks:
- Collections — group proteins and complexes into project views
- Binding sites with references — every prediction linked to its evidence
- PAE matrices for complexes — interface confidence, not just single chains
- Expression-system-aware Design — rank constructs for the system you'll actually use
Group proteins and complexes into Collections
Group proteins by project, target family, or campaign. However you want. Each Collection gets its own dashboard view.
- "GPCR program" — your active drug discovery targets
- "Q3 antibody panel" — your current characterization sprint
- "Kinase inhibitor leads" — anything you'll come back to
Find your protein in two seconds, not twenty.
Binding sites, now with references
Every binding-site prediction now links directly to its structural and/or chemical evidence:
- PDB entries for related complexes
- AlphaFill transplants of cofactors, substrates, and inhibitors
- ChEMBL bioactivity for any matched ligand
All linked under Residue-level Annotations — no more switching tabs to validate.
PAE matrices for complexes
Predicted Aligned Error is now visible across every chain pair in a complex — the confidence map you've had for single chains, now for the interface.
- High-confidence chain–chain contacts, flagged
- Flexible or low-confidence interface regions, exposed
- Subdomain-level reliability to plan cryo-EM or mutagenesis from
Tell a real interface from a confidently wrong one — before you commit to mutagenesis or cryo-EM time.
Expression-system-aware Design scoring
Design now ranks and filters constructs against your target expression system — E. coli, insect cells, yeast, mammalian, and cell-free.
- Solubility, codon bias, and tag compatibility, scored per system
- Different #1s for different systems — E. coli and HEK293 read codons very differently
- A ranked list your wet-lab scientist will actually use
Pick the system. Get the list the bench can run with.
What's the one thing we should ship next? Open your dashboard — and tell us. We read every reply.
