AI agents only knowthe famous proteins.
LLMs keep quoting the same few thousand entries, ignoring most of the PDB.
Famous proteins, which get cited often, are those your AI learns.
It can recall PDB IDs from memory, but cannot traverse the database by function or structure.
You need sharp answers to specific questions about a niche enzyme or organism, not generic responses.
We re-index the PDBfor AI agents.
Function, biology, ligands, structure, organism, experiments, literature.
Every entry distilled into a profile your AI can actually search.
Queryable text covering function, structure, and context.
What changes is not the data.It’s the text.
We synthesize metadata, annotations, and computed features into profiles optimized for search.
- Ligand appears as an internal code (A1CHN), not usable chemistry
- No pocket information (size, hydrophobicity, accessibility)
- No in-depth functional description
Human synaptic vesicle protein 2A (SV2A) is a multi-pass membrane protein that facilitates regulated neurotransmitter release and selectively enhances low‑frequency synaptic transmission. The structure shows Major Facilitator Superfamily / sugar-transporter–like transmembrane domains and an inhibitor, UCB7361, bound in the transmembrane region.
- Ligand identified as inhibitor
- Pocket characteristics described
- Biological role detailed
Ask questions a search form can’t express.
Each query below mixes biological roles, binding-site character, and experimental conditions in a single sentence. Click any card to see real results.
enzymes with buried hydrophobic active site pocket bound to noncovalent drug-like inhibitor
Pocket physics + ligand role — not searchable as structured fields.
CRISPR Cas9 or Cas12 endonuclease ternary complex with guide RNA and target DNA
Functional state: protein + RNA + DNA present together.
integral multi-pass transmembrane protein with enclosed intramembrane hydrophobic cavity bound to small-molecule drug
Membrane-embedded binding pockets — not just any hydrophobic cavity.
vertebrate hemoglobin tetramer oxygen transport protein excluding human
Function + assembly + organism exclusion in one query.
Why semantic search wins here
Structural-biology queries are multi-dimensional. Keyword search flattens them.
Multi-dimensional queries
Cross function, ligand, pocket, organism, and method in one shot.
Concept-level search
"Intersubunit pocket", "post-hydrolysis state" — not only IDs and gene names.
Every entry, equal footing
Obscure depositions rank alongside textbook ones. No citation bias.
Expanding thesearch surface
Today you can search by molecular role, binding character, and assembly. We’re adding dimensions that matter most for experimental reuse — so you can find structures worth building on, not just structures that match a keyword.
Get early access
Help shape this tool. leave your email and we’ll be in touch.
No spam. Unsubscribe anytime. We only email about Protplex.
Get in touch
Questions, feedback, or collaboration ideas. We welcome hearing from you.
Your email is used only for direct communication about Protplex. We do not share or sell contact information.