← all candidates · target GHSR1a
PRX-GHSR-002
very strong grade D gates 3/5
About this candidate
| Code | PRX-GHSR-002 |
|---|---|
| Target | GHSR1a · Growth hormone secretagogue receptor 1a (ghrelin receptor) |
| Lane | metabolic |
| UniProt | Q92847 |
| Mode | agonist |
| ortho | |
| Binding tier | very_strong |
| Drug-likeness | — |
| Half-life | medium |
| Regulatory | class_2_research_only |
| Evidence grade | D |
| Created | 2026-05-01 21:03 UTC |
Sequence
| Length | 13 residues |
|---|---|
| Hydrophobic | 7 (54%) |
| Polar | 2 (15%) |
| Positive (K/R/H) | 2 |
| Negative (D/E) | 2 |
Sequence redacted under public scope.
The raw amino-acid sequence is not exposed on this page or any
public-stats response. It is available to authenticated agents
with admin or licensed-review scope via
GET /api/candidates/PRX-GHSR-002 using
X-API-Key + X-Agent-ID headers.
Phase 4 metadata
| Design method | rfdiffusion_proteinmpnn_boltz2 |
|---|---|
| Structure confidence (ipTM) | 0.96 |
| Predicted affinity (kcal/mol) | -7.9 |
| PoP-Shield tier | green |
| Design origin | effect_driven |
ΔΔG regressor preview
Regressor not trained yet. Run
python scripts/train_affinity_regressor.py once
≥10 (target, peptide, ipTM) tuples are cached.
Target structure
AlphaFold v6 prediction of GHSR1a · cached on disk
The candidate peptide binds this structure. The per-candidate predicted complex PDB is not currently persisted — the on-target AlphaFold prediction is the proxy you can interrogate.
AlphaFold context
| pLDDT mean | 81.6 / Confident backbone model |
|---|---|
| Coverage | 100.0% (366/366 residues) |
| Source | alphafold db |
| Low-confidence intervals | res 1-36 / mean 35.02 res 341-366 / mean 56.57 res 284-294 / mean 65.33 |
Functional sites
Functional-site annotations are not cached for this target yet.
Binding partners / GO terms
Binding partners and GO terms are not cached for this target yet.
Proof spine
This candidate is addressable inside the Peptide Claim Graph: target, fold, AlphaFold metric receipt, candidate gates, evidence finding, timeline state, and export surface where available.
How this could be wrong
Cryptographic provenance
| PDA envelope | eb9d1888ba1cf2ce62a8beb1d2b96d91da7a71f03d9d97eefcbf01ca949a31f7 |
|---|
Recomputing the run with different weights produces a divergent envelope hash — that's the cryptographic guarantee binding this candidate to the exact pipeline that generated it.