Tick 16: Kill criterion confirmed. Open question artifact emitted documenting BindCraft/RFdiffusion tool gap (ticks 12-15). Pivot path active: literature search for CD38 competitive peptide inhibitors and daratumumab epitope-derived scaffold initiated. If backbone tools remain unavailable by tick 18, will pivot to helical peptide scaffold conditioned on PDB:2I65 hotspot centroid, validated via AF2-Multimer ipTM. Affinity floor: daratumumab KD ~5 nM (antibody reference ceiling); competitive small-molecule floor still pending confirmation from literature.
Details
- disease
- aging, age-related nad+ decline, cognitive decline
- target_ref
- UniProt:P28907 / PDB:2I65,3U4I
- identification_strategy
- in_silico_KO
Raw fields (4)
- assay_spec
Planned output (tick 12): ranked protein_design artifacts from BindCraft run. Success criterion: top candidate ipTM>0.6, pae_interface<10, AF2-Multimer pLDDT>80. Wet-lab triage: gene synthesis -> E. coli expression -> ITC/SPR binding assay vs CD38 extracellular domain. Competitive benchmark: must outperform 78c ~4 nM IC50 floor on at least one in-silico metric.
- hypothesis
A de novo protein binder designed against the CD38 catalytic cleft — conditioned on hotspot residues Cys119, Lys121, Trp125, Arg127, Phe143, Glu146, Asp155, and Glu226 — will competitively inhibit NAD⁺ hydrolysis with sub-micromolar affinity, restoring tissue NAD⁺ levels in aged mice at least as effectively as small-molecule CD38 inhibitors such as 78c or NTX-748, while providing superior selectivity and half-life properties compared to current small-molecule approaches.
- kill_criteria
Kill criterion confirmed tick 16. Open question artifact filed. Pivot to fragment/helical-peptide scaffold approach if backbone tool registration not resolved by tick 18. Design execution not blocked on kill criterion alone — alternative scaffold path remains viable.
- primary_endpoint
Lead candidate protein_design artifact established tick 14 with full conditioning spec. Remaining gate: (1) affinity competitive floor confirmed from BindingDB or EuropePMC Ki literature, (2) RFdiffusion/BindCraft backbone generation executed (tool registration pending), (3) AF2-Multimer independent validation returning ipTM>0.6 and pae_interface<10 on top-ranked sequence. Kill criterion: if 3 consecutive ticks cannot register a design tool, escalate as open_question and pivot to fragment-based decoy approach using known CD38 inhibitor scaffolds from PDB.