Peptiter / DiscoveryLab
Discovery
Discovery story

Why brute-force peptide search fails.

Framing the discovery problem before generating any candidate. The combinatorial sequence universe is unworkable without constraints.

The problem

Drug design fails when mechanism is asserted, not proven — and combinations are found by trial and error.

Two unsolved problems compound. Search spaces are astronomically large — even short peptides yield ~10¹³ candidates no sampler can cover — and there is no predictive theory of which combinations are stable or super-additive, so polypharmacology stays empirical. DiscoveryLab attacks both: it constrains the search and it machine-checks whether a cross-modal combination is a stable controller.

10-mer
20¹⁰
≈ 1.024 × 10¹³
natural amino acids
14-mer
20¹⁴
≈ 1.64 × 10¹⁸
natural amino acids
14-mer + ncAA
(20+k)¹⁴
intractable
with non-natural residues

DiscoveryLab treats drug design as a search-space and mechanism-verification problem, not a text-generation problem. Generation, ranking, rejection, pathway reachability, intervention blockade, and perturbation-evidence checks — and, for combinations, a machine-checked stability proof — happen before a candidate moves toward wet-lab validation. The peptide search-space figures below are our calibrated reference modality.