StructureMoa is a chemical-structure exploration tool. Structural similarity, clustering, and SAR relationships are intended for research exploration and do not establish shared potency, selectivity, safety, efficacy, or clinical performance.

Core structure exploration tiers

Rule-based Catalog Exploration Score from SMILES similarity, approved/clinical tags, family network size, and SAR metadata. It does not predict efficacy, safety, or development success.

StructureMoa is a chemical-structure exploration tool. Structural similarity, clustering, and SAR relationships are intended for research exploration and do not establish shared potency, selectivity, safety, efficacy, or clinical performance.

Core 후보 44파생 후보 15439 패밀리

MedMedium exploration priority1

Upper-mid Exploration Score percentile (X-A bin)

High exploration priority2DStructureClassExploration ScoreFactors
Med
Sotorasib
KRAS G12C inhibitor chemical landscape · Drug
KRAS inhibitor56패밀리 core · SMILES 확보

KRAS G12C inhibitor chemical landscape · Drug · KRAS inhibitor

패밀리 core · SMILES 확보

MedMedium exploration priority2

Mid-pool Exploration Score (X-B bin)

High exploration priority2DStructureClassExploration ScoreFactors
Med
Daraxonrasib
Daraxonrasib · RAS(ON) · Drug
KRAS inhibitor48패밀리 core · SMILES 확보
Med
MRTX1133
KRAS G12D / RAS(ON) inhibitor series · Drug
KRAS inhibitor48패밀리 core · SMILES 확보

Daraxonrasib · RAS(ON) · Drug · KRAS inhibitor

패밀리 core · SMILES 확보

KRAS G12D / RAS(ON) inhibitor series · Drug · KRAS inhibitor

패밀리 core · SMILES 확보

LowLow exploration priority2

Lower Exploration Score — still listed for inspection (X-C bin)

High exploration priority2DStructureClassExploration ScoreFactors
Low
SMCC
Daraxonrasib · RAS(ON) · Linker
KRAS inhibitor38Linker hub · SMILES 확보
Low
Sulfo-SPDB
Daraxonrasib · RAS(ON) · Linker
KRAS inhibitor38Linker hub · SMILES 확보
LowSMCC

Daraxonrasib · RAS(ON) · Linker · KRAS inhibitor

Linker hub · SMILES 확보

Daraxonrasib · RAS(ON) · Linker · KRAS inhibitor

Linker hub · SMILES 확보

Actual weights, formula and input limits

Exploration score method

StructureMoa ranks are a rule-based Exploration Score. An LLM does not predict activity or efficacy. Role, SMILES, and relation signals inside each family are summed and cut into High / Medium / Low exploration priority.

Core tab

  • scaffold / payload / linker role weights
  • family center-compound bonus
  • SMILES source (db · known · pubchem · inherited)
  • relation-edge count and derived-child count

Derivative tab

  • drug / derivative / adc candidates
  • similarity to family center via Morgan fingerprint, radius 2, 2048 bits, Tanimoto
  • Approval, clinical and lead tag bonuses have been removed.

Explore tab

  • sparse SAR, cross-scaffold pairs, expansion leaves
  • weakly linked nodes and catalog-coverage gaps
  • relation graph + stored Morgan fingerprints
  • category filters

Exploration-priority bins

  • High (X-S): top ~12%
  • Medium (X-A/X-B): ~12–62%
  • Low (X-C): remainder (pools under 5 compounds stay Medium)

Internal exploration metric based on catalog and structural signals. Not a prediction of efficacy, safety, potency, or development success.

Full methodology

평가 시각: 2026. 10. 28. 오후 7:04:43 · 캐시 5분 · 전체 패밀리