Molecular and Ligand Modeling
Molecular and ligand modeling은 small-molecule identity, representation, standardization, conformer, property prediction, retrieval, generation을 다룹니다. Ligand는 molecule과 다른 object type이 아니라 binding 또는 target context 안에 있는 molecule입니다.
이 페이지는 molecule 관련 작업을 어디로 보낼지 정하는 route page입니다. Chemical identity 자체는 Molecules 에서 고정하고, target, assay, pocket, pose가 들어오면 Interaction Modeling 또는 Structure-Based Modeling 으로 이동합니다.
r L = ϕ ( L , c ) 여기서 L c
Molecule note에서는 chemical identity를 먼저 고정하고, ligand note에서는 그 molecule이 어떤 target, pocket, assay, pose context 안에 들어갔는지 고정합니다.
Route Map
Question Start Watch 어떤 chemical object를 모델링하는가? Molecules , Molecule , Ligand salt, mixture, stereochemistry, tautomer, protonation model은 어떤 representation을 보는가? RDKit , Molecular graph , SMILES , Molecular fingerprint equivalent molecule이 다른 input이 될 수 있음 3D conformation이 포함되는가? Conformer , Force field conformer source와 minimization protocol이 지배할 수 있음 molecule이 target-conditioned인가? Interaction Modeling , Target-assay-label contract molecule-only claim은 target-specific activity를 포함하지 않음 structure 또는 docking이 중심인가? Structure-Based Modeling pose, pocket, scoring, geometry는 별도 check가 필요함
Main Subroutes
Area Use for Start Chemical identity standardization, salt stripping, stereo, tautomer, protonation Molecules 2D representation SMILES, molecular graph, fingerprint, scaffold, similarity RDKit , Molecular graph 3D representation conformer, force field, shape, geometry, docking input Conformer Property and retrieval property prediction, similarity search, candidate ranking Molecular property prediction Generation valid molecule samples, constrained generation, scaffold editing Molecular generation
Molecule vs Ligand
Object Meaning Typical question Molecule chemical identity after standardization property, similarity, scaffold, generation Ligand molecule in binding/target context pose, activity, interaction, docking input Assay compound record measured sample in an assay source endpoint, unit, censoring, replicate Conformer 3D state of a molecule shape, strain, docking preparation Pose ligand conformer placed in protein/pocket frame RMSD, clash, interaction, score
Chemical State Checklist
State Ask Salt / mixture modeled molecule이 parent, salt, mixture, assay record 중 무엇인가? Stereochemistry unspecified/specified stereocenter를 일관되게 처리했는가? Tautomer canonical tautomer 하나를 택했는가, 여러 state가 가능한가? Protonation / charge state가 pH, assay, docking, force field assumption과 compatible한가? Conformer 3D geometry가 generated, experimental, minimized, complex reuse 중 무엇인가? Deduplication equivalent molecule을 split construction 전에 collapse했는가?
RDKit Boundary
RDKit은 molecule parsing, standardization, fingerprint, descriptor, substructure search, conformer의 implementation layer인 경우가 많습니다. 이 setting은 invisible preprocessing이 아니라 method choice로 다룹니다.
RDKit use Record canonical SMILES standardization and stereo policy Morgan fingerprint radius, bit length, count/binary mode, chirality flag descriptor vector descriptor list and missing-value policy scaffold split standardized molecule used for scaffold extraction conformer generation seed, conformer count, force field, minimization and failure policy
Claim Boundary
Claim Needs molecule property prediction standardized molecule, split, endpoint, metric ligand activity prediction target/assay context and label contract molecular generation validity, uniqueness, novelty, filter denominator conformer quality conformer source, energy/minimization, geometry checks docking-ready ligand protonation, tautomer, charge, atom typing, conformer policy
Boundary
Molecule-only 또는 ligand-preparation question은 이 페이지를 사용합니다. Row가 molecule-target-assay relation이면 Interaction modeling 을 사용합니다. Pocket, pose, coordinate frame, protein-ligand complex가 claim의 일부이면 Structure-based modeling 을 사용합니다.
Page Owns Molecules chemical identity, standardization, representation contract Molecular and Ligand Modeling route choice across property, retrieval, generation, ligand preparation Interaction Modeling target-conditioned activity, affinity, selectivity, assay relation Structure-Based Modeling pocket, pose, docking, protein-ligand geometry
Checks
molecular standardization이 deduplication과 splitting 전에 수행되었는가?
stereochemistry, charge, protonation, tautomer, conformer policy가 explicit한가?
split이 scaffold-based, temporal, target-aware, assay-aware 중 무엇인가?
metric이 property prediction, retrieval, generation, docking, screening 중 무엇을 test하는가?