B ž3Rcó=ã@sšdZddlZddlZddlmZddlZddlmZddlmZddlm Z ddlm Z ddl Tddl Tddl TddlTddlTddlTddlTddlTddlTddlTddlTddlTddlTddlTdd lmZdd lmZdd lmZmZy ddl TWne!k rYnXe"e#_"e$e#_$eƒZd'dd„Z%d(dd„Z&d)dd„Z'd*dd„Z(d+dd„Z)d,dd„Z*dd de+fd d!„Z,d"d#„Z-d-d$d%„Z.e/d&kr–e.ƒdS).z% Import all RDKit chemistry modules éN)Ú namedtuple)Ú DataStructs)Ú ForceField)ÚRDConfig)ÚrdBase)Ú*)Ú rdGeometry)Úlogger)ÚStereoEnumerationOptionsÚEnumerateStereoisomerséÿÿÿÿFcCsj| |¡}t||ƒ|sf|dkr$d}dd„| ¡Dƒ}x|D]}||ks<| |¡q<sz TransformMol..N)Ú GetConformerZTransformConformerÚ GetConformersZRemoveConformerZSetId)ÚmolZtformÚconfIdZ keepConfsZrefConfZ allConfIdsZcidr r rÚ TransformMol3s   r©érrçà?cKs2tj|d|d|d|d}t|||f|Ž|S)z; returns a grid representation of the molecule's shape réé)Úspacing)rÚ UniformGrid3DÚ EncodeShape)rrZboxDimrÚkwargsZresr r rÚComputeMolShapeDsr!çš™™™™™É?ç@c CsØt |¡}| |¡}t|ddt|ƒ}|dj|djd||dj|djd||dj|djd|f}tj |d|d|d|d}t |||ddd|d }|  ¡} d d „| Dƒ} |t | ƒ} | S) aü Calculates the volume of a particular conformer of a molecule based on a grid-encoding of the molecular shape. A bit of demo as well as a test of github #1883: >>> from rdkit import Chem >>> from rdkit.Chem import AllChem >>> mol = Chem.AddHs(Chem.MolFromSmiles('C')) >>> AllChem.EmbedMolecule(mol) 0 >>> ComputeMolVolume(mol) 28... >>> mol = Chem.AddHs(Chem.MolFromSmiles('O')) >>> AllChem.EmbedMolecule(mol) 0 >>> ComputeMolVolume(mol) 20... F)ÚignoreHsrrr)rgð?)r$ZvdwScaleécSsg|]}|dkrd‘qS)r%rr )rÚxr r rrjsz$ComputeMolVolume..) ÚrdchemÚMolrZCanonicalizeConformerZComputeConfBoxr&ÚyÚzrrrZGetOccupancyVectÚlen) rrZ gridSpacingZ boxMarginÚconfZboxZsideLenÚshapeZvoxelVolZoccVectZvoxelsZvolr r rÚComputeMolVolumeLs   4 r.c Cs–|s,|rt|||g|dnt|||gd|j|d}|j|d}d}x6t| ¡ƒD]&}| |¡ | |¡¡} || | 7}qVW|| ¡}t |¡S)aš Returns the RMS between two conformations. By default, the conformers will be aligned to the first conformer before the RMS calculation and, as a side-effect, the second will be left in the aligned state. Arguments: - mol: the molecule - confId1: the id of the first conformer - confId2: the id of the second conformer - atomIds: (optional) list of atom ids to use a points for alingment - defaults to all atoms - prealigned: (optional) by default the conformers are assumed be unaligned and the second conformer be aligned to the first )ÚconfIdsÚatomIds)r/)Úidr)ÚAlignMolConformersrÚrangeÚ GetNumAtomsÚGetAtomPositionÚDistanceÚnumpyZsqrt) rZconfId1ZconfId2r0Ú prealignedZconf1Zconf2ZssrÚiÚdr r rÚGetConformerRMSos   r;c CsÚg}dd„| ¡Dƒ}|s<|r.t|||dqtt||dn8x6tdt|ƒƒD]$}| t||d||||d¡qLWg}x\tdt|ƒƒD]J}| ||d¡x2td|ƒD]$}| t||||||dd¡qªWqˆW|S) aÀ Returns the RMS matrix of the conformers of a molecule. As a side-effect, the conformers will be aligned to the first conformer (i.e. the reference) and will left in the aligned state. Arguments: - mol: the molecule - atomIds: (optional) list of atom ids to use a points for alingment - defaults to all atoms - prealigned: (optional) by default the conformers are assumed be unaligned and will therefore be aligned to the first conformer Note that the returned RMS matrix is symmetrical, i.e. it is the lower half of the matrix, e.g. for 5 conformers:: rmsmatrix = [ a, b, c, d, e, f, g, h, i, j] where a is the RMS between conformers 0 and 1, b is the RMS between conformers 0 and 2, etc. This way it can be directly used as distance matrix in e.g. Butina clustering. cSsg|] }| ¡‘qSr )r)rr,r r rr±sz)GetConformerRMSMatrix..)r0ÚRMSlist)r<rr)r0r8T)rr2r3r+Úappendr;)rr0r8Zrmsvalsr/r9ZcmatÚjr r rÚGetConformerRMSMatrix“s (r?c#s†t|ƒ| ¡kr(tdt|ƒ| ¡fƒ‚d‡fdd„ ‰tddƒ}x@ˆ|ƒD]4}| |¡}x$|D]}|rt|||ƒVq^|Vq^WqJWdS) a» Returns a generator for the virtual library defined by a reaction and a sequence of sidechain sets >>> from rdkit import Chem >>> from rdkit.Chem import AllChem >>> s1=[Chem.MolFromSmiles(x) for x in ('NC','NCC')] >>> s2=[Chem.MolFromSmiles(x) for x in ('OC=O','OC(=O)C')] >>> rxn = AllChem.ReactionFromSmarts('[O:2]=[C:1][OH].[N:3]>>[O:2]=[C:1][N:3]') >>> r = AllChem.EnumerateLibraryFromReaction(rxn,[s2,s1]) >>> [Chem.MolToSmiles(x[0]) for x in list(r)] ['CNC=O', 'CCNC=O', 'CNC(C)=O', 'CCNC(C)=O'] Note that this is all done in a lazy manner, so "infinitely" large libraries can be done without worrying about running out of memory. Your patience will run out first: Define a set of 10000 amines: >>> amines = (Chem.MolFromSmiles('N'+'C'*x) for x in range(10000)) ... a set of 10000 acids >>> acids = (Chem.MolFromSmiles('OC(=O)'+'C'*x) for x in range(10000)) ... now the virtual library (1e8 compounds in principle): >>> r = AllChem.EnumerateLibraryFromReaction(rxn,[acids,amines]) ... look at the first 4 compounds: >>> [Chem.MolToSmiles(next(r)[0]) for x in range(4)] ['NC=O', 'CNC=O', 'CCNC=O', 'CCCNC=O'] z#%d sidechains provided, %d requiredrc3sbx\||D]P}|dt|ƒkrRˆ||dƒ}x,|D]}|g}| |¡|Vq2Wq |gVq WdS)Nr)r+Úextend)ÚitemsZdepthÚitemÚvÚentryÚl)Ú_combiEnumeratorr rrFës   z6EnumerateLibraryFromReaction.._combiEnumeratorÚProductReactantszproducts,reactantsN)r)r+ZGetNumReactantTemplatesÚ ValueErrorrZ RunReactants)ZreactionZ sidechainSetsZreturnReactantsrGZchainsZprodSetsZprodsr )rFrÚEnumerateLibraryFromReactionÄs#   rITi& c Ks| |¡}|stdƒ‚i}| |¡} x&t|ƒD]\} } |  | ¡} | || <q.Wt|f||dœ|—Ž} | dkrrtdƒ‚dd„t|ƒDƒ}|s<||dd}x\t|ƒD]P\} } xFt| dt|ƒƒD]0}||}||  ||¡}|  | |||d ¡q¼Wq W|  ¡d }|  ¡}x"|r*|r*|  ¡}|d8}q Wt |||d }nÈt |||d }||dd}| ¡}xRt|  ¡ƒD]B} | | ¡}|j|j|j|jd d d}|  ||| ddd ¡qlW|  ¡d }|j ddd}x(|rô|rô|j ddd}|d8}qÎWt |||d }| dt|ƒ¡|S)aX generates an embedding of a molecule where part of the molecule is constrained to have particular coordinates Arguments - mol: the molecule to embed - core: the molecule to use as a source of constraints - useTethers: (optional) if True, the final conformation will be optimized subject to a series of extra forces that pull the matching atoms to the positions of the core atoms. Otherwise simple distance constraints based on the core atoms will be used in the optimization. - coreConfId: (optional) id of the core conformation to use - randomSeed: (optional) seed for the random number generator An example, start by generating a template with a 3D structure: >>> from rdkit.Chem import AllChem >>> template = AllChem.MolFromSmiles("c1nn(Cc2ccccc2)cc1") >>> AllChem.EmbedMolecule(template) 0 >>> AllChem.UFFOptimizeMolecule(template) 0 Here's a molecule: >>> mol = AllChem.MolFromSmiles("c1nn(Cc2ccccc2)cc1-c3ccccc3") Now do the constrained embedding >>> mol = AllChem.ConstrainedEmbed(mol, template) Demonstrate that the positions are nearly the same with template: >>> import math >>> molp = mol.GetConformer().GetAtomPosition(0) >>> templatep = template.GetConformer().GetAtomPosition(0) >>> all(math.isclose(v, 0.0, abs_tol=0.01) for v in molp-templatep) True >>> molp = mol.GetConformer().GetAtomPosition(1) >>> templatep = template.GetConformer().GetAtomPosition(1) >>> all(math.isclose(v, 0.0, abs_tol=0.01) for v in molp-templatep) True zmolecule doesn't match the core)ÚcoordMapZ randomSeedrzCould not embed molecule.cSsg|]\}}||f‘qSr r )rr9r>r r rr;sz$ConstrainedEmbed..)rrgY@é)ZatomMapT)Zfixedg-Cëâ6?gü©ñÒMbP?)Z energyTolZforceTolZEmbedRMS)ÚGetSubstructMatchrHrÚ enumerater5Z EmbedMoleculer3r+r6ZAddDistanceConstraintZ InitializeZMinimizeZAlignMolr4Z AddExtraPointr&r)r*ZSetPropÚstr)rZcoreZ useTethersZ coreConfIdZ randomseedZ getForceFieldr ÚmatchrJZcoreConfr9ZidxIZcorePtIZciZalgMapZffr>ZidxJr:ÚnZmoreZrmsr,ÚpZpIdxr r rÚConstrainedEmbedsT.       rRc CsÄt |¡}t |¡}| |¡}|sÀx4| ¡D](}| ¡tjkr.| tj¡| d¡q.Wx&| ¡D]}| tj¡| d¡qdWx|  ¡D]}|  d¡qŒWx|  ¡D]}|  d¡q¨W|j |dd}|r¸t |ƒdkrät  d¡|d}xP| ¡D]D}|| ¡}|| ¡}| ||¡} |  | ¡¡|  | ¡¡qöWx\|  ¡D]P}| || ¡¡} |  | ¡¡|  | ¡¡|  | ¡¡|   | ¡¡qHWt|ƒt|dƒrÀd|_ntdƒ‚|S) aO assigns bond orders to a molecule based on the bond orders in a template molecule Arguments - refmol: the template molecule - mol: the molecule to assign bond orders to An example, start by generating a template from a SMILES and read in the PDB structure of the molecule >>> import os >>> from rdkit.Chem import AllChem >>> template = AllChem.MolFromSmiles("CN1C(=NC(C1=O)(c2ccccc2)c3ccccc3)N") >>> mol = AllChem.MolFromPDBFile(os.path.join(RDConfig.RDCodeDir, 'Chem', 'test_data', '4DJU_lig.pdb')) >>> len([1 for b in template.GetBonds() if b.GetBondTypeAsDouble() == 1.0]) 8 >>> len([1 for b in mol.GetBonds() if b.GetBondTypeAsDouble() == 1.0]) 22 Now assign the bond orders based on the template molecule >>> newMol = AllChem.AssignBondOrdersFromTemplate(template, mol) >>> len([1 for b in newMol.GetBonds() if b.GetBondTypeAsDouble() == 1.0]) 8 Note that the template molecule should have no explicit hydrogens else the algorithm will fail. It also works if there are different formal charges (this was github issue 235): >>> template=AllChem.MolFromSmiles('CN(C)C(=O)Cc1ccc2c(c1)NC(=O)c3ccc(cc3N2)c4ccc(c(c4)OC)[N+](=O)[O-]') >>> mol = AllChem.MolFromMolFile(os.path.join(RDConfig.RDCodeDir, 'Chem', 'test_data', '4FTR_lig.mol')) >>> AllChem.MolToSmiles(mol) 'COC1CC(C2CCC3C(O)NC4CC(CC(O)N(C)C)CCC4NC3C2)CCC1N(O)O' >>> newMol = AllChem.AssignBondOrdersFromTemplate(template, mol) >>> AllChem.MolToSmiles(newMol) 'COc1cc(-c2ccc3c(c2)Nc2ccc(CC(=O)N(C)C)cc2NC3=O)ccc1[N+](=O)[O-]' Fr)Zuniquifyrz2More than one matching pattern found - picking oneÚ __sssAtomsNzNo matching found)r'r(rLZGetBondsZ GetBondTypeZBondTypeZSINGLEZ SetBondTypeZ SetIsAromaticZGetAtomsZSetFormalChargeZGetSubstructMatchesr+r ZwarningZGetBeginAtomIdxZ GetEndAtomIdxZGetBondBetweenAtomsZ GetIsAromaticZGetAtomWithIdxZGetIdxZSetHybridizationZGetHybridizationZSetNumExplicitHsZGetNumExplicitHsZGetFormalChargeZ SanitizeMolÚhasattrrSrH) ZrefmolrZrefmol2Zmol2ZmatchingÚbÚaZatom1Zatom2Zb2Za2r r rÚAssignBondOrdersFromTemplatebsJ(           rWcCs2ddl}ddl}|j|j|d\}}| |¡dS)Nr)Z optionflagsÚverbose)ÚsysÚdoctestZtestmodÚELLIPSISÚexit)rXrYrZZfailedÚ_r r rÚ _runDoctests¾sr^Ú__main__)r F)r rr)r r"r#)NF)NF)F)N)0Ú__doc__rYÚwarningsÚ collectionsrr7ZrdkitrrrrZ rdkit.ChemZrdkit.Chem.ChemicalFeaturesZrdkit.Chem.rdChemReactionsZrdkit.Chem.rdDepictorZrdkit.Chem.rdDistGeomZrdkit.Chem.rdForceFieldHelpersZrdkit.Chem.rdMolAlignZrdkit.Chem.rdMolDescriptorsZrdkit.Chem.rdMolTransformsZrdkit.Chem.rdPartialChargesZrdkit.Chem.rdReducedGraphsZrdkit.Chem.rdShapeHelpersZrdkit.Chem.rdqueriesZrdkit.Chem.rdMolEnumeratorZrdkit.GeometryrZrdkit.RDLoggerr Z!rdkit.Chem.EnumerateStereoisomersr r Zrdkit.Chem.rdSLNParseÚ ImportErrorZCompute2DCoordsr(ZComputeGasteigerChargesrr!r.r;r?rIZUFFGetMoleculeForceFieldrRrWr^Ú__name__r r r rÚ sX           # $ 1 < a\