def __init__( self, fdcd, fref, box=0, pdbCode=None, log=StdLog(), verbose=0): """ @param fdcd: path to input dcd file @type fdcd: str @param fref: PDB or pickled PDBModel or directly an open PDBModel instancewith same atom content and order @type fref: str or PDBModel @param box: expect line with box info at the end of each frame (default: 0) @type box: 1|0 @param pdbCode: pdb code to be put into the model (default: None) @type pdbCode: str @param log: LogFile instance [Biskit.StdLog] @type log: Biskit.LogFile @param verbose: print progress to log [0] @type verbose: int """ self.fdcd = T.absfile( fdcd ) self.dcd = open(self.fdcd, "r", 0) if isinstance(fref, str) : self.ref=PDBModel(T.absfile(fref), pdbCode=pdbCode) elif fref : self.ref = fref self.box = box self.n = self.ref.lenAtoms() self.log = log self.verbose = verbose self.readHeader() self.set_pointerInfo()
def test_Capping(self): """PDBCleaner.capTerminals test""" ## Loading PDB... self.model = PDBModel(t.testRoot() + '/rec/1A2P_rec_original.pdb') self.c = PDBCleaner(self.model, log=self.log, verbose=self.local) self.m2 = self.c.capTerminals(breaks=True) self.assert_(self.m2.atomNames() == self.model.atomNames()) self.m3 = self.model.clone() self.m3.removeRes([10, 11, 12, 13, 14, 15]) self.m4 = self.m3.clone() self.c = PDBCleaner(self.m3, log=self.log, verbose=self.local) self.m3 = self.c.capTerminals(breaks=True, capC=[0], capN=[0, 1]) self.assertEqual( self.m3.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN') if self.local: self.log.add('\nTesting automatic chain capping...\n') self.c = PDBCleaner(self.m4, log=self.log, verbose=self.local) self.m4 = self.c.capTerminals(auto=True) self.assertEqual( self.m4.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN')
def prepare_target(self, cluster, output_folder = None): """ Create the 'target.fasta' file for each template to validate @param cluster: name of the cluster which is used for the foldder name in which the validation is run. @type cluster: str @param output_folder: top output folder (default: None -> L{F_RESULT_FOLDER}) @type output_folder: str """ output_folder = output_folder or self.outFolder + \ self.F_RESULT_FOLDER + '/%s/'%cluster target = open("%s"%(output_folder + self.F_TEMPLATE_SEQUENCE),'w') target.write(">target\n") for pdb in self.pdb_path: if(cluster == os.path.split(pdb)[1][0:4]): model = PDBModel('%s'%pdb) sequence = model.sequence() sequence = MU.format_fasta(seq = sequence) target.write("%s"%sequence) target.close()
def prepare_target(self, cluster, output_folder = None): """ Create the 'target.fasta' file for each template to validate @param cluster: name of the cluster which is used for the foldder name in which the validation is run. @type cluster: str @param output_folder: top output folder (default: None -> L{F_RESULT_FOLDER}) @type output_folder: str """ output_folder = output_folder or self.outFolder + \ self.F_RESULT_FOLDER + '/%s/'%cluster target = open("%s"%(output_folder + self.F_TEMPLATE_SEQUENCE),'w') target.write(">target\n") for pdb in self.pdb_path: if(cluster == os.path.split(pdb)[1][0:4]): model = PDBModel('%s'%pdb) sequence = model.sequence() sequence = MU.format_fasta(seq = sequence) target.write("%s"%sequence) target.close()
def test_capping_extra(self): """PDBCleaner.capTerminals extra challenge""" self.m2 = PDBModel(t.testRoot() + '/pdbclean/foldx_citche.pdb') self.c = PDBCleaner(self.m2, verbose=self.local, log=self.log) self.assertRaises(CappingError, self.c.capTerminals, auto=True) if self.local: self.log.add('OK: CappingError has been raised indicating clash.') self.assertEqual(len(self.m2.takeChains([1]).chainBreaks()), 1)
def prepare(self): root = T.testRoot() + '/amber/' self.ref = PDBModel(T.testRoot() + '/amber/1HPT_0.pdb') self.refdry = root + '1HPT_0dry.pdb' self.dryparm = tempfile.mktemp('.parm', 'dry_') self.drycrd = tempfile.mktemp('.crd', 'dry_') self.drypdb = tempfile.mktemp('.pdb', 'dry_') self.wetparm = tempfile.mktemp('.parm', 'wet_') self.wetcrd = tempfile.mktemp('.crd', 'wet_') self.wetpdb = tempfile.mktemp('.pdb', 'wet_') self.leapout = tempfile.mktemp('.out', 'leap_')
def __init__(self, fpdb, log=None, verbose=True): """ @param fpdb: pdb file OR PDBModel instance @type fpdb: str OR Biskit.PDBModel @param log: Biskit.LogFile object (default: STDOUT) @type log: Biskit.LogFile @param verbose: log warnings and infos (default: True) @type verbose: bool """ self.model = PDBModel(fpdb) self.log = log or StdLog() self.verbose = verbose
def test_Capping( self ): """PDBCleaner.capTerminals test""" ## Loading PDB... self.model = PDBModel(t.testRoot() + '/rec/1A2P_rec_original.pdb') self.c = PDBCleaner( self.model, log=self.log, verbose=self.local ) self.m2 = self.c.capTerminals( breaks=True ) self.assert_( self.m2.atomNames() == self.model.atomNames() ) self.m3 = self.model.clone() self.m3.removeRes( [10,11,12,13,14,15] ) self.m4 = self.m3.clone() self.c = PDBCleaner( self.m3, log=self.log, verbose=self.local ) self.m3 = self.c.capTerminals( breaks=True, capC=[0], capN=[0,1]) self.assertEqual( self.m3.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN' ) if self.local: self.log.add( '\nTesting automatic chain capping...\n' ) self.c = PDBCleaner( self.m4, log=self.log, verbose=self.local ) self.m4 = self.c.capTerminals( auto=True ) self.assertEqual( self.m4.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN' )
def test_capIrregular(self): """AmberParmBuilder.capNME & capACE test""" gfp = PDBModel('1GFL') normal = gfp.takeResidues([10, 11]) chromo = gfp.takeResidues([64, 65]) self.a = AmberParmBuilder(normal) self.m4 = self.a.capACE(normal, 0) self.assertEqual(len(self.m4), 17) ## del chromo.residues['biomol'] self.m5 = self.a.capACE(chromo, 0) self.m5 = self.a.capNME(self.m5, 0) self.assertEqual(self.m5.sequence(), 'XSYX')
def test_capIrregular( self ): """AmberParmBuilder.capNME & capACE test""" gfp = PDBModel('1GFL') normal = gfp.takeResidues([10,11]) chromo = gfp.takeResidues([64,65]) self.a = AmberParmBuilder( normal ) self.m4 = self.a.capACE( normal, 0 ) self.assertEqual( len(self.m4), 17 ) ## del chromo.residues['biomol'] self.m5 = self.a.capACE( chromo, 0 ) self.m5 = self.a.capNME( self.m5, 0 ) self.assertEqual( self.m5.sequence(), 'XSYX' )
def go(self, output_folder = None, template_folder = None): """ Run analysis of models. @param output_folder: folder for result files (default: None S{->} outFolder/L{F_RESULT_FOLDER}) @type output_folder: str @param template_folder: folder with template structures (default: None S{->} outFolder/L{VS.F_RESULT_FOLDER}) @type template_folder: str """ ## pdb_list = T.load(self.outFolder + self.F_MODELS) model = PDBModel(pdb_list[0]) ## output_folder = output_folder or self.outFolder + self.F_RESULT_FOLDER template_folder = template_folder or self.outFolder +VS.F_RESULT_FOLDER templates = self.__listDir(template_folder) ## global_rmsd_aa_wo_if, global_rmsd_aa_if = self.global_rmsd_aa() global_rmsd_ca_wo_if, global_rmsd_ca_if = self.global_rmsd_ca() nb_templates = len(templates)-1 identities = self.get_identities(nb_templates) score = self.get_score() self.output_values(global_rmsd_aa_wo_if, global_rmsd_aa_if, global_rmsd_ca_wo_if, global_rmsd_ca_if, identities, score, nb_templates) ## aln_dic = self.get_aln_info(output_folder=self.outFolder) template_rmsd_dic = self.get_templates_rmsd(templates) templates_profiles = self.templates_profiles(templates, aln_dic, template_rmsd_dic) mean_rmsd = self.output_cross_val(aln_dic, templates_profiles, templates, model) ## mean_rmsd_atoms = model.res2atomProfile(mean_rmsd) self.updatePDBs_charge(mean_rmsd_atoms, model)
def go(self, output_folder=None, template_folder=None): """ Run analysis of models. @param output_folder: folder for result files (default: None S{->} outFolder/L{F_RESULT_FOLDER}) @type output_folder: str @param template_folder: folder with template structures (default: None S{->} outFolder/L{VS.F_RESULT_FOLDER}) @type template_folder: str """ ## pdb_list = T.load(self.outFolder + self.F_MODELS) model = PDBModel(pdb_list[0]) ## output_folder = output_folder or self.outFolder + self.F_RESULT_FOLDER template_folder = template_folder or self.outFolder + VS.F_RESULT_FOLDER templates = self.__listDir(template_folder) ## global_rmsd_aa_wo_if, global_rmsd_aa_if = self.global_rmsd_aa() global_rmsd_ca_wo_if, global_rmsd_ca_if = self.global_rmsd_ca() nb_templates = len(templates) - 1 identities = self.get_identities(nb_templates) score = self.get_score() self.output_values(global_rmsd_aa_wo_if, global_rmsd_aa_if, global_rmsd_ca_wo_if, global_rmsd_ca_if, identities, score, nb_templates) ## aln_dic = self.get_aln_info(output_folder=self.outFolder) template_rmsd_dic = self.get_templates_rmsd(templates) templates_profiles = self.templates_profiles(templates, aln_dic, template_rmsd_dic) mean_rmsd = self.output_cross_val(aln_dic, templates_profiles, templates, model) ## mean_rmsd_atoms = model.res2atomProfile(mean_rmsd) self.updatePDBs_charge(mean_rmsd_atoms, model)
def test_capping_extra( self ): """PDBCleaner.capTerminals extra challenge""" self.m2 = PDBModel( t.testRoot() + '/pdbclean/foldx_citche.pdb' ) self.c = PDBCleaner( self.m2, verbose=self.local, log=self.log) self.assertRaises(CappingError, self.c.capTerminals, auto=True) if self.local: self.log.add('OK: CappingError has been raised indicating clash.' ) self.assertEqual( len(self.m2.takeChains([1]).chainBreaks()), 1 )
def __init__(self, model, leap_template=F_leap_in, leaprc=None, leap_out=None, leap_in=None, leap_pdb=None, log=None, debug=0, verbose=0, **kw): """ @param model: model @type model: PDBModel or str @param leap_template: path to template file for leap input @type leap_template: str @param leaprc: forcefield parameter file or code (e.g. ff99) @type leaprc: str @param leap_out: target file for leap.log (default: discard) @type leap_out: str @param leap_in: target file for leap.in script (default: discard) @type leap_in: str @param kw: kw=value pairs for additional options in the leap_template @type kw: key=value """ self.m = PDBModel(model) self.leap_template = leap_template self.leaprc = leaprc self.leap_pdb = leap_pdb or tempfile.mktemp('_leap_pdb') self.keep_leap_pdb = leap_pdb is not None self.leap_in = leap_in self.leap_out = leap_out self.log = log or StdLog() self.output = None # last output of leap self.debug = debug self.verbose = verbose self.__dict__.update(kw)
def test_AmberParmMirror(self): """AmberParmBuilder.parmMirror test""" ref = self.ref mask = N.logical_not(ref.maskH2O()) ## keep protein and Na+ ion self.mdry = ref.compress(mask) self.a = AmberParmBuilder(self.mdry, verbose=self.local, leap_out=self.leapout, debug=self.DEBUG) self.a.parmMirror(f_out=self.dryparm, f_out_crd=self.drycrd) self.a.parm2pdb(self.dryparm, self.drycrd, self.drypdb) self.m1 = PDBModel(self.drypdb) self.m2 = PDBModel(self.refdry) eq = N.array(self.m1.xyz == self.m2.xyz) self.assert_(eq.all())
def prepare(self): root = T.testRoot() + '/amber/' self.ref = PDBModel( T.testRoot() + '/amber/1HPT_0.pdb') self.refdry = root + '1HPT_0dry.pdb' self.dryparm = tempfile.mktemp('.parm', 'dry_') self.drycrd = tempfile.mktemp('.crd', 'dry_') self.drypdb = tempfile.mktemp('.pdb', 'dry_') self.wetparm = tempfile.mktemp('.parm', 'wet_') self.wetcrd = tempfile.mktemp('.crd', 'wet_') self.wetpdb = tempfile.mktemp('.pdb', 'wet_') self.leapout = tempfile.mktemp('.out', 'leap_')
def __init__( self, fpdb, log=None, verbose=True ): """ @param fpdb: pdb file OR PDBModel instance @type fpdb: str OR Biskit.PDBModel @param log: Biskit.LogFile object (default: STDOUT) @type log: Biskit.LogFile @param verbose: log warnings and infos (default: True) @type verbose: bool """ self.model = PDBModel( fpdb ) self.log = log or StdLog() self.verbose = verbose
def prepareRef(self, fname): """ Prepare reference model. @param fname: file name @type fname: str @return: reference structure @rtype: PDBModel|Complex @raise EntropistError: if unknown reference type """ if not fname: return None if self.__splitFilenames(fname): f1, f2 = self.__splitFilenames(fname) m1, m2 = PDBModel( self.__getModel(f1) ), \ PDBModel( self.__getModel(f2) ) ref = Complex(m1, m2) else: ref = t.load(fname) if isinstance(ref, Trajectory): ref = ref.ref if isinstance(ref, PDBModel): return self.__cleanAtoms(ref) if isinstance(ref, Complex): self.__cleanAtoms(ref.rec_model) self.__cleanAtoms(ref.lig_model) ref.lig_model_transformed = None return ref raise EntropistError, 'unknown reference type'
def test_Benchmark(self): """Mod.Benchmark test""" from Biskit import Pymoler self.b = Benchmark(self.outfolder) self.b.go() pdb = T.load(self.outfolder + "/modeller/PDBModels.list")[0] reference = PDBModel(self.outfolder + "/reference.pdb") tmp_model = pdb.clone() reference = reference.compress(reference.maskCA()) pdb = pdb.compress(pdb.maskCA()) tmp_model = tmp_model.compress(tmp_model.maskCA()) tm = tmp_model.transformation(reference, n_it=0, profname="rms_outliers") pdb = pdb.transform(tm) if self.local: pm = Pymoler() pm.addPdb(pdb, "m") pm.addPdb(reference, "r") pm.colorAtoms("m", tmp_model.profile("rms_outliers")) pm.add('set ribbon_trace,1') pm.add('show ribbon') pm.show() if self.DEBUG: self.log.add( 'The result from the benchmarking is in %s/benchmark'%\ self.outfolder) globals().update(locals())
def test_Benchmark(self): """Mod.Benchmark test""" from Biskit import Pymoler self.b = Benchmark( self.outfolder ) self.b.go() pdb = T.load( self.outfolder + "/modeller/PDBModels.list" )[0] reference = PDBModel(self.outfolder + "/reference.pdb" ) tmp_model = pdb.clone() reference = reference.compress( reference.maskCA() ) pdb = pdb.compress( pdb.maskCA() ) tmp_model = tmp_model.compress(tmp_model.maskCA()) tm = tmp_model.transformation( reference, n_it=0, profname="rms_outliers") pdb = pdb.transform( tm ) if self.local: pm = Pymoler() pm.addPdb( pdb, "m" ) pm.addPdb( reference, "r" ) pm.colorAtoms( "m", tmp_model.profile("rms_outliers") ) pm.add('set ribbon_trace,1') pm.add('show ribbon') pm.show() if self.DEBUG: self.log.add( 'The result from the benchmarking is in %s/benchmark'%\ self.outfolder) globals().update( locals() )
def prepareSource(inFile, outFile, wat=1, sort=1, foldx=1, surf=1, dens=1, cons=1, dssp=1, delphi=0): """ Strip waters, add profiles and save as doped source model. """ source = PDBModel(inFile) if wat: source.remove(lambda a: a['residue_name'] in ['HOH', 'WAT', 'TIP3']) if sort: source = source.sort() doper = PDBDope(source) if surf: ## doper.addASA() ## doper.addSurfaceMask() doper.addSurfaceRacer(probe=1.4) if foldx: doper.addFoldX() if dens: doper.addDensity() if dssp: doper.addSecondaryStructure() if delphi: doper.addDelphi() try: if cons: doper.addConservation() except: errWriteln('\n ERROR: Conservation profile could not be added to '\ + str(sourceOut) + '\n' ) source.saveAs(outFile) return source
def test_AmberParmSolvated(self): """AmberParmBuilder.parmSolvated test""" ## remove waters and hydrogens self.mdry = self.ref.compress(self.ref.maskProtein()) self.mdry = self.mdry.compress(self.mdry.maskHeavy()) self.a = AmberParmBuilder(self.mdry, leap_out=self.leapout, verbose=self.local, debug=self.DEBUG) self.a.parmSolvated(self.wetparm, f_out_crd=self.wetcrd, f_out_pdb=self.wetpdb, box=2.5) self.m3 = PDBModel(self.wetpdb) m3prot = self.m3.compress(self.m3.maskProtein()) refprot = self.ref.compress(self.ref.maskProtein()) refprot.xplor2amber() self.assertEqual(self.ref.lenChains(), self.m3.lenChains()) self.assertEqual(refprot.atomNames(), m3prot.atomNames())
def prepare_templatesfasta(self, cluster_list, pdb_dictionary, output_folder=None): """ Create 'templates.fasta' file for each template to validate @param cluster_list: pdb codes of templates @type cluster_list: [str] @param pdb_dictionary: dictionary mapping pdb code to pdb files used by Modeller @type pdb_dictionary: {str:str} @param output_folder: top output folder (default: None -> L{F_RESULT_FOLDER}) @type output_folder: str """ output_folder = output_folder or self.outFolder + self.F_RESULT_FOLDER for cluster in cluster_list: folder = '%s/%s'%(output_folder, cluster + \ TemplateSearcher.F_RESULT_FOLDER) if not os.path.exists(folder): os.mkdir(folder) else: print 'Directory %s exists, skipping'%( cluster + \ TemplateSearcher.F_RESULT_FOLDER) pdb_path = pdb_dictionary["%s" % cluster] PDBModels_list = [] pdb_name = [] for pdb in pdb_path: PDBModels_list.append(PDBModel('%s' % pdb)) pdb_name.append(os.path.split(pdb)[1][:-4]) input_file = self.outFolder + self.F_RESULT_FOLDER + \ '/%s'%cluster + TemplateSearcher.F_RESULT_FOLDER \ + self.F_TEMPLATES_FASTA templatesfasta = open("%s" % input_file, 'w') for i in range(len(PDBModels_list)): templatesfasta.write(">%s\n" % pdb_name[i]) sequence = PDBModels_list[i].sequence() sequence = MU.format_fasta(seq=sequence) templatesfasta.write("%s\n" % sequence) templatesfasta.close()
def __init__( self, model, leap_template=F_leap_in, leaprc=None, leap_out=None, leap_in=None, leap_pdb=None, log=None, debug=0, verbose=0, **kw ): """ @param model: model @type model: PDBModel or str @param leap_template: path to template file for leap input @type leap_template: str @param leaprc: forcefield parameter file or code (e.g. ff99) @type leaprc: str @param leap_out: target file for leap.log (default: discard) @type leap_out: str @param leap_in: target file for leap.in script (default: discard) @type leap_in: str @param kw: kw=value pairs for additional options in the leap_template @type kw: key=value """ self.m = PDBModel( model ) self.leap_template = leap_template self.leaprc = leaprc self.leap_pdb = leap_pdb or tempfile.mktemp( '_leap_pdb' ) self.keep_leap_pdb = leap_pdb is not None self.leap_in = leap_in self.leap_out= leap_out self.log = log or StdLog() self.output = None # last output of leap self.debug = debug self.verbose = verbose self.__dict__.update( kw )
def test_AmberParmSolvated( self ): """AmberParmBuilder.parmSolvated test""" ## remove waters and hydrogens self.mdry = self.ref.compress( self.ref.maskProtein() ) self.mdry = self.mdry.compress( self.mdry.maskHeavy() ) self.a = AmberParmBuilder( self.mdry, leap_out=self.leapout, verbose=self.local, debug=self.DEBUG) self.a.parmSolvated( self.wetparm, f_out_crd=self.wetcrd, f_out_pdb=self.wetpdb, box=2.5 ) self.m3 = PDBModel( self.wetpdb ) m3prot = self.m3.compress( self.m3.maskProtein() ) refprot= self.ref.compress( self.ref.maskProtein() ) refprot.xplor2amber() self.assertEqual( self.ref.lenChains(), self.m3.lenChains() ) self.assertEqual( refprot.atomNames(), m3prot.atomNames() )
def prepareSource( inFile, outFile, wat=1, sort=1, surf=1, dens=1, cons=1, dssp=1, delphi=0 ): """ Strip waters, add profiles and save as doped source model. """ source = PDBModel( inFile ) if wat: source.remove( lambda a: a['residue_name'] in ['HOH','WAT','TIP3'] ) if sort: source = source.sort() doper = PDBDope( source ) if surf: ## doper.addASA() ## doper.addSurfaceMask() doper.addSurfaceRacer( probe=1.4 ) ## if foldx: ## doper.addFoldX() if dens: doper.addDensity() if dssp: doper.addSecondaryStructure() if delphi: doper.addDelphi() try: if cons: doper.addConservation( ) except: errWriteln('\n ERROR: Conservation profile could not be added to '\ + str(sourceOut) + '\n' ) source.saveAs( outFile ) return source
class Test(BT.BiskitTest): """Test class """ def prepare(self): from Biskit.LogFile import LogFile import tempfile def test_PDBCleaner( self ): """PDBCleaner general test""" ## Loading PDB... self.c = PDBCleaner( t.testRoot() + '/rec/1A2P_rec_original.pdb', log=self.log, verbose=self.local) self.m = self.c.process() self.assertAlmostEqual( self.m.mass(), 34029.0115499993, 7 ) def test_DNACleaning( self ): """PDBCleaner DNA test""" ## Loading PDB... self.c = PDBCleaner( t.testRoot() + 'amber/entropy/0_com.pdb', log=self.log, verbose=self.local ) self.dna = self.c.process(amber=True) self.assertAlmostEqual( self.dna.mass(), 26953.26, 1 ) def test_Capping( self ): """PDBCleaner.capTerminals test""" ## Loading PDB... self.model = PDBModel(t.testRoot() + '/rec/1A2P_rec_original.pdb') self.c = PDBCleaner( self.model, log=self.log, verbose=self.local ) self.m2 = self.c.capTerminals( breaks=True ) self.assert_( self.m2.atomNames() == self.model.atomNames() ) self.m3 = self.model.clone() self.m3.removeRes( [10,11,12,13,14,15] ) self.m4 = self.m3.clone() self.c = PDBCleaner( self.m3, log=self.log, verbose=self.local ) self.m3 = self.c.capTerminals( breaks=True, capC=[0], capN=[0,1]) self.assertEqual( self.m3.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN' ) if self.local: self.log.add( '\nTesting automatic chain capping...\n' ) self.c = PDBCleaner( self.m4, log=self.log, verbose=self.local ) self.m4 = self.c.capTerminals( auto=True ) self.assertEqual( self.m4.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN' ) def test_capping_extra( self ): """PDBCleaner.capTerminals extra challenge""" self.m2 = PDBModel( t.testRoot() + '/pdbclean/foldx_citche.pdb' ) self.c = PDBCleaner( self.m2, verbose=self.local, log=self.log) self.assertRaises(CappingError, self.c.capTerminals, auto=True) if self.local: self.log.add('OK: CappingError has been raised indicating clash.' ) self.assertEqual( len(self.m2.takeChains([1]).chainBreaks()), 1 )
class NamdDCDParser: def __init__( self, fdcd, fref, box=0, pdbCode=None, log=StdLog(), verbose=0): """ @param fdcd: path to input dcd file @type fdcd: str @param fref: PDB or pickled PDBModel or directly an open PDBModel instancewith same atom content and order @type fref: str or PDBModel @param box: expect line with box info at the end of each frame (default: 0) @type box: 1|0 @param pdbCode: pdb code to be put into the model (default: None) @type pdbCode: str @param log: LogFile instance [Biskit.StdLog] @type log: Biskit.LogFile @param verbose: print progress to log [0] @type verbose: int """ self.fdcd = T.absfile( fdcd ) self.dcd = open(self.fdcd, "r", 0) if isinstance(fref, str) : self.ref=PDBModel(T.absfile(fref), pdbCode=pdbCode) elif fref : self.ref = fref self.box = box self.n = self.ref.lenAtoms() self.log = log self.verbose = verbose self.readHeader() self.set_pointerInfo() def readHeader(self, verbose=False): """ Read NAMD DCD coordinates file Only for 32bit DCD with opposite endianness! AND only if no atom is fixed! """ f = self.dcd unpack = struct.unpack # Check we are in the beggining of the file # or move there to read the header if f.tell() != 0: f.seek(0) # First read header information and check correct file format header = struct.unpack(">I 4s 9I f 11I", f.read(92)) if header[0] == 84 and header[1] == 'CORD' and header[-1] == 84 and header[-2] != 0: #print "recognized 32 bit DCD file of opposite endianness" #Store the number of sets of coordinates (NSET). Frames self.nset = header[2] #Store ISTART, the starting timestep self.istart = header[3] #Store NSAVC, the number of steps between dcd saves self.nsavc = header[4] # Store NTOT, number of total simulation steps self.ntot = header[5] #Store DELTA, the time step of simulation self.delta = header[11] # Have box information? self.has_extrablock = bool(header[12]) # Have a 4th dimension? self.had_4dims = bool(header[13]) else: f.close() sys.exit("Bad DCD Format") # Read title information if (unpack('>I',f.read(4))[0] - 4) % 80 == 0: # Number of title lines self.ntitle = int(unpack('>I', f.read(4))[0]) self.title = [unpack(">80s",f.read(80)) for i in range(self.ntitle)] f.read(4) # Skip closing block number else: f.close() sys.exit("ERROR in title. Bad DCD format") # Read number of atoms atomBlock = unpack('>3I', f.read(12)) if atomBlock[-1] == 4: self.natoms = atomBlock[1] else: f.close() sys.exit("Bad DCD format") if verbose: print self.title print "Number of atoms:", self.natoms print "Number of frames:",self.nset print "Starting timestep", self.istart print "Final timestep:", self.ntot print "Steps between frames:", self.nsavc print "Time step of simulation:", self.delta def set_pointerInfo(self): """ Store sizes for browsing the file later """ # Header size is: 116 + 80* self.ntitle self.h_size = 116 + (80 * self.ntitle) # Frame size # 4 bytes because it's floats per 3 axis per total num of atoms # Add the enclosing integers (two for each axis) = 6 * 4 f_size = (3 * 4 * self.natoms) + 24 if self.has_extrablock: f_size += 56 self.f_size = f_size def read_charmm_extrablock(self): f = self.dcd unpack = struct.unpack # This block contains the box information if unpack('>I', f.read(4))[0] == 48: self.unitcell = npy.fromstring( f.read(48), dtype=">d") f.read(4) else: f.close() sys.exit("ERROR in read_charmm_extrablock(). Bad DCD Format") def read_dcdstep(self): f = self.dcd size = struct.calcsize # If there is box information if self.has_extrablock: self.read_charmm_extrablock() # Read coordinates # Each coordinates block is enclosed by one integer # that we will skip all the times xyz = npy.zeros([self.natoms, 3], dtype=">f4") f.read(4) xyz[:,0] = npy.fromstring(f.read(size('f')*self.natoms), dtype=">f4") f.read(8) xyz[:,1] = npy.fromstring(f.read(size('f')*self.natoms), dtype=">f4") f.read(8) xyz[:,2] = npy.fromstring(f.read(size('f')*self.natoms), dtype=">f4") f.read(4) return xyz def read_all(self): """ Read all snapshots """ # Go to the beggining of the frames f = self.dcd f.seek(self.h_size) # Read Frames all_snap = npy.zeros([self.nset, self.natoms, 3], dtype=">f4") for i in range(self.nset): all_snap[i,:] = self.read_dcdstep() return all_snap def close(self): self.dcd.close() def getFrame(self, i): """ Read specific frame. """ f = self.dcd # Calculate pointer position # for the desired frame and move there pointer = self.h_size + (self.f_size * i) f.seek(pointer) return self.read_dcdstep() def __getitem__(self, i): return self.getFrame(i)
def randomSurfaces( base_folder, label, mask ): """ calculate surfaces for all peptides and return the average and SD """ ## container for results and standard deviations MS, AS = {}, {} MS_sd, AS_sd = {}, {} ## loop over peptide directories for k in MOU.aaAtoms.keys(): dir = base_folder + 'GLY-%s-GLY_pcr/pcr_00'%(k) fLst = glob.glob( dir + '/*.pdb') msLst = [] asLst = [] ## loop over pdb files for each peptide T.flushPrint( '\nNow collecting data in %s'%dir ) for f in fLst: ## load peptide and remove waters and hydrogens m = PDBModel( f ) m = m.compress( m.maskProtein() * m.maskHeavy() ) T.flushPrint( '.') ## add surface data try: d = PDBDope( m ) d.addSurfaceRacer( probe=1.4 ) ## remove tailing GLY m = m.compress( m.res2atomMask(mask) ) ## collect surface data for each peptide msLst += [ m.profile('MS') ] asLst += [ m.profile('AS') ] except: print 'Failed calculating exposure for GLY-%s-GLY'%(k) print '\t and file %s'%f ## get result dictionary for peptide T.flushPrint('\nCollecting data ...\n') msDic = {} asDic = {} msDic_sd = {} asDic_sd = {} j = 0 #atoms = [ a['name'] for a in m.atoms ] for n in m['name']: msDic[n] = N.average(msLst)[j] asDic[n] = N.average(asLst)[j] msDic_sd[n] = MAU.SD( msLst )[j] asDic_sd[n] = MAU.SD( asLst )[j] j += 1 MS[ k ] = msDic AS[ k ] = asDic MS_sd[ k ] = msDic_sd AS_sd[ k ] = asDic_sd return MS, AS, MS_sd, AS_sd
def changeModel( inFile, prefix, sourceModel ): print '\nget ' + os.path.basename( inFile ) + '..', model = PDBModel( inFile ) model.update() model = model.sort() eq = model.equals( sourceModel ) if not eq[0] and eq[1]: raise ConvertError('source and other models are not equal: ' + str(eq)) # model.validSource() model.setSource( sourceModel.validSource() ) #model.atomsChanged = 0 for k in model.atoms: model.atoms[k,'changed'] = N0.all( model[k] == sourceModel[k] ) model.xyzChanged = ( 0 != N0.sum( N0.ravel( model.xyz - sourceModel.xyz)) ) model.update( updateMissing=1 ) if model.xyzChanged: doper = PDBDope( model ) if 'MS' in sourceModel.atoms.keys(): doper.addSurfaceRacer( probe=1.4 ) if 'density' in sourceModel.atoms.keys(): doper.addDensity() ## if 'foldX' in sourceModel.info.keys(): ## doper.addFoldX() if 'delphi' in sourceModel.info.keys(): doper.addDelphi() outFile = os.path.dirname( inFile ) + '/' + prefix +\ T.stripFilename( inFile ) + '.model' T.dump( model, outFile ) print '-> ' + os.path.basename( outFile )
class AmberParmBuilder: """ AmberParmBuilder ================ Create Amber topology and coordinate file from PDB. - parmMirror(): ...builds a fake parm that exactly mirrors a given PDB file. This parm can be used for ptraj but not for simulations. Currently, parmMirror only accepts amber-formatted PDBs as input. It should be possible to create topologies that have the same content and order of atoms as an xplor PDB but some atoms will have different names. - parmSolvated(): ...builds a solvated system for PME simulations (incl. closing of S-S bonds, capping of chain breaks). parmSolvated accepts both xplor and amber-formatted PDBs as input. Requires the amber programs C{tleap} and C{ambpdb}. Requires leap template files in C{biskit/Biskit/data/amber/leap/}. Note on forcefields: The default forcefield used is specified in exe_tleap and currently is ff10. This translates to loading amber11/dat/leap/cmd/leaprc.ff10 at the beginning of the leap run. As of 2011, ff10 is the recommended default forcefield for proteins and nucleic acids. Comment from Jason Swails on the Amber mailing list: " Try using ff99SB (which is the protein force field part of ff10, which is the version I would actually suggest using). Despite its label, it is actually a 2006 update of the ff99 force field which performs at least as well (if not better) as ff03." Unfortunately, ions are only "half" paramterized in ff10. Additional parameters need to be loaded from a frmod file, typically frcmod.ionsjc_tip3p. There are additional versions of this file optimized for other water models than TIP3. frcmod.ionsjc_tip3p is set as the default frmod file to include by parmSolvated and parmMirror. Please include it if you provide your own list of frmod files. @note: The design of AmberParmBuilder is less than elegant. It would make more sense to split it into two classes that are both derrived from Executor. """ ## script to create a parm that exactly mirrors a given PDB script_mirror_pdb = """ logFile %(f_out)s source %(leaprc)s %(fmod)s %(fprep)s p = loadPdb %(in_pdb)s %(delete_atoms)s saveAmberParm p %(out_parm)s %(out_crd)s quit """ ## tleap command to close a single S-S bond ss_bond = "bond p.%i.SG p.%i.SG\n" ## leap script for solvated topology F_leap_in = t.dataRoot() + '/amber/leap/solvate_box.leap' ## PDB with ACE capping residue F_ace_cap = t.dataRoot() + '/amber/leap/ace_cap.pdb' ## PDB with NME capping residue F_nme_cap = t.dataRoot() + '/amber/leap/nme_cap.pdb' def __init__(self, model, leap_template=F_leap_in, leaprc=None, leap_out=None, leap_in=None, leap_pdb=None, log=None, debug=0, verbose=0, **kw): """ @param model: model @type model: PDBModel or str @param leap_template: path to template file for leap input @type leap_template: str @param leaprc: forcefield parameter file or code (e.g. ff99) @type leaprc: str @param leap_out: target file for leap.log (default: discard) @type leap_out: str @param leap_in: target file for leap.in script (default: discard) @type leap_in: str @param kw: kw=value pairs for additional options in the leap_template @type kw: key=value """ self.m = PDBModel(model) self.leap_template = leap_template self.leaprc = leaprc self.leap_pdb = leap_pdb or tempfile.mktemp('_leap_pdb') self.keep_leap_pdb = leap_pdb is not None self.leap_in = leap_in self.leap_out = leap_out self.log = log or StdLog() self.output = None # last output of leap self.debug = debug self.verbose = verbose self.__dict__.update(kw) def __runLeap(self, in_script, in_pdb, norun=0, **kw): """ Create script file and run Leap. @param in_script: content of ptraj script with place holders @type in_script: str @param in_pdb: PDB file to load into tleap @type in_pdb: str @param norun: 1 - only create leap scrip (default: 0) @type norun: 1|0 @param kw: key=value pairs for filling place holders in script @type kw: key=value @raise AmberError: if missing option for leap input file or if could not create leap input file """ x = AmberLeap(in_script, in_pdb=in_pdb, log=self.log, verbose=self.verbose, debug=self.debug, catch_out=True, f_in=self.leap_in, f_out=self.leap_out, **kw) if norun: x.generateInp() else: x.run() self.output = x.output ## ## create leap script ## try: ## ## use own fields and given kw as parameters for leap script ## d = copy.copy( self.__dict__ ) ## d.update( kw ) ## in_script = in_script % d ## f = open( self.leap_in, 'w') ## f.write( in_script ) ## f.close() ## if self.verbose: ## self.log.add('leap-script: ') ## self.log.add( in_script ) ## except IOError: ## raise AmberError('Could not create leap input file') ## except: ## raise AmberError('missing option for leap input file\n'+\ ## 'available: %s' % (str( d.keys() ) )) ## ## run tleap ## args = '-f %s' % self.leap_in ## if not norun: ## self.exe = Executor('tleap', args, log=self.log,verbose=1, ## catch_out=0) ## self.output, self.error, self.status = self.exe.run() ## if not os.path.exists( kw['out_parm'] ): ## raise AmberError, "tleap failed" ## ## clean up ## if not self.keep_leap_in and not self.debug: ## t.tryRemove( self.leap_in ) ## if not self.keep_leap_out and not self.debug: ## t.tryRemove( self.leap_out) def parm2pdb(self, f_parm, f_crd, f_out, aatm=0): """ Use ambpdb to build PDB from parm and crd. @param f_parm: existing parm file @type f_parm: str @param f_crd: existing crd file @type f_crd: str @param f_out: target file name for PDB @type f_out: str @return: f_out, target file name for PDB @rtype: str @raise AmberError: if ambpdb fail """ ## cmd = '%s -p %s -aatm < %s > %s' % \ args = '-p %s %s' % (f_parm, '-aatm' * aatm) x = Executor('ambpdb', args, f_in=f_crd, f_out=f_out, log=self.log, verbose=1, catch_err=1) output, error, status = x.run() if not os.path.exists(f_out): raise AmberError, 'ambpdb failed.' return f_out def __ssBonds(self, model, cutoff=4.): """ Identify disulfide bonds. @param model: model @type model: PDBModel @param cutoff: distance cutoff for S-S distance (default: 4.0) @type cutoff: float @return: list with numbers of residue pairs forming S-S @rtype: [(int, int)] """ m = model.compress(model.mask(['SG'])) if len(m) < 2: return [] pw = MU.pairwiseDistances(m.xyz, m.xyz) pw = N.less(pw, cutoff) r = [] for i in range(len(pw)): for j in range(i + 1, len(pw)): if pw[i, j]: r += [(m.atoms['residue_number'][i], m.atoms['residue_number'][j])] return r def __cys2cyx(self, model, ss_residues): """ Rename all S-S bonded CYS into CYX. @param model: model @type model: PDBModel @param ss_residues: original residue numbers of S-S pairs @type ss_residues: [(int, int)] """ ss = [] for a, b in ss_residues: ss += [a, b] for a in model: if a['residue_number'] in ss: a['residue_name'] = 'CYX' def capACE(self, model, chain): """ Cap N-terminal of given chain. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int """ cleaner = PDBCleaner(model, log=self.log) return cleaner.capACE(model, chain, breaks=True) def capNME(self, model, chain): """ Cap C-terminal of given chain. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int """ cleaner = PDBCleaner(model, log=self.log) return cleaner.capNME(model, chain, breaks=True) def centerModel(self, model): """ Geometric centar of model. @param model: model @type model: PDBMode """ center = N.average(model.getXyz()) model.setXyz(model.xyz - center) def leapModel(self, hetatm=0, center=True): """ Get a clean PDBModel for input into leap. @param hetatm: keep HETATM records (default: 0) @type hetatm: 1|0 @return: model @rtype: PDBMod """ m = self.m.clone() m.xplor2amber() cleaner = PDBCleaner(m, log=self.log, verbose=self.verbose) m = cleaner.process(keep_hetatoms=hetatm, amber=1) m.renumberResidues(addChainId=1) if center: self.centerModel(m) return m def __fLines(self, template, values): if not type(values) is list: values = [values] return ''.join([template % v for v in values]) def parmSolvated(self, f_out, f_out_crd=None, f_out_pdb=None, hetatm=0, norun=0, cap=0, capN=[], capC=[], fmod=['frcmod.ionsjc_tip3p'], fprep=[], box=10.0, center=True, **kw): """ @param f_out: target file for parm (topology) @type f_out: str @param f_out_crd: target file for crd (coordinates) (default:|f_out_base|.crd) @type f_out_crd: str @param f_out_pdb: target file for pdb (default:|f_out_base|.pdb) @type f_out_pdb: str @param hetatm: keep hetero atoms (default: 0) @type hetatm: 1|0 @param cap: put ACE and NME capping residue on chain breaks (default: 0) @type cap: 1|0 @param capN: indices of chains that should get ACE cap (default: []) @type capN: [int] @param capC: indices of chains that should get NME cap (default: []) @type capC: [int] @param box: minimal distance of solute from box edge (default: 10.0) @type box: float @param center: re-center coordinates (default: True) @type center: bool @param fmod: list of files with amber parameter modifications to be loaded into leap with loadAmberParams (default:['frcmod.ionsjc_tip3p'] ... mod file needed for default Amber ff10 ions -- topology saving will fail if this one is missing) @type fmod: [str] @param fprep: list of files with amber residue definitions (to be loaded into leap with loadAmberPrep) (default: []) @type fprep: [str] @param kw: additional key=value pairs for leap input template @type kw: key=value @raise IOError: """ f_out = t.absfile(f_out) f_out_crd = t.absfile(f_out_crd) or t.stripSuffix(f_out) + '.crd' f_out_pdb = t.absfile( f_out_pdb ) or t.stripSuffix( f_out ) +\ '_leap.pdb' ## removed: (bugfix 3434136) #fmod = [ t.absfile( f ) for f in t.toList( fmod ) ] #fprep = [ t.absfile( f ) for f in t.toList( fprep ) ] try: if self.verbose: self.log.add('\nCleaning PDB file for Amber:') m = self.leapModel(hetatm=hetatm, center=center) if cap: end_broken = m.atom2chainIndices(m.chainBreaks()) capC = MU.union(capC, end_broken) capN = MU.union(capN, N.array(end_broken) + 1) for i in capN: if self.verbose: self.log.add('Adding ACE cap to chain %i' % i) m = self.capACE(m, i) for i in capC: if self.verbose: self.log.add('Adding NME cap to chain %i' % i) m = self.capNME(m, i) m.renumberResidues(addChainId=1) ## again, to accomodate capping template = open(self.leap_template).read() leap_mod = self.__fLines('m = loadAmberParams %s\n', fmod) leap_prep = self.__fLines('loadAmberPrep %s\n', fprep) ss = self.__ssBonds(m, cutoff=4.) self.__cys2cyx(m, ss) leap_ss = self.__fLines(self.ss_bond, ss) if self.verbose: self.log.add('Found %i disulfide bonds: %s' % (len(ss), str(ss))) if self.verbose: self.log.add('writing cleaned PDB to %s' % self.leap_pdb) m.writePdb(self.leap_pdb, ter=3) self.__runLeap(template, in_pdb=self.leap_pdb, out_parm=f_out, out_crd=f_out_crd, ss_bonds=leap_ss, fmod=leap_mod, fprep=leap_prep, norun=norun, box=box, **kw) if not norun: parm_pdb = self.parm2pdb(f_out, f_out_crd, f_out_pdb) if not self.keep_leap_pdb and not self.debug: t.tryRemove(self.leap_pdb) except IOError, why: raise IOError, why
def go(self, model_list=None, reference=None): """ Run benchmarking. @param model_list: list of models (default: None S{->} outFolder/L{F_PDBModels}) @type model_list: ModelList @param reference: reference model (default: None S{->} outFolder/L{F_INPUT_REFERENCE}) @type reference: PDBModel """ model_list = model_list or self.outFolder + self.F_PDBModels reference = reference or self.outFolder + self.F_INPUT_REFERENCE pdb_list = T.load('%s' % model_list) reference = PDBModel(reference) # check with python 2.4 iref, imodel = reference.compareAtoms(pdb_list[0]) mask_casting = N0.zeros(len(pdb_list[0])) N0.put(mask_casting, imodel, 1) reference = reference.take(iref) #reference_mask_CA = reference_rmsd.maskCA() atom_mask = N0.zeros(len(pdb_list[0])) N0.put(atom_mask, imodel, 1) rmask = pdb_list[0].profile2mask("n_templates", 1, 1000) amask = pdb_list[0].res2atomMask(rmask) mask_final_ref = N0.compress(mask_casting, amask) mask_final = mask_casting * amask reference = reference.compress(mask_final_ref) for i in range(len(pdb_list)): #self.cad(reference, pdb_list[i]) pdb_list[i], pdb_wo_if = self.output_fittedStructures(\ pdb_list[i], reference, i, mask_final) fitted_model_if = pdb_list[i].compress(mask_final) fitted_model_wo_if = pdb_wo_if.compress(mask_final) coord1 = reference.getXyz() coord2 = fitted_model_if.getXyz() aprofile = self.rmsd_res(coord1, coord2) self.calc_rmsd(fitted_model_if, fitted_model_wo_if, reference, pdb_list[i]) pdb_list[i].atoms.set('rmsd2ref_if', aprofile, mask=mask_final, default=-1, comment="rmsd to known reference structure") self.output_rmsd_aa(pdb_list) self.output_rmsd_ca(pdb_list) self.output_rmsd_res(pdb_list) self.write_PDBModels(pdb_list)
def go(self, model_list = None, reference = None): """ Run benchmarking. @param model_list: list of models (default: None S{->} outFolder/L{F_PDBModels}) @type model_list: ModelList @param reference: reference model (default: None S{->} outFolder/L{F_INPUT_REFERENCE}) @type reference: PDBModel """ model_list = model_list or self.outFolder + self.F_PDBModels reference = reference or self.outFolder + self.F_INPUT_REFERENCE pdb_list = T.load('%s'%model_list) reference = PDBModel(reference) # check with python 2.4 iref, imodel = reference.compareAtoms(pdb_list[0]) mask_casting = N.zeros(len(pdb_list[0])) N.put(mask_casting, imodel, 1) reference = reference.take(iref) #reference_mask_CA = reference_rmsd.maskCA() atom_mask = N.zeros(len(pdb_list[0])) N.put(atom_mask,imodel,1) rmask = pdb_list[0].profile2mask("n_templates", 1,1000) amask = pdb_list[0].res2atomMask(rmask) mask_final_ref = N.compress(mask_casting, amask) mask_final = mask_casting * amask reference = reference.compress(mask_final_ref) for i in range(len(pdb_list)): #self.cad(reference, pdb_list[i]) pdb_list[i], pdb_wo_if = self.output_fittedStructures(\ pdb_list[i], reference, i, mask_final) fitted_model_if = pdb_list[i].compress(mask_final) fitted_model_wo_if = pdb_wo_if.compress(mask_final) coord1 = reference.getXyz() coord2 = fitted_model_if.getXyz() aprofile = self.rmsd_res(coord1,coord2) self.calc_rmsd(fitted_model_if, fitted_model_wo_if, reference, pdb_list[i]) pdb_list[i].atoms.set('rmsd2ref_if', aprofile, mask=mask_final, default = -1, comment="rmsd to known reference structure") self.output_rmsd_aa(pdb_list) self.output_rmsd_ca(pdb_list) self.output_rmsd_res(pdb_list) self.write_PDBModels(pdb_list)
def capACE(self, model, chain, breaks=True): """ Cap N-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping @rtype : PDBModel """ if self.verbose: self.logWrite('Capping N-terminal of chain %i with ACE' % chain) c_start = model.chainIndex(breaks=breaks) c_end = model.chainEndIndex(breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() m_ace = PDBModel(self.F_ace_cap) chains_before = model.takeChains(range(chain), breaks=breaks) m_chain = model.takeChains([chain], breaks=breaks) chains_after = model.takeChains(range(chain + 1, len(c_start)), breaks=breaks) m_term = m_chain.resModels()[0] ## we need 3 atoms for superposition, CB might mess things up but ## could help if there is no HN ## if 'HN' in m_term.atomNames(): m_ace.remove(['CB']) ## use backbone 'C' rather than CB for fitting ## rename overhanging residue in cap PDB for a in m_ace: if a['residue_name'] != 'ACE': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0] - 1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## fit cap onto first residue of chain m_ace = m_ace.magicFit(m_term) cap = m_ace.resModels()[0] serial = m_term['serial_number'][0] - len(cap) cap['serial_number'] = range(serial, serial + len(cap)) ## concat cap on chain m_chain = cap.concat(m_chain, newChain=False) ## re-assemble whole model r = chains_before.concat(m_chain, newChain=not Nterm_is_break) r = r.concat(chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains(breaks=breaks): raise CappingError, 'Capping ACE would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r
def randomSurfaces( base_folder, label, mask ): """ calculate surfaces for all peptides and return the average and SD """ ## container for results and standard deviations MS, AS = {}, {} MS_sd, AS_sd = {}, {} ## loop over peptide directories for k in MOU.aaAtoms.keys(): dir = base_folder + 'GLY-%s-GLY_pcr/pcr_00'%(k) fLst = glob.glob( dir + '/*.pdb') msLst = [] asLst = [] ## loop over pdb files for each peptide T.flushPrint( '\nNow collecting data in %s'%dir ) for f in fLst: ## load peptide and remove waters and hydrogens m = PDBModel( f ) m = m.compress( m.maskProtein() * m.maskHeavy() ) T.flushPrint( '.') ## add surface data try: d = PDBDope( m ) d.addSurfaceRacer( probe=1.4 ) ## remove tailing GLY m = m.compress( m.res2atomMask(mask) ) ## collect surface data for each peptide msLst += [ m.profile('MS') ] asLst += [ m.profile('AS') ] except: print 'Failed calculating exposure for GLY-%s-GLY'%(k) print '\t and file %s'%f ## get result dictionary for peptide T.flushPrint('\nCollecting data ...\n') msDic = {} asDic = {} msDic_sd = {} asDic_sd = {} j = 0 #atoms = [ a['name'] for a in m.atoms ] for n in m['name']: msDic[n] = N0.average(msLst)[j] asDic[n] = N0.average(asLst)[j] msDic_sd[n] = MAU.SD( msLst )[j] asDic_sd[n] = MAU.SD( asLst )[j] j += 1 MS[ k ] = msDic AS[ k ] = asDic MS_sd[ k ] = msDic_sd AS_sd[ k ] = asDic_sd return MS, AS, MS_sd, AS_sd
def capNME(self, model, chain, breaks=True): """ Cap C-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping residue @rtype : PDBModel """ if self.verbose: self.logWrite('Capping C-terminal of chain %i with NME.' % chain) m_nme = PDBModel(self.F_nme_cap) c_start = model.chainIndex(breaks=breaks) c_end = model.chainEndIndex(breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() chains_before = model.takeChains(range(chain), breaks=breaks) m_chain = model.takeChains([chain], breaks=breaks) chains_after = model.takeChains(range(chain + 1, len(c_start)), breaks=breaks) m_term = m_chain.resModels()[-1] ## rename overhanging residue in cap PDB, renumber cap residue for a in m_nme: if a['residue_name'] != 'NME': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0] + 1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## chain should not have any terminal O after capping m_chain.remove(['OXT']) ## fit cap onto last residue of chain m_nme = m_nme.magicFit(m_term) cap = m_nme.resModels()[-1] serial = m_term['serial_number'][-1] + 1 cap['serial_number'] = range(serial, serial + len(cap)) ## concat cap on chain m_chain = m_chain.concat(cap, newChain=False) ## should be obsolete now if getattr(m_chain, '_PDBModel__terAtoms', []) != []: m_chain._PDBModel__terAtoms = [len(m_chain) - 1] assert m_chain.lenChains() == 1 ## re-assemble whole model r = chains_before.concat(m_chain, newChain=not Nterm_is_break) r = r.concat(chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains(breaks=breaks): raise CappingError, 'Capping NME would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r
def parmMirror(self, f_out, f_out_crd=None, fmod=['frcmod.ionsjc_tip3p'], fprep=[], **kw): """ Create a parm7 file whose atom content (and order) exactly mirrors the given PDBModel. This requires two leap runs. First we get a temporary topology, then we identify all atoms added by leap and build a final topology where these atoms are deleted. This parm is hence NOT suited for simulations but can be used to parse e.g. a trajectory or PDB into ptraj. @param f_out: target parm file @type f_out: str @param f_out_crd: target crd file (default: f_out but ending .crd) @type f_out_crd: str @param fmod : list of amber Mod files (loaded with loadAmberParams) @type fmod : [str] @param fmod : list of amber Prep files (loaded with loadAmberPrep) @type fmod : [str] """ f_out = t.absfile(f_out) f_out_crd = t.absfile(f_out_crd) or t.stripSuffix(f_out) + '.crd' ## if there are hydrogens, recast them to standard amber names aatm = 'HA' in self.m.atomNames() ## 'HB2' in self.m.atomNames() ## First leap round ## m_ref = self.m.clone() m_ref.xplor2amber(aatm=aatm, parm10=True) tmp_in = tempfile.mktemp('leap_in0.pdb') m_ref.writePdb(tmp_in, ter=3) tmp_parm = tempfile.mktemp('_parm0') tmp_crd = tempfile.mktemp('_crd0') leap_mod = self.__fLines('m = loadAmberParams %s\n', fmod) leap_prep = self.__fLines('loadAmberPrep %s\n', fprep) self.__runLeap(self.script_mirror_pdb, leaprc=self.leaprc, fmod=leap_mod, fprep=leap_prep, in_pdb=tmp_in, out_parm=tmp_parm, out_crd=tmp_crd, delete_atoms='') tmp_pdb = self.parm2pdb(tmp_parm, tmp_crd, tempfile.mktemp('leap_out.pdb'), aatm=aatm) if not self.debug: t.tryRemove(tmp_parm) t.tryRemove(tmp_crd) t.tryRemove(tmp_in) ## load model with missing atoms added by leap m_leap = PDBModel(tmp_pdb) ## compare atom content iLeap, iRef = m_leap.compareAtoms(m_ref) ## check that ref model doesn't need any change if iRef != range(len(m_ref)): uLeap, uRef = m_leap.unequalAtoms(m_ref, iLeap, iRef) atms = m_ref.reportAtoms(uRef, n=6) raise AmberError, "Cannot create exact mirror of %s.\n" % tmp_in +\ "Leap has renamed/deleted original atoms in %s:\n"% tmp_pdb+\ atms ## indices of atoms that were added by leap delStr = self.__deleteAtoms(m_leap, self.__inverseIndices(m_leap, iLeap)) ## Second leap round ## self.__runLeap(self.script_mirror_pdb, leaprc=self.leaprc, in_pdb=tmp_pdb, fmod=leap_mod, fprep=leap_prep, out_parm=f_out, out_crd=f_out_crd, delete_atoms=delStr) if not self.debug: t.tryRemove(tmp_pdb)
class PDBCleaner: """ PDBCleaner performs the following tasks: * remove HETAtoms from PDB * replace non-standard AA by its closest standard AA * remove non-standard atoms from standard AA residues * delete atoms that follow missing atoms (in a chain) * remove multiple occupancy atoms (except the one with highest occupancy) * add ACE and NME capping residues to C- and N-terminals or chain breaks (see capTerminals(), this is NOT done automatically in process()) Usage: ======= >>> c = PDBCleaner( model ) >>> c.process() >>> c.capTerminals( auto=True ) This will modify the model in-place and report changes to STDOUT. Alternatively, you can specify a log file instance for the output. PDBCleaner.process accepts several options to modify the processing. Capping ======= Capping will add N-methyl groups to free C-terminal carboxy ends or Acetyl groups to free N-terminal Amines and will thus 'simulate' the continuation of the protein chain -- a common practice in order to prevent fake terminal charges. The automatic discovery of missing residues is guess work at best. The more conservative approach is to use, for example: >>> c.capTerminals( breaks=1, capC=[0], capN=[2] ) In this case, only the chain break detection is used for automatic capping -- the last residue before a chain break is capped with NME and the first residue after the chain break is capped with ACE. Chain break detection relies on PDBModel.chainBreaks() (via PDBModel.chainIndex( breaks=1 )). The normal terminals to be capped are now specified explicitely. The first chain (not counting chain breaks) will receive a NME C-terminal cap and the third chain of the PDB will receive a N-terminal ACE cap. Note: Dictionaries with standard residues and atom content are defined in Biskit.molUtils. This is a duplicate effort with the new strategy to parse Amber prep files for very similar information (AmberResidueType, AmberResidueLibrary) and should change once we implement a real framework for better residue handling. """ #: these atoms always occur at the tip of of a chain or within a ring #: and, if missing, will not trigger the removal of other atoms TOLERATE_MISSING = [ 'O', 'CG2', 'CD1', 'CD2', 'OG1', 'OE1', 'NH1', 'OD1', 'OE1', 'H5T', "O5'", ] ## PDB with ACE capping residue F_ace_cap = t.dataRoot() + '/amber/leap/ace_cap.pdb' ## PDB with NME capping residue F_nme_cap = t.dataRoot() + '/amber/leap/nme_cap.pdb' def __init__(self, fpdb, log=None, verbose=True): """ @param fpdb: pdb file OR PDBModel instance @type fpdb: str OR Biskit.PDBModel @param log: Biskit.LogFile object (default: STDOUT) @type log: Biskit.LogFile @param verbose: log warnings and infos (default: True) @type verbose: bool """ self.model = PDBModel(fpdb) self.log = log or StdLog() self.verbose = verbose def logWrite(self, msg, force=1): if self.log: self.log.add(msg) else: if force: print msg def remove_multi_occupancies(self): """ Keep only atoms with alternate A field (well, or no alternate). """ if self.verbose: self.logWrite(self.model.pdbCode + ': Removing multiple occupancies of atoms ...') i = 0 to_be_removed = [] for a in self.model: if a['alternate']: try: str_id = "%i %s %s %i" % (a['serial_number'], a['name'], a['residue_name'], a['residue_number']) if a['alternate'].upper() == 'A': a['alternate'] = '' else: if float(a['occupancy']) < 1.0: to_be_removed += [i] if self.verbose: self.logWrite( 'removing %s (%s %s)' % (str_id, a['alternate'], a['occupancy'])) else: if self.verbose: self.logWrite(( 'keeping non-A duplicate %s because of 1.0 ' + 'occupancy') % str_id) except: self.logWrite("Error removing duplicate: " + t.lastError()) i += 1 try: self.model.remove(to_be_removed) if self.verbose: self.logWrite('Removed %i atoms' % len(to_be_removed)) except: if self.verbose: self.logWrite('No atoms with multiple occupancies to remove') def replace_non_standard_AA(self, amber=0, keep=[]): """ Replace amino acids with none standard names with standard amino acids according to L{MU.nonStandardAA} @param amber: don't rename HID, HIE, HIP, CYX, NME, ACE [0] @type amber: 1||0 @param keep: names of additional residues to keep @type keep: [ str ] """ standard = MU.atomDic.keys() + keep if amber: standard.extend(['HID', 'HIE', 'HIP', 'CYX', 'NME', 'ACE']) replaced = 0 if self.verbose: self.logWrite(self.model.pdbCode + ': Looking for non-standard residue names...') resnames = self.model['residue_name'] for i in self.model.atomRange(): resname = resnames[i].upper() if resname not in standard: if resname in MU.nonStandardAA: resnames[i] = MU.nonStandardAA[resname] if self.verbose: self.logWrite('renamed %s %i to %s' % \ (resname, i, MU.nonStandardAA[ resname ])) else: resnames[i] = 'ALA' self.logWrite('Warning: unknown residue name %s %i: ' \ % (resname, i ) ) if self.verbose: self.logWrite('\t->renamed to ALA.') replaced += 1 if self.verbose: self.logWrite('Found %i atoms with non-standard residue names.'% \ replaced ) def __standard_res(self, resname, amber=0): """ Check if resname is a standard residue (according to L{MU.atomDic}) if not return the closest standard residue (according to L{MU.nonStandardAA}). @param resname: 3-letter residue name @type resname: str @return: name of closest standard residue or resname itself @rtype: str """ if resname in MU.atomDic: return resname if resname in MU.nonStandardAA: return MU.nonStandardAA[resname] return resname def remove_non_standard_atoms(self): """ First missing standard atom triggers removal of standard atoms that follow in the standard order. All non-standard atoms are removed too. Data about standard atoms are taken from L{MU.atomDic} and symomym atom name is defined in L{MU.atomSynonyms}. @return: number of atoms removed @rtype: int """ mask = [] if self.verbose: self.logWrite("Checking content of standard amino-acids...") for res in self.model.resList(): resname = self.__standard_res(res[0]['residue_name']).upper() if resname == 'DC5': pass ## bugfix: ignore non-standard residues that have no matching ## standard residue if resname in MU.atomDic: standard = copy.copy(MU.atomDic[resname]) ## replace known synonyms by standard atom name for a in res: n = a['name'] if not n in standard and MU.atomSynonyms.get( n, 0) in standard: a['name'] = MU.atomSynonyms[n] if self.verbose: self.logWrite('%s: renaming %s to %s in %s %i' %\ ( self.model.pdbCode, n, a['name'], a['residue_name'], a['residue_number'])) anames = [a['name'] for a in res] keep = 1 ## kick out all standard atoms that follow a missing one rm = [] for n in standard: if (not n in anames) and not (n in self.TOLERATE_MISSING): keep = 0 if not keep: rm += [n] for n in rm: standard.remove(n) ## keep only atoms that are standard (and not kicked out above) for a in res: if a['name'] not in standard: mask += [1] if self.verbose: self.logWrite('%s: removing atom %s in %s %i '%\ ( self.model.pdbCode, a['name'], a['residue_name'], a['residue_number'])) else: mask += [0] self.model.remove(mask) if self.verbose: self.logWrite('Removed ' + str(N0.sum(mask)) + ' atoms because they were non-standard' + ' or followed a missing atom.') return N0.sum(mask) def capACE(self, model, chain, breaks=True): """ Cap N-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping @rtype : PDBModel """ if self.verbose: self.logWrite('Capping N-terminal of chain %i with ACE' % chain) c_start = model.chainIndex(breaks=breaks) c_end = model.chainEndIndex(breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() m_ace = PDBModel(self.F_ace_cap) chains_before = model.takeChains(range(chain), breaks=breaks) m_chain = model.takeChains([chain], breaks=breaks) chains_after = model.takeChains(range(chain + 1, len(c_start)), breaks=breaks) m_term = m_chain.resModels()[0] ## we need 3 atoms for superposition, CB might mess things up but ## could help if there is no HN ## if 'HN' in m_term.atomNames(): m_ace.remove(['CB']) ## use backbone 'C' rather than CB for fitting ## rename overhanging residue in cap PDB for a in m_ace: if a['residue_name'] != 'ACE': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0] - 1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## fit cap onto first residue of chain m_ace = m_ace.magicFit(m_term) cap = m_ace.resModels()[0] serial = m_term['serial_number'][0] - len(cap) cap['serial_number'] = range(serial, serial + len(cap)) ## concat cap on chain m_chain = cap.concat(m_chain, newChain=False) ## re-assemble whole model r = chains_before.concat(m_chain, newChain=not Nterm_is_break) r = r.concat(chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains(breaks=breaks): raise CappingError, 'Capping ACE would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r def capNME(self, model, chain, breaks=True): """ Cap C-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping residue @rtype : PDBModel """ if self.verbose: self.logWrite('Capping C-terminal of chain %i with NME.' % chain) m_nme = PDBModel(self.F_nme_cap) c_start = model.chainIndex(breaks=breaks) c_end = model.chainEndIndex(breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() chains_before = model.takeChains(range(chain), breaks=breaks) m_chain = model.takeChains([chain], breaks=breaks) chains_after = model.takeChains(range(chain + 1, len(c_start)), breaks=breaks) m_term = m_chain.resModels()[-1] ## rename overhanging residue in cap PDB, renumber cap residue for a in m_nme: if a['residue_name'] != 'NME': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0] + 1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## chain should not have any terminal O after capping m_chain.remove(['OXT']) ## fit cap onto last residue of chain m_nme = m_nme.magicFit(m_term) cap = m_nme.resModels()[-1] serial = m_term['serial_number'][-1] + 1 cap['serial_number'] = range(serial, serial + len(cap)) ## concat cap on chain m_chain = m_chain.concat(cap, newChain=False) ## should be obsolete now if getattr(m_chain, '_PDBModel__terAtoms', []) != []: m_chain._PDBModel__terAtoms = [len(m_chain) - 1] assert m_chain.lenChains() == 1 ## re-assemble whole model r = chains_before.concat(m_chain, newChain=not Nterm_is_break) r = r.concat(chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains(breaks=breaks): raise CappingError, 'Capping NME would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r def convertChainIdsNter(self, model, chains): """ Convert normal chain ids to chain ids considering chain breaks. """ if len(chains) == 0: return chains i = N0.take(model.chainIndex(), chains) ## convert back to chain indices but this time including chain breaks return model.atom2chainIndices(i, breaks=1) def convertChainIdsCter(self, model, chains): """ Convert normal chain ids to chain ids considering chain breaks. """ if len(chains) == 0: return chains ## fetch last atom of given chains index = N0.concatenate((model.chainIndex(), [len(model)])) i = N0.take(index, N0.array(chains) + 1) - 1 ## convert back to chain indices but this time including chain breaks return model.atom2chainIndices(i, breaks=1) def unresolvedTerminals(self, model): """ Autodetect (aka "guess") which N- and C-terminals are most likely not the real end of each chain. This guess work is based on residue numbering: * unresolved N-terminal: a protein residue with a residue number > 1 * unresolved C-terminal: a protein residue that does not contain either OXT or OT or OT1 or OT2 atoms @param model: PDBModel @return: chains with unresolved N-term, with unresolved C-term @rtype : ([int], [int]) """ c_first = model.chainIndex() c_last = model.chainEndIndex() capN = [ i for (i,pos) in enumerate(c_first)\ if model['residue_number'][pos] > 1 ] capN = [i for i in capN if model['residue_name'][c_first[i]] != 'ACE'] capN = self.filterProteinChains(model, capN, c_first) capC = [] for (i, pos) in enumerate(c_last): atoms = model.takeResidues(model.atom2resIndices([pos ])).atomNames() if not( 'OXT' in atoms or 'OT' in atoms or 'OT1' in atoms or \ 'OT2' in atoms ): capC += [i] capC = self.filterProteinChains(model, capC, c_last) return capN, capC #@todo filter for protein positions in breaks=1 def filterProteinChains(self, model, chains, chainindex): maskProtein = model.maskProtein() chains = [i for i in chains if maskProtein[chainindex[i]]] return chains def capTerminals(self, auto=False, breaks=False, capN=[], capC=[]): """ Add NME and ACE capping residues to chain breaks or normal N- and C-terminals. Note: these capping residues contain hydrogen atoms. Chain indices for capN and capC arguments can be interpreted either with or without chain break detection enabled. For example, let's assume we have a two-chain protein with some missing residues (chain break) in the first chain: A: MGSKVSK---FLNAGSK B: FGHLAKSDAK Then: capTerminals( breaks=False, capN=[1], capC=[1]) will add N-and C-terminal caps to chain B. However: capTerminals( breaks=True, capN=[1], capC=[1]) will add N- and C-terminal caps to the second fragment of chain A. Note: this operation *replaces* the internal model. @param auto: put ACE and NME capping residue on chain breaks and on suspected false N- and C-termini (default: False) @type auto: bool @param breaks: switch on chain break detection before interpreting capN and capC @type breaks: False @param capN: indices of chains that should get ACE cap (default: []) @type capN: [int] @param capC: indices of chains that should get NME cap (default: []) @type capC: [int] """ m = self.model c_len = m.lenChains() i_breaks = m.chainBreaks() if auto: if not breaks: capN = self.convertChainIdsNter(m, capN) capC = self.convertChainIdsCter(m, capC) breaks = True capN, capC = self.unresolvedTerminals(m) end_broken = m.atom2chainIndices(m.chainBreaks(), breaks=1) capC = M.union(capC, end_broken) capN = M.union(capN, N0.array(end_broken) + 1) capN = self.filterProteinChains(m, capN, m.chainIndex(breaks=breaks)) capC = self.filterProteinChains(m, capC, m.chainEndIndex(breaks=breaks)) for i in capN: m = self.capACE(m, i, breaks=breaks) assert m.lenChains() == c_len, '%i != %i' % \ (m.lenChains(), c_len) assert len(m.chainBreaks(force=True)) == len(i_breaks) assert m[ 'serial_number'].dtype == N0.Int32, 'serial_number not int' for i in capC: m = self.capNME(m, i, breaks=breaks) assert m.lenChains() == c_len assert len(m.chainBreaks(force=True)) == len(i_breaks) self.model = m return self.model def process(self, keep_hetatoms=0, amber=0, keep_xaa=[]): """ Remove Hetatoms, waters. Replace non-standard names. Remove non-standard atoms. @param keep_hetatoms: option @type keep_hetatoms: 0||1 @param amber: don't rename amber residue names (HIE, HID, CYX,..) @type amber: 0||1 @param keep_xaa: names of non-standard residues to be kept @type keep_xaa: [ str ] @return: PDBModel (reference to internal) @rtype: PDBModel @raise CleanerError: if something doesn't go as expected ... """ try: if not keep_hetatoms: self.model.remove(self.model.maskHetatm()) self.model.remove(self.model.maskH2O()) self.model.remove(self.model.maskH()) self.remove_multi_occupancies() self.replace_non_standard_AA(amber=amber, keep=keep_xaa) self.remove_non_standard_atoms() except KeyboardInterrupt, why: raise KeyboardInterrupt(why) except Exception, why: self.logWrite('Error: ' + t.lastErrorTrace()) raise CleanerError('Error cleaning model: %r' % why)
class Test(BT.BiskitTest): """Test class """ def prepare(self): from Biskit.LogFile import LogFile import tempfile def test_PDBCleaner(self): """PDBCleaner general test""" ## Loading PDB... self.c = PDBCleaner(t.testRoot() + '/rec/1A2P_rec_original.pdb', log=self.log, verbose=self.local) self.m = self.c.process() self.assertAlmostEqual(self.m.mass(), 34029.0115499993, 7) def test_DNACleaning(self): """PDBCleaner DNA test""" ## Loading PDB... self.c = PDBCleaner(t.testRoot() + 'amber/entropy/0_com.pdb', log=self.log, verbose=self.local) self.dna = self.c.process(amber=True) self.assertAlmostEqual(self.dna.mass(), 26953.26, 1) def test_Capping(self): """PDBCleaner.capTerminals test""" ## Loading PDB... self.model = PDBModel(t.testRoot() + '/rec/1A2P_rec_original.pdb') self.c = PDBCleaner(self.model, log=self.log, verbose=self.local) self.m2 = self.c.capTerminals(breaks=True) self.assert_(self.m2.atomNames() == self.model.atomNames()) self.m3 = self.model.clone() self.m3.removeRes([10, 11, 12, 13, 14, 15]) self.m4 = self.m3.clone() self.c = PDBCleaner(self.m3, log=self.log, verbose=self.local) self.m3 = self.c.capTerminals(breaks=True, capC=[0], capN=[0, 1]) self.assertEqual( self.m3.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN') if self.local: self.log.add('\nTesting automatic chain capping...\n') self.c = PDBCleaner(self.m4, log=self.log, verbose=self.local) self.m4 = self.c.capTerminals(auto=True) self.assertEqual( self.m4.takeChains([0]).sequence()[:18], 'XVINTFDGVADXXKLPDN') def test_capping_extra(self): """PDBCleaner.capTerminals extra challenge""" self.m2 = PDBModel(t.testRoot() + '/pdbclean/foldx_citche.pdb') self.c = PDBCleaner(self.m2, verbose=self.local, log=self.log) self.assertRaises(CappingError, self.c.capTerminals, auto=True) if self.local: self.log.add('OK: CappingError has been raised indicating clash.') self.assertEqual(len(self.m2.takeChains([1]).chainBreaks()), 1)
def capACE( self, model, chain, breaks=True ): """ Cap N-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping @rtype : PDBModel """ if self.verbose: self.logWrite('Capping N-terminal of chain %i with ACE' % chain ) c_start = model.chainIndex( breaks=breaks ) c_end = model.chainEndIndex( breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() m_ace = PDBModel( self.F_ace_cap ) chains_before = model.takeChains( range(chain), breaks=breaks ) m_chain = model.takeChains( [chain], breaks=breaks ) chains_after = model.takeChains( range(chain+1, len(c_start)), breaks=breaks ) m_term = m_chain.resModels()[0] ## we need 3 atoms for superposition, CB might mess things up but ## could help if there is no HN ## if 'HN' in m_term.atomNames(): m_ace.remove( ['CB'] ) ## use backbone 'C' rather than CB for fitting ## rename overhanging residue in cap PDB for a in m_ace: if a['residue_name'] != 'ACE': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0]-1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## fit cap onto first residue of chain m_ace = m_ace.magicFit( m_term ) cap = m_ace.resModels()[0] serial = m_term['serial_number'][0] - len(cap) cap['serial_number'] = range( serial, serial + len(cap) ) ## concat cap on chain m_chain = cap.concat( m_chain, newChain=False ) ## re-assemble whole model r = chains_before.concat( m_chain, newChain=not Nterm_is_break) r = r.concat( chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains( breaks=breaks ): raise CappingError, 'Capping ACE would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r
class PDBCleaner: """ PDBCleaner performs the following tasks: * remove HETAtoms from PDB * replace non-standard AA by its closest standard AA * remove non-standard atoms from standard AA residues * delete atoms that follow missing atoms (in a chain) * remove multiple occupancy atoms (except the one with highest occupancy) * add ACE and NME capping residues to C- and N-terminals or chain breaks (see capTerminals(), this is NOT done automatically in process()) Usage: ======= >>> c = PDBCleaner( model ) >>> c.process() >>> c.capTerminals( auto=True ) This will modify the model in-place and report changes to STDOUT. Alternatively, you can specify a log file instance for the output. PDBCleaner.process accepts several options to modify the processing. Capping ======= Capping will add N-methyl groups to free C-terminal carboxy ends or Acetyl groups to free N-terminal Amines and will thus 'simulate' the continuation of the protein chain -- a common practice in order to prevent fake terminal charges. The automatic discovery of missing residues is guess work at best. The more conservative approach is to use, for example: >>> c.capTerminals( breaks=1, capC=[0], capN=[2] ) In this case, only the chain break detection is used for automatic capping -- the last residue before a chain break is capped with NME and the first residue after the chain break is capped with ACE. Chain break detection relies on PDBModel.chainBreaks() (via PDBModel.chainIndex( breaks=1 )). The normal terminals to be capped are now specified explicitely. The first chain (not counting chain breaks) will receive a NME C-terminal cap and the third chain of the PDB will receive a N-terminal ACE cap. Note: Dictionaries with standard residues and atom content are defined in Biskit.molUtils. This is a duplicate effort with the new strategy to parse Amber prep files for very similar information (AmberResidueType, AmberResidueLibrary) and should change once we implement a real framework for better residue handling. """ #: these atoms always occur at the tip of of a chain or within a ring #: and, if missing, will not trigger the removal of other atoms TOLERATE_MISSING = ['O', 'CG2', 'CD1', 'CD2', 'OG1', 'OE1', 'NH1', 'OD1', 'OE1', 'H5T',"O5'", ] ## PDB with ACE capping residue F_ace_cap = t.dataRoot() + '/amber/leap/ace_cap.pdb' ## PDB with NME capping residue F_nme_cap = t.dataRoot() + '/amber/leap/nme_cap.pdb' def __init__( self, fpdb, log=None, verbose=True ): """ @param fpdb: pdb file OR PDBModel instance @type fpdb: str OR Biskit.PDBModel @param log: Biskit.LogFile object (default: STDOUT) @type log: Biskit.LogFile @param verbose: log warnings and infos (default: True) @type verbose: bool """ self.model = PDBModel( fpdb ) self.log = log or StdLog() self.verbose = verbose def logWrite( self, msg, force=1 ): if self.log: self.log.add( msg ) else: if force: print msg def remove_multi_occupancies( self ): """ Keep only atoms with alternate A field (well, or no alternate). """ if self.verbose: self.logWrite( self.model.pdbCode + ': Removing multiple occupancies of atoms ...') i = 0 to_be_removed = [] for a in self.model: if a['alternate']: try: str_id = "%i %s %s %i" % (a['serial_number'], a['name'], a['residue_name'], a['residue_number']) if a['alternate'].upper() == 'A': a['alternate'] = '' else: if float( a['occupancy'] ) < 1.0: to_be_removed += [ i ] if self.verbose: self.logWrite( 'removing %s (%s %s)' % (str_id,a['alternate'], a['occupancy'])) else: if self.verbose: self.logWrite( ('keeping non-A duplicate %s because of 1.0 '+ 'occupancy') % str_id ) except: self.logWrite("Error removing duplicate: "+t.lastError() ) i+=1 try: self.model.remove( to_be_removed ) if self.verbose: self.logWrite('Removed %i atoms' % len( to_be_removed ) ) except: if self.verbose: self.logWrite('No atoms with multiple occupancies to remove' ) def replace_non_standard_AA( self, amber=0, keep=[] ): """ Replace amino acids with none standard names with standard amino acids according to L{MU.nonStandardAA} @param amber: don't rename HID, HIE, HIP, CYX, NME, ACE [0] @type amber: 1||0 @param keep: names of additional residues to keep @type keep: [ str ] """ standard = MU.atomDic.keys() + keep if amber: standard.extend( ['HID', 'HIE', 'HIP', 'CYX', 'NME', 'ACE'] ) replaced = 0 if self.verbose: self.logWrite(self.model.pdbCode + ': Looking for non-standard residue names...') resnames = self.model['residue_name'] for i in self.model.atomRange(): resname = resnames[i].upper() if resname not in standard: if resname in MU.nonStandardAA: resnames[i] = MU.nonStandardAA[ resname ] if self.verbose: self.logWrite('renamed %s %i to %s' % \ (resname, i, MU.nonStandardAA[ resname ])) else: resnames[i] = 'ALA' self.logWrite('Warning: unknown residue name %s %i: ' \ % (resname, i ) ) if self.verbose: self.logWrite('\t->renamed to ALA.') replaced += 1 if self.verbose: self.logWrite('Found %i atoms with non-standard residue names.'% \ replaced ) def __standard_res( self, resname, amber=0 ): """ Check if resname is a standard residue (according to L{MU.atomDic}) if not return the closest standard residue (according to L{MU.nonStandardAA}). @param resname: 3-letter residue name @type resname: str @return: name of closest standard residue or resname itself @rtype: str """ if resname in MU.atomDic: return resname if resname in MU.nonStandardAA: return MU.nonStandardAA[ resname ] return resname def remove_non_standard_atoms( self ): """ First missing standard atom triggers removal of standard atoms that follow in the standard order. All non-standard atoms are removed too. Data about standard atoms are taken from L{MU.atomDic} and symomym atom name is defined in L{MU.atomSynonyms}. @return: number of atoms removed @rtype: int """ mask = [] if self.verbose: self.logWrite("Checking content of standard amino-acids...") for res in self.model.resList(): resname = self.__standard_res( res[0]['residue_name'] ).upper() if resname == 'DC5': pass ## bugfix: ignore non-standard residues that have no matching ## standard residue if resname in MU.atomDic: standard = copy.copy( MU.atomDic[ resname ] ) ## replace known synonyms by standard atom name for a in res: n = a['name'] if not n in standard and MU.atomSynonyms.get(n,0) in standard: a['name'] = MU.atomSynonyms[n] if self.verbose: self.logWrite('%s: renaming %s to %s in %s %i' %\ ( self.model.pdbCode, n, a['name'], a['residue_name'], a['residue_number'])) anames = [ a['name'] for a in res ] keep = 1 ## kick out all standard atoms that follow a missing one rm = [] for n in standard: if (not n in anames) and not (n in self.TOLERATE_MISSING): keep = 0 if not keep: rm += [ n ] for n in rm: standard.remove( n ) ## keep only atoms that are standard (and not kicked out above) for a in res: if a['name'] not in standard: mask += [1] if self.verbose: self.logWrite('%s: removing atom %s in %s %i '%\ ( self.model.pdbCode, a['name'], a['residue_name'], a['residue_number'])) else: mask += [0] self.model.remove( mask ) if self.verbose: self.logWrite('Removed ' + str(N.sum(mask)) + ' atoms because they were non-standard' + ' or followed a missing atom.' ) return N.sum( mask ) def capACE( self, model, chain, breaks=True ): """ Cap N-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping @rtype : PDBModel """ if self.verbose: self.logWrite('Capping N-terminal of chain %i with ACE' % chain ) c_start = model.chainIndex( breaks=breaks ) c_end = model.chainEndIndex( breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() m_ace = PDBModel( self.F_ace_cap ) chains_before = model.takeChains( range(chain), breaks=breaks ) m_chain = model.takeChains( [chain], breaks=breaks ) chains_after = model.takeChains( range(chain+1, len(c_start)), breaks=breaks ) m_term = m_chain.resModels()[0] ## we need 3 atoms for superposition, CB might mess things up but ## could help if there is no HN ## if 'HN' in m_term.atomNames(): m_ace.remove( ['CB'] ) ## use backbone 'C' rather than CB for fitting ## rename overhanging residue in cap PDB for a in m_ace: if a['residue_name'] != 'ACE': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0]-1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## fit cap onto first residue of chain m_ace = m_ace.magicFit( m_term ) cap = m_ace.resModels()[0] serial = m_term['serial_number'][0] - len(cap) cap['serial_number'] = range( serial, serial + len(cap) ) ## concat cap on chain m_chain = cap.concat( m_chain, newChain=False ) ## re-assemble whole model r = chains_before.concat( m_chain, newChain=not Nterm_is_break) r = r.concat( chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains( breaks=breaks ): raise CappingError, 'Capping ACE would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r def capNME( self, model, chain, breaks=True ): """ Cap C-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping residue @rtype : PDBModel """ if self.verbose: self.logWrite('Capping C-terminal of chain %i with NME.' % chain ) m_nme = PDBModel( self.F_nme_cap ) c_start = model.chainIndex( breaks=breaks ) c_end = model.chainEndIndex( breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() chains_before = model.takeChains( range(chain), breaks=breaks ) m_chain = model.takeChains( [chain], breaks=breaks ) chains_after = model.takeChains( range(chain+1, len(c_start)), breaks=breaks ) m_term = m_chain.resModels()[-1] ## rename overhanging residue in cap PDB, renumber cap residue for a in m_nme: if a['residue_name'] != 'NME': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0]+1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## chain should not have any terminal O after capping m_chain.remove( ['OXT'] ) ## fit cap onto last residue of chain m_nme = m_nme.magicFit( m_term ) cap = m_nme.resModels()[-1] serial = m_term['serial_number'][-1]+1 cap['serial_number'] = range( serial, serial + len(cap) ) ## concat cap on chain m_chain = m_chain.concat( cap, newChain=False ) ## should be obsolete now if getattr( m_chain, '_PDBModel__terAtoms', []) != []: m_chain._PDBModel__terAtoms = [ len( m_chain ) - 1 ] assert m_chain.lenChains() == 1 ## re-assemble whole model r = chains_before.concat( m_chain, newChain=not Nterm_is_break) r = r.concat( chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains( breaks=breaks ): raise CappingError, 'Capping NME would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r def convertChainIdsNter( self, model, chains ): """ Convert normal chain ids to chain ids considering chain breaks. """ if len(chains) == 0: return chains i = N.take( model.chainIndex(), chains ) ## convert back to chain indices but this time including chain breaks return model.atom2chainIndices( i, breaks=1 ) def convertChainIdsCter( self, model, chains ): """ Convert normal chain ids to chain ids considering chain breaks. """ if len(chains) == 0: return chains ## fetch last atom of given chains index = N.concatenate( (model.chainIndex(), [len(model)]) ) i = N.take( index, N.array( chains ) + 1 ) - 1 ## convert back to chain indices but this time including chain breaks return model.atom2chainIndices( i, breaks=1 ) def unresolvedTerminals( self, model ): """ Autodetect (aka "guess") which N- and C-terminals are most likely not the real end of each chain. This guess work is based on residue numbering: * unresolved N-terminal: a protein residue with a residue number > 1 * unresolved C-terminal: a protein residue that does not contain either OXT or OT or OT1 or OT2 atoms @param model: PDBModel @return: chains with unresolved N-term, with unresolved C-term @rtype : ([int], [int]) """ c_first = model.chainIndex() c_last = model.chainEndIndex() capN = [ i for (i,pos) in enumerate(c_first)\ if model['residue_number'][pos] > 1 ] capN = [i for i in capN if model['residue_name'][c_first[i]] != 'ACE'] capN = self.filterProteinChains( model, capN, c_first ) capC = [] for (i,pos) in enumerate(c_last): atoms = model.takeResidues(model.atom2resIndices([pos])).atomNames() if not( 'OXT' in atoms or 'OT' in atoms or 'OT1' in atoms or \ 'OT2' in atoms ): capC += [ i ] capC = self.filterProteinChains( model, capC, c_last ) return capN, capC #@todo filter for protein positions in breaks=1 def filterProteinChains( self, model, chains, chainindex ): maskProtein = model.maskProtein() chains = [ i for i in chains if maskProtein[ chainindex[i] ] ] return chains def capTerminals( self, auto=False, breaks=False, capN=[], capC=[] ): """ Add NME and ACE capping residues to chain breaks or normal N- and C-terminals. Note: these capping residues contain hydrogen atoms. Chain indices for capN and capC arguments can be interpreted either with or without chain break detection enabled. For example, let's assume we have a two-chain protein with some missing residues (chain break) in the first chain: A: MGSKVSK---FLNAGSK B: FGHLAKSDAK Then: capTerminals( breaks=False, capN=[1], capC=[1]) will add N-and C-terminal caps to chain B. However: capTerminals( breaks=True, capN=[1], capC=[1]) will add N- and C-terminal caps to the second fragment of chain A. Note: this operation *replaces* the internal model. @param auto: put ACE and NME capping residue on chain breaks and on suspected false N- and C-termini (default: False) @type auto: bool @param breaks: switch on chain break detection before interpreting capN and capC @type breaks: False @param capN: indices of chains that should get ACE cap (default: []) @type capN: [int] @param capC: indices of chains that should get NME cap (default: []) @type capC: [int] """ m = self.model c_len = m.lenChains() i_breaks = m.chainBreaks() if auto: if not breaks: capN = self.convertChainIdsNter( m, capN ) capC = self.convertChainIdsCter( m, capC ) breaks=True capN, capC = self.unresolvedTerminals( m ) end_broken = m.atom2chainIndices( m.chainBreaks(), breaks=1 ) capC = M.union( capC, end_broken ) capN = M.union( capN, N.array( end_broken ) + 1 ) capN = self.filterProteinChains(m, capN, m.chainIndex(breaks=breaks)) capC = self.filterProteinChains(m, capC, m.chainEndIndex(breaks=breaks)) for i in capN: m = self.capACE( m, i, breaks=breaks ) assert m.lenChains() == c_len, '%i != %i' % \ (m.lenChains(), c_len) assert len(m.chainBreaks(force=True)) == len(i_breaks) for i in capC: m = self.capNME( m, i, breaks=breaks ) assert m.lenChains() == c_len assert len(m.chainBreaks(force=True)) == len(i_breaks) self.model = m return self.model def process( self, keep_hetatoms=0, amber=0, keep_xaa=[] ): """ Remove Hetatoms, waters. Replace non-standard names. Remove non-standard atoms. @param keep_hetatoms: option @type keep_hetatoms: 0||1 @param amber: don't rename amber residue names (HIE, HID, CYX,..) @type amber: 0||1 @param keep_xaa: names of non-standard residues to be kept @type keep_xaa: [ str ] @return: PDBModel (reference to internal) @rtype: PDBModel @raise CleanerError: if something doesn't go as expected ... """ try: if not keep_hetatoms: self.model.remove( self.model.maskHetatm() ) self.model.remove( self.model.maskH2O() ) self.model.remove( self.model.maskH() ) self.remove_multi_occupancies() self.replace_non_standard_AA( amber=amber, keep=keep_xaa ) self.remove_non_standard_atoms() except KeyboardInterrupt, why: raise KeyboardInterrupt( why ) except Exception, why: self.logWrite('Error: '+t.lastErrorTrace()) raise CleanerError( 'Error cleaning model: %r' % why )
class AmberParmBuilder: """ AmberParmBuilder ================ Create Amber topology and coordinate file from PDB. - parmMirror(): ...builds a fake parm that exactly mirrors a given PDB file. This parm can be used for ptraj but not for simulations. Currently, parmMirror only accepts amber-formatted PDBs as input. It should be possible to create topologies that have the same content and order of atoms as an xplor PDB but some atoms will have different names. - parmSolvated(): ...builds a solvated system for PME simulations (incl. closing of S-S bonds, capping of chain breaks). parmSolvated accepts both xplor and amber-formatted PDBs as input. Requires the amber programs C{tleap} and C{ambpdb}. Requires leap template files in C{biskit/Biskit/data/amber/leap/}. Note on forcefields: The default forcefield used is specified in exe_tleap and currently is ff10. This translates to loading amber11/dat/leap/cmd/leaprc.ff10 at the beginning of the leap run. As of 2011, ff10 is the recommended default forcefield for proteins and nucleic acids. Comment from Jason Swails on the Amber mailing list: " Try using ff99SB (which is the protein force field part of ff10, which is the version I would actually suggest using). Despite its label, it is actually a 2006 update of the ff99 force field which performs at least as well (if not better) as ff03." Unfortunately, ions are only "half" paramterized in ff10. Additional parameters need to be loaded from a frmod file, typically frcmod.ionsjc_tip3p. There are additional versions of this file optimized for other water models than TIP3. frcmod.ionsjc_tip3p is set as the default frmod file to include by parmSolvated and parmMirror. Please include it if you provide your own list of frmod files. @note: The design of AmberParmBuilder is less than elegant. It would make more sense to split it into two classes that are both derrived from Executor. """ ## script to create a parm that exactly mirrors a given PDB script_mirror_pdb = """ logFile %(f_out)s source %(leaprc)s %(fmod)s %(fprep)s p = loadPdb %(in_pdb)s %(delete_atoms)s saveAmberParm p %(out_parm)s %(out_crd)s quit """ ## tleap command to close a single S-S bond ss_bond = "bond p.%i.SG p.%i.SG\n" ## leap script for solvated topology F_leap_in = t.dataRoot() + '/amber/leap/solvate_box.leap' ## PDB with ACE capping residue F_ace_cap = t.dataRoot() + '/amber/leap/ace_cap.pdb' ## PDB with NME capping residue F_nme_cap = t.dataRoot() + '/amber/leap/nme_cap.pdb' def __init__( self, model, leap_template=F_leap_in, leaprc=None, leap_out=None, leap_in=None, leap_pdb=None, log=None, debug=0, verbose=0, **kw ): """ @param model: model @type model: PDBModel or str @param leap_template: path to template file for leap input @type leap_template: str @param leaprc: forcefield parameter file or code (e.g. ff99) @type leaprc: str @param leap_out: target file for leap.log (default: discard) @type leap_out: str @param leap_in: target file for leap.in script (default: discard) @type leap_in: str @param kw: kw=value pairs for additional options in the leap_template @type kw: key=value """ self.m = PDBModel( model ) self.leap_template = leap_template self.leaprc = leaprc self.leap_pdb = leap_pdb or tempfile.mktemp( '_leap_pdb' ) self.keep_leap_pdb = leap_pdb is not None self.leap_in = leap_in self.leap_out= leap_out self.log = log or StdLog() self.output = None # last output of leap self.debug = debug self.verbose = verbose self.__dict__.update( kw ) def __runLeap( self, in_script, in_pdb, norun=0, **kw ): """ Create script file and run Leap. @param in_script: content of ptraj script with place holders @type in_script: str @param in_pdb: PDB file to load into tleap @type in_pdb: str @param norun: 1 - only create leap scrip (default: 0) @type norun: 1|0 @param kw: key=value pairs for filling place holders in script @type kw: key=value @raise AmberError: if missing option for leap input file or if could not create leap input file """ x = AmberLeap( in_script, in_pdb=in_pdb, log=self.log, verbose=self.verbose, debug=self.debug, catch_out=True, f_in=self.leap_in, f_out=self.leap_out, **kw ) if norun: x.generateInp() else: x.run() self.output = x.output ## ## create leap script ## try: ## ## use own fields and given kw as parameters for leap script ## d = copy.copy( self.__dict__ ) ## d.update( kw ) ## in_script = in_script % d ## f = open( self.leap_in, 'w') ## f.write( in_script ) ## f.close() ## if self.verbose: ## self.log.add('leap-script: ') ## self.log.add( in_script ) ## except IOError: ## raise AmberError('Could not create leap input file') ## except: ## raise AmberError('missing option for leap input file\n'+\ ## 'available: %s' % (str( d.keys() ) )) ## ## run tleap ## args = '-f %s' % self.leap_in ## if not norun: ## self.exe = Executor('tleap', args, log=self.log,verbose=1, ## catch_out=0) ## self.output, self.error, self.status = self.exe.run() ## if not os.path.exists( kw['out_parm'] ): ## raise AmberError, "tleap failed" ## ## clean up ## if not self.keep_leap_in and not self.debug: ## t.tryRemove( self.leap_in ) ## if not self.keep_leap_out and not self.debug: ## t.tryRemove( self.leap_out) def parm2pdb( self, f_parm, f_crd, f_out, aatm=0 ): """ Use ambpdb to build PDB from parm and crd. @param f_parm: existing parm file @type f_parm: str @param f_crd: existing crd file @type f_crd: str @param f_out: target file name for PDB @type f_out: str @return: f_out, target file name for PDB @rtype: str @raise AmberError: if ambpdb fail """ ## cmd = '%s -p %s -aatm < %s > %s' % \ args = '-p %s %s' % (f_parm, '-aatm'*aatm ) x = Executor('ambpdb', args, f_in=f_crd, f_out=f_out, log=self.log, verbose=1, catch_err=1) output,error,status = x.run() if not os.path.exists( f_out ): raise AmberError, 'ambpdb failed.' return f_out def __ssBonds( self, model, cutoff=4. ): """ Identify disulfide bonds. @param model: model @type model: PDBModel @param cutoff: distance cutoff for S-S distance (default: 4.0) @type cutoff: float @return: list with numbers of residue pairs forming S-S @rtype: [(int, int)] """ m = model.compress( model.mask( ['SG'] ) ) if len( m ) < 2: return [] pw = MU.pairwiseDistances( m.xyz, m.xyz ) pw = N.less( pw, cutoff ) r = [] for i in range( len( pw ) ): for j in range( i+1, len(pw) ): if pw[i,j]: r += [ (m.atoms['residue_number'][i], m.atoms['residue_number'][j]) ] return r def __cys2cyx( self, model, ss_residues ): """ Rename all S-S bonded CYS into CYX. @param model: model @type model: PDBModel @param ss_residues: original residue numbers of S-S pairs @type ss_residues: [(int, int)] """ ss = [] for a,b in ss_residues: ss += [a,b] for a in model: if a['residue_number'] in ss: a['residue_name'] = 'CYX' def capACE( self, model, chain ): """ Cap N-terminal of given chain. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int """ cleaner = PDBCleaner( model, log=self.log ) return cleaner.capACE( model, chain, breaks=True ) def capNME( self, model, chain ): """ Cap C-terminal of given chain. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int """ cleaner = PDBCleaner( model, log=self.log ) return cleaner.capNME( model, chain, breaks=True) def centerModel( self, model ): """ Geometric centar of model. @param model: model @type model: PDBMode """ center = N.average( model.getXyz() ) model.setXyz( model.xyz - center ) def leapModel( self, hetatm=0, center=True ): """ Get a clean PDBModel for input into leap. @param hetatm: keep HETATM records (default: 0) @type hetatm: 1|0 @return: model @rtype: PDBMod """ m = self.m.clone() m.xplor2amber() cleaner = PDBCleaner( m, log=self.log, verbose=self.verbose ) m = cleaner.process( keep_hetatoms=hetatm, amber=1 ) m.renumberResidues( addChainId=1 ) if center: self.centerModel( m ) return m def __fLines( self, template, values ): if not type( values ) is list: values = [ values ] return ''.join( [ template % v for v in values ] ) def parmSolvated( self, f_out, f_out_crd=None, f_out_pdb=None, hetatm=0, norun=0, cap=0, capN=[], capC=[], fmod=['frcmod.ionsjc_tip3p'], fprep=[], box=10.0, center=True, **kw ): """ @param f_out: target file for parm (topology) @type f_out: str @param f_out_crd: target file for crd (coordinates) (default:|f_out_base|.crd) @type f_out_crd: str @param f_out_pdb: target file for pdb (default:|f_out_base|.pdb) @type f_out_pdb: str @param hetatm: keep hetero atoms (default: 0) @type hetatm: 1|0 @param cap: put ACE and NME capping residue on chain breaks (default: 0) @type cap: 1|0 @param capN: indices of chains that should get ACE cap (default: []) @type capN: [int] @param capC: indices of chains that should get NME cap (default: []) @type capC: [int] @param box: minimal distance of solute from box edge (default: 10.0) @type box: float @param center: re-center coordinates (default: True) @type center: bool @param fmod: list of files with amber parameter modifications to be loaded into leap with loadAmberParams (default:['frcmod.ionsjc_tip3p'] ... mod file needed for default Amber ff10 ions -- topology saving will fail if this one is missing) @type fmod: [str] @param fprep: list of files with amber residue definitions (to be loaded into leap with loadAmberPrep) (default: []) @type fprep: [str] @param kw: additional key=value pairs for leap input template @type kw: key=value @raise IOError: """ f_out = t.absfile( f_out ) f_out_crd = t.absfile( f_out_crd ) or t.stripSuffix( f_out ) + '.crd' f_out_pdb = t.absfile( f_out_pdb ) or t.stripSuffix( f_out ) +\ '_leap.pdb' ## removed: (bugfix 3434136) #fmod = [ t.absfile( f ) for f in t.toList( fmod ) ] #fprep = [ t.absfile( f ) for f in t.toList( fprep ) ] try: if self.verbose: self.log.add( '\nCleaning PDB file for Amber:' ) m = self.leapModel( hetatm=hetatm, center=center ) if cap: end_broken = m.atom2chainIndices( m.chainBreaks() ) capC = MU.union( capC, end_broken ) capN = MU.union( capN, N.array( end_broken ) + 1 ) for i in capN: if self.verbose: self.log.add( 'Adding ACE cap to chain %i' % i ) m = self.capACE( m, i ) for i in capC: if self.verbose: self.log.add( 'Adding NME cap to chain %i' % i ) m = self.capNME( m, i ) m.renumberResidues( addChainId=1 ) ## again, to accomodate capping template = open( self.leap_template ).read() leap_mod = self.__fLines( 'm = loadAmberParams %s\n', fmod ) leap_prep= self.__fLines( 'loadAmberPrep %s\n', fprep ) ss = self.__ssBonds( m, cutoff=4. ) self.__cys2cyx( m, ss ) leap_ss = self.__fLines( self.ss_bond, ss ) if self.verbose: self.log.add('Found %i disulfide bonds: %s' % (len(ss),str(ss))) if self.verbose: self.log.add( 'writing cleaned PDB to %s' % self.leap_pdb ) m.writePdb( self.leap_pdb, ter=3 ) self.__runLeap( template, in_pdb=self.leap_pdb, out_parm=f_out, out_crd=f_out_crd, ss_bonds=leap_ss, fmod=leap_mod, fprep=leap_prep, norun=norun, box=box, **kw ) if not norun: parm_pdb = self.parm2pdb( f_out, f_out_crd, f_out_pdb ) if not self.keep_leap_pdb and not self.debug: t.tryRemove( self.leap_pdb ) except IOError, why: raise IOError, why
def changeModel(inFile, prefix, sourceModel): print '\nget ' + os.path.basename(inFile) + '..', model = PDBModel(inFile) model.update() model = model.sort() eq = model.equals(sourceModel) if not eq[0] and eq[1]: raise ConvertError('source and other models are not equal: ' + str(eq)) # model.validSource() model.setSource(sourceModel.validSource()) #model.atomsChanged = 0 for k in model.atoms: model.atoms[k, 'changed'] = N0.all(model[k] == sourceModel[k]) model.xyzChanged = (0 != N0.sum(N0.ravel(model.xyz - sourceModel.xyz))) model.update(updateMissing=1) if model.xyzChanged: doper = PDBDope(model) if 'MS' in sourceModel.atoms.keys(): doper.addSurfaceRacer(probe=1.4) if 'density' in sourceModel.atoms.keys(): doper.addDensity() if 'foldX' in sourceModel.info.keys(): doper.addFoldX() if 'delphi' in sourceModel.info.keys(): doper.addDelphi() outFile = os.path.dirname( inFile ) + '/' + prefix +\ T.stripFilename( inFile ) + '.model' T.dump(model, outFile) print '-> ' + os.path.basename(outFile)
def capNME( self, model, chain, breaks=True ): """ Cap C-terminal of given chain. Note: In order to allow the capping of chain breaks, the chain index is, by default, based on model.chainIndex(breaks=True), that means with chain break detection activated! This is not the default behaviour of PDBModel.chainIndex or takeChains or chainLength. Please use the wrapping method capTerminals() for more convenient handling of the index. @param model: model @type model: PDBMode @param chain: index of chain to be capped @type chain: int @param breaks: consider chain breaks when identifying chain boundaries @type breaks: bool @return: model with added NME capping residue @rtype : PDBModel """ if self.verbose: self.logWrite('Capping C-terminal of chain %i with NME.' % chain ) m_nme = PDBModel( self.F_nme_cap ) c_start = model.chainIndex( breaks=breaks ) c_end = model.chainEndIndex( breaks=breaks) Nterm_is_break = False Cterm_is_break = False if breaks: Nterm_is_break = c_start[chain] not in model.chainIndex() Cterm_is_break = c_end[chain] not in model.chainEndIndex() chains_before = model.takeChains( range(chain), breaks=breaks ) m_chain = model.takeChains( [chain], breaks=breaks ) chains_after = model.takeChains( range(chain+1, len(c_start)), breaks=breaks ) m_term = m_chain.resModels()[-1] ## rename overhanging residue in cap PDB, renumber cap residue for a in m_nme: if a['residue_name'] != 'NME': a['residue_name'] = m_term.atoms['residue_name'][0] else: a['residue_number'] = m_term.atoms['residue_number'][0]+1 a['chain_id'] = m_term.atoms['chain_id'][0] a['segment_id'] = m_term.atoms['segment_id'][0] ## chain should not have any terminal O after capping m_chain.remove( ['OXT'] ) ## fit cap onto last residue of chain m_nme = m_nme.magicFit( m_term ) cap = m_nme.resModels()[-1] serial = m_term['serial_number'][-1]+1 cap['serial_number'] = range( serial, serial + len(cap) ) ## concat cap on chain m_chain = m_chain.concat( cap, newChain=False ) ## should be obsolete now if getattr( m_chain, '_PDBModel__terAtoms', []) != []: m_chain._PDBModel__terAtoms = [ len( m_chain ) - 1 ] assert m_chain.lenChains() == 1 ## re-assemble whole model r = chains_before.concat( m_chain, newChain=not Nterm_is_break) r = r.concat( chains_after, newChain=not Cterm_is_break) if len(c_start) != r.lenChains( breaks=breaks ): raise CappingError, 'Capping NME would mask a chain break. '+\ 'This typically indicates a tight gap with high risk of '+\ 'clashes and other issues.' return r
def parmMirror( self, f_out, f_out_crd=None, fmod=['frcmod.ionsjc_tip3p'], fprep=[], **kw ): """ Create a parm7 file whose atom content (and order) exactly mirrors the given PDBModel. This requires two leap runs. First we get a temporary topology, then we identify all atoms added by leap and build a final topology where these atoms are deleted. This parm is hence NOT suited for simulations but can be used to parse e.g. a trajectory or PDB into ptraj. @param f_out: target parm file @type f_out: str @param f_out_crd: target crd file (default: f_out but ending .crd) @type f_out_crd: str @param fmod : list of amber Mod files (loaded with loadAmberParams) @type fmod : [str] @param fmod : list of amber Prep files (loaded with loadAmberPrep) @type fmod : [str] """ f_out = t.absfile( f_out ) f_out_crd = t.absfile( f_out_crd ) or t.stripSuffix( f_out ) + '.crd' ## if there are hydrogens, recast them to standard amber names aatm = 'HA' in self.m.atomNames() ## 'HB2' in self.m.atomNames() ## First leap round ## m_ref = self.m.clone() m_ref.xplor2amber( aatm=aatm, parm10=True ) tmp_in = tempfile.mktemp( 'leap_in0.pdb' ) m_ref.writePdb( tmp_in, ter=3 ) tmp_parm = tempfile.mktemp( '_parm0' ) tmp_crd = tempfile.mktemp( '_crd0' ) leap_mod = self.__fLines( 'm = loadAmberParams %s\n', fmod ) leap_prep= self.__fLines( 'loadAmberPrep %s\n', fprep ) self.__runLeap( self.script_mirror_pdb, leaprc=self.leaprc, fmod=leap_mod, fprep=leap_prep, in_pdb=tmp_in, out_parm=tmp_parm, out_crd=tmp_crd, delete_atoms='' ) tmp_pdb = self.parm2pdb( tmp_parm, tmp_crd, tempfile.mktemp( 'leap_out.pdb' ), aatm=aatm ) if not self.debug: t.tryRemove( tmp_parm ) t.tryRemove( tmp_crd ) t.tryRemove( tmp_in ) ## load model with missing atoms added by leap m_leap = PDBModel( tmp_pdb ) ## compare atom content iLeap, iRef = m_leap.compareAtoms( m_ref ) ## check that ref model doesn't need any change if iRef != range( len( m_ref ) ): uLeap, uRef = m_leap.unequalAtoms( m_ref, iLeap, iRef ) atms = m_ref.reportAtoms( uRef, n=6 ) raise AmberError, "Cannot create exact mirror of %s.\n" % tmp_in +\ "Leap has renamed/deleted original atoms in %s:\n"% tmp_pdb+\ atms ## indices of atoms that were added by leap delStr = self.__deleteAtoms( m_leap, self.__inverseIndices( m_leap, iLeap ) ) ## Second leap round ## self.__runLeap( self.script_mirror_pdb, leaprc=self.leaprc, in_pdb=tmp_pdb, fmod=leap_mod, fprep=leap_prep, out_parm=f_out, out_crd=f_out_crd, delete_atoms=delStr ) if not self.debug: t.tryRemove( tmp_pdb )
class Test(BT.BiskitTest): """Test AmberParmBuilder""" TAGS = [BT.EXE] def prepare(self): root = T.testRoot() + '/amber/' self.ref = PDBModel(T.testRoot() + '/amber/1HPT_0.pdb') self.refdry = root + '1HPT_0dry.pdb' self.dryparm = tempfile.mktemp('.parm', 'dry_') self.drycrd = tempfile.mktemp('.crd', 'dry_') self.drypdb = tempfile.mktemp('.pdb', 'dry_') self.wetparm = tempfile.mktemp('.parm', 'wet_') self.wetcrd = tempfile.mktemp('.crd', 'wet_') self.wetpdb = tempfile.mktemp('.pdb', 'wet_') self.leapout = tempfile.mktemp('.out', 'leap_') def cleanUp(self): if not self.DEBUG: T.tryRemove(self.dryparm) T.tryRemove(self.drycrd) T.tryRemove(self.drypdb) T.tryRemove(self.wetparm) T.tryRemove(self.wetcrd) T.tryRemove(self.wetpdb) T.tryRemove(self.leapout) def test_AmberParmMirror(self): """AmberParmBuilder.parmMirror test""" ref = self.ref mask = N0.logical_not(ref.maskH2O()) ## keep protein and Na+ ion self.mdry = ref.compress(mask) self.a = AmberParmBuilder(self.mdry, verbose=self.local, leap_out=self.leapout, debug=self.DEBUG) self.a.parmMirror(f_out=self.dryparm, f_out_crd=self.drycrd) self.a.parm2pdb(self.dryparm, self.drycrd, self.drypdb) self.m1 = PDBModel(self.drypdb) self.m2 = PDBModel(self.refdry) eq = N0.array(self.m1.xyz == self.m2.xyz) self.assert_(eq.all()) def test_AmberParmSolvated(self): """AmberParmBuilder.parmSolvated test""" ## remove waters and hydrogens self.mdry = self.ref.compress(self.ref.maskProtein()) self.mdry = self.mdry.compress(self.mdry.maskHeavy()) self.a = AmberParmBuilder(self.mdry, leap_out=self.leapout, verbose=self.local, debug=self.DEBUG) self.a.parmSolvated(self.wetparm, f_out_crd=self.wetcrd, f_out_pdb=self.wetpdb, box=2.5) self.m3 = PDBModel(self.wetpdb) m3prot = self.m3.compress(self.m3.maskProtein()) refprot = self.ref.compress(self.ref.maskProtein()) refprot.xplor2amber() self.assertEqual(self.ref.lenChains(), self.m3.lenChains()) self.assertEqual(refprot.atomNames(), m3prot.atomNames())
class Test(BT.BiskitTest): """Test AmberParmBuilder""" TAGS = [BT.EXE] def prepare(self): root = T.testRoot() + '/amber/' self.ref = PDBModel(T.testRoot() + '/amber/1HPT_0.pdb') self.refdry = root + '1HPT_0dry.pdb' self.dryparm = tempfile.mktemp('.parm', 'dry_') self.drycrd = tempfile.mktemp('.crd', 'dry_') self.drypdb = tempfile.mktemp('.pdb', 'dry_') self.wetparm = tempfile.mktemp('.parm', 'wet_') self.wetcrd = tempfile.mktemp('.crd', 'wet_') self.wetpdb = tempfile.mktemp('.pdb', 'wet_') self.leapout = tempfile.mktemp('.out', 'leap_') def cleanUp(self): if not self.DEBUG: T.tryRemove(self.dryparm) T.tryRemove(self.drycrd) T.tryRemove(self.drypdb) T.tryRemove(self.wetparm) T.tryRemove(self.wetcrd) T.tryRemove(self.wetpdb) T.tryRemove(self.leapout) def test_AmberParmMirror(self): """AmberParmBuilder.parmMirror test""" ref = self.ref mask = N.logical_not(ref.maskH2O()) ## keep protein and Na+ ion self.mdry = ref.compress(mask) self.a = AmberParmBuilder(self.mdry, verbose=self.local, leap_out=self.leapout, debug=self.DEBUG) self.a.parmMirror(f_out=self.dryparm, f_out_crd=self.drycrd) self.a.parm2pdb(self.dryparm, self.drycrd, self.drypdb) self.m1 = PDBModel(self.drypdb) self.m2 = PDBModel(self.refdry) eq = N.array(self.m1.xyz == self.m2.xyz) self.assert_(eq.all()) def test_AmberParmSolvated(self): """AmberParmBuilder.parmSolvated test""" ## remove waters and hydrogens self.mdry = self.ref.compress(self.ref.maskProtein()) self.mdry = self.mdry.compress(self.mdry.maskHeavy()) self.a = AmberParmBuilder(self.mdry, leap_out=self.leapout, verbose=self.local, debug=self.DEBUG) self.a.parmSolvated(self.wetparm, f_out_crd=self.wetcrd, f_out_pdb=self.wetpdb, box=2.5) self.m3 = PDBModel(self.wetpdb) m3prot = self.m3.compress(self.m3.maskProtein()) refprot = self.ref.compress(self.ref.maskProtein()) refprot.xplor2amber() self.assertEqual(self.ref.lenChains(), self.m3.lenChains()) self.assertEqual(refprot.atomNames(), m3prot.atomNames()) def test_capIrregular(self): """AmberParmBuilder.capNME & capACE test""" gfp = PDBModel('1GFL') normal = gfp.takeResidues([10, 11]) chromo = gfp.takeResidues([64, 65]) self.a = AmberParmBuilder(normal) self.m4 = self.a.capACE(normal, 0) self.assertEqual(len(self.m4), 17) ## del chromo.residues['biomol'] self.m5 = self.a.capACE(chromo, 0) self.m5 = self.a.capNME(self.m5, 0) self.assertEqual(self.m5.sequence(), 'XSYX')
class Test( BT.BiskitTest ): """Test AmberParmBuilder""" TAGS = [ BT.EXE ] def prepare(self): root = T.testRoot() + '/amber/' self.ref = PDBModel( T.testRoot() + '/amber/1HPT_0.pdb') self.refdry = root + '1HPT_0dry.pdb' self.dryparm = tempfile.mktemp('.parm', 'dry_') self.drycrd = tempfile.mktemp('.crd', 'dry_') self.drypdb = tempfile.mktemp('.pdb', 'dry_') self.wetparm = tempfile.mktemp('.parm', 'wet_') self.wetcrd = tempfile.mktemp('.crd', 'wet_') self.wetpdb = tempfile.mktemp('.pdb', 'wet_') self.leapout = tempfile.mktemp('.out', 'leap_') def cleanUp(self): if not self.DEBUG: T.tryRemove( self.dryparm ) T.tryRemove( self.drycrd ) T.tryRemove( self.drypdb ) T.tryRemove( self.wetparm ) T.tryRemove( self.wetcrd ) T.tryRemove( self.wetpdb ) T.tryRemove( self.leapout ) def test_AmberParmMirror(self): """AmberParmBuilder.parmMirror test""" ref = self.ref mask = N.logical_not( ref.maskH2O() ) ## keep protein and Na+ ion self.mdry = ref.compress( mask ) self.a = AmberParmBuilder( self.mdry, verbose=self.local, leap_out=self.leapout, debug=self.DEBUG ) self.a.parmMirror(f_out=self.dryparm, f_out_crd=self.drycrd ) self.a.parm2pdb( self.dryparm, self.drycrd, self.drypdb ) self.m1 = PDBModel(self.drypdb) self.m2 = PDBModel(self.refdry) eq = N.array( self.m1.xyz == self.m2.xyz ) self.assert_( eq.all() ) def test_AmberParmSolvated( self ): """AmberParmBuilder.parmSolvated test""" ## remove waters and hydrogens self.mdry = self.ref.compress( self.ref.maskProtein() ) self.mdry = self.mdry.compress( self.mdry.maskHeavy() ) self.a = AmberParmBuilder( self.mdry, leap_out=self.leapout, verbose=self.local, debug=self.DEBUG) self.a.parmSolvated( self.wetparm, f_out_crd=self.wetcrd, f_out_pdb=self.wetpdb, box=2.5 ) self.m3 = PDBModel( self.wetpdb ) m3prot = self.m3.compress( self.m3.maskProtein() ) refprot= self.ref.compress( self.ref.maskProtein() ) refprot.xplor2amber() self.assertEqual( self.ref.lenChains(), self.m3.lenChains() ) self.assertEqual( refprot.atomNames(), m3prot.atomNames() ) def test_capIrregular( self ): """AmberParmBuilder.capNME & capACE test""" gfp = PDBModel('1GFL') normal = gfp.takeResidues([10,11]) chromo = gfp.takeResidues([64,65]) self.a = AmberParmBuilder( normal ) self.m4 = self.a.capACE( normal, 0 ) self.assertEqual( len(self.m4), 17 ) ## del chromo.residues['biomol'] self.m5 = self.a.capACE( chromo, 0 ) self.m5 = self.a.capNME( self.m5, 0 ) self.assertEqual( self.m5.sequence(), 'XSYX' )