def guessFiles(refiner,homeDir):
if homeDir == None:
homeDir = os.getcwd( )
seqFile = os.path.join(homeDir,'sequence.seq')
if os.path.exists(seqFile):
refiner.readSequence(seqFile)
else:
pdbFile = os.path.join(homeDir,'*.pdb')
pdbFiles = glob.glob(pdbFile)
if (len(pdbFiles) > 0):
molio.readPDB(pdbFiles[0])
angleFiles = glob.glob(os.path.join(homeDir,'dihedral*.tbl'))
for file in angleFiles:
refiner.addAngleFile(file,'nv')
disFiles = glob.glob(os.path.join(homeDir,'distance*.tbl'))
for file in disFiles:
refiner.addDistanceFile(file,'nv')
angleFiles = glob.glob(os.path.join(homeDir,'dihedral*.txt'))
for file in angleFiles:
refiner.addAngleFile(file,'cyana')
disFiles = glob.glob(os.path.join(homeDir,'*.upl'))
for file in disFiles:
file = file[:-4]
refiner.addDistanceFile(file,'cyana')
def setupFile(fileName,disFile=''):
refiner = refine()
molecule = molio.readPDB(fileName)
if disFile != '':
refiner.addDistanceFile(disFile,mode='NV')
else:
disFile = 'distances'
refiner.addDistanceFile(disFile,mode='cyana')
seed = 0
refiner.setup(homeDir,seed)
refiner.outDir = outDir
refiner.energy()
return refiner
def predictWithRCFromPDB(pdbFile, refShifts, ringRatio, typeRCDist, atomNames, builtin):
"""
Reads an RNA molecule from PDB file and predict chemical shifts.
# Parameters:
pdbFile (str); the name of the PDB file
refShifts (dict); the reference shifts for each atom to be predicted.
ringRatio (float); A scale parameter to multiply the ring-current contributions by
typeRCDist (String): Type of analysis to perform, 'rc' (ring current) or 'dist' (distances)
# Returns:
shifts (dict) the chemical shifts that were predicted. The reference ppm
value of each atom is also updated with shift
"""
Molecule.removeAll()
pdb = PDBFile()
mol = molio.readPDB(pdbFile)
#pdb.readCoordinates(pdbFile,-1,False, False)
activeStructures = mol.getActiveStructures()
avgOverStructures = False
if not avgOverStructures:
if len(activeStructures) > 0:
repI = super.findRepresentative(mol)
iStruct = repI[0]
else:
iStruct=0
structList=[iStruct]
else:
if len(activeStructures) > 0:
structList = list(activeStructures)
else:
structList=[0]
if builtin:
shifts = rnapred.predictBuiltIn(mol, atomNames, typeRCDist, structList)
else:
if typeRCDist.lower() == 'rc':
shifts = rnapred.predictRCShifts(mol, structList, refShifts, ringRatio, ringTypes)
elif typeRCDist.lower() == 'dist':
alphas = ringRatio
shifts = rnapred.predictDistShifts(mol, rmax, structList, refShifts, alphas)
return mol, shifts
def genRCTrainingMatrix(outFileName, pdbFiles, shiftSources, atomNames, ringMode, typeRCDist):
"""
Generate the training data from a list of pdbFiles and dotBracket values.
Each file is predicted using the attribute method based on a specified
dot-bracket string and the output is appended to the training matrix
# Parameters:
outFileName (str); The output file name. File is deleted if present already.
pdbFiles (list); list of PDB Files to use
shiftSources (list); list of sources for shifts. Either dot-bracket values (vienna string) or bmrb id to use for each pdb file
atomNames (list): list of atom names
ringMode
typeRCDist (String): Type of analysis to perform, 'rc' (ring current) or 'dist' (distances)
# Returns:
_ (None); Training data is written to specified file
"""
try:
os.remove(outFileName)
except:
pass
with open(outFileName,'a') as f1:
for pdbID,shiftSource in zip(pdbFiles,shiftSources):
Molecule.removeAll()
pdbFile = 'pdbfiles/'+pdbID+'.pdb'
if not getPDBFile(pdbID):
print 'skip',pdbFile
continue
print 'train',pdbFile
pdb = PDBFile()
mol = molio.readPDB(pdbFile)
if shiftSource[0]=="." or shiftSource[0]=="(":
rnapred.predictFromSequence(mol,shiftSource)
else:
setRefShiftsFromBMRB(shiftSource, {})
genRCMat(mol,atomNames,f1, ringMode, typeRCDist)
input = FileInputStream(argFile)
yaml = Yaml()
data = yaml.load(input)
refiner=refine()
osfiles.setOutFiles(refiner,dataDir, 0)
refiner.rootName = "temp"
refiner.loadFromYaml(data,0)
mol = refiner.molecule
vienna = data['rna']['vienna']
mol.setDotBracket(vienna)
#rnapred.predictFromSequence(mol,vienna)
rnapred.predictFromSequence()
rnapred.dumpPredictions(mol)
elif argFile.endswith('.pdb') or argFile.endswith('.cif'):
if argFile.endswith('.pdb'):
mol = molio.readPDB(argFile)
elif argFile.endswith('.cif'):
mol = molio.readMMCIF(argFile)
else:
print 'Invalid file type'
exit(1)
# fixme, need to do this by polymer so you can have protein-rna complex
if isRNA(mol):
refiner=refine()
shifts = refiner.predictRNAShifts()
dumpPredictions(mol)
else:
predictProtein(mol)