def make_confspace(cfs_file, data, cfs_dir=DEFAULT_CFS_DIR):
"""make_confspace
Generate conf spaces from input file
Takes as input a cfs_file that contains chain definitions, flexible
residues, and mutable residues.
Returns a dictionary containing forcefield parameters and OSPREY
confspace objects for:
protein
ligand
complex
"""
# Get the pdb code for this design
data["pdb"] = cfs_file[:4]
print("pdb: %s" % data["pdb"])
# Load the design conformation space information
data["cfs"] = os.path.join(cfs_dir, cfs_file)
confspace = imp.load_source('confspace', data["cfs"])
print("CFS File: %s" % data["cfs"])
# Get the sequences if they exist
try:
data["sequences"] = confspace.sequences
except AttributeError:
data["sequences"] = [None]
ffparams = osprey.ForcefieldParams()
mol = osprey.readPdb(confspace.mol)
template_library = osprey.TemplateLibrary(ffparams.forcefld)
# Make sure we don't have 3 strands (some cfs files do)
assert len(confspace.strand_defs) == 2
# Define the protein strand
protein = osprey.Strand(mol,
templateLib=template_library,
residues=confspace.strand_defs["strand0"])
for resi, res_allowed in confspace.strand_flex["strand0"].iteritems():
protein.flexibility[resi]\
.setLibraryRotamers(osprey.WILD_TYPE, *res_allowed)\
.setContinuous()\
.addWildTypeRotamers()
# Define the ligand strand
ligand = osprey.Strand(mol,
templateLib=template_library,
residues=confspace.strand_defs["strand1"])
for resi, res_allowed in confspace.strand_flex["strand1"].iteritems():
ligand.flexibility[resi]\
.setLibraryRotamers(osprey.WILD_TYPE,*res_allowed)\
.setContinuous()\
.addWildTypeRotamers()
# Build spaces
return {
'protein': osprey.ConfSpace(protein),
'ligand': osprey.ConfSpace(ligand),
'complex': osprey.ConfSpace([protein, ligand]),
'ffparams': ffparams
}
import osprey
osprey.start()
# define a strand
strand = osprey.Strand('1CC8.ss.pdb')
strand.flexibility['A2'].setLibraryRotamers('ALA', 'GLY')
strand.flexibility['A3'].setLibraryRotamers(osprey.WILD_TYPE, 'VAL')
strand.flexibility['A4'].setLibraryRotamers(osprey.WILD_TYPE)
# make the conf space
confSpace = osprey.ConfSpace(strand)
# choose a forcefield
ffparams = osprey.ForcefieldParams()
# how should we compute energies of molecules?
ecalc = osprey.EnergyCalculator(confSpace, ffparams)
# how should we define energies of conformations?
confEcalc = osprey.ConfEnergyCalculator(confSpace, ecalc)
# compute the energy matrix
emat = osprey.EnergyMatrix(confEcalc)
# run DEE with just steric pruning
pmat = osprey.DEE(confSpace, emat, showProgress=True)
# or run DEE with Goldstein pruning
#i0 = 10.0 # kcal/mol
protein.flexibility['A8'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# define the ligand strand
ligand = osprey.Strand(mol, templateLib=templateLib, residues=['A38', 'A58'])
ligand.flexibility['A41'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A42'].setLibraryRotamers(
osprey.WILD_TYPE, 'THR', 'LYS').addWildTypeRotamers().setContinuous()
ligand.flexibility['A45'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A46'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# make the conf space for the protein+ligand complex
complexConfSpace = osprey.ConfSpace([protein, ligand])
# how should we compute energies of molecules?
# (give the complex conf space to the ecalc since it knows about all the templates and degrees of freedom)
parallelism = osprey.Parallelism(cpuCores=4)
ecalc = osprey.EnergyCalculator(complexConfSpace,
ffparams,
parallelism=parallelism)
# configure PAStE
paste = osprey.Paste(
complexConfSpace,
numPDBs=15,
epsilon=
0.68, # you proabably want something more precise in your real designs
useWindowCriterion=True,
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
design.flexibility['G651'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# define the ligand strand
ligand = osprey.Strand(mol, templateLib=templateLib, residues=['A155', 'A194'])
ligand.flexibility['A172'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A192'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A193'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# make a multi-state conf space
confSpace = osprey.MultiStateConfSpace([
osprey.StateMutable('design', osprey.ConfSpace(design)),
osprey.StateUnmutable('ligand', osprey.ConfSpace(ligand)),
osprey.StateMutable('complex', osprey.ConfSpace([design, ligand]))
])
# train a LUTE model for each state
for state in confSpace.states:
# how should we compute energies of molecules?
ecalc = osprey.EnergyCalculator(state.confSpace,
ffparams,
parallelism=parallelism)
# how should we define energies of conformations?
eref = osprey.ReferenceEnergies(state.confSpace, ecalc)
confEcalc = osprey.ConfEnergyCalculator(state.confSpace,
protein.flexibility['G654'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# define the ligand strand
ligand = osprey.Strand(mol, templateLib=templateLib, residues=['A155', 'A194'])
ligand.flexibility['A156'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A172'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A192'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A193'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# make the conf space for the protein
proteinConfSpace = osprey.ConfSpace(protein)
# make the conf space for the ligand
ligandConfSpace = osprey.ConfSpace(ligand)
# make the conf space for the protein+ligand complex
complexConfSpace = osprey.ConfSpace([protein, ligand])
# how should we compute energies of molecules?
# (give the complex conf space to the ecalc since it knows about all the templates and degrees of freedom)
parallelism = osprey.Parallelism(cpuCores=4)
minimizingEcalc = osprey.EnergyCalculator(complexConfSpace,
ffparams,
parallelism=parallelism,
isMinimizing=True)
ffparams = osprey.ForcefieldParams()
# read a PDB file for molecular info
mol = osprey.readPdb('2RL0.min.reduce.pdb')
# define the conformation space
strand = osprey.Strand(mol, residues=['G648', 'G654'])
for resNum in ['G649', 'G650', 'G651']:
strand.flexibility[resNum]\
.setLibraryRotamers(osprey.WILD_TYPE)\
.addWildTypeRotamers()\
.setContinuous()
confSpace = osprey.MultiStateConfSpace([
# yeah, we didn't define any mutations,
# but SOFEA still needs one "mutable" state to define the sequence space
osprey.StateMutable('protein', osprey.ConfSpace(strand))
])
# we only care about the "wild-type" sequence in this case
seq = confSpace.seqSpace.makeWildTypeSequence()
# how should we compute energies of molecules?
ecalc = osprey.EnergyCalculator(confSpace, ffparams, parallelism=parallelism)
# make a function that makes a SOFEA config object for a state
def config(state):
# how should we define energies of conformations?
confEcalc = osprey.ConfEnergyCalculator(
state.confSpace,
ecalc,
customizedTemplateLib = osprey.TemplateLibrary(
extraTemplates=[customTemplates],
extraTemplateCoords=[customTemplateCoords],
extraRotamers=[customRotamers])
# or read templates from files
# customizedTemplateLibFromFiles = osprey.TemplateLibrary(
# extraTemplates=['/path/to/templates/file']
# etc...
# )
# or completely replace default templates
# completelyCustomTemplateLib = osprey.TemplateLibrary(
# defaultTemplates=False,
# extraTemplates=['/path/to/all/templates'],
# etc...
# )
# load the molecule and make the strand using our custom template library
protein = osprey.Strand('1CC8.ss.pdb', templateLib=customizedTemplateLib)
# make the conf space
protein.flexibility['A2'].setLibraryRotamers('ALA', 'GLY')
protein.flexibility['A3'].setLibraryRotamers(osprey.WILD_TYPE, 'VAL',
'ARG').setContinuous(10)
protein.flexibility['A4'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers()
confSpace = osprey.ConfSpace(protein)
# continue design with confSpace
# define the ligand strand
ligandStrand = osprey.Strand(mol,
templateLib=templateLib,
residues=['A155', 'A194'])
ligandStrand.flexibility['A156'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligandStrand.flexibility['A172'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligandStrand.flexibility['A192'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligandStrand.flexibility['A193'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
complex = osprey.EwakstarDoer_State(
'PL', osprey.ConfSpace([proteinStrand, ligandStrand]))
# number of CPUs
cpus = 4
# configure EWAKStar
ewakstarDoer = osprey.EwakstarDoer(
state=complex,
smaNodes=0,
useSMA=False,
printPDBs=True,
# do we want to set our baseline at the wild-type sequence? as in: do we only want sequences better than wild-type?
useWtBenchmark=False,
logFile='ewakstar.sequences.tsv',
templateLib=templateLib,
residues=[startResL, endResL])
for i in range(0, len(resNumsL)):
ligand.flexibility[resNumsL[i]].setLibraryRotamers(
*AATypeOptions[posL[i]]).addWildTypeRotamers().setContinuous()
# define the ligand strand
protein = osprey.Strand(mol,
templateLib=templateLib,
residues=[startResP, endResP])
for i in range(0, len(resNumsP)):
protein.flexibility[resNumsP[i]].setLibraryRotamers(
*AATypeOptions[posP[i]]).addWildTypeRotamers().setContinuous()
# make the conf space for the protein
confSpaceP = osprey.ConfSpace(protein)
# make the conf space for the ligand
confSpaceL = osprey.ConfSpace(ligand)
# make the conf space for the protein+ligand complex
confSpace = osprey.ConfSpace([protein, ligand])
# how should we compute energies of molecules?
# (give the complex conf space to the ecalc since it knows about all the templates and degrees of freedom)
numCPUs = 40
parallelism = osprey.Parallelism(cpuCores=numCPUs)
ecalc = osprey.EnergyCalculator(confSpace, ffparams, parallelism=parallelism)
rigidEcalc = osprey.EnergyCalculator(confSpace,
ffparams,
parallelism=parallelism,
# define the ligand strand
ligandStrand = osprey.Strand(mol,
templateLib=templateLib,
residues=['A155', 'A194'])
ligandStrand.flexibility['A156'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligandStrand.flexibility['A172'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligandStrand.flexibility['A192'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligandStrand.flexibility['A193'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# define the MSK* states
protein = osprey.MSKStar_State('Protein', osprey.ConfSpace(proteinStrand))
ligand = osprey.MSKStar_State('Ligand', osprey.ConfSpace(ligandStrand))
complex = osprey.MSKStar_State('Complex',
osprey.ConfSpace([proteinStrand, ligandStrand]))
# configure COMETS
mskstar = osprey.MSKStar(
objective=osprey.MSKStar_LMFE({
complex: 1.0,
protein: -1.0,
ligand: -1.0
}),
logFile='mskstar.tsv',
# how precisely should we estimate partition function values?
# (in range (0,1], lower is more precise)
# define a strand
strand = osprey.Strand('1CC8.ss.pdb')
strand.flexibility['A38'].setLibraryRotamers('ILE', 'ALA',
'GLY').setContinuous()
strand.flexibility['A39'].setLibraryRotamers("Ser", "Ala",
"Gly").setContinuous()
strand.flexibility['A40'].setLibraryRotamers("Met", "Ser", "Ala", "Gly",
"Leu").setContinuous()
strand.flexibility['A41'].setLibraryRotamers("Glu", "Ala",
"Gly").setContinuous()
strand.flexibility['A42'].setLibraryRotamers("Ala", "Gly").setContinuous()
# make the conf space
bbflex = osprey.c.confspace.CATSStrandFlex(strand, 'A38', 'A42')
confSpace = osprey.ConfSpace([[strand, bbflex]])
# choose a forcefield
ffparams = osprey.ForcefieldParams()
ffparams.solvScale = 0.
# how should we compute energies of molecules?
ecalc = osprey.EnergyCalculator(confSpace, ffparams)
eref = osprey.ReferenceEnergies(confSpace, ecalc)
# how should we define energies of conformations?
confEcalc = osprey.ConfEnergyCalculator(confSpace,
ecalc,
referenceEnergies=eref)
def loadd_confspaces(config,
xtal_rotamers=True,
continuous=False,
force_wt=True):
"""Loads OSPREY ConfSpace objects from a config dictionary.
OPTIONS:
xtal_rotamers: if true, add the rotamers from the input pdb
continuous: if true, set continuous rotamers
force_wt: if true, always add the wild-type amino acid
"""
ffparams = osprey.ForcefieldParams()
mol = osprey.readPdb(join(PDB, config["molecule"]))
template_library = osprey.TemplateLibrary(ffparams.forcefld)
# Make sure we don't have 3 strands (some cfs files do)
if len(config["strand_definitions"]) > 2:
exit()
# Define the protein strand
protein = osprey.Strand(mol,
templateLib=template_library,
residues=config["strand_definitions"]["strand0"])
for resi, res_allowed in config["strand_mutations"]["strand0"].items():
# Add the osprey.WILD_TYPE object to the allowed residues if desired
if force_wt:
res_allowed.append(osprey.WILD_TYPE)
# Set the flexibility
protein.flexibility[resi]\
.setLibraryRotamers(*res_allowed)
if xtal_rotamers:
protein.flexibility[resi].addWildTypeRotamers()
if continuous:
protein.flexibility[resi].setContinuous()
# Define the ligand strand
ligand = osprey.Strand(mol,
templateLib=template_library,
residues=config["strand_definitions"]["strand1"])
for resi, res_allowed in config["strand_mutations"]["strand1"].items():
# Add the osprey.WILD_TYPE object to the allowed residues if desired
if force_wt:
res_allowed.append(osprey.WILD_TYPE)
# Set the flexibility
ligand.flexibility[resi]\
.setLibraryRotamers(*res_allowed)
if xtal_rotamers:
ligand.flexibility[resi].addWildTypeRotamers()
if continuous:
ligand.flexibility[resi].setContinuous()
# Build spaces
return {
'protein': osprey.ConfSpace(protein),
'ligand': osprey.ConfSpace(ligand),
'complex': osprey.ConfSpace([protein, ligand]),
'ffparams': ffparams
}
strand1.flexibility['A3'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers()
strand1.flexibility['A4'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers()
# define another strand
strand2 = osprey.Strand(mol, residues=['A21', 'A40'])
strand2.flexibility['A21'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers()
strand2.flexibility['A22'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers()
strand2.flexibility['A23'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers()
# make the conf space
confSpace = osprey.ConfSpace([strand1, strand2])
# choose a forcefield
ffparams = osprey.ForcefieldParams()
# how to compute the energy of a conformation?
ecalc = osprey.EnergyCalculator(confSpace, ffparams)
confEcalc = osprey.ConfEnergyCalculator(confSpace, ecalc)
# how should confs be ordered and searched?
emat = osprey.EnergyMatrix(confEcalc)
astar = osprey.AStarMPLP(emat, confSpace)
# find the best sequence and rotamers
gmec = osprey.GMECFinder(astar, confEcalc).find()
endResL = 'B161'
AATypeOptions = [['HID', 'ASP', 'GLU', 'SER', 'THR', 'ASN', 'GLN', 'ALA', 'VAL', 'ILE', 'LEU', 'CYS', 'GLY'], ['ALA', 'LEU', 'ILE', 'VAL', 'PHE', 'TYR', 'MET', 'GLU', 'ASP','HID', 'ASN', 'GLN', 'GLY'], ['CYS', 'SER', 'THR', 'ASN', 'GLN', 'HIE', 'ALA', 'VAL', 'ILE', 'LEU', 'PHE', 'TYR', 'GLY'], ['ILE', 'ALA', 'LEU', 'VAL', 'PHE', 'TYR', 'MET','GLU', 'ASP', 'HID', 'ASN', 'GLN', 'GLY'], ['GLU', 'ASP', 'PHE', 'TYR', 'ALA', 'VAL', 'ILE', 'LEU', 'HIE', 'HID', 'ASN', 'GLN', 'GLY'], ['TYR', 'PHE', 'ALA', 'VAL', 'ILE','LEU', 'GLU', 'ASP', 'HID', 'HIE', 'ASN', 'GLN', 'GLY'], ['PHE', 'TYR', 'ALA', 'VAL', 'ILE', 'LEU', 'GLU', 'ASP', 'HID', 'HIE', 'ASN', 'GLN', 'GLY'], ['TRP', 'ALA', 'VAL','ILE', 'LEU', 'PHE', 'TYR', 'MET', 'SER', 'THR', 'ASN', 'GLN', 'GLY'], ['TYR', 'PHE', 'ALA', 'VAL', 'ILE', 'LEU', 'GLU', 'ASP', 'HID', 'HIE', 'ASN', 'GLN', 'GLY']]
# make sure all strands share the same template library (including wild-type rotamers)
templateLib = osprey.TemplateLibrary(ffparams.forcefld, moleculesForWildTypeRotamers=[mol])
# define the protein strand
ligand = osprey.Strand(mol, templateLib=templateLib, residues=[startResL, endResL])
for i in range(0,len(resNumsL)):
ligand.flexibility[resNumsL[i]].setLibraryRotamers(*AATypeOptions[posL[i]]).addWildTypeRotamers().setContinuous()
# define the ligand strand
protein = osprey.Strand(mol, templateLib=templateLib, residues=[startResP, endResP])
for i in range(0,len(resNumsP)):
protein.flexibility[resNumsP[i]].setLibraryRotamers(*AATypeOptions[posP[i]]).addWildTypeRotamers().setContinuous()
complex = osprey.EwakstarDoer_State('PL', osprey.ConfSpace([protein, ligand]))
# number of CPUs
cpus = 40;
# configure EWAKStar
ewakstarDoer = osprey.EwakstarDoer(
state = complex,
# do we want to set our baseline at the wild-type sequence? as in: do we only want sequences better than wild-type?
useWtBenchmark = False,
# number of sequences we want to limit ourselves to during the "sequence filter" portion of ewakstarDoer
numEWAKStarSeqs = 10000,
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
design.flexibility['G651'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# define the ligand strand
ligand = osprey.Strand(mol, templateLib=templateLib, residues=['A155', 'A194'])
ligand.flexibility['A172'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A192'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A193'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# make a single-state conf space:
confSpace = osprey.MultiStateConfSpace(
[osprey.StateMutable('complex', osprey.ConfSpace([design, ligand]))])
complexState = confSpace.getState('complex')
# use the "wild-type" sequence
seq = confSpace.seqSpace.makeWildTypeSequence()
# (or any other sequence defined in your conf space)
#seq = confSpace.seqSpace.makeUnassignedSequence()\
# .set('G649', 'ALA')
# how should we compute energies of molecules?
ecalc = osprey.EnergyCalculator(confSpace, ffparams, parallelism=parallelism)
# make a function that makes a SOFEA config object for a state
def config(state):
protein.flexibility['G654'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# define the ligand strand
ligand = osprey.Strand(mol, templateLib=templateLib, residues=['A155', 'A194'])
ligand.flexibility['A156'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A172'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A192'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
ligand.flexibility['A193'].setLibraryRotamers(
osprey.WILD_TYPE).addWildTypeRotamers().setContinuous()
# define the COMETS states
bound = osprey.COMETS_State('Bound', osprey.ConfSpace([protein, ligand]))
unbound = osprey.COMETS_State('Unbound', osprey.ConfSpace(protein))
# configure COMETS
comets = osprey.COMETS(objective=osprey.COMETS_LME({
bound: 1.0,
unbound: -1.0
}),
logFile='comets.tsv')
# how should we compute energies of molecules?
# (give all the conf spaces to the ecalc)
confSpaces = [state.confSpace for state in comets.states]
parallelism = osprey.Parallelism(cpuCores=4)
ecalc = osprey.EnergyCalculator(confSpaces, ffparams, parallelism=parallelism)
