def calcClusterNumber(self,
min_clst=5,
max_clst=30,
rmsLimit=1.0,
weight=1.13,
converged=1e-11,
aMask=None,
force=0):
"""
Calculate the approximate number of clusters needed to pass
the average intra-cluster rmsd limit.
@param min_clst: lower limit for clusters (default: 5)
@type min_clst: int
@param max_clst: upper limit for clusters (default: 30 )
@type max_clst: int
@param rmsLimit: rmsd criteria that the average of all clusters
must meet in Angstrom (default: 1.0)
@type rmsLimit: float
@param weight: fuzziness weigth (default: 1.13)
@type weight: float
@param converged: stop iteration if min dist changes less than
converged (default: 1e-11)
@type converged: float
@param force: re-calculate even if parameters haven't changed
(default: 0)
@type force: 1|0
@return: number of clusters
@rtype: int
@raise ClusterError: if can't determining number of clusters
"""
pos = [min_clst, max_clst]
while 1:
clst = int(N0.average(pos))
self.cluster(clst, weight, converged, aMask, force=force)
rmsLst = [self.avgRmsd(i, aMask)[0] for i in range(clst)]
if N0.average(rmsLst) > rmsLimit:
pos[0] = clst
else:
pos[1] = clst
if pos[1] - pos[0] == 1:
if self.verbose:
T.flushPrint(
'Converged at %i clusters, current average cluster rmsd %.2f\n'
% (clst, N0.average(rmsLst)))
return pos[1]
if pos[1] - pos[0] != 1:
if self.verbose:
T.flushPrint(
'Current cluster setting %i, current average cluster rmsd %.2f\n'
% (clst, N0.average(rmsLst)))
if pos[1] - pos[0] <= 0 or pos[0] < min_clst or pos[1] > max_clst:
raise ClusterError("Error determining number of clusters")
def calcClusterNumber( self, min_clst=5, max_clst=30, rmsLimit=1.0,
weight=1.13, converged=1e-11, aMask=None, force=0 ):
"""
Calculate the approximate number of clusters needed to pass
the average intra-cluster rmsd limit.
@param min_clst: lower limit for clusters (default: 5)
@type min_clst: int
@param max_clst: upper limit for clusters (default: 30 )
@type max_clst: int
@param rmsLimit: rmsd criteria that the average of all clusters
must meet in Angstrom (default: 1.0)
@type rmsLimit: float
@param weight: fuzziness weigth (default: 1.13)
@type weight: float
@param converged: stop iteration if min dist changes less than
converged (default: 1e-11)
@type converged: float
@param force: re-calculate even if parameters haven't changed
(default: 0)
@type force: 1|0
@return: number of clusters
@rtype: int
@raise ClusterError: if can't determining number of clusters
"""
pos = [ min_clst, max_clst ]
while 1:
clst = int( N0.average(pos) )
self.cluster( clst, weight, converged, aMask, force=force )
rmsLst = [ self.avgRmsd(i, aMask)[0] for i in range(clst)]
if N0.average( rmsLst ) > rmsLimit:
pos[0] = clst
else:
pos[1] = clst
if pos[1]-pos[0] == 1:
if self.verbose:
T.flushPrint('Converged at %i clusters, current average cluster rmsd %.2f\n'%( clst, N0.average( rmsLst ) ))
return pos[1]
if pos[1]-pos[0] != 1:
if self.verbose:
T.flushPrint('Current cluster setting %i, current average cluster rmsd %.2f\n'%( clst, N0.average( rmsLst ) ))
if pos[1]-pos[0]<= 0 or pos[0]<min_clst or pos[1]>max_clst:
raise ClusterError("Error determining number of clusters")
def blockFit2ref( self, refModel=None, mask=None, conv=1e-6 ):
"""
Fit trajectory until convergence onto it's own average and then
transform the average of all frames onto the reference. To be used
with parallell trajectories.
:param refModel: Reference model (default: None)
:type refModel: PDBModel
:param mask: atom mask to apply before fitting
(default: None, all atoms)
:type mask: [1|0]
:param conv: convergence cutoff creterion (default: 1e-6)
:type conv: float
"""
self.fit( ref=self.ref )
m_avg = self.avgModel()
## fit on average until it's not getting better
d = 1.
dd= 1.
while dd >= conv:
self.fit( ref=m_avg, mask=mask )
m_new_avg = self.avgModel()
oldD, d = d, m_avg.rms( m_new_avg, mask=mask )
T.flushPrint( "rms difference: %f" % d )
dd = oldD - d
m_avg = m_new_avg
## transform trajectory en block onto reference
if refModel:
T.flushPrint('fitting trajectory en-block onto reference...')
if refModel.atomNames() != self.ref.atomNames():
ref_i, i = refModel.compareAtoms( m_avg )
refModel = refModel.take( ref_i )
m_avg = m_avg.take( i )
r, t = m_avg.transformation( refModel, mask )
self.transform( r, t )
out = T.absfile(o['o'])
inLst = T.toList(o['i'])
start = int(o.get('s', '0'))
end = o.get('e', None)
if end:
end = int(end)
step = int(o.get('step', 1))
prot = 'prot' in o
ref = o.get('ref', None)
if ref:
ref = PDBModel(T.absfile(ref))
if 'prot' in o:
ref = ref.compress(ref.maskProtein())
T.flushPrint("Loading and appending trajectories...")
r = loadTraj(inLst[0], 0, start, end, step, prot=prot)
for i in range(1, len(inLst)):
t = loadTraj(inLst[i], i, start, end, step)
if t.ref.equals(r.ref) != [1, 1]:
raise Exception('Incompatible reference structure.')
r = r.concat(t)
T.flushPrint('#')
print(" Done")
pdb - PDB code to be stored in trajectory
""")
sys.exit(0)
##########
## MAIN ##
use()
o = T.cmdDict( {'n':10} )
f_in = T.absfile( o['i'] )
f_out = T.absfile( o.get('o', f_in) )
n = int( o['n'] )
T.flushPrint("Loading...")
t = T.load( f_in )
T.flushPrint("Converting %i frames..." % len(t) )
if isinstance(t, EnsembleTraj ):
T.flushPrint( "Nothing to be done!\n")
sys.exit(0)
t = traj2ensemble( t, n )
if 'pdb' in o:
t.ref.pdbCode = o['pdb']
if f_in == f_out:
os.rename( f_in, f_in + '_backup')
########
## MAIN
########
syntax()
## get and clean up options
o = options()
o['step'] = int( o['step'] )
o['i'] = T.absfile( o['i'] )
o['o'] = o.get('o',
'%s/%s_rms.eps' % (osp.dirname(o['i']), T.stripFilename(o['i'])))
o['show'] = 'show' in o
T.flushPrint( "Loading..." )
t = T.load( o['i'] )
T.flushPrint( "done loading trajectory with %i frames." % len(t) )
if o['step'] != 1:
t = t.thin( o['step'] )
T.flushPrint( "Fitting ...")
calcRmsd( t )
T.flushPrint( "done." )
p = plot( t, o.get( 'title', T.stripFilename(o['i']) ) )
if o['show']:
p.show()
########
## MAIN
########
syntax()
## get and clean up options
o = options()
o['z'] = float(o['z'])
o['step'] = int(o['step'])
o['i'] = T.absfile(o['i'])
o['o'] = o.get(
'o', '%s/%s_rms.eps' % (osp.dirname(o['i']), T.stripFilename(o['i'])))
o['show'] = 'show' in o
T.flushPrint("Loading...")
t0 = T.load(o['i'])
T.flushPrint("done loading %i member trajectory." % t0.n_members)
t = t0
if o['step'] != 1:
t = t0.thin(o['step'])
T.flushPrint("Fitting %i members..." % t.n_members)
calcRmsd(t)
T.flushPrint("done.")
p = plot(t, o.get('title', T.stripFilename(o['i'])))
if o['z'] > 0.0:
T.flushPrint("Getting outliers ...")