def test_simplified_codes(self):
traj = pt.fetch_pdb('1l2y')
data_full = pt.dssp(traj)[1]
data_sim = pt.dssp(traj, simplified=True)[1]
expected_1st = ['C', 'H', 'H', 'H', 'H', 'H', 'H', 'H', 'C', 'C', 'H',
'H', 'H', 'H', 'C', 'C', 'C', 'C', 'C', 'C']
assert expected_1st == data_sim[0].tolist(
), 'test_simplified_codes: must equal'
def test_simplified_codes(self):
traj = pt.load(fn("1L2Y.pdb"))
pt.dssp(traj)[1]
data_sim = pt.dssp(traj, simplified=True)[1]
expected_1st = [
'C', 'H', 'H', 'H', 'H', 'H', 'H', 'H', 'C', 'C', 'H', 'H', 'H',
'H', 'C', 'C', 'C', 'C', 'C', 'C'
]
assert expected_1st == data_sim[0].tolist(
), 'test_simplified_codes: must equal'
def test_simplified_codes(self):
traj = pt.fetch_pdb('1l2y')
data_full = pt.dssp(traj)[1]
data_sim = pt.dssp(traj, simplified=True)[1]
expected_1st = [
'C', 'H', 'H', 'H', 'H', 'H', 'H', 'H', 'C', 'C', 'H', 'H', 'H',
'H', 'C', 'C', 'C', 'C', 'C', 'C'
]
assert expected_1st == data_sim[0].tolist(
), 'test_simplified_codes: must equal'
def test_vs_cpptraj(self):
data = pt.dssp(self.traj, "*", dtype='cpptraj_dataset')
data_int = np.array(
[d0.values for d0 in data if d0.dtype == 'integer'], dtype='i4')
# load cpptraj output
cpp_data = np.loadtxt("./data/dssp.Tc5b.dat", skiprows=1)[:, 1:].T
aa_eq(data_int.flatten(), cpp_data.flatten())
def test_vs_cpptraj(self):
traj = pt.iterload(fn('Tc5b.x'), fn('Tc5b.top'))
data = pt.dssp(traj, "*", dtype='cpptraj_dataset')
data_int = np.array(
[d0.values for d0 in data if d0.dtype == 'integer'], dtype='i4')
# load cpptraj output
cpp_data = np.loadtxt(fn("dssp.Tc5b.dat"), skiprows=1)[:, 1:].T
aa_eq(data_int.flatten(), cpp_data.flatten())
def test_vs_cpptraj(self):
data = pt.dssp(self.traj, "*", dtype='cpptraj_dataset')
data_int = np.array(
[d0.values for d0 in data if d0.dtype == 'integer'],
dtype='i4')
# load cpptraj output
cpp_data = np.loadtxt("./data/dssp.Tc5b.dat", skiprows=1)[:, 1:].T
aa_eq(data_int.flatten(), cpp_data.flatten())
def calculate_angles(traj, residues=[], angles='phi psi'):
if residues:
data = list(
pt.multidihedral(traj, dihedral_types=angles, resrange=residues))
else:
# get all residues numbers
residues = [int(l[4:]) for l in list(pt.dssp(traj)[0])]
data = pt.multidihedral(traj, dihedral_types=angles, resrange=residues)
return data
def set_resrange(residues, traj):
print("Selected residues:", residues)
# use all residues in topology
if not residues:
residues = [int(l[4:]) for l in list(pt.dssp(traj)[0])]
# range of residues
elif '-' in residues:
start, end = residues.split('-')
residues = list(np.arange(int(start), int(end)))
# list of residues
else:
residues = residues.split(',')
residues = [int(r) for r in residues]
return residues
def test_frame_indices(self):
from numpy.testing import assert_equal
traj = pt.iterload(fn('Tc5b.x'), fn('Tc5b.top'))
s_0 = pt.dssp(traj)[1]
s_1 = pt.dssp(traj, frame_indices=[0, 2, 5])[1]
assert_equal(s_0[[0, 2, 5]], s_1)
def compare_traj2PDB(traj, ref_pdb, tot_res, dih, residue_array=[1,0]):
'''
traj is your pytraj trajectory to compare to the pdb structure
generated in the python script as traj = pt.iterload([crd], top=top)
tot_res is total residue
ref_pdb is pdb path e.g ./PDB/1hhp.pdb
residue_array is an array has the minimum and maximum residue number
if residue_array is not given then function use from residue 1 to total residue
'''
#traj = traj[0:99] # 1 ns
r0 = residue_array[0] # the lowest residue #
# the last or highest residue is total residue if the residue_array is not specify
if residue_array[1] == 0:
r1 = tot_res
else:
r1 = residue_array[1] # the highest residue #
# reference traj which is the pdb, if you are using a trajectory frame, use cpptraj and make that
# frame into a pdb
ref_traj = pt.load(ref_pdb)
# length of trajectory
traj_len = len(traj)
# set up a dataframe that frames on the y axis (index) and dihedral, dih, of a residue# on x axis (columns)
data_df = pd.DataFrame(index = np.arange(1, traj_len+1, 1))
# dataframe that holds the reference dihedral value
data_df_ref = pd.DataFrame()
# for each residue in ....
for k in np.arange(r0, r1+1, 1):
# for phi
if (dih == 'phi' and k != 1): # no phi for first residue
# cheap method to get the residues, a hack but works
# res is the residue name and k is the residue number
residues, ss, _ = pt.dssp(ref_traj, ":%s" %(k))
res = residues[0][0:3]
# get the index for phi
indx = '%s' %(dihmask(int(k), dih, res))
# calculate the phi dihedral value for residue k in the trajectory and store in data1
data1 = cal_dih(traj, indx)
# create a column name for the residue. e.g gly17_phi and store values in column
data_df['%s%s_%s'%(res,k,dih)] = data1
# repeat for reference, should remove out of the loop?
data_ref1 = cal_dih(ref_traj, indx)
data_df_ref['%s%s_%s'%(res,k,dih)] = data_ref1
# set to nan for the 1st residue if it is a phi dihedral
elif (dih == 'phi' and k ==1):
# res is the residue name and k is the residue number
residues, ss, _ = pt.dssp(ref_traj, ":%s" %(k))
res = residues[0][0:3]
data_df['%s%s_%s'%(res,k,dih)] = "Nan"
data_df_ref['%s%s_%s'%(res,k,dih)] = "Nan"
# for psi
elif (dih == 'psi' and k != tot_res): # no psi for last residue
# cheap method to get the residues, a hack but works
# res is the residue name and k is the residue number
residues, ss, _ = pt.dssp(ref_traj, ":%s" %(k))
res = residues[0][0:3]
# get the index for psi
indx = '%s' %(dihmask(int(k), dih, res))
# calculate the psi dihedral value for residue k in the trajectory and store in data1
data1 = cal_dih(traj, indx)
# create a column name for the residue. e.g gly17_psi and store values in column
data_df['%s%s_%s'%(res,k,dih)] = data1
# repeat for reference, should remove out of the loop?
data_ref1 = cal_dih(ref_traj, indx)
data_df_ref['%s%s_%s'%(res,k,dih)] = data_ref1
# set to nan for the last residue if it is a psi dihedral
elif (dih == 'phi' and k == tot_res):
# res is the residue name and k is the residue number
residues, ss, _ = pt.dssp(ref_traj, ":%s" %(k))
res = residues[0][0:3]
data_df['%s%s_%s'%(res,k,dih)] = "Nan"
data_df_ref['%s%s_%s'%(res,k,dih)] = "Nan"
# dihedral is a chi dihedral
else:
# cheap method to get the residues, a hack but works
# res is the residue name and k is the residue number
residues, ss, _ = pt.dssp(ref_traj, ":%s" %(k))
res = residues[0][0:3]
# get the index for psi
indx = '%s' %(dihmask(int(k), dih, res))
# calculate the chi dihedral value for residue k in the trajectory and store in data1
data1 = cal_dih(traj, indx)
if ( (len(data1)) != 0): # if there is values in data1 then there is a chi
# create a column name for the residue. e.g pro79_chi1 and store values in column
data_df['%s%s_%s'%(res,k,dih)] = data1
# repeat for reference, should remove out of the loop?
data_ref1 = cal_dih(ref_traj, indx)
data_df_ref['%s%s_%s'%(res,k,dih)] = data_ref1
else: # enter Nan for the missing chi
data_df['%s%s_%s'%(res,k,dih)] = "Nan"
data_df_ref['%s%s_%s'%(res,k,dih)] = "Nan"
# create dih data frame by appending data_df with data_df_ref
dih_val = data_df.append(data_df_ref)
# name the index frame
dih_val.index.name = "Frame"
# rename the reference row ref
dih_val.rename(index={0:'ref'},inplace=True)
# write out dihedral value to 4 decimal place
os.system("touch %s_values.dat" %(dih))
dih_val.to_csv("%s_values.dat" %(dih), float_format='%.4f')
# now let us get the deviation
dev_df = pd.DataFrame(index = np.arange(1, traj_len+1, 1)) # deviation dataframe
# for values in given residue calculate the deviation to a ref value
for k in np.arange(r0, r1+1, 1):
diff=[]
residues, ss, _ = pt.dssp(ref_traj, ":%s" %(k))
res = residues[0][0:3]
val1 = data_df_ref['%s%s_%s'%(res,k,dih)].values # reference for dihedral of residue k
# for value at each frame, calculate the deviation to the reference value val1
for val in data_df['%s%s_%s'%(res,k,dih)].values:
if (val != 'Nan'): #check one, either reference or traj val should be always Nan
else: # sign is different ,to deal with values on other end of the -180 to 180 spectrum
if np.sign(val) == np.sign(val1): # so they have the same sign
diff1 = abs(val) - abs(val1)
diff1 = abs(diff1)
diff.append(diff1[0])
else: # sign is different ,to deal with values on other end of the -180 to 180 spectrum
# sum the distance from 180 for each number
diff1 = (180 - abs(val)) + (180 - abs(val1))
# sum the distance from 0 for each number
diff2 = (abs(val) - 0) + (abs(val1 - 0))
if diff1 <= diff2:
diff.append(diff1[0])
else:
diff.append(diff2[0])
else:
diff.append('nan')
dev_df['%s%s_%s'%(res,k,dih)] = diff
diff = 0.0 # zero diff since we are appending
def create_PDBdf(pdb_list, tot_res):
'''
pdb_list is a list of pdb e.g ['2qnp', '1hhp']. They must be stored in a directory name PDB
tot_res is the total number of residues
function return a dictionary name pdb. it can be called as pdb[pdbname]
e.g pdb['1hhp']
'''
# populate a dataframe for each PDB that contains all the dihedrals
pdb = {} #dictionary
#get each pdb
# strip the less using comma delimiter
list_pdb = np.genfromtxt(pdb_list, dtype=str, delimiter=',')
for i in list_pdb:
# convert to lower case
i = i.lower()
i = i.strip()
# get the PDB path
#print (i)
j = 'PDB/%s.pdb' %(i)
# define PDB as topology using pytraj
top = pt.load_topology(j)
traj = pt.iterload(j)
# set up a dataframe that have residue#, residue name, and the dihedrals up to chi5
data_df = pd.DataFrame(index = np.arange(1, tot_res, 1), columns=["res#", "resname", "phi", "psi", "chi1", "chi2", "chi3", "chi4", "chi5"])
# populate the column residue #
data_df['res#'] = np.arange(1, tot_res, 1)
# we will calculate the dihedrals for each residued and make a map, k is residue #
for k in np.arange(1, tot_res, 1):
# cheap method to get the residues, a hack but works
residues, ss, _ = pt.dssp(traj, ":%s" %(k))
res = residues[0][0:3]
#print (res)
# put the residue name in the dataframe
data_df.at[k, 'resname'] = '%s' %(res)
# do phi first
if k != 1 : #no phi for first residue
# get the index for phi
indx = '%s' %(dihmask(int(k), 'phi', res))
# calculate the dihedral value, and store in data
data = cal_dih(traj, indx)
data_df.at[k, 'phi'] = data[0]
#print (data)
# do psi after
if k != tot_res: # no psi for last residue
# get the index for pytraj dihedral function
indx = '%s' %(dihmask(int(k), 'psi', res))
# calculate the dihedral value
data = pt.dihedral(traj, indx)
data_df.at[k, 'psi'] = data[0]
#print (data)
# now do chi's
chis = ['chi1', 'chi2', 'chi3', 'chi4', 'chi5']
for chi in chis:
indx = '%s' %(dihmask(int(k), chi, res))
# calculate the dihedral value
if indx != 'empty':
data = pt.dihedral(traj, indx)
data_df.at[k, chi] = data[0]
#make a directory to store reference value
os.system("mkdir reference_values")
data_df.to_csv("./reference_values/%s_ref.dat" %(i), float_format='%.4f')
pdb[i] = data_df # set pdb dataframes
return pdb
def test_frame_indices(self):
from numpy.testing import assert_equal
s_0 = pt.dssp(self.traj)[1]
s_1 = pt.dssp(self.traj, frame_indices=[0, 2, 5])[1]
assert_equal(s_0[[0, 2, 5]], s_1)
import pytraj as pt
pdb = pt.load_pdb_rcsb("1l2y")
out = pt.dssp(pdb)
print(out)
def test_frame_indices(self):
from numpy.testing import assert_equal
s_0 = pt.dssp(self.traj)[1]
s_1 = pt.dssp(self.traj, frame_indices=[0, 2, 5])[1]
assert_equal(s_0[[0, 2, 5]], s_1)
