def add_dssp(pdb_file):
'This is a small function that makes a dictionary with secondary structure information'
from Bio.Seq import Seq
from Bio import SeqIO
from Bio.Alphabet import IUPAC
io = PDBIO()
import pandas as pd
from Bio.PDB.DSSP import DSSP
pdbl = PDBList()
parser = PDBParser()
ppb = PPBuilder()
'If the structure hasnt been downloaed then it will - else parse it'
structure = parser.get_structure(pdb_file,pdb_file)
model = structure[0]
chain = model['A']
'path to the dssp excecutable:'
dssp_exc = '/Users/thorn/dssp-2.2.1/mkdssp'
dssp = DSSP(model, pdb_file, dssp_exc)
sec_dict = {}
for i in range(len(dssp)):
a_key = dssp.keys()[i]
index = a_key[1][1]
sec_structure = dssp[a_key][1]
sec_dict[index] = sec_structure
return sec_dict
def get_surface_residues(filename, my_acc_array, my_threshold):
"""
Given a pdb file, finds the residues exposed to the solvent (not buried)
according to the ASA (accessible surface area) value given by DSSP module.
The user can select a threshold of ASA. Default is 0.2.
"""
p = PDBParser(PERMISSIVE=1)
s = p.get_structure("code.pdb", filename)
model = s[0]
d = DSSP(model, filename, dssp='mkdssp', acc_array=my_acc_array)
sys.stderr.write("\nHandled %i residues\n" % len(d))
residue_number = set()
for element in sorted(d):
if type(element[3]) is not str: #Sometimes the element[3] is NA
if element[3] >= my_threshold:
# foreach aa in the surface (according to threshold) store residue_number
try:
residue_number.add(
str(list(d.keys())[element[0] - 1][1][1]) +
list(d.keys())[element[0] - 1][0])
except IndexError:
sys.stderr.write("Element " + str(d.keys()[0]) +
" index out of range\n")
return residue_number
def calc_dssp(model, chain_sites: dict, pdb_name: str) -> None:
# DSSP
# ============ ===
# Tuple Index Value
# ============ ===
# 0 DSSP index
# 1 Amino acid
# 2 Secondary structure
# 3 Relative ASA
# 4 Phi
# 5 Psi
# 6 NH-->O_1_relidx
# 7 NH-->O_1_energy
# 8 O-->NH_1_relidx
# 9 O-->NH_1_energy
# 10 NH-->O_2_relidx
# 11 NH-->O_2_energy
# 12 O-->NH_2_relidx
# 13 O-->NH_2_energy
# ============ ===
try:
dssp = DSSP(model, pdb_name, dssp="mkdssp")
except:
dssp = {}
print("dssp failed!")
for residue in dssp.keys():
(
dssp_i,
aa,
sec_struct,
sasa_r,
phi,
psi,
nh_o1_relidx,
nh_o1_e,
o_nh1_relidx,
o_nh1_e,
nh_o2_relidx,
nh_o2_e,
o_nh2_relidx,
o_nh2_e,
) = dssp[residue]
chain_id, res_id = residue
_, resnumb, _ = res_id
if chain_id in chain_sites and resnumb in chain_sites[chain_id]:
resid = chain_sites[chain_id][resnumb][1]
new_res = chain_sites[chain_id][resnumb][2]
new_res.sec_struct = sec_struct
new_res.sasa_r = sasa_r
new_res.phi = phi
new_res.psi = psi
session.commit()
def _run(self, pdbcode, chains):
pdb_file = os.path.join(PDB_PATH, 'pdb' + pdbcode + '.ent')
p = PDBParser()
structure = p.get_structure(pdbcode, pdb_file)
model = structure[0]
dssp = DSSP(model, pdb_file, dssp=dssp_route)
valid_keys = [key for key in dssp.keys() if key[0] in chains]
chain = dssp.keys()[0][0]
return [dssp[key] for key in valid_keys], chain
def calc_features(PATH, PDB_id, OUTPATH):
#Loading the files
parser = PDBParser(PERMISSIVE=1)
filename = os.path.join(PATH, PDB_id + ".pdb")
structure = parser.get_structure(PDB_id, filename)
model = structure[0]
#DSSP Analysis for SS, PHI, PSI
dssp = DSSP(model, filename)
#NACCESS Analysis for SASA
rsa, asa = run_naccess(model, filename)
rsa = process_rsa_data(rsa)
#Feature mapping to each atomic coordinate
dssp_present, dssp_not_present = 0, 0
feature = dict() #The feature dictionary
for model in structure:
for chain in model:
for residue in chain:
for atom in residue:
print(atom.get_full_id())
ID = (atom.get_full_id()[2], atom.get_full_id()[3])
if (ID in list(dssp.keys())):
if (rsa[ID]["all_atoms_abs"] > Threshold):
rsa_label = 1
else:
rsa_label = 0
feat = (SS_Labels[dssp[ID][2]], dssp[ID][4] / 360,
dssp[ID][5] / 360, rsa_label)
feature[tuple(atom.get_coord())] = feat
print(ID, atom.get_coord(), feat)
dssp_present += 1
else:
print("==> ID not present : ", atom.get_full_id())
dssp_not_present += 1
#Printing the Stats
print(
"==> STATS : PDBID : %s , DSSP PRESENT : %s , DSSP NOT PRESENT : %s" %
(PDB_id, dssp_present, dssp_not_present))
#Saving the feature to each PDB file
with open(os.path.join(OUTPATH, PDB_id + ".dat"), "wb+") as f:
pickle.dump(feature, f)
print("==> Dump completed")
def getSecondaryStructure(pdbID):
pdb.fetchInPDBFormat(pdbID)
p = PDBParser()
structure = p.get_structure(pdbID, pdb.getTemporaryPDBPath())
model = structure[0]
dssp = DSSP(model, pdb.getTemporaryPDBPath())
seq = ""
secStructure = ""
for key in dssp.keys():
seq = seq + dssp[key][1]
secStructure = secStructure + dssp[key][2]
# print("Secuencia: ", seq + " de longitud " + str(len(seq)))
# print("Estructura secundaria: ", secStructure + " de longitud " + str(len(secStructure)))
return secStructure
def accessible_surface_area(PDB_file):
# Calcul de la surface accessible au solvant pour chaque residus
ASA_dict = {}
parser = PDBParser()
structure_id = PDB_file.split(".")[0]
structure = parser.get_structure(structure_id, PDB_file)
model = structure[0]
dssp = DSSP(model, PDB_file, dssp='mkdssp')
id_CA = 0
for CA in list(dssp.keys()):
if dssp[CA][1] != 'X':
ASA_dict[id_CA] = dssp[CA][3]
id_CA += 1
return ASA_dict
def calculate_asa(model, filename, AROM_LIST, chain_list):
"""Returns a list of surface exposed residues as determined by relative solvent accessibility.
Only standard protein residues are currently supported. Non-protein and user specified custom residues cannot be
classified as surface exposed using this criteria.
Parameters
---------
model: :class:`Bio.PDB.Model.Model`
Model which contains chains and residues of protein strucutre
filename: str
Name of pdb file to be analyzed
AROM_LIST : list of str
List containing which standard residues are included in analysis
chain_list: list of str
Chains are included in analysis
Notes
-----
The relative accessible surface area (RSA) of each residue is calculated using the Bio.PDB.DSSP module. A residue
with an RSA value of 0.05 or higher is classified as surface exposed.
References
---------
Tien, M. Z.; Meyer, A. G.; Sydykova, D. K.; Spielman, S. J.; Wilke, C. O. PLoS ONE 2013, 8 (11).
Reference for relative solvent accessibility cutoff of 0.05, and for MaxASA values
"""
cutoff = .05
surface_exposed_res = []
letter_list = []
for res_name in AROM_LIST:
if res_name_to_char.get(res_name):
letter_list.append(res_name_to_char.get(res_name))
try:
dssp = DSSP(model, filename, acc_array="Wilke")
keys = list(dssp.keys())
for key in keys:
if key[0] in chain_list and dssp[key][3] >= cutoff and dssp[key][
1] in letter_list:
goal_str = dssp[key][1] + \
str(key[1][1]) + "(" + str(key[0]) + ")"
surface_exposed_res.append(goal_str)
except Exception as e:
warnings.warn(
"Unable to calculate solvent accessibility. Check that DSSP is installed.",
RuntimeWarning,
stacklevel=2)
return surface_exposed_res
def get_SS(structure, path_structure):
"""
given a structure
return a list of 3 features coresspondig to the secondary structure
for each residue
Helix : (0,0,1)
Sheet : (0,1,0)
Coil : (1,0,0)
"""
list_SS = []
dssp = DSSP(structure[0], path_structure)
list_dssp = list(dssp)
list_dssp_features = []
for i in range(len(list_dssp)):
value_dssp = list(list_dssp[i])
if value_dssp[2] == "G" or\
value_dssp[2] == "H" or\
value_dssp[2] == "I":
list_dssp_features.append([1,0,0])
elif value_dssp[2] or value_dssp[2] == "B" or value_dssp[2] =="E":
list_dssp_features.append([0,1,0])
elif value_dssp[2] or value_dssp[2] == "T" or\
value_dssp[2] =="S" or value_dssp[2] =="-":
list_dssp_features.append([0,1,0])
return(list_dssp_features)
def getDSSP(pdb_file):
parser = PDBParser()
name = os.path.splitext(os.path.basename(pdb_file))[0]
structure = parser.get_structure(name, pdb_file)
model = structure[0]
dssp = DSSP(model, pdb_file, dssp='mkdssp')
return dssp
def make_a_dssp_model(model, pdb_filepath, dssp='dssp'): #ticked
"""
Parameters
----------
model : TYPE
The Structure class contains a collection of Model instances.
pdb_filepath : str
DESCRIPTION.
dssp : str, optional
DESCRIPTION. The default is 'dssp'.
Returns
-------
dssp_object : Bio.PDB.DSSP.DSSP
parsed secondary structure and accessibility.
"""
#DSSP class, which maps Residue objects to their secondary structure
try:
dssp_object = DSSP(model, pdb_filepath, dssp)
except Exception as e:
if type(e) is not Exception:
raise
print('oops')
return dssp_object
def secondary_torsions_casp(domain): #, start, end):
"""Extract Secondary structure and torsion angles using the DSSP package"""
structure = PDBParser().get_structure('',
f'../../data/pdbfiles/{domain}.pdb')
try:
raw = DSSP(structure[0], f'../../data/pdbfiles/{domain}.pdb')
except:
print('PDBException. Nothing we can do')
return None, None
dssp = np.array(raw.property_list, dtype='O')
sequence = ''.join(dssp[:, 1])
sec_torsions = dssp[:, [2, 4, 5]]
# translating torsion angles to range (-180, 180)
for i in range(sec_torsions.shape[0]):
for j in range(1, 3):
if sec_torsions[i, j] > 180:
sec_torsions[i, j] = sec_torsions[i, j] - 360
elif sec_torsions[i, j] < -180:
sec_torsions[i, j] = 360 - sec_torsions[i, j]
return sec_torsions
def __init__(self, pdbfile, fastafile):
names = {'HIS':'H','ASP':'D','ARG':'R','PHE':'F','ALA':'A','CYS':'C','GLY':'G',\
'GLN':'Q','GLU':'E','LYS':'K','LEU':'L','MET':'M','ASN':'N','SER':'S',\
'TYR':'Y','THR':'T','ILE':'I','TRP':'W','PRO':'P','VAL':'V','SER':'S'}
# Load fasta residue sequence
f = open(fastafile)
ff = [line.rstrip("\n") for line in f]
f.close()
p_id = ff[0][1:]
self.seq = ff[1]
# Load pdb information
p = PDBParser(PERMISSIVE=1)
st = p.get_structure(p_id, pdbfile)
model = st[0]
tag = p_id[-1]
chain = model[tag]
residues = chain.get_residues()
self.residues = [res for res in residues]
## sequence info from pdb
self.pdbseq = "".join([names[res.get_resname()] for res in self.residues if \
names.has_key(res.get_resname())])
## 3-state sse info from pdb
dssp = DSSP(model, pdbfile)
to3_dict = {'-':'C', 'G':'H', 'H':'H', 'I':'H', 'E':'E', 'B':'E', 'T':'C', \
'S':'C', 'L':'C'}
keys = list(dssp.keys())
self.pdbss3seq = "".join([to3_dict[dssp[k][2]]for k in keys])
# Align the pdb sequence(always missing some residues) to fasta sequence
alignment = AlignNW(self.seq, self.pdbseq)
self.re_index = alignment['j']
self.re_index = [i-1 for i in alignment['j']] # minus 1 for indexing
# generate sequence alignment between pdb sequence and fasta sequence
self.alignment = "".join([self.pdbseq[i] if i > -1 else "-" for i in self.re_index])
self.alignment = "\n".join([self.seq, self.alignment])
# generate full lenght of 3-state SSE sequence according to re-index
self.ss3seq = "".join([self.pdbss3seq[i] if i > -1 else "C" for i in self.re_index])
# generate full lenght of distance matrix(distance=-1 when disappear in pdbseq)
self.dist_matrix = self.generate_dist_matrix()
#np.savetxt("test.txt", self.dist_matrix)
# generate full lenght of angle matrix(distance=None when disappear in pdbseq)
self.angle_matrix = self.generate_angle_matrix()
def getDSSP(
struct,
fname,
dsspPath=os.path.expanduser(
"~/Tesis/rriPredMethod/dependencies/bioinformaticTools/dssp/mkdssp")):
dssp = DSSP(struct[0], fname, dssp=dsspPath)
chains = struct[0].child_list
dsspDict = {
chain.get_id(): {symbol: []
for symbol in DSSP_SYMBOLS}
for chain in chains
}
for chainId, resId in dssp.keys():
secStruct = dssp[(chainId, resId)][2]
dsspDict[chainId][secStruct].append(resId)
return dsspDict
def _phi_psi_dic(model, pdb_filename):
dssp = DSSP(model, pdb_filename, dssp='mkdssp')
phi_psi_dic = {}
for x in dssp.property_dict:
phi_psi_dic[x[1][1]] = str(round(
dssp.property_dict[x][4], -1))[:-2] + "_" + str(
round(dssp.property_dict[x][5], -1))[:-2]
return phi_psi_dic
def __applyDssp(self):
import Bio.PDB as bio
print('PSU: applying dssp')
from Bio.PDB.DSSP import DSSP
p = bio.PDBParser()
pdbFile = self.pdbDataPath + 'pdb' + self.pdbCode + '.ent'
structure = p.get_structure(self.pdbCode, pdbFile)
model = structure[0]
dssp = DSSP(model, pdbFile)
for akey in list(dssp.keys()):
chain = akey[0]
res_no = akey[1][1]
row = dssp[akey]
ss = row[2]
for atom in self.atoms:
if atom.values['rid'] == res_no and atom.values[
'chain'] == chain:
atom.setDsspInfo(ss)
print('PSU: applied dssp successfully')
def get_dssp_amino_acid_sequence(pdb_path, dssp_path):
structure = Bio.PDB.PDBParser(QUIET=True).get_structure(
pdb_path[:-4], pdb_path)
model = structure[0]
dssp = DSSP.DSSP(model, dssp_path)
aa = ''
for key in list(dssp.keys()):
#chain_id = key[0]
#if chain_id == letters[chain_index]:#'A': # first chain
aa += dssp[key][1]
return aa
def get_dssp_torsion_angles(pdb_path, dssp_path):
structure = Bio.PDB.PDBParser(QUIET=True).get_structure(
pdb_path[:-4], pdb_path)
model = structure[0]
dssp = DSSP.DSSP(model, dssp_path)
phi = ''
psi = ''
for key in list(dssp.keys()):
phi += dssp[key][4]
psi += dssp[key][5]
return phi, psi
def getDSSP(model,
dssp_map=None,
feature_name='secondary_structure',
formatstr="{}({})",
clean=True):
try:
from Bio.PDB.DSSP import DSSP
except ModuleNotFoundError:
raise ModuleNotFoundError(
"The module 'Bio.PDB.DSSP' is required for this functionality!")
if dssp_map is None:
# map eight ss types to three
dssp_map = {
"H": formatstr.format(feature_name, "H"),
"G": formatstr.format(feature_name, "H"),
"I": formatstr.format(feature_name, "H"),
"E": formatstr.format(feature_name, "S"),
"B": formatstr.format(feature_name, "L"),
"T": formatstr.format(feature_name, "L"),
"S": formatstr.format(feature_name, "L"),
"-": formatstr.format(feature_name, "L")
}
# Write a PDB file
pdb_file = model.save()
# run DSSP using the DSSP class from BioPython
dssp = DSSP(model, pdb_file)
# store secondary structure in each atom property dict
keys = list(sorted(set(dssp_map.values())))
for chain in model:
cid = chain.get_id()
for residue in chain:
rid = residue.get_id()
dkey = (cid, rid)
if dkey in dssp:
ss = dssp_map[dssp[dkey][2]]
else:
ss = dssp_map['-']
for atom in residue:
atom.xtra[keys[0]] = 0.0
atom.xtra[keys[1]] = 0.0
atom.xtra[keys[2]] = 0.0
atom.xtra[ss] = 1.0
if clean:
os.remove(pdb_file)
return keys
def load_data(self):
"""
Load all the pdb files in the folder. Create BIO.PDB.DSSP.DSSP object to stock each data's file
"""
stock = []
pdb = self.m_folder + self.m_id + ".pdb"
p = PDBParser()
structure = p.get_structure("prot", pdb)
model = structure[0]
dssp = DSSP(model, pdb)
stock.append(dssp)
return stock
def get_dssp_secondary_structure(pdb_path, dssp_path):
structure = Bio.PDB.PDBParser(QUIET=True).get_structure(
pdb_path[:-4], pdb_path)
model = structure[0]
dssp = DSSP.DSSP(model, dssp_path)
q8 = ''
for key in list(dssp.keys()):
#chain_id = key[0]
#if chain_id == letters[chain_index]:#'A': # first chain
#aa += dssp[key][1]
q8 += dssp[key][2]
return q8
def get_dssp_absolute_surface_area(pdb_path, dssp_path):
structure = Bio.PDB.PDBParser(QUIET=True).get_structure(
pdb_path[:-4], pdb_path)
model = structure[0]
dssp = DSSP.DSSP(model, dssp_path)
asa = ''
#letters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' # chain_id is in letters
for key in list(dssp.keys()):
#chain_id = key[0]
#if chain_id == letters[chain_index]:#'A': # first chain
asa += dssp[key][3]
return asa
def get_mali_structure_stats(root):
""" Reads in the manual alignments and obtains stats.
Parameters
----------
root : path
Path to root directory
Returns
-------
pd.DataFrame
alpha residues
beta residues
"""
from Bio.PDB import PDBParser
from Bio.PDB.DSSP import DSSP
res = []
tool = 'manual'
for path, directories, files in os.walk(root):
for f in files:
if '.pdb' in f and tool in f:
fname = os.path.join(path, f)
parser = PDBParser()
# ids = os.path.basename(fname).split('_')
structs = parser.get_structure('', fname)
dssp1 = DSSP(structs[0], fname, dssp='mkdssp')
classes1 = list(map(lambda x: x[2], dssp1))
len1 = len(classes1)
classes1 = pd.Series(Counter(classes1))
classes1.index = list(map(lambda x: 'x' + x, classes1.index))
pdb_name = os.path.basename(f).split('.')[0]
# stats = pd.concat((classes1, classes2))
stats = classes1
stats['pdb'] = pdb_name
stats['path'] = fname
stats['xlen'] = len1
# dssp2 = DSSP(structs[1], fname, dssp='mkdssp')
# classes2 = list(map(lambda x: x[2], dssp2))
# len2 = len(classes2)
# classes2 = pd.Series(Counter(classes2))
# classes2.index = list(map(lambda x: 'y' + x, classes2.index))
# stats['ylen'] = len2
res.append(stats)
res = pd.DataFrame(res)
return res
def Make_dssp():
ref = {
'A': 'ALA',
'R': 'ARG',
'N': 'ASN',
'D': 'ASP',
'B': 'ASX',
'C': 'CYS',
'E': 'GLU',
'Q': 'GLN',
'Z': 'GLX',
'G': 'GLY',
'H': 'HIS',
'I': 'ILE',
'L': 'LEU',
'K': 'LYS',
'M': 'MET',
'F': 'PHE',
'P': 'PRO',
'S': 'SER',
'T': 'THR',
'W': 'TRP',
'Y': 'TYR',
'V': 'VAL',
'X': '---'
}
dssp_dict = {}
dssp_dict['X'] = np.NaN
p = PDBParser()
structure = p.get_structure('model', './current_pdb.txt')
mod = structure[0]
dssp = DSSP(mod, './current_pdb.txt')
for i in range(len(dssp)):
a_key = list(dssp.keys())[i]
dssp_dict[str(a_key[0]) + str(a_key[1][1]) +
str(ref[dssp[a_key][1]])] = dssp[a_key][2]
return (dssp_dict)
def get_dssp_amino_acid_sequences(pdb_path, dssp_path):
structure = Bio.PDB.PDBParser(QUIET=True).get_structure(
pdb_path[:-4], pdb_path)
model = structure[0]
dssp = DSSP.DSSP(model, dssp_path)
aa = ''
aas = []
prev_chain_id = ''
for key in list(dssp.keys()):
chain_id = key[0]
if chain_id != prev_chain_id:
prev_chain_id = chain_id
if aa != '':
aas.append(aa)
aa = ''
#if chain_id == letters[chain_index]:#'A': # first chain
aa += dssp[key][1]
return aas
def recup_acc_solvant(fichier_pdb):
"""
entrée le fichier pdb
sortie une liste de l'accesibilité relative (entre 0 et 1) des résidues
utilise Biopython et le programme dssp
"""
acces_solvant = []
p = PDBParser()
id_structure = fichier_pdb.split(".")[0]
structure = p.get_structure(id_structure, fichier_pdb)
model = structure[0]
dssp = DSSP(model, fichier_pdb, dssp = 'mkdssp')
for i in range(len(list(dssp))):
acces_solvant.append(round((list(dssp)[i][3]),3)) #ajout de chaque élément arrondi au millième
return acces_solvant
def SecStr(pdb_id, chain_id, start, stop):
#Change pdb_id to lower cases - as in local pdb db.
pdb_id = pdb_id.lower()
#Read pdb structure if it exists.
p = PDBParser()
try:
structure = p.get_structure(pdb_id, f'/home/m.pak/pdb/pdb{pdb_id}.pdb')
except FileNotFoundError:
print(f'File not found, proceed... {pdb_id}')
return None, None, None
model = structure[0]
#Run DSSP.
try:
dssp = DSSP(model, f'/home/m.pak/pdb/pdb{pdb_id}.pdb')
except:
print(f'DSSP unable to process the structure {pdb_id}, proceed...')
return None, None, None
#Keep annotation of secondaty structure elements, Phi and Psi angles for defined region of structure.
sec_str = ''
phi_lst = []
psi_lst = []
#INCLUDES STOP!!!!
for num in range(start, stop+1):
try:
res_key = (chain_id, (' ', num, ' ')) #Can not deal with hetero-flag and insertion code
res = dssp[res_key]
except:
print(f'{res_key} not found in {pdb_id}, proceed...')
continue
sec_str += res[2]
phi_lst.append(res[4])
psi_lst.append(res[5])
return sec_str, phi_lst, psi_lst
def secondary_torsions(domain): #, start, end):
"""Extract Secondary structure and torsion angles using the DSSP package"""
domain_id = domain[:4]
chain_id = domain[4]
structure = PDBParser().get_structure(
'', f'../../data/pdbfiles/{domain_id}.pdb')
try:
raw = DSSP(structure[0], f'../../data/pdbfiles/{domain_id}.pdb')
except:
print('PDBException. Nothing we can do')
return None, None
dssp = np.array(raw.property_list, dtype='O')
# extract chain
#keys = np.array([i[0] for i in raw.keys()])
#positions = np.array([int(i[1][1]) for i in raw.keys()])
#positions = positions[keys == chain_id]
#dssp = dssp[keys == chain_id]
sequence = ''.join(dssp[:, 1])
sec_torsions = dssp[:, [2, 4, 5]]
# translating torsion angles to range (-180, 180)
for i in range(sec_torsions.shape[0]):
for j in range(1, 3):
if sec_torsions[i, j] > 180:
sec_torsions[i, j] = sec_torsions[i, j] - 360
elif sec_torsions[i, j] < -180:
sec_torsions[i, j] = 360 - sec_torsions[i, j]
#try:
# dssp_start, dssp_end = np.where(positions == start)[0][0], np.where(positions == end)[0][0]
#except IndexError:
# print(domain, 'positions not found')
# return None, None
return sec_torsions, sequence #sec_torsions[dssp_start:(dssp_end + 1)], sequence[dssp_start:(dssp_end + 1)]
def retrieve_secondary_struc(chain, input_path):
# Uses biopython's built-in DSSP Function to retrieve the secondary structure
# note! STRIDE and DSSP agree in 95,4% of the cases. DSSP tends to assign shorter secondary structures
# Must have DSSP installed --> see tutorial for instructions
# conda install -c salilab dssp
# https://en.wikipedia.org/wiki/STRIDE
# H - Alpha-Helix
# B - Isolated Beta-Bridge
# E - Strand
# G - 3-10 Helix
# I - Pi helix
# T - Turn
# S - Bend
model = chain.get_parent()
res_list = Selection.unfold_entities(chain, "R")
chain_len = len(res_list)
dssp = DSSP(model, input_path)
dssplist = list(dssp)[:chain_len]
seq = [row[1] for row in dssplist]
seq = ''.join(seq)
struc = [row[2] for row in dssplist]
struc = ''.join(struc)
struc = struc.replace('-', ' ')
if len(struc) > len(seq):
struc = struc[0:len(seq)]
if len(seq) != len(res_list):
warnings.warn(
f'PDB file and Secondary structure map do not match!\n {chain.get_parent().get_parent().id} - PDB: {len(res_list)} Residues VS. SS: {len(seq)} Residues. '
)
return seq, struc
def getDSSP(model,
PDBFileName,
dssp_map=None,
feature_name='secondary_structure',
formatstr="{}({})"):
if (dssp_map is None):
# map eight ss types to three
dssp_map = {
"H": formatstr.format(feature_name, "H"),
"G": formatstr.format(feature_name, "H"),
"I": formatstr.format(feature_name, "H"),
"E": formatstr.format(feature_name, "S"),
"B": formatstr.format(feature_name, "L"),
"T": formatstr.format(feature_name, "L"),
"S": formatstr.format(feature_name, "L"),
"-": formatstr.format(feature_name, "L")
}
# run DSSP using the DSSP class from BioPython
dssp = DSSP(model, PDBFileName)
# store secondary structure in each atom property dict
for chain in model:
cid = chain.get_id()
for residue in chain:
rid = residue.get_id()
dkey = (cid, rid)
if (dkey in dssp):
ss = dssp_map[dssp[dkey][2]]
else:
ss = dssp_map['-']
for atom in residue:
atom.xtra[ss] = 1.0
return list(set(dssp_map.values()))
ftp.retrbinary("RETR %s" % filename, callback=fp.write)
print("processing: %s" % filename)
p = PDBParser()
with gzip.open(filename, 'rt') as f:
structure = p.get_structure("", f)
pdb_id = structure.header["idcode"]
assert pdb_id, "no PDB ID for %s" % filename
model = structure[0]
try:
dssp = DSSP(model, filename, dssp="/Users/luis/dssp-2.3.0/mkdssp")
except Exception as e:
print(e)
print()
os.remove(filename)
continue
valid_aa = [
'A', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'K', 'L', 'M', 'N', 'P',
'Q', 'R', 'S', 'T', 'V', 'W', 'Y'
]
current_chain = ""
chain = ""
phi_psis = []
dihedrals = []
chains = []
