def get_saxs_factors(pdb):
factors = []
pdb = rmsdlib.read_pdb(pdb)
for a in pdb.atoms():
atomtype = int(a.line[57:59])
w = 0.0
if a.name in skipped: continue
e = a.name[0]
if e == "H": continue
f = saxs_factors[e]
factors.append(f)
return np.array(factors,dtype="float32")
def get_coordinates(pdb, ignore_weightless):
pdb = rmsdlib.read_pdb(pdb)
if ignore_weightless:
coors = []
for a in pdb.atoms():
if a.name in skipped: continue
e = a.name[0]
if e == "H": continue
coors.append((a.x, a.y, a.z))
else:
coors = pdb.coordinates()
return np.array(coors, dtype="float32")
def get_dummy_factors(pdb):
factors = []
pdb = rmsdlib.read_pdb(pdb)
for a in pdb.atoms():
atomtype = int(a.line[57:59])
w = 0.0
if a.name in skipped: continue
e = a.name[0]
if e == "H": continue
hnum = get_hydrogens(a.resname, a.name)
f = saxs_factors[e][1] + hnum * saxs_factors["H"][
1] #dummy solvent form factors
factors.append(f)
return np.array(factors, dtype="float32")
def get_weights(pdb, ignore_weightless):
weights = []
pdb = rmsdlib.read_pdb(pdb)
for a in pdb.atoms():
atomtype = int(a.line[57:59])
w = 0.0
if atomtype in (32,99) and ignore_weightless:
continue
elif atomtype < 32:
w = aweights[atomtype]
elif atomtype in (32,99):
pass
else:
if a.name in skipped: continue
e = a.name[0]
if e == "H": continue
w = element_masses[e]
weights.append(w)
return np.array(weights,dtype="float32")
import argparse
a = argparse.ArgumentParser(prog="rmsd-matrix-pdb.py")
a.add_argument("pdblist")
a.add_argument("--np", type=int)
a.add_argument("--allatoms", action="store_true")
a.add_argument("--ca", action="store_true")
args = a.parse_args()
#structures
pdbfiles = [
l.strip().strip("\n") for l in open(args.pdblist)
if len(l.strip().strip("\n"))
]
#read atoms
pdbs = [rmsdlib.read_pdb(f) for f in pdbfiles]
for pdb in pdbs:
rmsdlib.check_pdbs((pdb, ), pdbs)
coors = np.array([list(pdb.coordinates()) for pdb in pdbs])
#select backbone
if args.allatoms:
assert not args.ca
else:
if args.ca:
atomnames = ("CA", )
else:
atomnames = ("CA", "C", "O", "N")
amask = np.array([a.name in atomnames for a in pdbs[0].atoms()])
coors = coors[:, amask]
raise ImportError("Parsing CNS .tbl files requires the Lark (lark-parser) library")
from lark import Lark
import os
currdir = os.path.dirname(__file__)
parser = Lark(
open(os.path.abspath(currdir) + "/tbl_grammar.ebnf").read(),
parser="earley", start="assign_statements"
)
from rmsdlib import read_pdb
pdbs = []
for p in args.pdbs:
pdbs.append(read_pdb(p))
resmaps = []
for m in args.mappings:
resmap = {}
for l in open(m):
ll = l.split()
if not len(ll) == 2: continue
resmap[ll[0]] = int(ll[1])
resmaps.append(resmap)
import numpy
pdblen = [len(list(p.atoms())) for p in pdbs]
pdbcumlen = [0]
for plen in pdblen:
if arg.startswith("--"): raise Exception("Unknown option '%s'" % arg)
if len(sys.argv) < 4 or len(sys.argv) % 2:
raise Exception(
"Please supply an even number of PDB files (unbound, bound)")
unboundfiles = []
boundfiles = []
for n in range(2, len(sys.argv), 2):
unboundfiles.append(sys.argv[n])
boundfiles.append(sys.argv[n + 1])
if len(boundfiles) == 1 and opt_allresidues == False:
raise Exception("Cannot determine the interface for a single PDB")
bounds = [rmsdlib.read_pdb(f) for f in boundfiles]
unbounds = [rmsdlib.read_pdb(f) for f in unboundfiles]
struc_header, structures = read_struc(sys.argv[1])
pivots = []
for hnr, h in enumerate(struc_header):
if not h.startswith("#pivot"): continue
hh = h.split()
assert len(hh) == 5 and hh[1] == str(hnr + 1), h
pivot = numpy.array([float(v) for v in hh[2:5]])
pivots.append(pivot)
initargs = [sys.argv[1]] + unboundfiles
if modefile: initargs += ["--modes", modefile]
if imodefile: initargs += ["--imodes", imodefile]
for nr, ensfile in ensfiles:
cf = open(clustfile, "w")
for cnr, c in enumerate(clust):
print >> cf, "Cluster %d ->" % (cnr + 1),
for cc in c:
print >> cf, cc,
print >> cf, ""
rootclusters = read_clustfile(clustfile)
superclust = []
subclust = []
maxstruc = 30000
coor = rmsdlib.read_pdb(pdbfiles[0]).coordinates()
coor = numpy.array(coor)
natom = len(coor)
lim = cutoff * cutoff * natom
coor = coor.flatten()
coorsize = len(coor)
clust_struc = numpy.zeros(dtype=float, shape=(maxstruc, coorsize))
for rootclustnr, rootclust in enumerate(rootclusters):
print >> sys.stderr, rootclustnr + 1
if len(rootclust) == 1:
subclust.append(rootclust)
superclust.append([len(subclust)])
continue
leafclust = []
for cnr, c in enumerate(rootclust):
#fills a PDB B-factor (temperature) column from per-residue numerical data
#input: PDB file, numerical data
#column 1-6 of the numerical data is the full resid (chainID + 5 resid columns; chainID _ is interpreted as " ")
#the rest of the line is interpreted as a numerical value (float), V
# B-factor = V / max(V) * 100
# i.e. the largest numerical value (max(V)) gets a B-factor of 100
# missing numerical data is interpreted as zero
import sys, rmsdlib
pdb = rmsdlib.read_pdb(sys.argv[1])
resids = [res[0].resid for res in pdb.residues()]
values = {}
for l in open(sys.argv[2]):
resid = l[:6].replace("_", " ")
assert resid in resids, resid
v = float(l[6:])
values[resid] = v
maxv = max(values.values())
for res in pdb.residues():
resid = res[0].resid
v = values.get(resid, 0.0)
b = v / maxv * 100.0
for atom in res:
l = atom.line
ll = l[:60] + "%6.2f" % b + l[66:]
sys.stdout.write(ll)
