def find_sec_str(self, pdb_hierarchy):
if self.params.secondary_structure.protein.search_method == "ksdssp":
pdb_str = pdb_hierarchy.as_pdb_string()
print >> self.log, " running ksdssp..."
(records, stderr) = run_ksdssp_direct(pdb_str)
return iotbx.pdb.secondary_structure.annotation.from_records(
records=records,
log=self.log)
elif self.params.secondary_structure.protein.search_method == "mmtbx_dssp":
from mmtbx.secondary_structure import dssp
print >> self.log, " running mmtbx.dssp..."
return dssp.dssp(
pdb_hierarchy=pdb_hierarchy,
pdb_atoms=self.pdb_atoms,
out=null_out()).get_annotation()
elif self.params.secondary_structure.protein.search_method == "from_ca":
from mmtbx.secondary_structure import find_ss_from_ca
print >> self.log, " running find_ss_from_ca..."
fss = find_ss_from_ca.find_secondary_structure(
hierarchy=pdb_hierarchy,
out=null_out())
return fss.get_annotation()
elif self.params.secondary_structure.protein.search_method == "cablam":
from mmtbx.validation import cablam
print >> self.log, " running cablam..."
cablam_results = cablam.cablamalyze(
pdb_hierarchy = pdb_hierarchy,
outliers_only=False,
out=null_out(),
quiet=False)
return cablam_results.as_secondary_structure()
else:
print >> self.log, " WARNING: Unknown search method for SS. No SS found."
return iotbx.pdb.secondary_structure.annotation.from_records()
def find_sec_str(self, pdb_hierarchy):
if (pdb_hierarchy.atoms_size() > 99999
and self.params.secondary_structure.protein.search_method
== "ksdssp"):
print("Warning!!! ksdssp method is not applicable for" + \
"structures with more than 99999 atoms!\nSwitching to from_ca.", file=self.log)
self.params.secondary_structure.protein.search_method = "from_ca"
if self.params.secondary_structure.protein.search_method == "ksdssp":
pdb_str = pdb_hierarchy.as_pdb_string()
print(" running ksdssp...", file=self.log)
(records, stderr) = run_ksdssp_direct(pdb_str)
return iotbx.pdb.secondary_structure.annotation.from_records(
records=records, log=self.log)
elif self.params.secondary_structure.protein.search_method == "mmtbx_dssp":
from mmtbx.secondary_structure import dssp
print(" running mmtbx.dssp...", file=self.log)
return dssp.dssp(pdb_hierarchy=pdb_hierarchy,
pdb_atoms=self.pdb_atoms,
out=null_out()).get_annotation()
elif self.params.secondary_structure.protein.search_method == "from_ca":
from mmtbx.secondary_structure import find_ss_from_ca
from_ca_args = []
ca_label = "liberal"
if self.params.secondary_structure.from_ca_conservative:
from_ca_args = [
"alpha.rise_tolerance=0.13",
"beta.rise_tolerance=0.3",
"alpha.dot_min=0.94",
]
ca_label = "conservative"
print(" running find_ss_from_ca %s..." % ca_label, file=self.log)
fss = find_ss_from_ca.find_secondary_structure(
hierarchy=pdb_hierarchy,
args = from_ca_args,
ss_by_chain=self.params.secondary_structure.ss_by_chain,
max_rmsd=self.params.secondary_structure.max_rmsd,
use_representative_chains=\
self.params.secondary_structure.use_representative_chains,
max_representative_chains=\
self.params.secondary_structure.max_representative_chains,
out=null_out())
return fss.get_annotation()
elif self.params.secondary_structure.protein.search_method == "cablam":
from mmtbx.validation import cablam
print(" running cablam...", file=self.log)
cablam_results = cablam.cablamalyze(pdb_hierarchy=pdb_hierarchy,
outliers_only=False,
out=null_out(),
quiet=False)
return cablam_results.as_secondary_structure()
else:
print(" WARNING: Unknown search method for SS. No SS found.",
file=self.log)
return iotbx.pdb.secondary_structure.annotation.from_records()
def find_sec_str(self, pdb_hierarchy):
if (pdb_hierarchy.atoms_size() > 99999
and self.params.secondary_structure.protein.search_method
== "ksdssp"):
print >> self.log, "Warning!!! ksdssp method is not applicable for" + \
"structures with more than 99999 atoms!\nSwitching to from_ca."
self.params.secondary_structure.protein.search_method = "from_ca"
if self.params.secondary_structure.protein.search_method == "ksdssp":
pdb_str = pdb_hierarchy.as_pdb_string()
print >> self.log, " running ksdssp..."
(records, stderr) = run_ksdssp_direct(pdb_str)
return iotbx.pdb.secondary_structure.annotation.from_records(
records=records, log=self.log)
elif self.params.secondary_structure.protein.search_method == "mmtbx_dssp":
from mmtbx.secondary_structure import dssp
print >> self.log, " running mmtbx.dssp..."
return dssp.dssp(pdb_hierarchy=pdb_hierarchy,
pdb_atoms=self.pdb_atoms,
out=null_out()).get_annotation()
elif self.params.secondary_structure.protein.search_method == "from_ca":
from mmtbx.secondary_structure import find_ss_from_ca
print >> self.log, " running find_ss_from_ca..."
fss = find_ss_from_ca.find_secondary_structure(
hierarchy=pdb_hierarchy,
ss_by_chain=self.params.secondary_structure.ss_by_chain,
max_rmsd=self.params.secondary_structure.max_rmsd,
use_representative_chains=\
self.params.secondary_structure.use_representative_chains,
max_representative_chains=\
self.params.secondary_structure.max_representative_chains,
out=null_out())
return fss.get_annotation()
elif self.params.secondary_structure.protein.search_method == "cablam":
from mmtbx.validation import cablam
print >> self.log, " running cablam..."
cablam_results = cablam.cablamalyze(pdb_hierarchy=pdb_hierarchy,
outliers_only=False,
out=null_out(),
quiet=False)
return cablam_results.as_secondary_structure()
else:
print >> self.log, " WARNING: Unknown search method for SS. No SS found."
return iotbx.pdb.secondary_structure.annotation.from_records()
def stats(model, prefix):
# Get rid of H, multi-model, no-protein and single-atom residue models
if (model.percent_of_single_atom_residues() > 20):
return None
sel = model.selection(string="protein")
if (sel.count(True) == 0):
return None
ssr = "protein and not (element H or element D or resname UNX or resname UNK or resname UNL)"
sel = model.selection(string=ssr)
model = model.select(sel)
if (len(model.get_hierarchy().models()) > 1):
return None
# Add H; this looses CRYST1 !
rr = run_reduce_with_timeout(
stdin_lines=model.get_hierarchy().as_pdb_string().splitlines(),
file_name=None,
parameters="-oh -his -flip -keep -allalt -pen9999 -",
override_auto_timeout_with=None)
# Create model; this is a single-model pure protein with new H added
pdb_inp = iotbx.pdb.input(source_info=None, lines=rr.stdout_lines)
model = mmtbx.model.manager(model_input=None,
build_grm=True,
pdb_hierarchy=pdb_inp.construct_hierarchy(),
process_input=True,
log=null_out())
box = uctbx.non_crystallographic_unit_cell_with_the_sites_in_its_center(
sites_cart=model.get_sites_cart(), buffer_layer=5)
model.set_sites_cart(box.sites_cart)
model._crystal_symmetry = box.crystal_symmetry()
# Get SS annotations
SS = find_ss_from_ca.find_secondary_structure(
hierarchy=model.get_hierarchy(),
#ss_by_chain = False, # enabling will make it slow.
out=null_out())
# Convert SS annotations into bool selections
alpha_sel = SS.annotation.overall_helix_selection().strip()
beta_sel = SS.annotation.overall_sheet_selection().strip()
if (len(alpha_sel) == 0 or alpha_sel == "()"): alpha_sel = None
if (len(beta_sel) == 0 or beta_sel == "()"): beta_sel = None
if ([alpha_sel, beta_sel].count(None) == 0):
alpha_sel = model.selection(string="%s" % alpha_sel)
beta_sel = model.selection(string="%s" % beta_sel)
loop_sel = ~(alpha_sel | beta_sel)
elif (alpha_sel is not None):
alpha_sel = model.selection(string="%s" % alpha_sel)
loop_sel = ~alpha_sel
elif (beta_sel is not None):
beta_sel = model.selection(string="%s" % beta_sel)
loop_sel = ~beta_sel
else:
loop_sel = model.selection(string="all")
# Get individual stats
def get_selected(sel):
result = None
if (type(sel) == str and sel == "all"):
return find(model=model, a_DHA_cutoff=90).get_params_as_arrays()
elif (sel is not None and sel.count(True) > 0):
result = find(model=model.select(sel),
a_DHA_cutoff=90).get_params_as_arrays()
return result
result_dict = {}
result_dict["all"] = get_selected(sel="all")
result_dict["alpha"] = get_selected(sel=alpha_sel)
result_dict["beta"] = get_selected(sel=beta_sel)
result_dict["loop"] = get_selected(sel=loop_sel)
# Load histograms for reference high-resolution d_HA and a_DHA
pkl_fn = libtbx.env.find_in_repositories(
relative_path="mmtbx") + "/nci/d_HA_and_a_DHA_high_res.pkl"
assert os.path.isfile(pkl_fn)
ref = easy_pickle.load(pkl_fn)
#
import matplotlib as mpl
mpl.use('Agg')
import matplotlib.pyplot as plt
fig = plt.figure(figsize=(10, 10))
kwargs = dict(histtype='bar', bins=20, range=[1.6, 3.0], alpha=.8)
for j, it in enumerate([["alpha", 1], ["beta", 3], ["loop", 5]]):
key, i = it
ax = plt.subplot(int("32%d" % i))
if (j in [0, 1]):
ax.tick_params(bottom=False)
ax.set_xticklabels([])
ax.tick_params(axis="x", labelsize=12)
ax.tick_params(axis="y", labelsize=12, left=False, pad=-2)
ax.text(0.98,
0.92,
key,
size=12,
horizontalalignment='right',
transform=ax.transAxes)
HB = result_dict[key]
if HB is None: continue
w1 = np.ones_like(HB.d_HA) / HB.d_HA.size()
ax.hist(HB.d_HA, color="orangered", weights=w1, rwidth=0.3, **kwargs)
#
start, end1, end2 = 0, max(ref.distances[key].vals), \
round(max(ref.distances[key].vals),2)
plt.yticks([0.01, end1], ["0", end2],
visible=True,
rotation="horizontal")
#
if (j == 0): ax.set_title("Distance", size=15)
bins = list(flex.double(ref.distances[key].bins))
ax.bar(bins, ref.distances[key].vals, alpha=.3, width=0.07)
#
kwargs = dict(histtype='bar', bins=20, range=[90, 180], alpha=.8)
for j, it in enumerate([["alpha", 2], ["beta", 4], ["loop", 6]]):
key, i = it
ax = plt.subplot(int("32%d" % i))
if (j in [0, 1]):
ax.tick_params(bottom=False)
ax.set_xticklabels([])
ax.tick_params(axis="x", labelsize=12)
ax.tick_params(axis="y", labelsize=12, left=False, pad=-2)
ax.text(0.98,
0.92,
key,
size=12,
horizontalalignment='right',
transform=ax.transAxes)
ax.text(0.98,
0.92,
key,
size=12,
horizontalalignment='right',
transform=ax.transAxes)
#if(j in [0,1]): ax.plot_params(bottom=False)
HB = result_dict[key]
if HB is None: continue
w1 = np.ones_like(HB.a_DHA) / HB.a_DHA.size()
ax.hist(HB.a_DHA, color="orangered", weights=w1, rwidth=0.3, **kwargs)
#
start, end1, end2 = 0, max(ref.angles[key].vals), \
round(max(ref.angles[key].vals),2)
plt.yticks([0.01, end1], ["0", end2],
visible=True,
rotation="horizontal")
#
if (j == 0): ax.set_title("Angle", size=15)
ax.bar(ref.angles[key].bins, ref.angles[key].vals, width=4.5, alpha=.3)
plt.subplots_adjust(wspace=0.12, hspace=0.025)
fig.savefig("%s.png" % prefix, dpi=1000)