def exercise_2 () :
hkl_file = libtbx.env.find_in_repositories(
relative_path="phenix_regression/wizards/p9_se_w2.sca",
test=os.path.isfile)
if (hkl_file is None) :
warnings.warn("phenix_regression not available, skipping test")
return
hkl_in = file_reader.any_file(hkl_file).assert_file_type("hkl")
i_obs_raw = hkl_in.file_object.as_miller_arrays(
merge_equivalents=False,
crystal_symmetry=crystal.symmetry(
space_group_symbol="I4",
unit_cell=(113.949,113.949,32.474,90,90,90)))[0]
i_obs = i_obs_raw.merge_equivalents().array()
# completeness and data strength
cstats = ds.i_sigi_completeness_stats(i_obs)
d_min_cut = cstats.resolution_cut
assert approx_equal(d_min_cut, 2.150815)
ws = ds.wilson_scaling(
miller_array=i_obs,
n_residues=120)
# outliers - this shouldn't actually work, since it requires additional
# processing steps on the input data
try :
outliers = ds.possible_outliers(i_obs)
except AssertionError :
pass
else :
raise Exception_expected
######################################################################
# OVERALL ANALYSIS
pdb_file = libtbx.env.find_in_repositories(
relative_path="phenix_examples/p9-build/p9.pdb",
test=os.path.isfile)
f_calc = None
if (pdb_file is not None) :
pdb_in = file_reader.any_file(pdb_file).assert_file_type("pdb")
hierarchy = pdb_in.file_object.hierarchy
xrs = pdb_in.file_object.xray_structure_simple(
crystal_symmetry=i_obs)
f_calc = xrs.structure_factors(d_min=i_obs.d_min()).f_calc()
f_calc = abs(f_calc).generate_bijvoet_mates()
f_calc = f_calc.set_observation_type_xray_amplitude()
i_obs, f_calc = i_obs.common_sets(other=f_calc)
open("tmp_xtriage.pdb", "w").write(hierarchy.as_pdb_string(
crystal_symmetry=i_obs))
pdb_file = "tmp_xtriage.pdb"
params = xtriage.master_params.extract()
params.scaling.input.asu_contents.n_residues = 141
result = xtriage.xtriage_analyses(
miller_obs=i_obs,
miller_calc=f_calc,
params=params,
unmerged_obs=i_obs_raw,
text_out=open("logfile3.log", "w"))#sys.stdout)
# XXX there appears to be some system-dependence here, hence sloppy limits
assert (15.5 < result.aniso_b_min < 15.9)
assert (10 < result.aniso_range_of_b < 11)
# check relative Wilson
if (pdb_file is not None) :
assert (result.relative_wilson is not None)
# FIXME
#assert (result.relative_wilson.n_outliers() == 34)
#show_pickled_object_sizes(result)
test_pickle_consistency_and_size(result)
# XXX PDB validation server
assert approx_equal(result.iso_b_wilson, 18.33, eps=0.1)
assert approx_equal(result.aniso_b_ratio, 0.546, eps=0.1)
assert (result.number_of_wilson_outliers == 0)
assert approx_equal(result.l_test_mean_l, 0.493, eps=0.1)
assert approx_equal(result.l_test_mean_l_squared, 0.326, eps=0.1)
assert approx_equal(result.i_over_sigma_outer_shell, 3.25, eps=0.1)
assert ("No significant pseudotranslation is detected" in
result.patterson_verdict)
# test consistency of output after pickling and unpickling
try :
from phenix_dev.phenix_cloud import xtriage_json
except ImportError :
pass
else :
json_out = xtriage_json.json_output("p9.sca")
result.show(out=json_out)
open("xtriage.json", "w").write(json_out.export())
# unmerged data
assert result.merging_stats is not None
out = StringIO()
result.merging_stats.show(out=out)
assert ("R-merge: 0.073" in out.getvalue())
assert approx_equal(result.estimate_d_min(min_i_over_sigma=10), 1.9645,
eps=0.001)
# FIXME PDB doesn't actually have unit cell!
# test detection of symmetry in reference file
if (pdb_file is not None) :
args = [hkl_file, pdb_file]
result = xtriage.run(args=args, out=null_out())
def exercise_2():
hkl_file = libtbx.env.find_in_repositories(
relative_path="phenix_regression/wizards/data/p9_se_w2.sca",
test=os.path.isfile)
if (hkl_file is None):
warnings.warn("phenix_regression not available, skipping test")
return
hkl_in = file_reader.any_file(hkl_file).assert_file_type("hkl")
i_obs_raw = hkl_in.file_object.as_miller_arrays(
merge_equivalents=False,
crystal_symmetry=crystal.symmetry(space_group_symbol="I4",
unit_cell=(113.949, 113.949, 32.474,
90, 90, 90)))[0]
i_obs = i_obs_raw.merge_equivalents().array()
# completeness and data strength
cstats = ds.i_sigi_completeness_stats(i_obs)
d_min_cut = cstats.resolution_cut
assert approx_equal(d_min_cut, 2.150815)
ws = ds.wilson_scaling(miller_array=i_obs, n_residues=120)
# outliers - this shouldn't actually work, since it requires additional
# processing steps on the input data
try:
outliers = ds.possible_outliers(i_obs)
except AssertionError:
pass
else:
raise Exception_expected
######################################################################
# OVERALL ANALYSIS
pdb_file = libtbx.env.find_in_repositories(
relative_path="phenix_examples/p9-build/p9.pdb", test=os.path.isfile)
f_calc = None
if (pdb_file is not None):
pdb_in = file_reader.any_file(pdb_file).assert_file_type("pdb")
hierarchy = pdb_in.file_object.hierarchy
xrs = pdb_in.file_object.xray_structure_simple(crystal_symmetry=i_obs)
f_calc = xrs.structure_factors(d_min=i_obs.d_min()).f_calc()
f_calc = abs(f_calc).generate_bijvoet_mates()
f_calc = f_calc.set_observation_type_xray_amplitude()
i_obs, f_calc = i_obs.common_sets(other=f_calc)
open("tmp_xtriage.pdb",
"w").write(hierarchy.as_pdb_string(crystal_symmetry=i_obs))
pdb_file = "tmp_xtriage.pdb"
params = xtriage.master_params.extract()
params.scaling.input.asu_contents.n_residues = 141
result = xtriage.xtriage_analyses(miller_obs=i_obs,
miller_calc=f_calc,
params=params,
unmerged_obs=i_obs_raw,
text_out=open("logfile3.log",
"w")) #sys.stdout)
# XXX there appears to be some system-dependence here, hence sloppy limits
assert (15.5 < result.aniso_b_min < 15.9)
assert (10 < result.aniso_range_of_b < 11)
# check relative Wilson
if (pdb_file is not None):
assert (result.relative_wilson is not None)
# FIXME
#assert (result.relative_wilson.n_outliers() == 34)
#show_pickled_object_sizes(result)
test_pickle_consistency_and_size(result)
# XXX PDB validation server
assert approx_equal(result.iso_b_wilson, 18.33, eps=0.1)
assert approx_equal(result.aniso_b_ratio, 0.546, eps=0.1)
assert (result.number_of_wilson_outliers == 0)
assert approx_equal(result.l_test_mean_l, 0.493, eps=0.1)
assert approx_equal(result.l_test_mean_l_squared, 0.326, eps=0.1)
assert approx_equal(result.i_over_sigma_outer_shell, 3.25, eps=0.1)
assert approx_equal(result.overall_i_sig_i, 10.34, eps=0.1)
assert approx_equal(
result.anomalous_info.plan_sad_experiment_stats.get_overall(
item="i_over_sigma_dict"),
10.61,
eps=0.1)
assert approx_equal(
result.anomalous_info.plan_sad_experiment_stats.get_overall(
item="anom_signal_dict"),
15.35,
eps=0.1)
assert ("No significant pseudotranslation is detected"
in result.patterson_verdict)
# test consistency of output after pickling and unpickling
try:
from phenix_dev.phenix_cloud import xtriage_json
except ImportError:
pass
else:
json_out = xtriage_json.json_output("p9.sca")
result.show(out=json_out)
open("xtriage.json", "w").write(json_out.export())
# unmerged data
assert result.merging_stats is not None
out = StringIO()
result.merging_stats.show(out=out)
assert ("R-merge: 0.073" in out.getvalue())
assert approx_equal(result.estimate_d_min(min_i_over_sigma=10),
1.9645,
eps=0.001)
# FIXME PDB doesn't actually have unit cell!
# test detection of symmetry in reference file
if (pdb_file is not None):
args = [hkl_file, pdb_file]
result = xtriage.run(args=args, out=null_out())
def __init__(self,
miller_array,
pre_scaling_protocol,
basic_info,
out=None):
## Make deep copy of the miller array of interest
self.x1 = miller_array.deep_copy()
self.options=pre_scaling_protocol
self.basic_info= basic_info
## Determine unit_cell contents
print(file=out)
print("Matthews analyses", file=out)
print("-----------------", file=out)
print(file=out)
print("Inspired by: Kantardjieff and Rupp. Prot. Sci. 12(9): 1865-1871 (2003).", file=out)
matthews_analyses = matthews.matthews_rupp(
crystal_symmetry = self.x1,
n_residues = self.basic_info.n_residues,
n_bases = self.basic_info.n_bases,
out=out, verbose=1)
n_residues=matthews_analyses[0]
n_bases=matthews_analyses[1]
n_copies_solc=matthews_analyses[2]
if (self.basic_info.n_residues==None):
self.basic_info.n_residues = n_residues
if (self.basic_info.n_bases == None):
self.basic_info.n_bases = n_bases
## apply resolution cut
print(file=out)
print("Applying resolution cut", file=out)
print("-----------------------", file=out)
if self.options.low_resolution is None:
if self.options.high_resolution is None:
print("No resolution cut is made", file=out)
low_cut=float(1e6)
if self.options.low_resolution is not None:
low_cut = self.options.low_resolution
print("Specified low resolution limit: %3.2f"%(
float(self.options.low_resolution) ), file=out)
high_cut = 0
if self.options.high_resolution is not None:
high_cut = self.options.high_resolution
print("Specified high resolution limit: %3.2f"%(
float(self.options.high_resolution) ), file=out)
## perform outlier analyses
##
## Do a simple outlier analyses please
print(file=out)
print("Wilson statistics based outlier analyses", file=out)
print("----------------------------------------", file=out)
print(file=out)
native_outlier = data_statistics.possible_outliers(
miller_array = self.x1,
prob_cut_ex = self.options.outlier_level_extreme,
prob_cut_wil = self.options.outlier_level_wilson )
native_outlier.show(out=out)
self.x1 = native_outlier.remove_outliers(
self.x1 )
## apply anisotropic scaling (final B-value will be set to b_add)!
if self.options.aniso_correction:
b_final = self.options.b_add
if b_final is None:
b_final = 0.0
print(file=out)
print("Anisotropic absolute scaling of data", file=out)
print("--------------------------------------", file=out)
print(file=out)
aniso_correct = absolute_scaling.ml_aniso_absolute_scaling(
miller_array = self.x1,
n_residues = n_residues*\
self.x1.space_group().order_z()*n_copies_solc,
n_bases = n_bases*\
self.x1.space_group().order_z()*n_copies_solc)
aniso_correct.show(out=out,verbose=1)
print(file=out)
print(" removing anisotropy for native ", file=out)
print(file=out)
u_star_correct_nat = aniso_correct.u_star
self.x1 = absolute_scaling.anisotropic_correction(
self.x1,
aniso_correct.p_scale,
u_star_correct_nat )
def __init__(self,
miller_array,
pre_scaling_protocol,
basic_info,
out=None):
## Make deep copy of the miller array of interest
self.x1 = miller_array.deep_copy()
self.options=pre_scaling_protocol
self.basic_info= basic_info
## Determine unit_cell contents
print >> out
print >> out, "Matthews analyses"
print >> out, "-----------------"
print >> out
print >> out, "Inspired by: Kantardjieff and Rupp. Prot. Sci. 12(9): 1865-1871 (2003)."
matthews_analyses = matthews.matthews_rupp(
crystal_symmetry = self.x1,
n_residues = self.basic_info.n_residues,
n_bases = self.basic_info.n_bases,
out=out, verbose=1)
n_residues=matthews_analyses[0]
n_bases=matthews_analyses[1]
n_copies_solc=matthews_analyses[2]
if (self.basic_info.n_residues==None):
self.basic_info.n_residues = n_residues
if (self.basic_info.n_bases == None):
self.basic_info.n_bases = n_bases
## apply resolution cut
print >> out
print >> out, "Applying resolution cut"
print >> out, "-----------------------"
if self.options.low_resolution is None:
if self.options.high_resolution is None:
print >> out, "No resolution cut is made"
low_cut=float(1e6)
if self.options.low_resolution is not None:
low_cut = self.options.low_resolution
print >> out, "Specified low resolution limit: %3.2f"%(
float(self.options.low_resolution) )
high_cut = 0
if self.options.high_resolution is not None:
high_cut = self.options.high_resolution
print >> out, "Specified high resolution limit: %3.2f"%(
float(self.options.high_resolution) )
## perform outlier analyses
##
## Do a simple outlier analyses please
print >> out
print >> out, "Wilson statistics based outlier analyses"
print >> out, "----------------------------------------"
print >> out
native_outlier = data_statistics.possible_outliers(
miller_array = self.x1,
prob_cut_ex = self.options.outlier_level_extreme,
prob_cut_wil = self.options.outlier_level_wilson )
native_outlier.show(out=out)
self.x1 = native_outlier.remove_outliers(
self.x1 )
## apply anisotropic scaling (final B-value will be set to b_add)!
if self.options.aniso_correction:
b_final = self.options.b_add
if b_final is None:
b_final = 0.0
print >> out
print >> out, "Anisotropic absolute scaling of data"
print >> out, "--------------------------------------"
print >> out
aniso_correct = absolute_scaling.ml_aniso_absolute_scaling(
miller_array = self.x1,
n_residues = n_residues*\
self.x1.space_group().order_z()*n_copies_solc,
n_bases = n_bases*\
self.x1.space_group().order_z()*n_copies_solc)
aniso_correct.show(out=out,verbose=1)
print >> out
print >> out, " removing anisotropy for native "
print >> out
u_star_correct_nat = aniso_correct.u_star
self.x1 = absolute_scaling.anisotropic_correction(
self.x1,
aniso_correct.p_scale,
u_star_correct_nat )