def print_validation(log, results, debug, pdb_hierarchy_selected):
box_1 = results.box_1
box_2 = results.box_2
box_3 = results.box_3
sites_cart_box = box_1.xray_structure_box.sites_cart()
sel = maptbx.grid_indices_around_sites(
unit_cell=box_1.xray_structure_box.unit_cell(),
fft_n_real=box_1.map_box.focus(),
fft_m_real=box_1.map_box.all(),
sites_cart=sites_cart_box,
site_radii=flex.double(sites_cart_box.size(), 2.0))
b1 = box_1.map_box.select(sel).as_1d()
b2 = box_2.map_box.select(sel).as_1d()
b3 = box_3.map_box.select(sel).as_1d()
print >> log, "Map 1: calculated Fobs with ligand"
print >> log, "Map 2: calculated Fobs without ligand"
print >> log, "Map 3: real Fobs data"
cc12 = flex.linear_correlation(x=b1, y=b2).coefficient()
cc13 = flex.linear_correlation(x=b1, y=b3).coefficient()
cc23 = flex.linear_correlation(x=b2, y=b3).coefficient()
print >> log, "CC(1,2): %6.4f" % cc12
print >> log, "CC(1,3): %6.4f" % cc13
print >> log, "CC(2,3): %6.4f" % cc23
#### D-function
b1 = maptbx.volume_scale_1d(map=b1, n_bins=10000).map_data()
b2 = maptbx.volume_scale_1d(map=b2, n_bins=10000).map_data()
b3 = maptbx.volume_scale_1d(map=b3, n_bins=10000).map_data()
print >> log, "Peak CC:"
print >> log, "CC(1,2): %6.4f" % flex.linear_correlation(
x=b1, y=b2).coefficient()
print >> log, "CC(1,3): %6.4f" % flex.linear_correlation(
x=b1, y=b3).coefficient()
print >> log, "CC(2,3): %6.4f" % flex.linear_correlation(
x=b2, y=b3).coefficient()
cutoffs = flex.double([i / 10. for i in range(1, 10)] +
[i / 100 for i in range(91, 100)])
d12 = maptbx.discrepancy_function(map_1=b1, map_2=b2, cutoffs=cutoffs)
d13 = maptbx.discrepancy_function(map_1=b1, map_2=b3, cutoffs=cutoffs)
d23 = maptbx.discrepancy_function(map_1=b2, map_2=b3, cutoffs=cutoffs)
print >> log, "q D(1,2) D(1,3) D(2,3)"
for c, d12_, d13_, d23_ in zip(cutoffs, d12, d13, d23):
print >> log, "%4.2f %6.4f %6.4f %6.4f" % (c, d12_, d13_, d23_)
###
if (debug):
#box_1.write_ccp4_map(file_name="box_1_polder.ccp4")
#box_2.write_ccp4_map(file_name="box_2_polder.ccp4")
#box_3.write_ccp4_map(file_name="box_3_polder.ccp4")
write_map_box(box=box_1, filename="box_1_polder.ccp4")
write_map_box(box=box_2, filename="box_2_polder.ccp4")
write_map_box(box=box_3, filename="box_3_polder.ccp4")
pdb_hierarchy_selected.adopt_xray_structure(box_1.xray_structure_box)
pdb_hierarchy_selected.write_pdb_file(
file_name="box_polder.pdb",
crystal_symmetry=box_1.box_crystal_symmetry)
#
print >> log, '*' * 79
message = result_message(cc12=cc12, cc13=cc13, cc23=cc23)
print >> log, message
return message
def exercise_cc_peak():
def get_map():
av = [random.random() for i in xrange(10*20*30)]
m = flex.double(av)
m = m-flex.min(m)
m = m/flex.max(m)
m.resize(flex.grid((10,20,30)))
return m
m1 = get_map()
m2 = get_map()
for t in range(0,11):
t=t/10.
ccp=maptbx.cc_peak(map_1=m1, map_2=m2, cutoff=t)
#
sites_frac = flex.vec3_double([
(0.50,0.50,0.50)])
from cctbx import xray
xray_structure = xray.structure(
crystal_symmetry=crystal.symmetry(
unit_cell=(5,5,5,90,90,90),
space_group_symbol="P1"),
scatterers=flex.xray_scatterer([
xray.scatterer(label=str(i), scattering_type="C", site=site_frac)
for i,site_frac in enumerate(sites_frac)]))
fc1 = xray_structure.structure_factors(d_min=1.6).f_calc()
fc2 = xray_structure.structure_factors(d_min=1.7).f_calc()
for t in range(0,11):
t=t/10.
ccp=maptbx.cc_peak(map_coeffs_1=fc1, map_coeffs_2=fc2, cutoff=t)
#
m1_he = maptbx.volume_scale(map = m1, n_bins = 10000).map_data()
m2_he = maptbx.volume_scale(map = m2, n_bins = 10000).map_data()
cutoffs = flex.double([i/20. for i in range(1,20)])
df = maptbx.discrepancy_function(map_1=m1_he, map_2=m2_he, cutoffs=cutoffs)
#
fc1 = xray_structure.structure_factors(d_min=2.2).f_calc()
fc2 = xray_structure.structure_factors(d_min=2.2).f_calc()
for t in range(0,10):
t=t/10.
ccp=maptbx.cc_peak(map_coeffs_1=fc1, map_coeffs_2=fc2, cutoff=t)
assert approx_equal(ccp, 1)
# 1D case
m1_he_1d = maptbx.volume_scale_1d(map = m1.as_1d(), n_bins = 10000).map_data()
m2_he_1d = maptbx.volume_scale_1d(map = m2.as_1d(), n_bins = 10000).map_data()
df_1d = maptbx.discrepancy_function(
map_1=m1_he_1d, map_2=m2_he_1d, cutoffs=cutoffs)
assert approx_equal(df, df_1d)
def cmd_run(args, validated=False, out=sys.stdout):
if (len(args) == 0):
print >> out, "-" * 79
print >> out, " phenix.polder"
print >> out, "-" * 79
print >> out, legend
print >> out, "-" * 79
master_params.show(out=out)
return
log = multi_out()
log.register("stdout", out)
log_file_name = "polder.log"
logfile = open(log_file_name, "w")
log.register("logfile", logfile)
print >> log, "phenix.polder is running..."
print >> log, "input parameters:\n", args
parsed = master_params
inputs = mmtbx.utils.process_command_line_args(args=args,
master_params=parsed)
#inputs.params.show() #check
params = inputs.params.extract()
# check model file
if len(inputs.pdb_file_names) == 0:
if (params.model_file_name is None):
raise Sorry("No model file found.")
elif (len(inputs.pdb_file_names) == 1):
params.model_file_name = inputs.pdb_file_names[0]
else:
raise Sorry("Only one model file should be given")
# check reflection file
reflection_files = inputs.reflection_files
if (len(reflection_files) == 0):
if (params.reflection_file_name is None):
raise Sorry("No reflection file found.")
else:
hkl_in = file_reader.any_file(params.reflection_file_name,
force_type="hkl")
hkl_in.assert_file_type("hkl")
reflection_files = [hkl_in.file_object]
# crystal symmetry
crystal_symmetry = None
crystal_symmetry = inputs.crystal_symmetry
if (crystal_symmetry is None):
crystal_symmetries = []
for f in [
str(params.model_file_name),
str(params.reflection_file_name)
]:
cs = crystal_symmetry_from_any.extract_from(f)
if (cs is not None): crystal_symmetries.append(cs)
if (len(crystal_symmetries) == 1):
crystal_symmetry = crystal_symmetries[0]
elif (len(crystal_symmetries) == 0):
raise Sorry("No crystal symmetry found.")
else:
if (not crystal_symmetries[0].is_similar_symmetry(
crystal_symmetries[1])):
raise Sorry(
"Crystal symmetry mismatch between different files.")
crystal_symmetry = crystal_symmetries[0]
f_obs, r_free_flags = None, None
rfs = reflection_file_utils.reflection_file_server(
crystal_symmetry=crystal_symmetry,
force_symmetry=True,
reflection_files=reflection_files,
err=StringIO())
parameters = mmtbx.utils.data_and_flags_master_params().extract()
if (params.data_labels is not None):
parameters.labels = params.data_labels
if (params.r_free_flags_labels is not None):
parameters.r_free_flags.label = params.r_free_flags_labels
determined_data_and_flags = mmtbx.utils.determine_data_and_flags(
reflection_file_server=rfs,
parameters=parameters,
keep_going=True,
log=StringIO())
f_obs = determined_data_and_flags.f_obs
if (params.data_labels is None):
params.data_labels = f_obs.info().label_string()
if (params.reflection_file_name is None):
params.reflection_file_name = parameters.file_name
r_free_flags = determined_data_and_flags.r_free_flags
assert f_obs is not None
print >> log, "Input data:"
print >> log, " Iobs or Fobs:", f_obs.info().labels
if (r_free_flags is not None):
print >> log, " Free-R flags:", r_free_flags.info().labels
params.r_free_flags_labels = r_free_flags.info().label_string()
else:
print >> log, " Free-R flags: Not present"
model_basename = os.path.basename(params.model_file_name.split(".")[0])
if (len(model_basename) > 0 and params.output_file_name_prefix is None):
params.output_file_name_prefix = model_basename
print params.output_file_name_prefix
new_params = master_params.format(python_object=params)
new_params.show()
if (not validated):
validate_params(params)
pdb_input = iotbx.pdb.input(file_name=params.model_file_name)
pdb_hierarchy = pdb_input.construct_hierarchy()
xray_structure = pdb_hierarchy.extract_xray_structure(
crystal_symmetry=crystal_symmetry)
# DON'T USE:
# xray_structure = pdb_input.xray_structure_simple()
# atom order might be wrong
mmtbx.utils.setup_scattering_dictionaries(
scattering_table=params.scattering_table,
xray_structure=xray_structure,
d_min=f_obs.d_min())
#if f_obs is not None:
f_obs = f_obs.resolution_filter(d_min=params.high_resolution,
d_max=params.low_resolution)
if (r_free_flags is not None):
r_free_flags = r_free_flags.resolution_filter(
d_min=params.high_resolution, d_max=params.low_resolution)
# Grab case that data are anomalous
if (f_obs.anomalous_flag()):
f_obs, r_free_flags = prepare_f_obs_and_flags(
f_obs=f_obs, r_free_flags=r_free_flags)
cpm_obj = compute_polder_map(f_obs=f_obs,
r_free_flags=r_free_flags,
xray_structure=xray_structure,
pdb_hierarchy=pdb_hierarchy,
params=params,
log=log)
# Significance check
fmodel = mmtbx.f_model.manager(f_obs=f_obs,
r_free_flags=r_free_flags,
xray_structure=xray_structure)
fmodel.update_all_scales(remove_outliers=False, fast=True)
f_obs_1 = abs(fmodel.f_model())
fmodel.update_xray_structure(
xray_structure=cpm_obj.xray_structure_noligand,
update_f_calc=True,
update_f_mask=True,
force_update_f_mask=True)
# PVA: do we need it? fmodel.update_all_scales(remove_outliers=False)
f_obs_2 = abs(fmodel.f_model())
xrs_selected = cpm_obj.pdb_hierarchy_selected.extract_xray_structure(
crystal_symmetry=f_obs.crystal_symmetry())
f_calc = f_obs.structure_factors_from_scatterers(
xray_structure=cpm_obj.xray_structure_noligand).f_calc()
f_mask = f_obs.structure_factors_from_map(map=cpm_obj.mask_polder,
use_scale=True,
anomalous_flag=False,
use_sg=False)
def get_poler_diff_map(f_obs):
fmodel = mmtbx.f_model.manager(f_obs=f_obs,
r_free_flags=r_free_flags,
f_calc=f_calc,
f_mask=f_mask)
fmodel.update_all_scales(remove_outliers=False)
mc_diff = map_tools.electron_density_map(
fmodel=fmodel).map_coefficients(map_type="mFo-DFc",
isotropize=True,
fill_missing=False)
fft_map = miller.fft_map(crystal_gridding=cpm_obj.crystal_gridding,
fourier_coefficients=mc_diff)
fft_map.apply_sigma_scaling()
map_data = fft_map.real_map_unpadded()
return mmtbx.utils.extract_box_around_model_and_map(
xray_structure=xrs_selected, map_data=map_data, box_cushion=2.1)
box_1 = get_poler_diff_map(f_obs=f_obs_1)
box_2 = get_poler_diff_map(f_obs=f_obs_2)
box_3 = get_poler_diff_map(f_obs=f_obs)
sites_cart_box = box_1.xray_structure_box.sites_cart()
sel = maptbx.grid_indices_around_sites(
unit_cell=box_1.xray_structure_box.unit_cell(),
fft_n_real=box_1.map_box.focus(),
fft_m_real=box_1.map_box.all(),
sites_cart=sites_cart_box,
site_radii=flex.double(sites_cart_box.size(), 2.0))
b1 = box_1.map_box.select(sel).as_1d()
b2 = box_2.map_box.select(sel).as_1d()
b3 = box_3.map_box.select(sel).as_1d()
print >> log, "Map 1: calculated Fobs with ligand"
print >> log, "Map 2: calculated Fobs without ligand"
print >> log, "Map 3: real Fobs data"
print >> log, "CC(1,2): %6.4f" % flex.linear_correlation(
x=b1, y=b2).coefficient()
print >> log, "CC(1,3): %6.4f" % flex.linear_correlation(
x=b1, y=b3).coefficient()
print >> log, "CC(2,3): %6.4f" % flex.linear_correlation(
x=b2, y=b3).coefficient()
### D-function
b1 = maptbx.volume_scale_1d(map=b1, n_bins=10000).map_data()
b2 = maptbx.volume_scale_1d(map=b2, n_bins=10000).map_data()
b3 = maptbx.volume_scale_1d(map=b3, n_bins=10000).map_data()
print >> log, "Peak CC:"
print >> log, "CC(1,2): %6.4f" % flex.linear_correlation(
x=b1, y=b2).coefficient()
print >> log, "CC(1,3): %6.4f" % flex.linear_correlation(
x=b1, y=b3).coefficient()
print >> log, "CC(2,3): %6.4f" % flex.linear_correlation(
x=b2, y=b3).coefficient()
cutoffs = flex.double([i / 10. for i in range(1, 10)] +
[i / 100 for i in range(91, 100)])
d12 = maptbx.discrepancy_function(map_1=b1, map_2=b2, cutoffs=cutoffs)
d13 = maptbx.discrepancy_function(map_1=b1, map_2=b3, cutoffs=cutoffs)
d23 = maptbx.discrepancy_function(map_1=b2, map_2=b3, cutoffs=cutoffs)
print >> log, "q D(1,2) D(1,3) D(2,3)"
for c, d12_, d13_, d23_ in zip(cutoffs, d12, d13, d23):
print >> log, "%4.2f %6.4f %6.4f %6.4f" % (c, d12_, d13_, d23_)
###
if (params.debug):
box_1.write_ccp4_map(file_name="box_1_polder.ccp4")
box_2.write_ccp4_map(file_name="box_2_polder.ccp4")
box_3.write_ccp4_map(file_name="box_3_polder.ccp4")
cpm_obj.pdb_hierarchy_selected.adopt_xray_structure(
box_1.xray_structure_box)
cpm_obj.pdb_hierarchy_selected.write_pdb_file(
file_name="box_polder.pdb",
crystal_symmetry=box_1.box_crystal_symmetry)
#
polder_file_name = "polder_map_coeffs.mtz"
if (params.output_file_name_prefix is not None):
polder_file_name = params.output_file_name_prefix + "_" + polder_file_name
#
print >> log, '*' * 79
print >> log, 'File %s was written.' % polder_file_name
print >> log, "Finished."
return True
def run(args, out=sys.stdout, validated=False):
show_citation(out=out)
if (len(args) == 0):
master_phil.show(out=out)
print('\nUsage: phenix.map_comparison <CCP4> <CCP4>\n',\
' phenix.map_comparison <CCP4> <MTZ> mtz_label_1=<label>\n',\
' phenix.map_comparison <MTZ 1> mtz_label_1=<label 1> <MTZ 2> mtz_label_2=<label 2>\n', file=out)
sys.exit()
# process arguments
params = None
input_attributes = ['map_1', 'mtz_1', 'map_2', 'mtz_2']
try: # automatic parsing
params = phil.process_command_line_with_files(
args=args, master_phil=master_phil).work.extract()
except Exception: # map_file_def only handles one map phil
from libtbx.phil.command_line import argument_interpreter
arg_int = argument_interpreter(master_phil=master_phil)
command_line_args = list()
map_files = list()
for arg in args:
if (os.path.isfile(arg)):
map_files.append(arg)
else:
command_line_args.append(arg_int.process(arg))
params = master_phil.fetch(sources=command_line_args).extract()
# check if more files are necessary
n_defined = 0
for attribute in input_attributes:
if (getattr(params.input, attribute) is not None):
n_defined += 1
# matches files to phil scope, stops once there is sufficient data
for map_file in map_files:
if (n_defined < 2):
current_map = file_reader.any_file(map_file)
if (current_map.file_type == 'ccp4_map'):
n_defined += 1
if (params.input.map_1 is None):
params.input.map_1 = map_file
elif (params.input.map_2 is None):
params.input.map_2 = map_file
elif (current_map.file_type == 'hkl'):
n_defined += 1
if (params.input.mtz_1 is None):
params.input.mtz_1 = map_file
elif (params.input.mtz_2 is None):
params.input.mtz_2 = map_file
else:
print('WARNING: only the first two files are used', file=out)
break
# validate arguments (GUI sets validated to true, no need to run again)
assert (params is not None)
if (not validated):
validate_params(params)
# ---------------------------------------------------------------------------
# check if maps need to be generated from mtz
n_maps = 0
maps = list()
map_names = list()
for attribute in input_attributes:
filename = getattr(params.input, attribute)
if (filename is not None):
map_names.append(filename)
current_map = file_reader.any_file(filename)
maps.append(current_map)
if (current_map.file_type == 'ccp4_map'):
n_maps += 1
# construct maps, if necessary
crystal_gridding = None
m1 = None
m2 = None
# 1 map, 1 mtz file
if (n_maps == 1):
for current_map in maps:
if (current_map.file_type == 'ccp4_map'):
uc = current_map.file_object.unit_cell()
sg_info = space_group_info(
current_map.file_object.space_group_number)
n_real = current_map.file_object.unit_cell_grid
crystal_gridding = maptbx.crystal_gridding(
uc, space_group_info=sg_info, pre_determined_n_real=n_real)
m1 = current_map.file_object.map_data()
if (crystal_gridding is not None):
label = None
for attribute in [('mtz_1', 'mtz_label_1'),
('mtz_2', 'mtz_label_2')]:
filename = getattr(params.input, attribute[0])
label = getattr(params.input, attribute[1])
if ((filename is not None) and (label is not None)):
break
# labels will match currently open mtz file
for current_map in maps:
if (current_map.file_type == 'hkl'):
m2 = miller.fft_map(
crystal_gridding=crystal_gridding,
fourier_coefficients=current_map.file_server.
get_miller_array(
label)).apply_sigma_scaling().real_map_unpadded()
else:
raise Sorry('Gridding is not defined.')
# 2 mtz files
elif (n_maps == 0):
crystal_symmetry = get_crystal_symmetry(maps[0])
d_min = min(get_d_min(maps[0]), get_d_min(maps[1]))
crystal_gridding = maptbx.crystal_gridding(
crystal_symmetry.unit_cell(),
d_min=d_min,
resolution_factor=params.options.resolution_factor,
space_group_info=crystal_symmetry.space_group_info())
m1 = miller.fft_map(
crystal_gridding=crystal_gridding,
fourier_coefficients=maps[0].file_server.get_miller_array(
params.input.mtz_label_1)).apply_sigma_scaling(
).real_map_unpadded()
m2 = miller.fft_map(
crystal_gridding=crystal_gridding,
fourier_coefficients=maps[1].file_server.get_miller_array(
params.input.mtz_label_2)).apply_sigma_scaling(
).real_map_unpadded()
# 2 maps
else:
m1 = maps[0].file_object.map_data()
m2 = maps[1].file_object.map_data()
# ---------------------------------------------------------------------------
# analyze maps
assert ((m1 is not None) and (m2 is not None))
# show general statistics
s1 = maptbx.more_statistics(m1)
s2 = maptbx.more_statistics(m2)
show_overall_statistics(out=out, s=s1, header="Map 1 (%s):" % map_names[0])
show_overall_statistics(out=out, s=s2, header="Map 2 (%s):" % map_names[1])
cc_input_maps = flex.linear_correlation(x=m1.as_1d(),
y=m2.as_1d()).coefficient()
print("CC, input maps: %6.4f" % cc_input_maps, file=out)
# compute CCpeak
cc_peaks = list()
m1_he = maptbx.volume_scale(map=m1, n_bins=10000).map_data()
m2_he = maptbx.volume_scale(map=m2, n_bins=10000).map_data()
cc_quantile = flex.linear_correlation(x=m1_he.as_1d(),
y=m2_he.as_1d()).coefficient()
print("CC, quantile rank-scaled (histogram equalized) maps: %6.4f" % \
cc_quantile, file=out)
print("Peak correlation:", file=out)
print(" cutoff CCpeak", file=out)
cutoffs = [i / 100.
for i in range(1, 90)] + [i / 1000 for i in range(900, 1000)]
for cutoff in cutoffs:
cc_peak = maptbx.cc_peak(map_1=m1_he, map_2=m2_he, cutoff=cutoff)
print(" %3.2f %7.4f" % (cutoff, cc_peak), file=out)
cc_peaks.append((cutoff, cc_peak))
# compute discrepancy function (D-function)
discrepancies = list()
cutoffs = flex.double(cutoffs)
df = maptbx.discrepancy_function(map_1=m1_he, map_2=m2_he, cutoffs=cutoffs)
print("Discrepancy function:", file=out)
print(" cutoff D", file=out)
for c, d in zip(cutoffs, df):
print(" %3.2f %7.4f" % (c, d), file=out)
discrepancies.append((c, d))
# compute and output histograms
h1 = maptbx.histogram(map=m1, n_bins=10000)
h2 = maptbx.histogram(map=m2, n_bins=10000)
print("Map histograms:", file=out)
print("Map 1 (%s) Map 2 (%s)"%\
(params.input.map_1,params.input.map_2), file=out)
print("(map_value,cdf,frequency) <> (map_value,cdf,frequency)", file=out)
for a1, c1, v1, a2, c2, v2 in zip(h1.arguments(), h1.c_values(),
h1.values(), h2.arguments(),
h2.c_values(), h2.values()):
print("(%9.5f %9.5f %9.5f) <> (%9.5f %9.5f %9.5f)"%\
(a1,c1,v1, a2,c2,v2), file=out)
# store results
s1_dict = create_statistics_dict(s=s1)
s2_dict = create_statistics_dict(s=s2)
results = dict()
inputs = list()
for attribute in input_attributes:
filename = getattr(params.input, attribute)
if (filename is not None):
inputs.append(filename)
assert (len(inputs) == 2)
results['map_files'] = inputs
results['map_statistics'] = (s1_dict, s2_dict)
results['cc_input_maps'] = cc_input_maps
results['cc_quantile'] = cc_quantile
results['cc_peaks'] = cc_peaks
results['discrepancies'] = discrepancies
# TODO, verify h1,h2 are not dicts, e.g. .values is py2/3 compat. I assume it is here
results['map_histograms'] = ((h1.arguments(), h1.c_values(), h1.values()),
(h2.arguments(), h2.c_values(), h2.values()))
return results
def validate_polder_map(self,
selection_bool,
xray_structure_noligand,
mask_data_polder,
box_cushion=2.1):
'''
The parameter box_cushion is hardcoded to be 2.1
The value is related to the site_radii used for CC calculation (box_cushion - 0.1)
Ideally the site_radii are calculated according to resolution, atom type and B factor for each atom
However, for the purpose of polder map validation, it is a reasonable approximation
to use 2.0.
If this value is changed, it will affect the values of the CCs and therefore also the
output messages (see mmtbx/programs/polder.py --> result_message)
So modify this value with caution.
'''
# Significance check
fmodel = mmtbx.f_model.manager(f_obs=self.f_obs,
r_free_flags=self.r_free_flags,
xray_structure=self.xray_structure)
fmodel.update_all_scales(remove_outliers=False, fast=True)
f_obs_1 = abs(fmodel.f_model())
fmodel.update_xray_structure(xray_structure=xray_structure_noligand,
update_f_calc=True,
update_f_mask=True,
force_update_f_mask=True)
## PVA: do we need it? fmodel.update_all_scales(remove_outliers=False)
f_obs_2 = abs(fmodel.f_model())
pdb_hierarchy_selected = self.pdb_hierarchy.select(selection_bool)
xrs_selected = pdb_hierarchy_selected.extract_xray_structure(
crystal_symmetry=self.cs)
f_calc = fmodel.f_obs().structure_factors_from_scatterers(
xray_structure=xray_structure_noligand).f_calc()
f_mask = fmodel.f_obs().structure_factors_from_map(
map=mask_data_polder,
use_scale=True,
anomalous_flag=False,
use_sg=False)
box_1 = self.get_polder_diff_map(f_obs=f_obs_1,
r_free_flags=fmodel.r_free_flags(),
f_calc=f_calc,
f_mask=f_mask,
xrs_selected=xrs_selected,
box_cushion=box_cushion)
box_2 = self.get_polder_diff_map(f_obs=f_obs_2,
r_free_flags=fmodel.r_free_flags(),
f_calc=f_calc,
f_mask=f_mask,
xrs_selected=xrs_selected,
box_cushion=box_cushion)
box_3 = self.get_polder_diff_map(f_obs=fmodel.f_obs(),
r_free_flags=fmodel.r_free_flags(),
f_calc=f_calc,
f_mask=f_mask,
xrs_selected=xrs_selected,
box_cushion=box_cushion)
sites_cart_box = box_1.xray_structure_box.sites_cart()
sel = maptbx.grid_indices_around_sites(
unit_cell=box_1.xray_structure_box.unit_cell(),
fft_n_real=box_1.map_box.focus(),
fft_m_real=box_1.map_box.all(),
sites_cart=sites_cart_box,
site_radii=flex.double(sites_cart_box.size(), box_cushion - 0.1))
b1 = box_1.map_box.select(sel).as_1d()
b2 = box_2.map_box.select(sel).as_1d()
b3 = box_3.map_box.select(sel).as_1d()
# Map 1: calculated Fobs with ligand
# Map 2: calculated Fobs without ligand
# Map 3: real Fobs data
cc12 = flex.linear_correlation(x=b1, y=b2).coefficient()
cc13 = flex.linear_correlation(x=b1, y=b3).coefficient()
cc23 = flex.linear_correlation(x=b2, y=b3).coefficient()
#### D-function
b1 = maptbx.volume_scale_1d(map=b1, n_bins=10000).map_data()
b2 = maptbx.volume_scale_1d(map=b2, n_bins=10000).map_data()
b3 = maptbx.volume_scale_1d(map=b3, n_bins=10000).map_data()
cc12_peak = flex.linear_correlation(x=b1, y=b2).coefficient()
cc13_peak = flex.linear_correlation(x=b1, y=b3).coefficient()
cc23_peak = flex.linear_correlation(x=b2, y=b3).coefficient()
#### Peak CC:
cutoffs = flex.double([i / 10. for i in range(1, 10)] +
[i / 100 for i in range(91, 100)])
d12 = maptbx.discrepancy_function(map_1=b1, map_2=b2, cutoffs=cutoffs)
d13 = maptbx.discrepancy_function(map_1=b1, map_2=b3, cutoffs=cutoffs)
d23 = maptbx.discrepancy_function(map_1=b2, map_2=b3, cutoffs=cutoffs)
pdb_hierarchy_selected.adopt_xray_structure(box_1.xray_structure_box)
return group_args(box_1=box_1,
box_2=box_2,
box_3=box_3,
cc12=cc12,
cc13=cc13,
cc23=cc23,
cc12_peak=cc12_peak,
cc13_peak=cc13_peak,
cc23_peak=cc23_peak,
d12=d12,
d13=d13,
d23=d23,
cutoffs=cutoffs,
ph_selected=pdb_hierarchy_selected)
def run(args, out=sys.stdout, validated=False):
show_citation(out=out)
if (len(args) == 0):
master_phil.show(out=out)
print >> out,\
'\nUsage: phenix.map_comparison <CCP4> <CCP4>\n',\
' phenix.map_comparison <CCP4> <MTZ> mtz_label_1=<label>\n',\
' phenix.map_comparison <MTZ 1> mtz_label_1=<label 1> <MTZ 2> mtz_label_2=<label 2>\n'
sys.exit()
# process arguments
params = None
input_attributes = ['map_1', 'mtz_1', 'map_2', 'mtz_2']
try: # automatic parsing
params = phil.process_command_line_with_files(
args=args, master_phil=master_phil).work.extract()
except Exception: # map_file_def only handles one map phil
from libtbx.phil.command_line import argument_interpreter
arg_int = argument_interpreter(master_phil=master_phil)
command_line_args = list()
map_files = list()
for arg in args:
if (os.path.isfile(arg)):
map_files.append(arg)
else:
command_line_args.append(arg_int.process(arg))
params = master_phil.fetch(sources=command_line_args).extract()
# check if more files are necessary
n_defined = 0
for attribute in input_attributes:
if (getattr(params.input, attribute) is not None):
n_defined += 1
# matches files to phil scope, stops once there is sufficient data
for map_file in map_files:
if (n_defined < 2):
current_map = file_reader.any_file(map_file)
if (current_map.file_type == 'ccp4_map'):
n_defined += 1
if (params.input.map_1 is None):
params.input.map_1 = map_file
elif (params.input.map_2 is None):
params.input.map_2 = map_file
elif (current_map.file_type == 'hkl'):
n_defined += 1
if (params.input.mtz_1 is None):
params.input.mtz_1 = map_file
elif (params.input.mtz_2 is None):
params.input.mtz_2 = map_file
else:
print >> out, 'WARNING: only the first two files are used'
break
# validate arguments (GUI sets validated to true, no need to run again)
assert (params is not None)
if (not validated):
validate_params(params)
# ---------------------------------------------------------------------------
# check if maps need to be generated from mtz
n_maps = 0
maps = list()
map_names = list()
for attribute in input_attributes:
filename = getattr(params.input, attribute)
if (filename is not None):
map_names.append(filename)
current_map = file_reader.any_file(filename)
maps.append(current_map)
if (current_map.file_type == 'ccp4_map'):
n_maps += 1
# construct maps, if necessary
crystal_gridding = None
m1 = None
m2 = None
# 1 map, 1 mtz file
if (n_maps == 1):
for current_map in maps:
if (current_map.file_type == 'ccp4_map'):
uc = current_map.file_object.unit_cell()
sg_info = space_group_info(current_map.file_object.space_group_number)
n_real = current_map.file_object.unit_cell_grid
crystal_gridding = maptbx.crystal_gridding(
uc, space_group_info=sg_info, pre_determined_n_real=n_real)
m1 = current_map.file_object.map_data()
if (crystal_gridding is not None):
label = None
for attribute in [('mtz_1', 'mtz_label_1'),
('mtz_2', 'mtz_label_2')]:
filename = getattr(params.input, attribute[0])
label = getattr(params.input, attribute[1])
if ( (filename is not None) and (label is not None) ):
break
# labels will match currently open mtz file
for current_map in maps:
if (current_map.file_type == 'hkl'):
m2 = miller.fft_map(
crystal_gridding=crystal_gridding,
fourier_coefficients=current_map.file_server.get_miller_array(
label)).apply_sigma_scaling().real_map_unpadded()
else:
raise Sorry('Gridding is not defined.')
# 2 mtz files
elif (n_maps == 0):
crystal_symmetry = get_crystal_symmetry(maps[0])
d_min = min(get_d_min(maps[0]), get_d_min(maps[1]))
crystal_gridding = maptbx.crystal_gridding(
crystal_symmetry.unit_cell(), d_min=d_min,
resolution_factor=params.options.resolution_factor,
space_group_info=crystal_symmetry.space_group_info())
m1 = miller.fft_map(
crystal_gridding=crystal_gridding,
fourier_coefficients=maps[0].file_server.get_miller_array(
params.input.mtz_label_1)).apply_sigma_scaling().real_map_unpadded()
m2 = miller.fft_map(
crystal_gridding=crystal_gridding,
fourier_coefficients=maps[1].file_server.get_miller_array(
params.input.mtz_label_2)).apply_sigma_scaling().real_map_unpadded()
# 2 maps
else:
m1 = maps[0].file_object.map_data()
m2 = maps[1].file_object.map_data()
# ---------------------------------------------------------------------------
# analyze maps
assert ( (m1 is not None) and (m2 is not None) )
# show general statistics
s1 = maptbx.more_statistics(m1)
s2 = maptbx.more_statistics(m2)
show_overall_statistics(out=out, s=s1, header="Map 1 (%s):"%map_names[0])
show_overall_statistics(out=out, s=s2, header="Map 2 (%s):"%map_names[1])
cc_input_maps = flex.linear_correlation(x = m1.as_1d(),
y = m2.as_1d()).coefficient()
print >> out, "CC, input maps: %6.4f" % cc_input_maps
# compute CCpeak
cc_peaks = list()
m1_he = maptbx.volume_scale(map = m1, n_bins = 10000).map_data()
m2_he = maptbx.volume_scale(map = m2, n_bins = 10000).map_data()
cc_quantile = flex.linear_correlation(x = m1_he.as_1d(),
y = m2_he.as_1d()).coefficient()
print >> out, "CC, quantile rank-scaled (histogram equalized) maps: %6.4f" % \
cc_quantile
print >> out, "Peak correlation:"
print >> out, " cutoff CCpeak"
cutoffs = [i/100. for i in range(1,90)]+ [i/1000 for i in range(900,1000)]
for cutoff in cutoffs:
cc_peak = maptbx.cc_peak(map_1=m1_he, map_2=m2_he, cutoff=cutoff)
print >> out, " %3.2f %7.4f" % (cutoff, cc_peak)
cc_peaks.append((cutoff, cc_peak))
# compute discrepancy function (D-function)
discrepancies = list()
cutoffs = flex.double(cutoffs)
df = maptbx.discrepancy_function(map_1=m1_he, map_2=m2_he, cutoffs=cutoffs)
print >> out, "Discrepancy function:"
print >> out, " cutoff D"
for c, d in zip(cutoffs, df):
print >> out, " %3.2f %7.4f" % (c,d)
discrepancies.append((c, d))
# compute and output histograms
h1 = maptbx.histogram(map=m1, n_bins=10000)
h2 = maptbx.histogram(map=m2, n_bins=10000)
print >> out, "Map histograms:"
print >> out, "Map 1 (%s) Map 2 (%s)"%\
(params.input.map_1,params.input.map_2)
print >> out, "(map_value,cdf,frequency) <> (map_value,cdf,frequency)"
for a1,c1,v1, a2,c2,v2 in zip(h1.arguments(), h1.c_values(), h1.values(),
h2.arguments(), h2.c_values(), h2.values()):
print >> out, "(%9.5f %9.5f %9.5f) <> (%9.5f %9.5f %9.5f)"%\
(a1,c1,v1, a2,c2,v2)
# store results
s1_dict = create_statistics_dict(s=s1)
s2_dict = create_statistics_dict(s=s2)
results = dict()
inputs = list()
for attribute in input_attributes:
filename = getattr(params.input,attribute)
if (filename is not None):
inputs.append(filename)
assert (len(inputs) == 2)
results['map_files'] = inputs
results['map_statistics'] = (s1_dict, s2_dict)
results['cc_input_maps'] = cc_input_maps
results['cc_quantile'] = cc_quantile
results['cc_peaks'] = cc_peaks
results['discrepancies'] = discrepancies
results['map_histograms'] = ( (h1.arguments(), h1.c_values(), h1.values()),
(h2.arguments(), h2.c_values(), h2.values()) )
return results
def cmd_run(args, validated=False, out=sys.stdout):
if (len(args) == 0):
print >> out, "-"*79
print >> out, " phenix.polder"
print >> out, "-"*79
print >> out, legend
print >> out, "-"*79
master_params.show(out=out)
return
log = multi_out()
log.register("stdout", out)
log_file_name = "polder.log"
logfile = open(log_file_name, "w")
log.register("logfile", logfile)
print >> log, "phenix.polder is running..."
print >> log, "input parameters:\n", args
parsed = master_params
inputs = mmtbx.utils.process_command_line_args(args = args,
master_params = parsed)
#inputs.params.show() #check
params = inputs.params.extract()
# check model file
if len(inputs.pdb_file_names) == 0:
if (params.model_file_name is None):
raise Sorry("No model file found.")
elif (len(inputs.pdb_file_names) == 1):
params.model_file_name = inputs.pdb_file_names[0]
else:
raise Sorry("Only one model file should be given")
# check reflection file
reflection_files = inputs.reflection_files
if (len(reflection_files) == 0):
if (params.reflection_file_name is None):
raise Sorry("No reflection file found.")
else:
hkl_in = file_reader.any_file(params.reflection_file_name,
force_type="hkl")
hkl_in.assert_file_type("hkl")
reflection_files = [ hkl_in.file_object ]
# crystal symmetry
crystal_symmetry = None
crystal_symmetry = inputs.crystal_symmetry
if (crystal_symmetry is None):
crystal_symmetries = []
for f in [str(params.model_file_name), str(params.reflection_file_name)]:
cs = crystal_symmetry_from_any.extract_from(f)
if(cs is not None): crystal_symmetries.append(cs)
if(len(crystal_symmetries) == 1): crystal_symmetry = crystal_symmetries[0]
elif(len(crystal_symmetries) == 0):
raise Sorry("No crystal symmetry found.")
else:
if(not crystal_symmetries[0].is_similar_symmetry(crystal_symmetries[1])):
raise Sorry("Crystal symmetry mismatch between different files.")
crystal_symmetry = crystal_symmetries[0]
f_obs, r_free_flags = None, None
rfs = reflection_file_utils.reflection_file_server(
crystal_symmetry = crystal_symmetry,
force_symmetry = True,
reflection_files = reflection_files,
err = StringIO())
parameters = mmtbx.utils.data_and_flags_master_params().extract()
if (params.data_labels is not None):
parameters.labels = params.data_labels
if (params.r_free_flags_labels is not None):
parameters.r_free_flags.label = params.r_free_flags_labels
determined_data_and_flags = mmtbx.utils.determine_data_and_flags(
reflection_file_server = rfs,
parameters = parameters,
keep_going = True,
log = StringIO())
f_obs = determined_data_and_flags.f_obs
if (params.data_labels is None):
params.data_labels = f_obs.info().label_string()
if (params.reflection_file_name is None):
params.reflection_file_name = parameters.file_name
r_free_flags = determined_data_and_flags.r_free_flags
assert f_obs is not None
print >> log, "Input data:"
print >> log, " Iobs or Fobs:", f_obs.info().labels
if (r_free_flags is not None):
print >> log, " Free-R flags:", r_free_flags.info().labels
params.r_free_flags_labels = r_free_flags.info().label_string()
else:
print >> log, " Free-R flags: Not present"
model_basename = os.path.basename(params.model_file_name.split(".")[0])
if (len(model_basename) > 0 and
params.output_file_name_prefix is None):
params.output_file_name_prefix = model_basename
print params.output_file_name_prefix
new_params = master_params.format(python_object=params)
new_params.show()
if (not validated):
validate_params(params)
pdb_input = iotbx.pdb.input(file_name = params.model_file_name)
pdb_hierarchy = pdb_input.construct_hierarchy()
xray_structure = pdb_hierarchy.extract_xray_structure(
crystal_symmetry = crystal_symmetry)
# DON'T USE:
# xray_structure = pdb_input.xray_structure_simple()
# atom order might be wrong
mmtbx.utils.setup_scattering_dictionaries(
scattering_table = params.scattering_table,
xray_structure = xray_structure,
d_min = f_obs.d_min())
#if f_obs is not None:
f_obs = f_obs.resolution_filter(
d_min = params.high_resolution,
d_max = params.low_resolution)
if (r_free_flags is not None):
r_free_flags = r_free_flags.resolution_filter(
d_min = params.high_resolution,
d_max = params.low_resolution)
# Grab case that data are anomalous
if (f_obs.anomalous_flag()):
f_obs, r_free_flags = prepare_f_obs_and_flags(
f_obs = f_obs,
r_free_flags = r_free_flags)
cpm_obj = compute_polder_map(
f_obs = f_obs,
r_free_flags = r_free_flags,
xray_structure = xray_structure,
pdb_hierarchy = pdb_hierarchy,
params = params,
log = log)
# Significance check
fmodel = mmtbx.f_model.manager(
f_obs = f_obs,
r_free_flags = r_free_flags,
xray_structure = xray_structure)
fmodel.update_all_scales(remove_outliers=False, fast=True)
f_obs_1 = abs(fmodel.f_model())
fmodel.update_xray_structure(xray_structure=cpm_obj.xray_structure_noligand,
update_f_calc=True, update_f_mask=True, force_update_f_mask=True)
# PVA: do we need it? fmodel.update_all_scales(remove_outliers=False)
f_obs_2 = abs(fmodel.f_model())
xrs_selected = cpm_obj.pdb_hierarchy_selected.extract_xray_structure(
crystal_symmetry = f_obs.crystal_symmetry())
f_calc = f_obs.structure_factors_from_scatterers(
xray_structure = cpm_obj.xray_structure_noligand).f_calc()
f_mask = f_obs.structure_factors_from_map(
map = cpm_obj.mask_polder,
use_scale = True,
anomalous_flag = False,
use_sg = False)
def get_poler_diff_map(f_obs):
fmodel = mmtbx.f_model.manager(
f_obs = f_obs,
r_free_flags = r_free_flags,
f_calc = f_calc,
f_mask = f_mask)
fmodel.update_all_scales(remove_outliers=False)
mc_diff = map_tools.electron_density_map(
fmodel = fmodel).map_coefficients(
map_type = "mFo-DFc",
isotropize = True,
fill_missing = False)
fft_map = miller.fft_map(
crystal_gridding = cpm_obj.crystal_gridding,
fourier_coefficients = mc_diff)
fft_map.apply_sigma_scaling()
map_data = fft_map.real_map_unpadded()
return mmtbx.utils.extract_box_around_model_and_map(
xray_structure = xrs_selected,
map_data = map_data,
box_cushion = 2.1)
box_1=get_poler_diff_map(f_obs = f_obs_1)
box_2=get_poler_diff_map(f_obs = f_obs_2)
box_3=get_poler_diff_map(f_obs = f_obs)
sites_cart_box = box_1.xray_structure_box.sites_cart()
sel = maptbx.grid_indices_around_sites(
unit_cell = box_1.xray_structure_box.unit_cell(),
fft_n_real = box_1.map_box.focus(),
fft_m_real = box_1.map_box.all(),
sites_cart = sites_cart_box,
site_radii = flex.double(sites_cart_box.size(), 2.0))
b1 = box_1.map_box.select(sel).as_1d()
b2 = box_2.map_box.select(sel).as_1d()
b3 = box_3.map_box.select(sel).as_1d()
print >> log, "Map 1: calculated Fobs with ligand"
print >> log, "Map 2: calculated Fobs without ligand"
print >> log, "Map 3: real Fobs data"
print >>log, "CC(1,2): %6.4f"%flex.linear_correlation(x=b1,y=b2).coefficient()
print >>log, "CC(1,3): %6.4f"%flex.linear_correlation(x=b1,y=b3).coefficient()
print >>log, "CC(2,3): %6.4f"%flex.linear_correlation(x=b2,y=b3).coefficient()
### D-function
b1 = maptbx.volume_scale_1d(map=b1, n_bins=10000).map_data()
b2 = maptbx.volume_scale_1d(map=b2, n_bins=10000).map_data()
b3 = maptbx.volume_scale_1d(map=b3, n_bins=10000).map_data()
print >> log, "Peak CC:"
print >>log, "CC(1,2): %6.4f"%flex.linear_correlation(x=b1,y=b2).coefficient()
print >>log, "CC(1,3): %6.4f"%flex.linear_correlation(x=b1,y=b3).coefficient()
print >>log, "CC(2,3): %6.4f"%flex.linear_correlation(x=b2,y=b3).coefficient()
cutoffs = flex.double(
[i/10. for i in range(1,10)]+[i/100 for i in range(91,100)])
d12 = maptbx.discrepancy_function(map_1=b1, map_2=b2, cutoffs=cutoffs)
d13 = maptbx.discrepancy_function(map_1=b1, map_2=b3, cutoffs=cutoffs)
d23 = maptbx.discrepancy_function(map_1=b2, map_2=b3, cutoffs=cutoffs)
print >> log, "q D(1,2) D(1,3) D(2,3)"
for c,d12_,d13_,d23_ in zip(cutoffs,d12,d13,d23):
print >> log, "%4.2f %6.4f %6.4f %6.4f"%(c,d12_,d13_,d23_)
###
if(params.debug):
box_1.write_ccp4_map(file_name="box_1_polder.ccp4")
box_2.write_ccp4_map(file_name="box_2_polder.ccp4")
box_3.write_ccp4_map(file_name="box_3_polder.ccp4")
cpm_obj.pdb_hierarchy_selected.adopt_xray_structure(
box_1.xray_structure_box)
cpm_obj.pdb_hierarchy_selected.write_pdb_file(file_name="box_polder.pdb",
crystal_symmetry=box_1.box_crystal_symmetry)
#
print >> log, "Finished."
return True
def validate_polder_map(self):
# Significance check
fmodel = mmtbx.f_model.manager(f_obs=self.f_obs,
r_free_flags=self.r_free_flags,
xray_structure=self.xray_structure)
fmodel.update_all_scales(remove_outliers=False, fast=True)
f_obs_1 = abs(fmodel.f_model())
fmodel.update_xray_structure(
xray_structure=self.xray_structure_noligand,
update_f_calc=True,
update_f_mask=True,
force_update_f_mask=True)
## PVA: do we need it? fmodel.update_all_scales(remove_outliers=False)
f_obs_2 = abs(fmodel.f_model())
pdb_hierarchy_selected = self.pdb_hierarchy.select(self.selection_bool)
xrs_selected = pdb_hierarchy_selected.extract_xray_structure(
crystal_symmetry=self.f_obs.crystal_symmetry())
f_calc = fmodel.f_obs().structure_factors_from_scatterers(
xray_structure=self.xray_structure_noligand).f_calc()
f_mask = fmodel.f_obs().structure_factors_from_map(
map=self.mask_data_polder,
use_scale=True,
anomalous_flag=False,
use_sg=False)
box_1 = self.get_polder_diff_map(f_obs=f_obs_1,
r_free_flags=fmodel.r_free_flags(),
f_calc=f_calc,
f_mask=f_mask,
xrs_selected=xrs_selected)
box_2 = self.get_polder_diff_map(f_obs=f_obs_2,
r_free_flags=fmodel.r_free_flags(),
f_calc=f_calc,
f_mask=f_mask,
xrs_selected=xrs_selected)
box_3 = self.get_polder_diff_map(f_obs=fmodel.f_obs(),
r_free_flags=fmodel.r_free_flags(),
f_calc=f_calc,
f_mask=f_mask,
xrs_selected=xrs_selected)
sites_cart_box = box_1.xray_structure_box.sites_cart()
sel = maptbx.grid_indices_around_sites(
unit_cell=box_1.xray_structure_box.unit_cell(),
fft_n_real=box_1.map_box.focus(),
fft_m_real=box_1.map_box.all(),
sites_cart=sites_cart_box,
site_radii=flex.double(sites_cart_box.size(), 2.0))
b1 = box_1.map_box.select(sel).as_1d()
b2 = box_2.map_box.select(sel).as_1d()
b3 = box_3.map_box.select(sel).as_1d()
# Map 1: calculated Fobs with ligand
# Map 2: calculated Fobs without ligand
# Map 3: real Fobs data
cc12 = flex.linear_correlation(x=b1, y=b2).coefficient()
cc13 = flex.linear_correlation(x=b1, y=b3).coefficient()
cc23 = flex.linear_correlation(x=b2, y=b3).coefficient()
#### D-function
b1 = maptbx.volume_scale_1d(map=b1, n_bins=10000).map_data()
b2 = maptbx.volume_scale_1d(map=b2, n_bins=10000).map_data()
b3 = maptbx.volume_scale_1d(map=b3, n_bins=10000).map_data()
cc12_peak = flex.linear_correlation(x=b1, y=b2).coefficient()
cc13_peak = flex.linear_correlation(x=b1, y=b3).coefficient()
cc23_peak = flex.linear_correlation(x=b2, y=b3).coefficient()
#### Peak CC:
cutoffs = flex.double([i / 10. for i in range(1, 10)] +
[i / 100 for i in range(91, 100)])
d12 = maptbx.discrepancy_function(map_1=b1, map_2=b2, cutoffs=cutoffs)
d13 = maptbx.discrepancy_function(map_1=b1, map_2=b3, cutoffs=cutoffs)
d23 = maptbx.discrepancy_function(map_1=b2, map_2=b3, cutoffs=cutoffs)
pdb_hierarchy_selected.adopt_xray_structure(box_1.xray_structure_box)
self.validation_results = group_args(
box_1=box_1,
box_2=box_2,
box_3=box_3,
cc12=cc12,
cc13=cc13,
cc23=cc23,
cc12_peak=cc12_peak,
cc13_peak=cc13_peak,
cc23_peak=cc23_peak,
d12=d12,
d13=d13,
d23=d23,
cutoffs=cutoffs,
ph_selected=pdb_hierarchy_selected)
def run(args, validated=False):
show_citation()
if ( (len(args) == 0) or (len(args) > 2) ):
print '\nUsage: phenix.map_comparison map_1=<first map> map_2=<second map>\n'
sys.exit()
# process arguments
try: # automatic parsing
params = phil.process_command_line_with_files(
args=args, master_phil=master_phil).work.extract()
except Exception: # map_file_def only handles one map phil
from libtbx.phil.command_line import argument_interpreter
arg_int = argument_interpreter(master_phil=master_phil)
command_line_args = list()
map_files = list()
for arg in args:
if (os.path.isfile(arg)):
map_files.append(arg)
else:
command_line_args.append(arg_int.process(arg))
params = master_phil.fetch(sources=command_line_args).extract()
for map_file in map_files:
if (params.input.map_1 is None):
params.input.map_1 = map_file
else:
params.input.map_2 = map_file
# validate arguments (GUI sets validated to true, no need to run again)
if (not validated):
validate_params(params)
# ---------------------------------------------------------------------------
# map 1
ccp4_map_1 = iotbx.ccp4_map.map_reader(file_name=params.input.map_1)
cs_1 = crystal.symmetry(ccp4_map_1.unit_cell().parameters(),
ccp4_map_1.space_group_number)
m1 = ccp4_map_1.map_data()
# map 2
ccp4_map_2 = iotbx.ccp4_map.map_reader(file_name=params.input.map_2)
cs_2 = crystal.symmetry(ccp4_map_2.unit_cell().parameters(),
ccp4_map_2.space_group_number)
m2 = ccp4_map_2.map_data()
# show general statistics
s1 = maptbx.more_statistics(m1)
s2 = maptbx.more_statistics(m2)
show_overall_statistics(s=s1, header="Map 1 (%s):"%params.input.map_1)
show_overall_statistics(s=s2, header="Map 2 (%s):"%params.input.map_2)
cc_input_maps = flex.linear_correlation(x = m1.as_1d(),
y = m2.as_1d()).coefficient()
print "CC, input maps: %6.4f" % cc_input_maps
# compute CCpeak
cc_peaks = list()
m1_he = maptbx.volume_scale(map = m1, n_bins = 10000).map_data()
m2_he = maptbx.volume_scale(map = m2, n_bins = 10000).map_data()
cc_quantile = flex.linear_correlation(x = m1_he.as_1d(),
y = m2_he.as_1d()).coefficient()
print "CC, quantile rank-scaled (histogram equalized) maps: %6.4f" % \
cc_quantile
print "Peak correlation:"
print " cutoff CCpeak"
for cutoff in [i/100. for i in range(0,100,5)]+[0.99, 1.0]:
cc_peak = maptbx.cc_peak(map_1=m1_he, map_2=m2_he, cutoff=cutoff)
print " %3.2f %7.4f" % (cutoff, cc_peak)
cc_peaks.append((cutoff, cc_peak))
# compute discrepancy function (D-function)
discrepancies = list()
cutoffs = flex.double([i/20. for i in range(1,20)])
df = maptbx.discrepancy_function(map_1=m1_he, map_2=m2_he, cutoffs=cutoffs)
print "Discrepancy function:"
print " cutoff D"
for c, d in zip(cutoffs, df):
print " %3.2f %7.4f" % (c,d)
discrepancies.append((c, d))
# compute and output histograms
h1 = maptbx.histogram(map=m1, n_bins=10000)
h2 = maptbx.histogram(map=m2, n_bins=10000)
print "Map histograms:"
print "Map 1 (%s) Map 2 (%s)"%(params.input.map_1,params.input.map_2)
print "(map_value,cdf,frequency) <> (map_value,cdf,frequency)"
for a1,c1,v1, a2,c2,v2 in zip(h1.arguments(), h1.c_values(), h1.values(),
h2.arguments(), h2.c_values(), h2.values()):
print "(%9.5f %9.5f %9.5f) <> (%9.5f %9.5f %9.5f)"%(a1,c1,v1, a2,c2,v2)
# store results
s1_dict = create_statistics_dict(s1)
s2_dict = create_statistics_dict(s2)
results = dict()
results['map_files'] = (params.input.map_1, params.input.map_2)
results['map_statistics'] = (s1_dict, s2_dict)
results['cc_input_maps'] = cc_input_maps
results['cc_quantile'] = cc_quantile
results['cc_peaks'] = cc_peaks
results['discrepancies'] = discrepancies
results['map_histograms'] = ( (h1.arguments(), h1.c_values(), h1.values()),
(h2.arguments(), h2.c_values(), h2.values()) )
return results
