def run_biojava_alignment(cmd_class,
code1,
code2,
alignment,
transformed,
superposed,
pdb1=None,
pdb2=None,
log=None):
try:
exec_dir = tempfile.mkdtemp()
pdbid1, chain1 = extract_ident(code1)
pdbid2, chain2 = extract_ident(code2)
pdb1 = pdb1 or '{0}.pdb'.format(pdbid1)
pdb2 = pdb2 or '{0}.pdb'.format(pdbid2)
# Setup command based on parameters
cmd = _biojava_base_cmd.bake(cmd_class, "-pdb1", code1, "-pdb2", code2,
"-pdbFilePath", exec_dir,
*BIOJAVA_COMMON_ARGS)
# BioJava Structure is very picky about names
for pdb, code in [(pdb1, pdbid1), (pdb2, pdbid2)]:
entfile = "pdb{0}.ent".format(code.lower())
os.symlink(pdb, os.path.join(exec_dir, entfile))
logging.info("Running biojava: {0!s} in {1}".format(cmd, exec_dir))
alignment_result = cmd(_cwd=exec_dir, _err=log)
with open(alignment, 'w') as f:
f.write(str(alignment_result.stdout))
superposed_pdb = os.path.join(exec_dir, ALIGNMENT_TEMP_PDB)
pdb2_transformed = get_pdb_selection(
pdbid2,
chain=chain2,
model=1,
root=superposed_pdb,
)
shutil.move(os.path.join(exec_dir, ALIGNMENT_TEMP_PDB), superposed)
shutil.move(pdb2_transformed, transformed)
except Exception as e:
raise
else:
shutil.rmtree(exec_dir)
def get_charged_pdb(code,
chain=None,
model=0,
alignment_id=None,
force=False,
log=None):
selection_path = get_pdb_selection(code,
chain=chain,
model=model,
alignment_id=alignment_id)
base, ext = split_ext_gz(selection_path)
protonated = base + '.H'
if force or not os.path.exists(protonated):
run_reduce(selection_path, protonated, log=log)
return protonated
def create_context_from_pdb_code(code, chain):
pdb = get_pdb_selection(code, chain)
params = ResidueContextParams(source=pdb, chain=chain, name=code)
builder = ContextBuilder(params)
context = builder.run()
return context
def run_residuecontext_alignment(code1,
code2,
alignment,
translation,
transformed,
pdb1=None,
pdb2=None,
log=None):
"""
:param cmd:
:param code1:
:param code2:
:param alignment:
:param transformed:
:param superposed:
:param code1:
:param code2:
:param log:
:return:
"""
try:
exec_dir = tempfile.mkdtemp()
working_dir = os.path.join(exec_dir, 'working')
os.mkdir(working_dir)
ident1 = code1.split()
ident2 = code2.split()
pdbid1 = ident1[0]
pdbid2 = ident2[0]
if len(ident1) > 1:
chain1 = ident1[1]
else:
chain1 = 'A'
if len(ident2) > 1:
chain2 = ident2[1]
else:
chain2 = 'A'
# Setup command based on parameters
cmd = _residue_context_cmd.bake(pdbid1.upper(), pdbid2.upper(), chain1,
chain2, "working",
*RESIDUE_CONTEXT_COMMON_ARGS)
# ResidueContext is picky about names
# And also requires all PDB record types to be 6 characters
# But biopython writes them out without padding
for (pdb, code) in [(pdb1, ident1[0]), (pdb2, ident2[0])]:
entfile = "{}.pdb".format(code.upper())
with open(pdb) as src, \
open(os.path.join(working_dir, entfile), 'w') as dst:
for line in src:
if line == 'END\n':
dst.write('END \n')
elif line == 'TER\n':
dst.write('TER \n')
else:
dst.write(line)
logging.info("Running residue context: {0!s} in {1}".format(
cmd, exec_dir))
cmd(_cwd=exec_dir, _err=log, _env=RESIDUE_CONTEXT_ENV)
pdb2_transformed = get_pdb_selection(
pdbid2.upper(),
chain=chain2,
root=working_dir,
)
shutil.move(os.path.join(exec_dir, ALIGNMENT_OUTPUT), alignment)
shutil.move(
os.path.join(working_dir,
"{0}.pdb-transform.txt".format(ident2[0].upper())),
translation)
shutil.move(pdb2_transformed, transformed)
except Exception as e:
raise
else:
shutil.rmtree(exec_dir)
def run_alignment_comparisons(identifier, code1, code2):
print("starting...")
working_dir = werkzeug.security.safe_join(ALIGNMENT_JOB_DIR,
str(identifier))
progress_file = os.path.join(working_dir, 'status')
if os.path.exists(working_dir):
shutil.rmtree(working_dir)
os.mkdir(working_dir)
with open(progress_file, 'w') as f:
print("running", file=f)
pdbid1, chain1 = extract_ident(code1)
pdbid2, chain2 = extract_ident(code2)
with open(os.path.join(working_dir, ALIGNMENT_DATA_FILE), 'w') as f:
json.dump(
{
'structure1': {
'ident': code1,
'pdb': pdbid1,
'chain': chain1
},
'structure2': {
'ident': code2,
'pdb': pdbid2,
'chain': chain2
}
}, f)
pdb1src = get_pdb_selection(pdbid1, chain=chain1)
pdb2src = get_pdb_selection(pdbid2, chain=chain2)
pdb1 = os.path.join(working_dir, "{0}.pdb".format(pdbid1))
pdb2 = os.path.join(working_dir, "{0}.pdb".format(pdbid2))
shutil.copy(pdb1src, pdb1)
shutil.copy(pdb2src, pdb2)
rcontext_dir = os.path.join(working_dir, "rcontext")
ce_dir = os.path.join(working_dir, "ce")
fatcat_dir = os.path.join(working_dir, "fatcat")
dali_dir = os.path.join(working_dir, "dali")
alignment_basename = 'alignment.txt'
translated_basename = 'transformed-{0}.pdb'.format(pdbid2)
superposed_basename = 'superposed.pdb'
transform_basename = '{0}.pdb-transform.txt'.format(pdbid2)
biojava_code1 = ident_to_biojava(pdbid1, chain1)
biojava_code2 = ident_to_biojava(pdbid2, chain2)
os.mkdir(ce_dir)
run_biojava_alignment(BIOJAVA_CE_CLASS,
biojava_code1,
biojava_code2,
alignment=os.path.join(ce_dir, alignment_basename),
transformed=os.path.join(ce_dir,
translated_basename),
superposed=os.path.join(ce_dir, superposed_basename),
pdb1=pdb1,
pdb2=pdb2)
os.symlink(pdb1, os.path.join(ce_dir, "{0}.pdb".format(pdbid1)))
os.symlink(pdb2, os.path.join(ce_dir, "{0}.pdb".format(pdbid2)))
phi = get_electrostatics_grid(pdbid1, chain=chain1, alignment_id=ce_dir)
get_vanderderwaals_grids(pdbid1,
chain=chain1,
alignment_id=ce_dir,
box=phi,
scale=1 / phi.scale)
phi = get_electrostatics_grid(pdbid2, chain=chain2, alignment_id=ce_dir)
get_vanderderwaals_grids(pdbid2,
chain=chain2,
alignment_id=ce_dir,
box=phi,
scale=1 / phi.scale)
os.mkdir(fatcat_dir)
run_biojava_alignment(BIOJAVA_FATCAT_CLASS,
biojava_code1,
biojava_code2,
alignment=os.path.join(fatcat_dir,
alignment_basename),
transformed=os.path.join(fatcat_dir,
translated_basename),
superposed=os.path.join(fatcat_dir,
superposed_basename),
pdb1=pdb1,
pdb2=pdb2)
os.symlink(pdb1, os.path.join(fatcat_dir, "{0}.pdb".format(pdbid1)))
os.symlink(pdb2, os.path.join(fatcat_dir, "{0}.pdb".format(pdbid2)))
phi = get_electrostatics_grid(pdbid1,
chain=chain1,
alignment_id=fatcat_dir)
get_vanderderwaals_grids(pdbid1,
chain=chain1,
alignment_id=fatcat_dir,
box=phi,
scale=1 / phi.scale)
phi = get_electrostatics_grid(pdbid2,
chain=chain2,
alignment_id=fatcat_dir)
get_vanderderwaals_grids(pdbid2,
chain=chain2,
alignment_id=fatcat_dir,
box=phi,
scale=1 / phi.scale)
os.mkdir(rcontext_dir)
run_residuecontext_alignment(
ident_to_rcontext(pdbid1, chain1),
ident_to_rcontext(pdbid2, chain2),
alignment=os.path.join(rcontext_dir, alignment_basename),
transformed=os.path.join(rcontext_dir, translated_basename),
translation=os.path.join(rcontext_dir, transform_basename),
pdb1=pdb1,
pdb2=pdb2)
os.symlink(pdb1, os.path.join(rcontext_dir, "{0}.pdb".format(pdbid1)))
os.symlink(pdb2, os.path.join(rcontext_dir, "{0}.pdb".format(pdbid2)))
phi = get_electrostatics_grid(pdbid1,
chain=chain1,
alignment_id=rcontext_dir)
get_vanderderwaals_grids(pdbid1,
chain=chain1,
alignment_id=rcontext_dir,
box=phi,
scale=1 / phi.scale)
phi = get_electrostatics_grid(pdbid2,
chain=chain2,
alignment_id=rcontext_dir)
get_vanderderwaals_grids(pdbid2,
chain=chain2,
alignment_id=rcontext_dir,
box=phi,
scale=1 / phi.scale)
os.mkdir(dali_dir)
run_dalilite_alignment(ident_to_rcontext(pdbid1, chain1),
ident_to_rcontext(pdbid2, chain2),
alignment=os.path.join(dali_dir,
alignment_basename),
translation=os.path.join(dali_dir,
transform_basename),
transformed=os.path.join(dali_dir,
translated_basename),
pdb1=pdb1,
pdb2=pdb2)
os.symlink(pdb1, os.path.join(dali_dir, "{0}.pdb".format(pdbid1)))
os.symlink(pdb2, os.path.join(dali_dir, "{0}.pdb".format(pdbid2)))
phi = get_electrostatics_grid(pdbid1, chain=chain1, alignment_id=dali_dir)
get_vanderderwaals_grids(pdbid1,
chain=chain1,
alignment_id=dali_dir,
box=phi,
scale=1 / phi.scale)
phi = get_electrostatics_grid(pdbid2, chain=chain2, alignment_id=dali_dir)
get_vanderderwaals_grids(pdbid2,
chain=chain2,
alignment_id=dali_dir,
box=phi,
scale=1 / phi.scale)
with open(progress_file, 'w') as f:
print("done", file=f)
print("Finished.")