def main(hhblitsdb, jackhmmerdb, seqfile, n_cores=1, n_jobs_plm=1, n_jobs_psi=1, layers=5, pconsc1_flag=False):
# Create input alignments for evalue thresholds 1, 10^-4, 10^-10,
# and 10^-40
prep.run_alignments(hhblitsdb, jackhmmerdb, seqfile, n_cores=n_cores)
# Run PSICOV and plmDCA on every alignment and collect output
# filepaths
psicovnames = prep.run_contact_pred(seqfile, 'psicov', n_cores=n_cores, n_jobs=n_jobs_psi)
plmdcanames = prep.run_contact_pred(seqfile, 'plmdca', n_cores=n_cores, n_jobs=n_jobs_plm)
predictionnames = dict(psicovnames.items() + plmdcanames.items())
# Generate PconsC/2 prediction command with all arguments in
# correct order
l = [root + '/src/predict2.py']
names = ['jhE4', 'jhE0', 'jhE10', 'jhE40', 'hhE4', 'hhE0', 'hhE10', 'hhE40']
for key in names:
l.append(predictionnames[key + 'psicov'])
l.append(predictionnames[key + 'plmdca'])
# Run PconsC if flag is given
if pconsc1_flag:
sys.stderr.write("Running PconsC1...\n")
sys.stderr.write("Predicting...\n")
result_name = seqfile + '.pconsc.out'
l[0] = root + '/src/predict.py'
results = check_output(l)
f = open(result_name, 'w')
f.write(results)
f.close()
# Run PconsC2 otherwise (default)
else:
sys.stderr.write("Running PconsC2...\n")
netsurfpredictionname, sspredictionname, pssmaliname = prep.run_pconsc2_dependencies(hhblitsdb, seqfile, n_cores=1)
sys.stderr.write("Predicting...\n")
result_name = seqfile + '.pconsc2.out'
l.extend([netsurfpredictionname, sspredictionname, pssmaliname, result_name])
check_output(l)
# Plot top L*1 contacts in a contact map (where L is the length of
# the input sequence).
# Those contacts are later used during protein folding
if plot_flag:
if os.path.exists('native.pdb') and os.path.exists(seqfile + '.horiz'):
plot_map(seqfile, result_name, 1., pdb_filename='native.pdb', psipred_filename=seqfile + '.horiz')
elif os.path.exists('native.pdb'):
plot_map(seqfile, result_name, 1., pdb_filename='native.pdb')
elif os.path.exists(seqfile + '.horiz'):
plot_map(seqfile, result_name, 1., psipred_filename=seqfile + '.horiz')
else:
plot_map(seqfile, result_name, 1.)
def main(hhblitsdb,
jackhmmerdb,
seqfile,
n_cores=1,
n_jobs_plm=1,
n_jobs_psi=1,
layers=5,
pconsc1_flag=False):
# Create input alignments for evalue thresholds 1, 10^-4, 10^-10,
# and 10^-40
prep.run_alignments(hhblitsdb, jackhmmerdb, seqfile, n_cores=n_cores)
# Run PSICOV and plmDCA on every alignment and collect output
# filepaths
psicovnames = prep.run_contact_pred(seqfile,
'psicov',
n_cores=n_cores,
n_jobs=n_jobs_psi)
plmdcanames = prep.run_contact_pred(seqfile,
'plmdca',
n_cores=n_cores,
n_jobs=n_jobs_plm)
predictionnames = dict(psicovnames.items() + plmdcanames.items())
# Generate PconsC/2 prediction command with all arguments in
# correct order
l = [root + '/src/predict2.py']
names = [
'jhE4', 'jhE0', 'jhE10', 'jhE40', 'hhE4', 'hhE0', 'hhE10', 'hhE40'
]
for key in names:
l.append(predictionnames[key + 'psicov'])
l.append(predictionnames[key + 'plmdca'])
# Run PconsC if flag is given
if pconsc1_flag:
sys.stderr.write("Running PconsC1...\n")
sys.stderr.write("Predicting...\n")
result_name = seqfile + '.pconsc.out'
l[0] = root + '/src/predict.py'
results = check_output(l)
f = open(result_name, 'w')
f.write(results)
f.close()
# Run PconsC2 otherwise (default)
else:
sys.stderr.write("Running PconsC2...\n")
netsurfpredictionname, sspredictionname, pssmaliname = prep.run_pconsc2_dependencies(
hhblitsdb, seqfile, n_cores=1)
sys.stderr.write("Predicting...\n")
result_name = seqfile + '.pconsc2.out'
l.extend([
netsurfpredictionname, sspredictionname, pssmaliname, result_name
])
check_output(l)
# Plot top L*1 contacts in a contact map (where L is the length of
# the input sequence).
# Those contacts are later used during protein folding
if plot_flag:
if os.path.exists('native.pdb') and os.path.exists(seqfile + '.horiz'):
plot_map(seqfile,
result_name,
1.,
pdb_filename='native.pdb',
psipred_filename=seqfile + '.horiz')
elif os.path.exists('native.pdb'):
plot_map(seqfile, result_name, 1., pdb_filename='native.pdb')
elif os.path.exists(seqfile + '.horiz'):
plot_map(seqfile,
result_name,
1.,
psipred_filename=seqfile + '.horiz')
else:
plot_map(seqfile, result_name, 1.)
def main(hhblitsdb, jackhmmerdb, seqfile, n_cores=1):
rundir = seqfile.rfind('/')
if rundir < 0:
rundir = '.'
else:
rundir = seqfile[:rundir]
if hhblitsdb.endswith('_a3m_db'):
hhblitsdb = hhblitsdb[:-7]
if not os.path.exists(hhblitsdb + '_a3m_db'):
sys.stderr.write('\n' + hhblitsdb + '_a3m_db' + 'does not exist\n')
sys.exit(1)
if not os.path.exists(jackhmmerdb):
sys.stderr.write('\n' + jackhmmerdb + 'does not exist\n')
sys.exit(1)
if not os.path.exists(seqfile):
sys.stderr.write('\n' + seqfile + 'does not exist\n')
sys.exit(0)
f = open(seqfile).read()
if os.path.exists(seqfile + '.fasta'):
subprocess.call(['mv', seqfile + '.fasta', seqfile +'.bak'])
f2 = open(seqfile +'.fasta', 'w')
if f[0] != '>':
f2.write('>target\n' + f +'\n')
else:
x = f.split('\n')
if len(x[0]) > 6:
target = x[0][1:5] + x[0][6]
f2.write('>target\n' + "".join(x[1:]) + '\n')
f2.close()
names = ['E4', 'E0', 'E10', 'E40']
cutoffs = ['1e-4', '1', '1e-10', '1e-40']
jhpredictionnames = []
hhpredictionnames = []
failed = []
for i in range(4):
exists_jh = os.path.exists(seqfile + '.jh' + names[i] + '.a3m')
exists_jh_psicov = os.path.exists(seqfile + '.jh' + names[i] + '.psicov')
exists_jh_plmdca = os.path.exists(seqfile + '.jh' + names[i] + '.plmdca')
exists_hh = os.path.exists(seqfile + '.hh' + names[i] + '.a3m')
exists_hh_psicov = os.path.exists(seqfile + '.hh' + names[i] + '.psicov')
exists_hh_plmdca = os.path.exists(seqfile + '.hh' + names[i] + '.plmdca')
# only create alignment file if at least one of the contact maps is missing
if not exists_jh and (not exists_jh_psicov or not exists_jh_plmdca):
sys.stderr.write(str(datetime.now()) + ' jackhmmer ' + names[i] + ': generating alignment\nThis may take quite a few minutes!\n ')
t = check_output([jackhmmer, '--cpu', str(n_cores), '-N', '5', '-E', cutoffs[i], '-A', seqfile +'.jh' + names[i] + '.ali', seqfile + '.fasta', jackhmmerdb])
check_output([reformat, 'sto', 'a3m', seqfile + '.jh' + names[i] + '.ali', seqfile + '.jh' + names[i] + '.a3m'])
check_output(['rm', seqfile + '.jh' + names[i] + '.ali'])
if not exists_jh_psicov:
#t = check_output([trim, seqfile + '.jh' + names[i] + '.fas'])
t = check_output([trim2jones, seqfile + '.jh' + names[i] + '.a3m'])
f = open(seqfile + '.jh' + names[i] + '.jones', 'w')
f.write(t)
f.close()
t = ''
sys.stderr.write(str(datetime.now()) + ' jackhmmer ' + names[i] + ': running PSICOV\nThis may take more than an hour.\n')
try:
# Joel @ NSC: Added -o flag, in case the psicov binary has not
# been compiled with MINEFSEQS=0.
t = check_output([psicov, '-o', seqfile + '.jh' + names[i] + '.jones'])
except:
t = ''
f = open(seqfile + '.jh' + names[i] + '.psicov', 'w')
f.write(t)
f.close()
jhpredictionnames.append(seqfile + '.jh' + names[i] + '.psicov')
if not exists_jh_plmdca:
t = check_output([trim2trimmed, seqfile + '.jh' + names[i] + '.a3m'])
f = open(seqfile + '.jh' + names[i] + '.trimmed', 'w')
f.write(t)
f.close()
sys.stderr.write(str(datetime.now()) + ' jackhmmer ' + names[i] + ': running plmDCA\nThis may take more than an hour.\n')
if plmdca:
#t = check_output([plmdca, matlabdir, seqfile + '.jh' + names[i] + ".trimmed", seqfile + '.jh' + names[i] + ".plmdca", "0.01", "0.01", "0.1", str(n_cores)])
t = check_output([plmdca, seqfile + '.jh' + names[i] + ".trimmed", seqfile + '.jh' + names[i] + ".plmdca", "0.01", "0.01", "0.1", str(n_cores)])
else:
t = check_output([matlab, '-nodesktop', '-nosplash', '-r', "path(path, '" + plmdcapath + "'); path(path, '" + plmdcapath + "/functions'); path(path, '" + plmdcapath + "/3rd_party_code/minFunc/'); plmDCA_symmetric ( '" + seqfile + '.jh' + names[i] + ".trimmed', '" + seqfile + '.jh' + names[i] + ".plmdca', 0.01, 0.01, 0.1, " + str(n_cores) + "); exit"])
jhpredictionnames.append(seqfile + '.jh' + names[i] + '.plmdca')
# only create alignment file if at least one of the contact maps is missing
if not exists_hh and (not exists_hh_psicov or not exists_hh_plmdca):
sys.stderr.write(str(datetime.now()) + ' HHblits' + names[i] + ': generating alignment\nThis may take quite a few minutes!\n ')
t = check_output([hhblits, '-all', '-oa3m', seqfile + '.hh' + names[i] + '.a3m', '-e', cutoffs[i], '-cpu', str(n_cores), '-i', seqfile + '.fasta', '-d', hhblitsdb])
#check_output([reformat, 'a3m', 'fas', seqfile + '.hh' + names[i] + '.a3m', seqfile + '.hh' + names[i] + '.fas'])
if not exists_hh_psicov:
#t = check_output([trim, seqfile + '.hh' + names[i] + '.fas'])
t = check_output([trim2jones, seqfile + '.hh' + names[i] + '.a3m'])
f = open(seqfile + '.hh' + names[i] + '.jones', 'w')
f.write(t)
f.close()
sys.stderr.write(str(datetime.now()) + ' HHblits ' + names[i] + ': running PSICOV\nThis may take more than an hour.\n')
t = ''
try:
# Joel @ NSC: Added -o flag, in case the psicov binary has not
# been compiled with MINEFSEQS=0.
t = check_output([psicov, '-o', seqfile + '.hh' + names[i] + '.jones'])
except:
t = ''
f = open(seqfile + '.hh' + names[i] + '.psicov', 'w')
f.write(t)
f.close()
hhpredictionnames.append(seqfile + '.hh' + names[i] + '.psicov')
if not exists_hh_plmdca:
#t = check_output([trim2, seqfile + '.hh' + names[i] + '.fas'])
t = check_output([trim2trimmed, seqfile + '.hh' + names[i] + '.a3m'])
f = open(seqfile + '.hh' + names[i] + '.trimmed', 'w')
f.write(t)
f.close()
sys.stderr.write(str(datetime.now()) + ' HHblits ' + names[i] + ': running plmDCA\nThis may take more than an hour.\n')
if plmdca:
#t = check_output([plmdca, matlabdir, seqfile + '.hh' + names[i] + ".trimmed", seqfile + '.hh' + names[i] + ".plmdca", "0.01", "0.01", "0.1", str(n_cores)])
t = check_output([plmdca, seqfile + '.hh' + names[i] + ".trimmed", seqfile + '.hh' + names[i] + ".plmdca", "0.01", "0.01", "0.1", str(n_cores)])
else:
t = check_output([matlab, '-nodesktop', '-nosplash', '-r', "path(path, '" + plmdcapath + "'); path(path, '" + plmdcapath + "/functions'); path(path, '" + plmdcapath + "/3rd_party_code/minFunc/'); plmDCA_symmetric ( '" + seqfile + '.hh' + names[i] + ".trimmed', '" + seqfile + '.hh' + names[i] + ".plmdca', 0.01, 0.01, 0.1, " + str(n_cores) + "); exit"])
hhpredictionnames.append(seqfile + '.hh' + names[i] + '.plmdca')
sys.stderr.write("Predicting...\n")
#l = [os.path.dirname(os.path.abspath(sys.argv[0])) + '/predict.py']
l = [root + '/src/predict.py']
l.extend(jhpredictionnames)
l.extend(hhpredictionnames)
results = check_output(l)
f = open(seqfile + '.pconsc.out', 'w')
f.write(results)
f.close()
# plot the top L*1 contacts in a contact map
# those contacts are later used during protein folding
if plot_flag:
if os.path.exists('native.pdb') and os.path.exists(seqfile + '.horiz'):
plot_map(seqfile, seqfile + '.pconsc.out', 1.0, pdb_filename='native.pdb', psipred_filename=seqfile + '.horiz')
elif os.path.exists('native.pdb'):
plot_map(seqfile, seqfile + '.pconsc.out', 1.0, pdb_filename='native.pdb')
elif os.path.exists(seqfile + '.horiz'):
plot_map(seqfile, seqfile + '.pconsc.out', 1.0, psipred_filename=seqfile + '.horiz')
else:
plot_map(seqfile, seqfile + '.pconsc.out', 1.0)
