def flip(blockades, model_file):
"""
Flips blockades
"""
blockade_model = load_model(model_file)
identifier = Identifier(blockade_model)
peptide = blockades[0].peptide
clusters = sp.preprocess_blockades(blockades, cluster_size=1,
min_dwell=0.0, max_dwell=1000)
print("Num\tFwd_dst\tRev_dst\t\tNeeds_flip", file=sys.stderr)
num_reversed = 0
new_blockades = []
for num, cluster in enumerate(clusters):
discr_signal = sp.discretize(cluster.consensus, len(peptide))
fwd_dist = identifier.signal_protein_distance(discr_signal, peptide)
rev_dist = identifier.signal_protein_distance(discr_signal,
peptide[::-1])
print("{0}\t{1:5.2f}\t{2:5.2f}\t\t{3}"
.format(num + 1, fwd_dist, rev_dist, fwd_dist > rev_dist),
file=sys.stderr)
new_blockades.append(cluster.blockades[0])
if fwd_dist > rev_dist:
new_blockades[-1].eventTrace = new_blockades[-1].eventTrace[::-1]
num_reversed += 1
print("Reversed:", num_reversed, "of", len(blockades), file=sys.stderr)
return new_blockades
def pvalues_test(blockades_file, cluster_size, blockade_model, db_file,
single_blockades, ostream):
"""
Performs protein identification and report results
"""
RANDOM_DB_SIZE = 10000
identifier = Identifier(blockade_model)
blockades = read_mat(blockades_file)
true_peptide = blockades[0].peptide
if db_file is None:
identifier.random_database(true_peptide, RANDOM_DB_SIZE)
target_id = "target"
db_len = RANDOM_DB_SIZE
else:
database, target_id = _make_database(db_file, true_peptide)
identifier.set_database(database)
db_len = len(database)
clusters = sp.preprocess_blockades(blockades,
cluster_size=cluster_size,
min_dwell=0.5,
max_dwell=20)
ostream.write("\nNo\tSize\tBest_id\t\tBest_dst\tTrg_dst\t\tTrg_rank\t"
"Trg_pval\n")
p_values = []
ranks = []
for num, cluster in enumerate(clusters):
db_ranking = identifier.rank_db_proteins(cluster.consensus)
target_rank = None
target_dist = None
for rank, (prot_id, prot_dist) in enumerate(db_ranking):
if prot_id == target_id:
target_rank = rank
target_dist = prot_dist
p_value = float(target_rank) / db_len
p_values.append(p_value)
ranks.append(target_rank)
ostream.write(
"{0}\t{1}\t{2:10}\t{3:5.2f}\t\t{4:5.2f}\t\t{5}\t\t{6:6.4}\n".
format(num + 1, len(cluster.blockades), db_ranking[0][0],
db_ranking[0][1], target_dist, target_rank + 1, p_value))
if single_blockades:
_detalize_cluster(identifier, cluster, db_ranking[0][0], target_id,
ostream)
ostream.write("\nMedian p-value: {0:7.4f}\n".format(np.median(p_values)))
ostream.write("Median target rank: {0:d}\n".format(int(np.median(ranks))))
return np.median(p_values), int(np.median(ranks))
def pvalues_test(blockades_file, cluster_size, blockade_model, db_file,
single_blockades, ostream):
"""
Performs protein identification and report results
"""
RANDOM_DB_SIZE = 10000
identifier = Identifier(blockade_model)
blockades = read_mat(blockades_file)
true_peptide = blockades[0].peptide
if db_file is None:
identifier.random_database(true_peptide, RANDOM_DB_SIZE)
target_id = "target"
db_len = RANDOM_DB_SIZE
else:
database, target_id = _make_database(db_file, true_peptide)
identifier.set_database(database)
db_len = len(database)
clusters = sp.preprocess_blockades(blockades, cluster_size=cluster_size,
min_dwell=0.5, max_dwell=20)
ostream.write("\nNo\tSize\tBest_id\t\tBest_dst\tTrg_dst\t\tTrg_rank\t"
"Trg_pval\n")
p_values = []
ranks = []
for num, cluster in enumerate(clusters):
db_ranking = identifier.rank_db_proteins(cluster.consensus)
target_rank = None
target_dist = None
for rank, (prot_id, prot_dist) in enumerate(db_ranking):
if prot_id == target_id:
target_rank = rank
target_dist = prot_dist
p_value = float(target_rank) / db_len
p_values.append(p_value)
ranks.append(target_rank)
ostream.write("{0}\t{1}\t{2:10}\t{3:5.2f}\t\t{4:5.2f}\t\t{5}\t\t{6:6.4}\n"
.format(num + 1, len(cluster.blockades), db_ranking[0][0],
db_ranking[0][1], target_dist, target_rank + 1, p_value))
if single_blockades:
_detalize_cluster(identifier, cluster, db_ranking[0][0],
target_id, ostream)
ostream.write("\nMedian p-value: {0:7.4f}\n".format(np.median(p_values)))
ostream.write("Median target rank: {0:d}\n".format(int(np.median(ranks))))
return np.median(p_values), int(np.median(ranks))
