def test_instance_from_file(filename, paramsconfig):
'''
THis function parses an instance from a file and then builds the model
and predicts the alignment for
one instance.
AN instance here means 1 ref sequence with 1 or more query
sequences
'''
instance = helpers.parse_instance(filename)
modelname = filename[0:len(filename)-4]
t1 = time.time()
test_instance(modelname, instance, paramsconfig)
t2 =time.time()
print("total processing time: " + str(t2-t1) + " seconds.")
def test_instance_from_file(filename, paramsconfig):
'''
THis function parses an instance from a file and then builds the model
and predicts the alignment for
one instance.
AN instance here means 1 ref sequence with 1 or more query
sequences
'''
instance = helpers.parse_instance(filename)
modelname = filename[0:len(filename) - 4]
t1 = time.time()
test_instance(modelname, instance, paramsconfig)
t2 = time.time()
print("total processing time: " + str(t2 - t1) + " seconds.")
def test_batch(folder_path, output_name, paramstuple):
if folder_path[-1] != "/":
folder_path += "/"
stats = [0] * 11
score_ratios = []
f = file(output_name, 'a')
f.write("--------------------------------")
f.write("\n\nCurrently processing folder " + folder_path + "\n\n")
file_names = os.listdir(folder_path)
for filename in file_names:
instance = helpers.parse_instance(folder_path + filename)
print("Currently testing: "+ filename)
print("length of reference sequence: " + str(len(instance[2])))
print("number of query sequences: " + str(len(instance[3])))
#modelname = filename[0:len(filename)-4]
modelname = filename
results, m = test_instance(modelname, instance, paramstuple)
f.write("\n\n\n" + filename + ":\n")
match_results = []
for query_index, result in enumerate(results):
frame_shifts = locate_frameshifts(instance[3][query_index],
result[0])
if frame_shifts == []:
stats[0] += 1
elif len(frame_shifts) == 1:
stats[1] += 1
elif len(frame_shifts) == 2:
stats[2] += 1
elif len(frame_shifts) == 3:
stats[3] += 1
else: #4 or more frame-shifts
stats[4] += 1
'''
print("\nVITERBI PATH: "+ str(result[0]))
print("\nPREDICTED ALIGNMENT:")
print("\n" + result[1][0])
print("\n" + result[1][1])
print("\nORIGINAL ALIGNMENT:")
print("\n" + instance[0])
print("\n" + instance[1][query_index])
'''
alignments_match = _alignments_match((result[1][0], result[1][1]),
(instance[0], instance[1][query_index]))
match_results.append(alignments_match)
if alignments_match:
stats[6] += 1
else:
stats[7] += 1
#print("in analytics: result[1] " + str(result[1]))
path_decoded_from_original = helpers.pair_to_path(m, result[1])
score_original_alignment = 0
#( model.score_alignment_against_model(m,(instance[0],
# instance[1][query_index]))
if score_original_alignment == 0:
score_ratio = 0
else:
score_ratio = result[2]/score_original_alignment
score_ratios.append(score_ratio)
acc_site_index_original = helpers.get_exon_start(instance[1][query_index])
acc_site_index_predicted = helpers.get_exon_start(result[1][1])
do_site_index_original = helpers.get_exon_end(instance[1][query_index])
do_site_index_predicted = helpers.get_exon_end(result[1][1])
acc_site_correct = (acc_site_index_original ==
acc_site_index_predicted)
do_site_correct = (do_site_index_original ==
do_site_index_predicted)
if acc_site_correct:
stats[9] += 1
if do_site_correct:
stats[10] += 1
_write_query_details(query_index, f, instance, result,
frame_shifts, alignments_match,
score_original_alignment, score_ratio,
path_decoded_from_original,
acc_site_correct,
do_site_correct)
f.write("\n\nmixed results: ")
if sum(match_results) == 0 or sum(match_results) == len(match_results):
f.write("False")
else:
f.write("True")
stats[8] += 1
f.flush()
aggregated_ratios = helpers.aggregate_ratios(score_ratios)
return stats, aggregated_ratios, score_ratios
def test_batch(folder_path, output_name, paramstuple):
if folder_path[-1] != "/":
folder_path += "/"
stats = [0] * 11
score_ratios = []
f = file(output_name, 'a')
f.write("--------------------------------")
f.write("\n\nCurrently processing folder " + folder_path + "\n\n")
file_names = os.listdir(folder_path)
for filename in file_names:
instance = helpers.parse_instance(folder_path + filename)
print("Currently testing: " + filename)
print("length of reference sequence: " + str(len(instance[2])))
print("number of query sequences: " + str(len(instance[3])))
#modelname = filename[0:len(filename)-4]
modelname = filename
results, m = test_instance(modelname, instance, paramstuple)
f.write("\n\n\n" + filename + ":\n")
match_results = []
for query_index, result in enumerate(results):
frame_shifts = locate_frameshifts(instance[3][query_index],
result[0])
if frame_shifts == []:
stats[0] += 1
elif len(frame_shifts) == 1:
stats[1] += 1
elif len(frame_shifts) == 2:
stats[2] += 1
elif len(frame_shifts) == 3:
stats[3] += 1
else: #4 or more frame-shifts
stats[4] += 1
'''
print("\nVITERBI PATH: "+ str(result[0]))
print("\nPREDICTED ALIGNMENT:")
print("\n" + result[1][0])
print("\n" + result[1][1])
print("\nORIGINAL ALIGNMENT:")
print("\n" + instance[0])
print("\n" + instance[1][query_index])
'''
alignments_match = _alignments_match(
(result[1][0], result[1][1]),
(instance[0], instance[1][query_index]))
match_results.append(alignments_match)
if alignments_match:
stats[6] += 1
else:
stats[7] += 1
#print("in analytics: result[1] " + str(result[1]))
path_decoded_from_original = helpers.pair_to_path(m, result[1])
score_original_alignment = 0
#( model.score_alignment_against_model(m,(instance[0],
# instance[1][query_index]))
if score_original_alignment == 0:
score_ratio = 0
else:
score_ratio = result[2] / score_original_alignment
score_ratios.append(score_ratio)
acc_site_index_original = helpers.get_exon_start(
instance[1][query_index])
acc_site_index_predicted = helpers.get_exon_start(result[1][1])
do_site_index_original = helpers.get_exon_end(
instance[1][query_index])
do_site_index_predicted = helpers.get_exon_end(result[1][1])
acc_site_correct = (
acc_site_index_original == acc_site_index_predicted)
do_site_correct = (
do_site_index_original == do_site_index_predicted)
if acc_site_correct:
stats[9] += 1
if do_site_correct:
stats[10] += 1
_write_query_details(query_index, f, instance, result,
frame_shifts, alignments_match,
score_original_alignment, score_ratio,
path_decoded_from_original, acc_site_correct,
do_site_correct)
f.write("\n\nmixed results: ")
if sum(match_results) == 0 or sum(match_results) == len(match_results):
f.write("False")
else:
f.write("True")
stats[8] += 1
f.flush()
aggregated_ratios = helpers.aggregate_ratios(score_ratios)
return stats, aggregated_ratios, score_ratios
". FILE: " + path_prefix + params.BLOSUM_FILE)
else:
print("substitution matrix = " + str(args.matrix) +
". FILE: " + path_prefix + params.ETH_FILE)
if args.is_first_exon:
print(path_prefix + params.FIRST_CODON_PROFILE + " used instead of the usual acceptor profile")
if args.is_last_exon:
print(path_prefix + params.LAST_CODON_PROFILE + " used instead of donor profile")
if args.has_AC_acceptor:
print(path_prefix + params.U12_ACC_PROFILE + " used instead of the usual acceptor profile")
if args.has_AT_donor:
print(path_prefix + params.U12_DONOR_PROFILE + " used instead of the usual donor profile")
#parsing the FASTA input file:
instance = helpers.parse_instance(args.input_file)
params.CLADE = args.clade
paramsconfig = params.Params_config()
paramsconfig.set_params((args.fs_prob, args.ci_prob, args.ci2_prob, args.total_cd_prob, args.matrix.upper()))
paramsconfig.set_sequence(instance[2], instance[4], instance[5],
args.is_first_exon, args.is_last_exon,
args.has_AC_acceptor, args.has_AT_donor)
if args.verbosity> 0:
print("\nReference sequene (raw): " + instance[0])
print("\nCodons in the reference sequence: " + str(paramsconfig.seq_as_codons))
print("\nUpstream intron phase: " + str(instance[4]))
print("\nDownstream intron phase: " + str(instance[5]))
print("\nBuilding a profile HMM for the reference sequence...")
m = model.HMM(args.input_file, paramsconfig)
path_prefix + params.ETH_FILE)
if args.is_first_exon:
print(path_prefix + params.FIRST_CODON_PROFILE +
" used instead of the usual acceptor profile")
if args.is_last_exon:
print(path_prefix + params.LAST_CODON_PROFILE +
" used instead of donor profile")
if args.has_AC_acceptor:
print(path_prefix + params.U12_ACC_PROFILE +
" used instead of the usual acceptor profile")
if args.has_AT_donor:
print(path_prefix + params.U12_DONOR_PROFILE +
" used instead of the usual donor profile")
#parsing the FASTA input file:
instance = helpers.parse_instance(args.input_file)
params.CLADE = args.clade
paramsconfig = params.Params_config()
paramsconfig.set_params(
(args.fs_prob, args.ci_prob, args.ci2_prob, args.total_cd_prob,
args.matrix.upper()))
paramsconfig.set_sequence(instance[2], instance[4], instance[5],
args.is_first_exon, args.is_last_exon,
args.has_AC_acceptor, args.has_AT_donor)
if args.verbosity > 0:
print("\nReference sequene (raw): " + instance[0])
print("\nCodons in the reference sequence: " +
str(paramsconfig.seq_as_codons))
print("\nUpstream intron phase: " + str(instance[4]))
print("\nDownstream intron phase: " + str(instance[5]))
