def evaluate_free_param(f_integration, f_bench, bin_size, median_y_n,
f_ens_red):
# define log
logger = logging.getLogger('evaluation free parameter')
logger.setLevel(logging.DEBUG)
ch = logging.StreamHandler()
ch.setLevel(logging.INFO)
formatter = logging.Formatter(
'%(asctime)s - %(name)s - %(levelname)s - %(message)s')
ch.setFormatter(formatter)
logger.addHandler(ch)
# determine the number of free parameter to evaluate
f = open(f_integration, 'r')
head = f.readline()
nb_d = len(head.rstrip().split('\t')) - 2
logger.debug("number of free parameter", nb_d)
# gene pair in bench
sem_sim = data.read_sem_sim(f_bench)
# comparison phenotypic bench vs integrated datase
for d in range(1, nb_d + 1):
# keep only gene pairs with phenotypic semantic similarity score
f_out1 = f_integration + "_d" + str(d)
data.report_pair_with_sem_sim(f_integration, f_out1, sem_sim, d)
# sort gene pair by value
f_out2 = f_integration + "_d" + str(d) + ".ord"
list_arg = "python $AP_PLN_HOME/src/scripts_python/sort_gene_pairs_by_value/sort_pair_value.py %s %s" % (
f_out1, f_out2)
proc = subprocess.Popen(list_arg, stdout=subprocess.PIPE, shell=True)
(out, err) = proc.communicate()
[logger.debug(val) for val in out.split("\n")]
if err is not None: logger.error(err)
os.system("rm %s" % f_out1)
# scale dataset
f_out3 = f_integration + "_d" + str(d) + ".ord.scale"
list_arg = "python $AP_PLN_HOME/src/scripts_python/scale_dataset/scale_dataset.py %s %s" % (
f_out2, f_out3)
proc = subprocess.Popen(list_arg, stdout=subprocess.PIPE, shell=True)
(out, err) = proc.communicate()
[logger.debug(val) for val in out.split("\n")]
if err is not None: logger.error(err)
os.system("rm %s" % f_out2)
# gene pair value vs benchmark
f_out4 = f_integration + "_d" + str(d) + "_bench"
list_arg = "python $AP_PLN_HOME/src/scripts_python/bench_versus_dataset/eval_with_scale.py %s %s %s %s %s %s" % (
f_ens_red, f_bench, f_out3, f_out4, bin_size, median_y_n)
proc = subprocess.Popen(list_arg, stdout=subprocess.PIPE, shell=True)
(out, err) = proc.communicate()
[logger.debug(val) for val in out.split("\n")]
if err is not None: logger.error(err)
def evaluate_free_param(f_integration,f_bench,bin_size,median_y_n,f_ens_red):
# define log
logger = logging.getLogger('evaluation free parameter')
logger.setLevel(logging.DEBUG)
ch = logging.StreamHandler()
ch.setLevel(logging.INFO)
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
ch.setFormatter(formatter)
logger.addHandler(ch)
# determine the number of free parameter to evaluate
f=open(f_integration,'r')
head=f.readline()
nb_d=len(head.rstrip().split('\t'))-2
logger.debug("number of free parameter",nb_d)
# gene pair in bench
sem_sim=data.read_sem_sim(f_bench)
# comparison phenotypic bench vs integrated datase
for d in range(1,nb_d+1):
# keep only gene pairs with phenotypic semantic similarity score
f_out1=f_integration+"_d"+str(d)
data.report_pair_with_sem_sim(f_integration,f_out1,sem_sim,d)
# sort gene pair by value
f_out2=f_integration+"_d"+str(d)+".ord"
list_arg="python $AP_PLN_HOME/src/scripts_python/sort_gene_pairs_by_value/sort_pair_value.py %s %s" % (f_out1,f_out2)
proc = subprocess.Popen(list_arg, stdout=subprocess.PIPE, shell=True)
(out, err) = proc.communicate()
[logger.debug(val) for val in out.split("\n")]
if err is not None: logger.error(err)
os.system("rm %s" % f_out1)
# scale dataset
f_out3=f_integration+"_d"+str(d)+".ord.scale"
list_arg="python $AP_PLN_HOME/src/scripts_python/scale_dataset/scale_dataset.py %s %s" % (f_out2,f_out3)
proc = subprocess.Popen(list_arg, stdout=subprocess.PIPE, shell=True)
(out, err) = proc.communicate()
[logger.debug(val) for val in out.split("\n")]
if err is not None: logger.error(err)
os.system("rm %s" % f_out2)
# gene pair value vs benchmark
f_out4=f_integration+"_d"+str(d)+"_bench"
list_arg="python $AP_PLN_HOME/src/scripts_python/bench_versus_dataset/eval_with_scale.py %s %s %s %s %s %s" % (f_ens_red,f_bench,f_out3,f_out4,bin_size,median_y_n)
proc = subprocess.Popen(list_arg, stdout=subprocess.PIPE, shell=True)
(out, err) = proc.communicate()
[logger.debug(val) for val in out.split("\n")]
if err is not None: logger.error(err)
f_ens_rd = sys.argv[1]
f_sem_sim = sys.argv[2]
f_dataset = sys.argv[3]
f_out = sys.argv[4]
bin_size = sys.argv[5]
bin_size = int(bin_size)
median_y_n = sys.argv[6]
# Redundant ens annotation
print "1) Read ens redundant annotations"
gene_convert = data.read_ens_rd(f_ens_rd)
print "Number of gene", len(gene_convert.keys())
# Read semantic similarity score (benchmark, y)
print "2) Read MGI similarity..."
sem_sim = data.read_sem_sim(f_sem_sim)
# Read genomic dataset (x) with gene pair benchmark value (y)
print "3) Read Score for pair with MGI Score..."
score_pair = data.read_score(f_dataset, gene_convert, sem_sim)
print "Number of gene pairs", len(score_pair.keys())
# sort pair according to score
print "4) Sort pair according Score..."
list_pair = functions.sort_by_val(score_pair)
print "Number of Pair", len(list_pair)
# Report dataset by bin versus semantic similarity
print "5) read and look at distribution mgi/hpo..."
data.report_data_ben_bin(f_out, score_pair, list_pair, sem_sim, bin_size,
median_y_n)
f_integration=sys.argv[1]
f_bench=sys.argv[2]
bin_size=sys.argv[3]
median_y_n=sys.argv[4]
# fixed parameters
f_ens_red="../../data/others/ensg_63symb_redundancy"
# determine the number of free parameter to evaluate
f=open(f_integration,'r')
head=f.readline()
nb_d=len(head.rstrip().split('\t'))-2
print "number of free parameter",nb_d
# gene pair in bench
sem_sim=data.read_sem_sim(f_bench)
# comparison phenotypic bench vs integrated datase
for d in range(1,nb_d+1):
# keep only gene pairs with phenotypic semantic similarity score
f_out1=f_integration+"_d"+str(d)
data.report_pair_with_sem_sim(f_integration,f_out1,sem_sim,d)
# sort gene pair by value
f_out2=f_integration+"_d"+str(d)+".ord"
os.system("python ../sort_gene_pairs_by_value/sort_pair_value.py %s %s" % (f_out1,f_out2))
os.system("rm %s" % f_out1)
# scale dataset
f_out3=f_integration+"_d"+str(d)+".ord.scale"
f_sem_sim=sys.argv[2] f_dataset=sys.argv[3] f_out=sys.argv[4] bin_size=sys.argv[5] bin_size=int(bin_size) median_y_n=sys.argv[6] # Redundant ens annotation print "1) Read ens redundant annotations" gene_convert=data.read_ens_rd(f_ens_rd) print "Number of gene",len(gene_convert.keys()) # Read semantic similarity score (benchmark, y) print "2) Read MGI similarity..." sem_sim=data.read_sem_sim(f_sem_sim) # Read genomic dataset (x) with gene pair benchmark value (y) print "3) Read Score for pair with MGI Score..." score_pair=data.read_score(f_dataset,gene_convert,sem_sim) print "Number of gene pairs",len(score_pair.keys()) # sort pair according to score print "4) Sort pair according Score..." list_pair=functions.sort_by_val(score_pair) print "Number of Pair",len(list_pair) # Report dataset by bin versus semantic similarity print "5) read and look at distribution mgi/hpo..." data.report_data_ben_bin(f_out,score_pair,list_pair,sem_sim,bin_size,median_y_n)
f_integration = sys.argv[1]
f_bench = sys.argv[2]
bin_size = sys.argv[3]
median_y_n = sys.argv[4]
# fixed parameters
f_ens_red = "../../data/others/ensg_63symb_redundancy"
# determine the number of free parameter to evaluate
f = open(f_integration, 'r')
head = f.readline()
nb_d = len(head.rstrip().split('\t')) - 2
print "number of free parameter", nb_d
# gene pair in bench
sem_sim = data.read_sem_sim(f_bench)
# comparison phenotypic bench vs integrated datase
for d in range(1, nb_d + 1):
# keep only gene pairs with phenotypic semantic similarity score
f_out1 = f_integration + "_d" + str(d)
data.report_pair_with_sem_sim(f_integration, f_out1, sem_sim, d)
# sort gene pair by value
f_out2 = f_integration + "_d" + str(d) + ".ord"
os.system("python ../sort_gene_pairs_by_value/sort_pair_value.py %s %s" %
(f_out1, f_out2))
os.system("rm %s" % f_out1)
# scale dataset