def run_permutations(self, temp_uuid):
"""Runs permutations and gets significant and suggestive LOD scores"""
top_lod_scores = []
#logger.debug("self.num_perm:", self.num_perm)
for permutation in range(self.num_perm):
pheno_vector = np.array([val == "x" and np.nan or float(val) for val in self.vals])
np.random.shuffle(pheno_vector)
key = "pylmm:input:" + temp_uuid
if self.dataset.group.species == "human":
p_values, t_stats = self.gen_human_results(pheno_vector, key, temp_uuid)
else:
genotype_data = [marker['genotypes'] for marker in self.dataset.group.markers.markers]
no_val_samples = self.identify_empty_samples()
trimmed_genotype_data = self.trim_genotypes(genotype_data, no_val_samples)
genotype_matrix = np.array(trimmed_genotype_data).T
params = dict(pheno_vector = pheno_vector.tolist(),
genotype_matrix = genotype_matrix.tolist(),
restricted_max_likelihood = True,
refit = False,
temp_uuid = temp_uuid,
# meta data
timestamp = datetime.datetime.now().isoformat(),
)
json_params = json.dumps(params)
Redis.set(key, json_params)
Redis.expire(key, 60*60)
command = PYLMM_COMMAND+' --key {} --species {}'.format(key,"other")
shell(command)
json_results = Redis.blpop("pylmm:results:" + temp_uuid, 45*60)
results = json.loads(json_results[1])
p_values = [float(result) for result in results['p_values']]
lowest_p_value = 1
for p_value in p_values:
if p_value < lowest_p_value:
lowest_p_value = p_value
#logger.debug("lowest_p_value:", lowest_p_value)
top_lod_scores.append(-math.log10(lowest_p_value))
#logger.debug("top_lod_scores:", top_lod_scores)
self.suggestive = np.percentile(top_lod_scores, 67)
self.significant = np.percentile(top_lod_scores, 95)
def run_permutations(self, temp_uuid):
"""Runs permutations and gets significant and suggestive LOD scores"""
top_lod_scores = []
#logger.debug("self.num_perm:", self.num_perm)
for permutation in range(self.num_perm):
pheno_vector = np.array([val == "x" and np.nan or float(val) for val in self.vals])
np.random.shuffle(pheno_vector)
key = "pylmm:input:" + temp_uuid
if self.dataset.group.species == "human":
p_values, t_stats = self.gen_human_results(pheno_vector, key, temp_uuid)
else:
genotype_data = [marker['genotypes'] for marker in self.dataset.group.markers.markers]
no_val_samples = self.identify_empty_samples()
trimmed_genotype_data = self.trim_genotypes(genotype_data, no_val_samples)
genotype_matrix = np.array(trimmed_genotype_data).T
params = dict(pheno_vector = pheno_vector.tolist(),
genotype_matrix = genotype_matrix.tolist(),
restricted_max_likelihood = True,
refit = False,
temp_uuid = temp_uuid,
# meta data
timestamp = datetime.datetime.now().isoformat(),
)
json_params = json.dumps(params)
Redis.set(key, json_params)
Redis.expire(key, 60*60)
command = PYLMM_COMMAND+' --key {} --species {}'.format(key,"other")
shell(command)
json_results = Redis.blpop("pylmm:results:" + temp_uuid, 45*60)
results = json.loads(json_results[1])
p_values = [float(result) for result in results['p_values']]
lowest_p_value = 1
for p_value in p_values:
if p_value < lowest_p_value:
lowest_p_value = p_value
#logger.debug("lowest_p_value:", lowest_p_value)
top_lod_scores.append(-math.log10(lowest_p_value))
#logger.debug("top_lod_scores:", top_lod_scores)
self.suggestive = np.percentile(top_lod_scores, 67)
self.significant = np.percentile(top_lod_scores, 95)
def gen_data(self, temp_uuid):
"""Generates p-values for each marker"""
logger.debug("self.vals is:", self.vals)
pheno_vector = np.array([(val == "x" or val == "") and np.nan or float(val) for val in self.vals])
#lmm_uuid = str(uuid.uuid4())
key = "pylmm:input:" + temp_uuid
logger.debug("key is:", pf(key))
#with Bench("Loading cache"):
# result = Redis.get(key)
if self.dataset.group.species == "human":
p_values, t_stats = self.gen_human_results(pheno_vector, key, temp_uuid)
#p_values = self.trim_results(p_values)
else:
logger.debug("NOW CWD IS:", os.getcwd())
genotype_data = [marker['genotypes'] for marker in self.dataset.group.markers.markers]
no_val_samples = self.identify_empty_samples()
trimmed_genotype_data = self.trim_genotypes(genotype_data, no_val_samples)
genotype_matrix = np.array(genotype_data).T
#logger.debug("pheno_vector: ", pf(pheno_vector))
#logger.debug("genotype_matrix: ", pf(genotype_matrix))
#logger.debug("genotype_matrix.shape: ", pf(genotype_matrix.shape))
#params = {"pheno_vector": pheno_vector,
# "genotype_matrix": genotype_matrix,
# "restricted_max_likelihood": True,
# "refit": False,
# "temp_data": tempdata}
# logger.debug("genotype_matrix:", str(genotype_matrix.tolist()))
# logger.debug("pheno_vector:", str(pheno_vector.tolist()))
params = dict(pheno_vector = pheno_vector.tolist(),
genotype_matrix = genotype_matrix.tolist(),
restricted_max_likelihood = True,
refit = False,
temp_uuid = temp_uuid,
# meta data
timestamp = datetime.datetime.now().isoformat(),
)
json_params = json.dumps(params)
#logger.debug("json_params:", json_params)
Redis.set(key, json_params)
Redis.expire(key, 60*60)
logger.debug("before printing command")
command = PYLMM_COMMAND + ' --key {} --species {}'.format(key, "other")
logger.debug("command is:", command)
logger.debug("after printing command")
shell(command)
#t_stats, p_values = lmm.run(key)
#lmm.run(key)
json_results = Redis.blpop("pylmm:results:" + temp_uuid, 45*60)
results = json.loads(json_results[1])
p_values = [float(result) for result in results['p_values']]
#logger.debug("p_values:", p_values[:10])
#p_values = self.trim_results(p_values)
t_stats = results['t_stats']
#t_stats, p_values = lmm.run(
# pheno_vector,
# genotype_matrix,
# restricted_max_likelihood=True,
# refit=False,
# temp_data=tempdata
#)
#logger.debug("p_values:", p_values)
self.dataset.group.markers.add_pvalues(p_values)
#self.get_lod_score_cutoff()
return self.dataset.group.markers.markers
def gen_data(self, temp_uuid):
"""Generates p-values for each marker"""
logger.debug("self.vals is:", self.vals)
pheno_vector = np.array([(val == "x" or val == "") and np.nan
or float(val) for val in self.vals])
#lmm_uuid = str(uuid.uuid4())
key = "pylmm:input:" + temp_uuid
logger.debug("key is:", pf(key))
#with Bench("Loading cache"):
# result = Redis.get(key)
if self.dataset.group.species == "human":
p_values, t_stats = self.gen_human_results(pheno_vector, key,
temp_uuid)
#p_values = self.trim_results(p_values)
else:
logger.debug("NOW CWD IS:", os.getcwd())
genotype_data = [
marker['genotypes']
for marker in self.dataset.group.markers.markers
]
no_val_samples = self.identify_empty_samples()
trimmed_genotype_data = self.trim_genotypes(
genotype_data, no_val_samples)
genotype_matrix = np.array(genotype_data).T
#logger.debug("pheno_vector: ", pf(pheno_vector))
#logger.debug("genotype_matrix: ", pf(genotype_matrix))
#logger.debug("genotype_matrix.shape: ", pf(genotype_matrix.shape))
#params = {"pheno_vector": pheno_vector,
# "genotype_matrix": genotype_matrix,
# "restricted_max_likelihood": True,
# "refit": False,
# "temp_data": tempdata}
# logger.debug("genotype_matrix:", str(genotype_matrix.tolist()))
# logger.debug("pheno_vector:", str(pheno_vector.tolist()))
params = dict(
pheno_vector=pheno_vector.tolist(),
genotype_matrix=genotype_matrix.tolist(),
restricted_max_likelihood=True,
refit=False,
temp_uuid=temp_uuid,
# meta data
timestamp=datetime.datetime.now().isoformat(),
)
json_params = json.dumps(params)
#logger.debug("json_params:", json_params)
Redis.set(key, json_params)
Redis.expire(key, 60 * 60)
logger.debug("before printing command")
command = PYLMM_COMMAND + ' --key {} --species {}'.format(
key, "other")
logger.debug("command is:", command)
logger.debug("after printing command")
shell(command)
#t_stats, p_values = lmm.run(key)
#lmm.run(key)
json_results = Redis.blpop("pylmm:results:" + temp_uuid, 45 * 60)
results = json.loads(json_results[1])
p_values = [float(result) for result in results['p_values']]
#logger.debug("p_values:", p_values[:10])
#p_values = self.trim_results(p_values)
t_stats = results['t_stats']
#t_stats, p_values = lmm.run(
# pheno_vector,
# genotype_matrix,
# restricted_max_likelihood=True,
# refit=False,
# temp_data=tempdata
#)
#logger.debug("p_values:", p_values)
self.dataset.group.markers.add_pvalues(p_values)
#self.get_lod_score_cutoff()
return self.dataset.group.markers.markers