def setup(test=False, order=1, learn_options=None, data_file=None, pam_audit=True, length_audit=True):
num_proc = shared_setup(learn_options, order, test)
assert "testing_non_binary_target_name" in learn_options.keys(), "need this in order to get metrics, though used to be not needed, so you may newly see this error"
if learn_options["testing_non_binary_target_name"] not in ['ranks', 'raw', 'thrs']:
raise Exception('learn_otions["testing_non_binary_target_name"] must be in ["ranks", "raw", "thrs"]')
Xdf, Y, gene_position, target_genes = azimuth.load_data.from_file(data_file, learn_options)
learn_options['all_genes'] = target_genes
if test:
learn_options["order"] = 1
if 'convert_30mer_to_31mer' in learn_options and learn_options['convert_30mer_to_31mer'] is True:
print "WARNING!!! converting 30 mer to 31 mer (and then cutting off first nucleotide to go back to 30mer with a right shift)"
for i in range(Xdf.shape[0]):
Xdf['30mer'].iloc[i] = azimuth.util.convert_to_thirty_one(Xdf.iloc[i]["30mer"], Xdf.index.values[i][1], Xdf.iloc[i]["Strand"])
# to_keep = Xdf['30mer'].isnull() == False
# Xdf = Xdf[to_keep]
# gene_position = gene_position[to_keep]
# Y = Y[to_keep]
Xdf["30mer"] = Xdf["30mer"].apply(lambda x: x[1:]) # chop the first nucleotide
if learn_options.has_key('left_right_guide_ind') and learn_options['left_right_guide_ind'] is not None:
seq_start, seq_end, expected_length = learn_options['left_right_guide_ind']
Xdf['30mer'] = Xdf['30mer'].apply(lambda seq: seq[seq_start:seq_end])
feature_sets = feat.featurize_data(Xdf, learn_options, Y, gene_position, pam_audit=pam_audit, length_audit=length_audit)
np.random.seed(learn_options['seed'])
return Y, feature_sets, target_genes, learn_options, num_proc
def setup(test=False, order=1, learn_options=None, data_file=None, pam_audit=True):
num_proc = shared_setup(learn_options, order, test)
assert "testing_non_binary_target_name" in learn_options.keys(), "need this in order to get metrics, though used to be not needed, so you may newly see this error"
if learn_options["testing_non_binary_target_name"] not in ['ranks', 'raw', 'thrs']:
raise Exception('learn_otions["testing_non_binary_target_name"] must be in ["ranks", "raw", "thrs"]')
Xdf, Y, gene_position, target_genes = azimuth.load_data.from_file(data_file, learn_options)
learn_options['all_genes'] = target_genes
if test:
learn_options["order"] = 1
if 'convert_30mer_to_31mer' in learn_options and learn_options['convert_30mer_to_31mer'] is True:
print "WARNING!!! converting 30 mer to 31 mer (and then cutting off first nucleotide to go back to 30mer with a right shift)"
for i in range(Xdf.shape[0]):
Xdf['30mer'].iloc[i] = azimuth.util.convert_to_thirty_one(Xdf.iloc[i]["30mer"], Xdf.index.values[i][1], Xdf.iloc[i]["Strand"])
# to_keep = Xdf['30mer'].isnull() == False
# Xdf = Xdf[to_keep]
# gene_position = gene_position[to_keep]
# Y = Y[to_keep]
Xdf["30mer"] = Xdf["30mer"].apply(lambda x: x[1:]) # chop the first nucleotide
feature_sets = feat.featurize_data(Xdf, learn_options, Y, gene_position, pam_audit)
np.random.seed(learn_options['seed'])
return Y, feature_sets, target_genes, learn_options, num_proc
def predict(seq,
aa_cut=-1,
percent_peptide=-1,
model=None,
model_file=None,
pam_audit=True):
"""
if pam_audit==False, then it will not check for GG in the expected position
this is useful if predicting on PAM mismatches, such as with off-target
"""
# assert not (model is None and model_file is None), "you have to specify either a model or a model_file"
print aa_cut, percent_peptide
if model_file is None:
azimuth_saved_model_dir = os.path.join(
os.path.dirname(azimuth.__file__), 'saved_models')
if np.any(percent_peptide == -1) or (percent_peptide is None
and aa_cut is None):
print("No model file specified, using V3_model_nopos")
model_name = 'V3_model_nopos.pickle'
else:
print("No model file specified, using V3_model_full")
model_name = 'V3_model_full.pickle'
model_file = os.path.join(azimuth_saved_model_dir, model_name)
if model is None:
with open(model_file, 'rb') as f:
model, learn_options = pickle.load(f)
else:
model, learn_options = model
learn_options["V"] = 2
# Y, feature_sets, target_genes, learn_options, num_proc = setup(test=False, order=2, learn_options=learn_options, data_file=test_filename)
# inputs, dim, dimsum, feature_names = pd.concatenate_feature_sets(feature_sets)
Xdf = pandas.DataFrame(columns=[u'30mer', u'Strand'],
data=zip(seq, ['NA' for x in range(len(seq))]))
if np.all(percent_peptide != -1) and (percent_peptide is not None
and aa_cut is not None):
gene_position = pandas.DataFrame(
columns=[u'Percent Peptide', u'Amino Acid Cut position'],
data=zip(percent_peptide, aa_cut))
else:
gene_position = pandas.DataFrame(
columns=[u'Percent Peptide', u'Amino Acid Cut position'],
data=zip(np.ones(seq.shape[0]) * -1,
np.ones(seq.shape[0]) * -1))
feature_sets = feat.featurize_data(Xdf, learn_options, pandas.DataFrame(),
gene_position, pam_audit)
inputs, dim, dimsum, feature_names = azimuth.util.concatenate_feature_sets(
feature_sets)
# call to scikit-learn, returns a vector of predicted values
return model.predict(inputs)
def predict(seq, aa_cut=-1, percent_peptide=-1, model=None, model_file=None, pam_audit=True):
"""
if pam_audit==False, then it will not check for GG in the expected position
this is useful if predicting on PAM mismatches, such as with off-target
"""
# assert not (model is None and model_file is None), "you have to specify either a model or a model_file"
assert isinstance(seq, (str, np.ndarray)), "Please ensure seq is a numpy array"
if isinstance(seq, np.ndarray) and len(seq) > 0:
assert isinstance(seq[0], str) or isinstance(seq[0], unicode), "Please ensure input sequences are in string format, i.e. 'AGAG' rather than ['A' 'G' 'A' 'G'] or alternate representations"
assert isinstance(aa_cut, (int, long, np.ndarray)), "Please ensure aa_cut is a numpy array"
if isinstance(aa_cut, np.ndarray) and len(aa_cut) > 0:
assert isinstance(aa_cut[0], (int, long))
assert isinstance(percent_peptide, (int, long, np.ndarray)), "Please ensure percent_peptide is a numpy array"
if isinstance(percent_peptide, np.ndarray) and len(percent_peptide) > 0:
assert isinstance(percent_peptide[0], (int, long))
print aa_cut, percent_peptide
if model_file is None:
azimuth_saved_model_dir = os.path.join(os.path.dirname(azimuth.__file__), 'saved_models')
if np.any(percent_peptide == -1) or (percent_peptide is None and aa_cut is None):
print("No model file specified, using V3_model_nopos")
model_name = 'V3_model_nopos.pickle'
else:
print("No model file specified, using V3_model_full")
model_name = 'V3_model_full.pickle'
model_file = os.path.join(azimuth_saved_model_dir, model_name)
if model is None:
with open(model_file, 'rb') as f:
model, learn_options = pickle.load(f)
else:
model, learn_options = model
learn_options["V"] = 2
# Y, feature_sets, target_genes, learn_options, num_proc = setup(test=False, order=2, learn_options=learn_options, data_file=test_filename)
# inputs, dim, dimsum, feature_names = pd.concatenate_feature_sets(feature_sets)
Xdf = pandas.DataFrame(columns=[u'30mer', u'Strand'], data=zip(seq, ['NA' for x in range(len(seq))]))
if np.all(percent_peptide != -1) and (percent_peptide is not None and aa_cut is not None):
gene_position = pandas.DataFrame(columns=[u'Percent Peptide', u'Amino Acid Cut position'], data=zip(percent_peptide, aa_cut))
else:
gene_position = pandas.DataFrame(columns=[u'Percent Peptide', u'Amino Acid Cut position'], data=zip(np.ones(seq.shape[0])*-1, np.ones(seq.shape[0])*-1))
feature_sets = feat.featurize_data(Xdf, learn_options, pandas.DataFrame(), gene_position, pam_audit)
inputs, dim, dimsum, feature_names = azimuth.util.concatenate_feature_sets(feature_sets)
# call to scikit-learn, returns a vector of predicted values
return model.predict(inputs)
def extract_features(Xdf, Y, gene_position, conservation_scores, order=2):
learn_options = {
'nuc_features': True,
'num_proc': 1,
'order': order,
'gc_features': True,
'include_pi_nuc_feat': True,
"include_gene_position": True,
"include_NGGX_interaction": True,
"include_Tm": True,
'include_known_pairs': False,
'include_microhomology': False,
'ignore_gene_level_for_inner_loop': True, # <- what?
"include_strand": False,
"include_gene_feature": False,
"include_gene_guide_feature": 0,
"include_gene_effect": False,
"include_drug": False,
"include_sgRNAscore": False,
"normalize_features": False
}
features = featurize_data(Xdf,
learn_options,
Y,
gene_position,
pam_audit=True,
length_audit=True)
conservation_scores.index = features['_nuc_pd_Order1'].index
features['conservation_scores'] = conservation_scores
y = Y['score_drug_gene_rank'].astype('float32').as_matrix()
# we need the genes associated to the features to do cv data selection
genes = features['conservation_scores'].index.get_level_values(
1).to_series().reset_index(drop=True)
combined_features, dim, dimsum, feature_names = concatenate_feature_sets(
features)
combined_features = combined_features.astype('float32')
return combined_features, y, genes, feature_names
def predict(seq,
aa_cut=None,
percent_peptide=None,
model=None,
model_file=None,
pam_audit=True,
length_audit=False,
learn_options_override=None):
"""
Args:
seq: numpy array of 30 nt sequences.
aa_cut: numpy array of amino acid cut positions (optional).
percent_peptide: numpy array of percent peptide (optional).
model: model instance to use for prediction (optional).
model_file: file name of pickled model to use for prediction (optional).
pam_audit: check PAM of each sequence.
length_audit: check length of each sequence.
learn_options_override: a dictionary indicating which learn_options to override (optional).
Returns: a numpy array of predictions.
"""
# assert not (model is None and model_file is None), "you have to specify either a model or a model_file"
assert isinstance(seq, (np.ndarray)), "Please ensure seq is a numpy array"
assert len(seq[0]) > 0, "Make sure that seq is not empty"
assert isinstance(
seq[0], str
), "Please ensure input sequences are in string format, i.e. 'AGAG' rather than ['A' 'G' 'A' 'G'] or alternate representations"
if aa_cut is not None:
assert len(aa_cut) > 0, "Make sure that aa_cut is not empty"
assert isinstance(
aa_cut, (np.ndarray)), "Please ensure aa_cut is a numpy array"
assert np.all(np.isreal(
aa_cut)), "amino-acid cut position needs to be a real number"
if percent_peptide is not None:
assert len(
percent_peptide) > 0, "Make sure that percent_peptide is not empty"
assert isinstance(
percent_peptide,
(np.ndarray)), "Please ensure percent_peptide is a numpy array"
assert np.all(np.isreal(
percent_peptide)), "percent_peptide needs to be a real number"
if model_file is None:
if np.any(percent_peptide == -1) or (percent_peptide is None
and aa_cut is None):
print("No model file specified, using V3_model_nopos")
model_name = 'V3_model_nopos.pickle'
else:
print("No model file specified, using V3_model_full")
model_name = 'V3_model_full.pickle'
model_file = os.path.join('saved_models', model_name)
print(model_file)
with pkg_resources.resource_stream(__package__, model_file) as f:
model = pickle.load(f, encoding='bytes')
if model is None:
with open(model_file, 'rb') as f:
model, learn_options = pickle.load(f, encoding='bytes')
else:
model, learn_options = model
learn_options["V"] = 2
learn_options = override_learn_options(learn_options_override,
learn_options)
# Y, feature_sets, target_genes, learn_options, num_proc = setup(test=False, order=2, learn_options=learn_options, data_file=test_filename)
# inputs, dim, dimsum, feature_names = pd.concatenate_feature_sets(feature_sets)
Xdf = pandas.DataFrame(columns=['30mer', 'Strand'],
data=list(zip(seq,
['NA' for x in range(len(seq))])))
if np.all(percent_peptide != -1) and (percent_peptide is not None
and aa_cut is not None):
gene_position = pandas.DataFrame(
columns=['Percent Peptide', 'Amino Acid Cut position'],
data=list(zip(percent_peptide, aa_cut)))
else:
gene_position = pandas.DataFrame(
columns=['Percent Peptide', 'Amino Acid Cut position'],
data=list(
zip(np.ones(seq.shape[0]) * -1,
np.ones(seq.shape[0]) * -1)))
feature_sets = feat.featurize_data(Xdf,
learn_options,
pandas.DataFrame(),
gene_position,
pam_audit=pam_audit,
length_audit=length_audit)
inputs, dim, dimsum, feature_names = azimuth.util.concatenate_feature_sets(
feature_sets)
# print "CRISPR"
# pandas.DataFrame(inputs).to_csv("CRISPR.inputs.test.csv")
# import ipdb; ipdb.set_trace()
# call to scikit-learn, returns a vector of predicted values
preds = model.predict(inputs)
# also check that predictions are not 0/1 from a classifier.predict() (instead of predict_proba() or decision_function())
unique_preds = np.unique(preds)
ok = False
for pr in preds:
if pr not in [0, 1]:
ok = True
assert ok, "model returned only 0s and 1s"
return preds
def predict(seq,
aa_cut=-1,
percent_peptide=-1,
model=None,
model_file=None,
pam_audit=True,
length_audit=False,
learn_options_override=None):
"""
if pam_audit==False, then it will not check for GG in the expected position
this is useful if predicting on PAM mismatches, such as with off-target
"""
# assert not (model is None and model_file is None), "you have to specify either a model or a model_file"
assert isinstance(seq, (np.ndarray)), "Please ensure seq is a numpy array"
assert len(seq[0]) > 0, "Make sure that seq is not empty"
assert isinstance(
seq[0], str
), "Please ensure input sequences are in string format, i.e. 'AGAG' rather than ['A' 'G' 'A' 'G'] or alternate representations"
if aa_cut is not None:
assert len(aa_cut) > 0, "Make sure that aa_cut is not empty"
assert isinstance(
aa_cut, (np.ndarray)), "Please ensure aa_cut is a numpy array"
assert np.all(np.isreal(
aa_cut)), "amino-acid cut position needs to be a real number"
if percent_peptide is not None:
assert len(
percent_peptide) > 0, "Make sure that percent_peptide is not empty"
assert isinstance(
percent_peptide,
(np.ndarray)), "Please ensure percent_peptide is a numpy array"
assert np.all(np.isreal(
percent_peptide)), "percent_peptide needs to be a real number"
if model_file is None:
azimuth_saved_model_dir = os.path.join(
os.path.dirname(azimuth.__file__), 'saved_models')
if np.any(percent_peptide == -1) or (percent_peptide is None
and aa_cut is None):
print("No model file specified, using V3_model_nopos")
model_name = 'V3_model_nopos.pickle'
else:
print("No model file specified, using V3_model_full")
model_name = 'V3_model_full.pickle'
model_file = os.path.join(azimuth_saved_model_dir, model_name)
if model is None:
with open(model_file, 'rb') as f:
model, learn_options = pickle.load(f)
else:
model, learn_options = model
learn_options["V"] = 2
learn_options = override_learn_options(learn_options_override,
learn_options)
# Y, feature_sets, target_genes, learn_options, num_proc = setup(test=False, order=2, learn_options=learn_options, data_file=test_filename)
# inputs, dim, dimsum, feature_names = pd.concatenate_feature_sets(feature_sets)
Xdf = pandas.DataFrame(columns=[u'30mer', u'Strand'],
data=zip(seq, ['NA' for x in range(len(seq))]))
if np.all(percent_peptide != -1) and (percent_peptide is not None
and aa_cut is not None):
gene_position = pandas.DataFrame(
columns=[u'Percent Peptide', u'Amino Acid Cut position'],
data=zip(percent_peptide, aa_cut))
else:
gene_position = pandas.DataFrame(
columns=[u'Percent Peptide', u'Amino Acid Cut position'],
data=zip(np.ones(seq.shape[0]) * -1,
np.ones(seq.shape[0]) * -1))
feature_sets = feat.featurize_data(Xdf,
learn_options,
pandas.DataFrame(),
gene_position,
pam_audit=pam_audit,
length_audit=length_audit)
inputs, dim, dimsum, feature_names = azimuth.util.concatenate_feature_sets(
feature_sets)
# call to scikit-learn, returns a vector of predicted values
preds = model.predict(inputs)
# also check that predictions are not 0/1 from a classifier.predict() (instead of predict_proba() or decision_function())
unique_preds = np.unique(preds)
ok = False
for pr in preds:
if pr not in [0, 1]:
ok = True
assert ok, "model returned only 0s and 1s"
return preds
def predict(seq, aa_cut=-1, percent_peptide=-1, model=None, model_file=None, pam_audit=True, length_audit=False, learn_options_override=None):
"""
if pam_audit==False, then it will not check for GG in the expected position
this is useful if predicting on PAM mismatches, such as with off-target
"""
# assert not (model is None and model_file is None), "you have to specify either a model or a model_file"
assert isinstance(seq, (np.ndarray)), "Please ensure seq is a numpy array"
assert len(seq[0]) > 0, "Make sure that seq is not empty"
assert isinstance(seq[0], str), "Please ensure input sequences are in string format, i.e. 'AGAG' rather than ['A' 'G' 'A' 'G'] or alternate representations"
if aa_cut is not None:
assert len(aa_cut) > 0, "Make sure that aa_cut is not empty"
assert isinstance(aa_cut, (np.ndarray)), "Please ensure aa_cut is a numpy array"
assert np.all(np.isreal(aa_cut)), "amino-acid cut position needs to be a real number"
if percent_peptide is not None:
assert len(percent_peptide) > 0, "Make sure that percent_peptide is not empty"
assert isinstance(percent_peptide, (np.ndarray)), "Please ensure percent_peptide is a numpy array"
assert np.all(np.isreal(percent_peptide)), "percent_peptide needs to be a real number"
if model_file is None:
azimuth_saved_model_dir = os.path.join(os.path.dirname(azimuth.__file__), 'saved_models')
if np.any(percent_peptide == -1) or (percent_peptide is None and aa_cut is None):
print("No model file specified, using V3_model_nopos")
model_name = 'V3_model_nopos.pickle'
else:
print("No model file specified, using V3_model_full")
model_name = 'V3_model_full.pickle'
model_file = os.path.join(azimuth_saved_model_dir, model_name)
if model is None:
with open(model_file, 'rb') as f:
model, learn_options = pickle.load(f)
else:
model, learn_options = model
learn_options["V"] = 2
learn_options = override_learn_options(learn_options_override, learn_options)
# Y, feature_sets, target_genes, learn_options, num_proc = setup(test=False, order=2, learn_options=learn_options, data_file=test_filename)
# inputs, dim, dimsum, feature_names = pd.concatenate_feature_sets(feature_sets)
Xdf = pandas.DataFrame(columns=[u'30mer', u'Strand'], data=zip(seq, ['NA' for x in range(len(seq))]))
if np.all(percent_peptide != -1) and (percent_peptide is not None and aa_cut is not None):
gene_position = pandas.DataFrame(columns=[u'Percent Peptide', u'Amino Acid Cut position'], data=zip(percent_peptide, aa_cut))
else:
gene_position = pandas.DataFrame(columns=[u'Percent Peptide', u'Amino Acid Cut position'], data=zip(np.ones(seq.shape[0])*-1, np.ones(seq.shape[0])*-1))
feature_sets = feat.featurize_data(Xdf, learn_options, pandas.DataFrame(), gene_position, pam_audit=pam_audit, length_audit=length_audit)
inputs, dim, dimsum, feature_names = azimuth.util.concatenate_feature_sets(feature_sets)
# call to scikit-learn, returns a vector of predicted values
preds = model.predict(inputs)
# also check that predictions are not 0/1 from a classifier.predict() (instead of predict_proba() or decision_function())
unique_preds = np.unique(preds)
ok = False
for pr in preds:
if pr not in [0,1]:
ok = True
assert ok, "model returned only 0s and 1s"
return preds