def calc_statistics(df, exp, alldf_dict):
# Calculate statistics on df, saving to alldf_dict
# Deletion positions
# Denominator is ins
if sum(_lib.crispr_subset(df)['Count']) <= 1000:
return
ins_criteria = (df['Category'] == 'ins') & (df['Length'] == 1) & (
df['Indel with Mismatches'] != 'yes')
ins_count = sum(df[ins_criteria]['Count'])
del_criteria = (df['Category']
== 'del') & (df['Indel with Mismatches'] != 'yes')
del_count = sum(df[del_criteria]['Count'])
if del_count == 0:
return
alldf_dict['Ins1bp/Del Ratio'].append(ins_count / (del_count + ins_count))
mhdel_crit = (df['Category']
== 'del') & (df['Indel with Mismatches'] !=
'yes') & (df['Microhomology-Based'] == 'yes')
mhdel_count = sum(df[mhdel_crit]['Count'])
try:
alldf_dict['Ins1bp/MHDel Ratio'].append(ins_count /
(mhdel_count + ins_count))
except ZeroDivisionError:
alldf_dict['Ins1bp/MHDel Ratio'].append(0)
ins_ratio = ins_count / sum(_lib.crispr_subset(df)['Count'])
alldf_dict['Ins1bp Ratio'].append(ins_ratio)
seq, cutsite = _lib.get_sequence_cutsite(df)
alldf_dict['Sequence Context'].append(seq[-55:-30] + 'NNNN' + seq[-26:])
alldf_dict['Fourbp'].append(seq[cutsite - 2:cutsite + 2])
alldf_dict['Base1'].append(seq[cutsite - 2])
alldf_dict['Base2'].append(seq[cutsite - 1])
alldf_dict['Base3'].append(seq[cutsite])
alldf_dict['Base4'].append(seq[cutsite + 1])
_predict2.init_model()
del_score = _predict2.total_deletion_score(seq, cutsite)
alldf_dict['Del Score'].append(del_score)
dlpred = _predict2.deletion_length_distribution(seq, cutsite)
from scipy.stats import entropy
norm_entropy = entropy(dlpred) / np.log(len(dlpred))
alldf_dict['Entropy'].append(norm_entropy)
alldf_dict['_Experiment'].append(exp)
return alldf_dict
def calc_statistics(orig_df, exp, alldf_dict):
# Calculate statistics on df, saving to alldf_dict
# Deletion positions
df = _lib.mh_del_subset(orig_df)
df = _lib.indels_without_mismatches_subset(df)
if sum(df['Count']) <= 1000:
return
df['Frequency'] = _lib.normalize_frequency(df)
_predict2.init_model()
seq, cutsite = _lib.get_sequence_cutsite(df)
pred_df = _predict2.predict_mhdel(seq, cutsite)
join_cols = ['Category', 'Genotype Position', 'Length']
mdf = df.merge(pred_df, how = 'outer', on = join_cols)
mdf['Frequency'].fillna(value = 0, inplace = True)
mdf['Predicted_Frequency'].fillna(value = 0, inplace = True)
obs = mdf['Frequency']
pred = mdf['Predicted_Frequency']
obs_entropy = entropy(obs) / np.log(len(obs))
pred_entropy = entropy(pred) / np.log(len(pred))
alldf_dict['obs gt entropy'].append(obs_entropy)
alldf_dict['pred gt entropy'].append(pred_entropy)
df = orig_df[orig_df['Category'] == 'del']
df = df[df['Length'] <= 28]
df['Frequency'] = _lib.normalize_frequency(df)
obs_dl = []
for del_len in range(1, 28+1):
freq = sum(df[df['Length'] == del_len]['Frequency'])
obs_dl.append(freq)
pred_dl = _predict2.deletion_length_distribution(seq, cutsite)
obs_entropy = entropy(obs_dl) / np.log(len(obs_dl))
pred_entropy = entropy(pred_dl) / np.log(len(pred_dl))
alldf_dict['obs dl entropy'].append(obs_entropy)
alldf_dict['pred dl entropy'].append(pred_entropy)
alldf_dict['_Experiment'].append(exp)
return alldf_dict
def calc_statistics(orig_df, exp, rate_model, bp_model, alldf_dict, rs):
# Calculate statistics on df, saving to alldf_dict
# Deletion positions
df = _lib.mh_del_subset(orig_df)
df = _lib.indels_without_mismatches_subset(df)
if sum(df['Count']) <= 1000:
return
ins_crit = (orig_df['Category'] == 'ins') & (orig_df['Length'] == 1)
ins_df = orig_df[ins_crit]
truncated_ins_d = defaultdict(list)
for ins_base in list('ACGT'):
crit = (ins_df['Inserted Bases'] == ins_base)
tot_count = sum(ins_df[crit]['Count'])
truncated_ins_d['Count'].append(tot_count)
truncated_ins_d['Inserted Bases'].append(ins_base)
truncated_ins_d['Category'].append('ins')
truncated_ins_d['Length'].append(1)
ins_df = pd.DataFrame(truncated_ins_d)
df = df.append(ins_df, ignore_index=True)
df['Frequency'] = _lib.normalize_frequency(df)
_predict2.init_model()
seq, cutsite = _lib.get_sequence_cutsite(orig_df)
pred_df = _predict2.predict_mhdel(seq, cutsite)
# Predict rate of 1 bp insertions
# Featurize first
del_score = _predict2.total_deletion_score(seq, cutsite)
dlpred = _predict2.deletion_length_distribution(seq, cutsite)
norm_entropy = entropy(dlpred) / np.log(len(dlpred))
ohmapper = {
'A': [1, 0, 0, 0],
'C': [0, 1, 0, 0],
'G': [0, 0, 1, 0],
'T': [0, 0, 0, 1]
}
fivebase = seq[cutsite - 1]
onebp_features = ohmapper[fivebase] + [norm_entropy] + [del_score]
onebp_features = np.array(onebp_features).reshape(1, -1)
rate_1bpins = float(rate_model.predict(onebp_features))
# Predict 1 bp genotype frequencies
pred_1bpins_d = defaultdict(list)
for ins_base in bp_model[fivebase]:
freq = bp_model[fivebase][ins_base]
freq *= rate_1bpins / (1 - rate_1bpins)
pred_1bpins_d['Category'].append('ins')
pred_1bpins_d['Length'].append(1)
pred_1bpins_d['Inserted Bases'].append(ins_base)
pred_1bpins_d['Predicted_Frequency'].append(freq)
pred_1bpins_df = pd.DataFrame(pred_1bpins_d)
pred_df = pred_df.append(pred_1bpins_df, ignore_index=True)
pred_df['Predicted_Frequency'] /= sum(pred_df['Predicted_Frequency'])
join_cols = ['Category', 'Genotype Position', 'Length', 'Inserted Bases']
mdf = df.merge(pred_df, how='outer', on=join_cols)
mdf['Frequency'].fillna(value=0, inplace=True)
mdf['Predicted_Frequency'].fillna(value=0, inplace=True)
obs = mdf['Frequency']
pred = mdf['Predicted_Frequency']
r = pearsonr(obs, pred)[0]
alldf_dict['gt_r'].append(r)
obs_entropy = entropy(obs) / np.log(len(obs))
pred_entropy = entropy(pred) / np.log(len(pred))
alldf_dict['obs entropy'].append(obs_entropy)
alldf_dict['pred entropy'].append(pred_entropy)
alldf_dict['_Experiment'].append(exp)
alldf_dict['rs'].append(rs)
return alldf_dict
def calc_statistics(df, exp, alldf_dict):
# Calculate statistics on df, saving to alldf_dict
# Deletion positions
# Denominator is ins
if sum(_lib.crispr_subset(df)['Count']) <= 1000:
return
editing_rate = sum(_lib.crispr_subset(df)['Count']) / sum(
_lib.notnoise_subset(df)['Count'])
alldf_dict['Editing Rate'].append(editing_rate)
ins_criteria = (df['Category'] == 'ins') & (df['Length'] == 1) & (
df['Indel with Mismatches'] != 'yes')
ins_count = sum(df[ins_criteria]['Count'])
del_criteria = (df['Category']
== 'del') & (df['Indel with Mismatches'] != 'yes')
del_count = sum(df[del_criteria]['Count'])
if del_count == 0:
return
alldf_dict['Ins1bp/Del Ratio'].append(ins_count / (del_count + ins_count))
mhdel_crit = (df['Category']
== 'del') & (df['Indel with Mismatches'] !=
'yes') & (df['Microhomology-Based'] == 'yes')
mhdel_count = sum(df[mhdel_crit]['Count'])
try:
alldf_dict['Ins1bp/MHDel Ratio'].append(ins_count /
(mhdel_count + ins_count))
except ZeroDivisionError:
alldf_dict['Ins1bp/MHDel Ratio'].append(0)
ins_ratio = ins_count / sum(_lib.crispr_subset(df)['Count'])
alldf_dict['Ins1bp Ratio'].append(ins_ratio)
seq, cutsite = _lib.get_sequence_cutsite(df)
fivebase = seq[cutsite - 1]
alldf_dict['Fivebase'].append(fivebase)
_predict2.init_model()
del_score = _predict2.total_deletion_score(seq, cutsite)
alldf_dict['Del Score'].append(del_score)
dlpred = _predict2.deletion_length_distribution(seq, cutsite)
from scipy.stats import entropy
norm_entropy = entropy(dlpred) / np.log(len(dlpred))
alldf_dict['Entropy'].append(norm_entropy)
local_seq = seq[cutsite - 4:cutsite + 4]
gc = (local_seq.count('C') + local_seq.count('G')) / len(local_seq)
alldf_dict['GC'].append(gc)
if fivebase == 'A':
fivebase_oh = np.array([1, 0, 0, 0])
if fivebase == 'C':
fivebase_oh = np.array([0, 1, 0, 0])
if fivebase == 'G':
fivebase_oh = np.array([0, 0, 1, 0])
if fivebase == 'T':
fivebase_oh = np.array([0, 0, 0, 1])
alldf_dict['Fivebase_OH'].append(fivebase_oh)
threebase = seq[cutsite]
alldf_dict['Threebase'].append(threebase)
if threebase == 'A':
threebase_oh = np.array([1, 0, 0, 0])
if threebase == 'C':
threebase_oh = np.array([0, 1, 0, 0])
if threebase == 'G':
threebase_oh = np.array([0, 0, 1, 0])
if threebase == 'T':
threebase_oh = np.array([0, 0, 0, 1])
alldf_dict['Threebase_OH'].append(threebase_oh)
alldf_dict['_Experiment'].append(exp)
return alldf_dict
def calc_statistics(orig_df, exp, rate_model, bp_model, alldf_dict, rs):
# Calculate statistics on df, saving to alldf_dict
# Deletion positions
df = _lib.mh_del_subset(orig_df)
df = _lib.indels_without_mismatches_subset(df)
if sum(df['Count']) <= 1000:
return
df = orig_df
# Get observed frameshift rates
obs_fs = {'+0': 0, '+1': 0, '+2': 0}
all_ins_lens = set(df[df['Category'].isin(['ins',
'ins_notatcut'])]['Length'])
for ins_len in all_ins_lens:
crit = (df['Category'].isin(['ins', 'ins_notatcut'])) & (df['Length']
== ins_len)
fs = ins_len % 3
count = sum(df[crit]['Count'])
key = '+%s' % (int(fs))
obs_fs[key] += count
all_del_lens = set(df[df['Category'].isin(['del',
'del_notatcut'])]['Length'])
for del_len in all_del_lens:
crit = (df['Category'].isin(['del', 'del_notatcut'])) & (df['Length']
== del_len)
fs = (-1 * del_len) % 3
count = sum(df[crit]['Count'])
key = '+%s' % (int(fs))
obs_fs[key] += count
tot = sum(obs_fs.values())
for key in obs_fs:
obs_fs[key] /= tot
# Predict
_predict2.init_model()
seq, cutsite = _lib.get_sequence_cutsite(orig_df)
# Predict rate of 1 bp insertions
# Featurize first
del_score = _predict2.total_deletion_score(seq, cutsite)
dlpred = _predict2.deletion_length_distribution(seq, cutsite)
norm_entropy = entropy(dlpred) / np.log(len(dlpred))
ohmapper = {
'A': [1, 0, 0, 0],
'C': [0, 1, 0, 0],
'G': [0, 0, 1, 0],
'T': [0, 0, 0, 1]
}
fivebase = seq[cutsite - 1]
onebp_features = ohmapper[fivebase] + [norm_entropy] + [del_score]
onebp_features = np.array(onebp_features).reshape(1, -1)
rate_1bpins = float(rate_model.predict(onebp_features))
# Predict 1 bp frequency
freq = rate_1bpins / (1 - rate_1bpins)
pred = list(dlpred)
pred.insert(0, freq)
pred = np.array(pred) / sum(pred)
pred_fs = {'+0': 0, '+1': 0, '+2': 0}
pred_fs['+1'] += pred[0]
for idx in range(1, len(pred)):
del_len = idx
fs = (-1 * del_len) % 3
key = '+%s' % (int(fs))
pred_fs[key] += pred[idx]
# Bae predict
bae_fs = {'+0': 0, '+1': 0, '+2': 0}
bae_dlpred = bae_prediction(seq, cutsite)
for idx in range(len(bae_dlpred)):
del_len = idx + 1
fs = (-1 * del_len) % 3
key = '+%s' % (int(fs))
bae_fs[key] += bae_dlpred[idx]
for fs in ['+0', '+1', '+2']:
alldf_dict['Frame'].append(fs)
alldf_dict['Bae'].append(bae_fs[fs])
alldf_dict['inDelphi'].append(pred_fs[fs])
alldf_dict['Obs'].append(obs_fs[fs])
alldf_dict['_Experiment'].append(exp)
alldf_dict['rs'].append(rs)
return alldf_dict
def calc_statistics(orig_df, exp, rate_model, bp_model, alldf_dict, rs):
# Calculate statistics on df, saving to alldf_dict
# Deletion positions
df = _lib.mh_del_subset(orig_df)
df = _lib.indels_without_mismatches_subset(df)
if sum(df['Count']) <= 1000:
return
df = orig_df
obs_dellens = []
# 1 bp insertion
crit = (df['Category'] == 'ins') & (df['Indel with Mismatches'] !=
'yes') & (df['Length'] == 1)
obs_dellens.append(sum(df[crit]['Count']))
# 1 - 28 bp deletions
for del_len in range(1, 28 + 1):
crit = (df['Category'] == 'del') & (df['Indel with Mismatches'] !=
'yes') & (df['Length'] == del_len)
obs_dellens.append(sum(df[crit]['Count']))
obs_dellens = np.array(obs_dellens) / sum(obs_dellens)
obs = obs_dellens
# Predict
_predict2.init_model()
seq, cutsite = _lib.get_sequence_cutsite(orig_df)
# Predict rate of 1 bp insertions
# Featurize first
del_score = _predict2.total_deletion_score(seq, cutsite)
dlpred = _predict2.deletion_length_distribution(seq, cutsite)
norm_entropy = entropy(dlpred) / np.log(len(dlpred))
ohmapper = {
'A': [1, 0, 0, 0],
'C': [0, 1, 0, 0],
'G': [0, 0, 1, 0],
'T': [0, 0, 0, 1]
}
fivebase = seq[cutsite - 1]
onebp_features = ohmapper[fivebase] + [norm_entropy] + [del_score]
onebp_features = np.array(onebp_features).reshape(1, -1)
rate_1bpins = float(rate_model.predict(onebp_features))
# Predict 1 bp frequency
freq = rate_1bpins / (1 - rate_1bpins)
pred = list(dlpred)
pred.insert(0, freq)
pred = np.array(pred) / sum(pred)
r = pearsonr(obs, pred)[0]
alldf_dict['dl_r'].append(r)
obs_entropy = entropy(obs) / np.log(len(obs))
pred_entropy = entropy(pred) / np.log(len(pred))
alldf_dict['obs entropy'].append(obs_entropy)
alldf_dict['pred entropy'].append(pred_entropy)
alldf_dict['_Experiment'].append(exp)
alldf_dict['rs'].append(rs)
return alldf_dict