def show_predcition_matrix(prediction: dict) -> None:
obs_pre = {0: {0: 0, 1: 2}, 1: {0: 3, 1: 1}}
binding_ = binding_data()
df = pd.DataFrame(data=0, index=list(binding_.values())[0].keys(), columns=binding_.keys())
for coh, doc_ in binding_.items():
for doc, obs in doc_.items():
df[coh][doc] = obs_pre[binding_[coh][doc]][prediction[coh][doc]]
plt.figure()
axis = plt.gca()
cmap = colors.ListedColormap(['white', 'cornflowerblue', 'red', 'darkorange'])
bounds = [-0.5, 0.5, 1.5, 2.5, 3.5]
norm = colors.BoundaryNorm(bounds, cmap.N)
heatmap = plt.pcolor(array(df), cmap=cmap, norm=norm, edgecolors='k', linewidth=2)
for y in range(array(df.shape)[0]):
for x in range(array(df.shape)[1]):
if array(df)[y, x] >= 0:
plt.text(x+0.5, y+0.5, array(df)[y, x], horizontalalignment='center', verticalalignment='center')
plt.yticks(arange(0.5, len(df.index), 1), df.index)
plt.xticks(arange(0.5, len(df.columns), 1), df.columns, rotation=70)
plt.xlabel('Cohesin name', style='oblique')
plt.ylabel('Dockerin name', style='oblique')
axis.set_aspect('equal')
plt.title('Cohesin dockerin cross binding')
plt.suptitle('0: obs no pred no, 1: obs yes, pred yes\n2: obs no pred yes, 3: obs yes pred no')
plt.show()
def postdictions_summary(args):
from matplotlib import colors
from _binding_data import binding_data
obs_pre = {False: {False: 0, True: 2}, True: {False: 3, True: 1}}
binding_dict = binding_data()
results_root = '/home/labs/fleishman/jonathaw/no_backup/postdiction_new/results/'
# run_filters = generate_run_filters(args={'ddg': -16, 'sasa': 1200, 'shape': 0.5, 'packstat': 0.5, 'buried_2': 30,
# 'hbonds': 12})
# args_ = {'ddg': 12, 'sasa': 1400, 'shape': 0.45, 'packstat': 0.45, 'buried_2': 2, 'hbonds': 4}
# args_ = {'ddg': 14, 'sasa': 1400, 'shape': 0.6, 'packstat': 0.6, 'buried_2': 2, 'hbonds': 4} # harsh
args_ = {'ddg': 16, 'sasa': 1400, 'shape': 0.6, 'packstat': 0.6, 'buried_2': 2, 'hbonds': 6}
run_filters = generate_run_filters(args_)
sc_files = [a for a in os.listdir(results_root) if '.score' in a]
results, cohs, docs = {}, [], []
bar_width = len(sc_files)
sys.stdout.write("{%s}" % (" " * bar_width))
sys.stdout.flush()
sys.stdout.write("\b" * (bar_width+1))
for sc_file in sc_files:
sc_dict = score2dict(results_root+sc_file)
passed, failed = all_who_pass_run_filters(args, sc_dict, run_filters)
coh_name = sc_file.split('all_')[1].split('_on_')[0]
doc_name = re.split(pattern='_[0-9]{1,2}\.[0-9]{1,2}\.score', string=sc_file.split('_on_')[1])[0]
# print(len(passed), coh_name, doc_name, binding_dict[coh_name][doc_name])
if coh_name not in results.keys():
results[coh_name] = {}
cohs.append(coh_name)
if doc_name not in docs:
docs.append(doc_name)
results[coh_name][doc_name] = obs_pre[binding_dict[coh_name][doc_name]][len(passed) >= 10]
sys.stdout.write("%s" % random.choice(['!', '@', '#', '$', '%', '^', '&', '*', '(', ')']))
sys.stdout.flush()
sys.stdout.write("}\n")
# move data to data frame
df = pd.DataFrame(columns=sorted(docs), index=sorted(cohs), data=-100)
for coh, doc_dict in results.items():
for doc, res in doc_dict.items():
df[doc][coh] = res
df = df.transpose()
print(df)
axis = plt.gca()
cmap = colors.ListedColormap(['white', 'red', 'blue', 'green', 'yellow'])
bounds = [-101, -0.5, 0.5, 1.5, 2.5, 3.5]
norm = colors.BoundaryNorm(bounds, cmap.N)
heatmap = plt.pcolor(np.array(df), cmap=cmap, norm=norm, edgecolors='k', linewidth=2)
plt.yticks(np.arange(0.5, len(df.index), 1), [official_names(a) for a in df.index])
plt.xticks(np.arange(0.5, len(df.columns), 1), [official_names(a) for a in df.columns], rotation=70)
axis.set_aspect('equal')
legend = plt.colorbar(heatmap)
# legend.ax.set_yticklabels(['NA', 'TN', 'TP', 'FP', 'FN'])
legend.ax.get_yaxis().set_ticks([])
for j, lab in enumerate(['NA', 'TN', 'TP', 'FP', 'FN']):
legend.ax.text(.5, (2 * j + 1) / 10.0, lab, ha='center', va='center')
legend.ax.get_yaxis().labelpad = 15
plt.suptitle(str(args_))
plt.show()
def parse_binding_data() -> pd.DataFrame:
"""
:return: data frame 'coh_name', 'doc_name', 'coh_seq', 'doc_seq', 'binders' for Rachel's data
"""
from _binding_data import binding_data
rachel_root = '/home/labs/fleishman/jonathaw/decision_tree/'
cohs = read_multi_fastas(rachel_root+'cohesins_from_rachel_and_vered.fasta_aln', suffix_to_remove='/', lower=True)
docs = read_multi_fastas(rachel_root+'dockerins_from_rachel_and_vered.fasta_aln', suffix_to_remove='/', lower=True)
rachel_bind = binding_data()
vered_bind = parse_vered_binding()
result = pd.DataFrame(columns=['coh_name', 'doc_name', 'coh_seq', 'doc_seq', 'binders'])
i = 1
for coh, docs_dict in rachel_bind.items():
for doc, res in docs_dict.items():
result.loc[i] = [coh, doc, cohs[coh], docs[doc], rachel_bind[coh][doc]]
i += 1
for coh, docs_dict in vered_bind.items():
for doc, res in docs_dict.items():
result.loc[i] = [coh, doc, cohs[coh], docs[doc], vered_bind[coh][doc] == 1]
i += 1
for name in ['1ohz', '2b59', '2ozn', '2vn5', '2y3n', '3ul4', '4fl4', '4fl5', '4dh2', '4uyp', '5new']:
result.loc[i] = [name, name, cohs[name], docs[name], True]
i += 1
print('there are %i rows in the data' % (i-1))
return result
def parse_binding_data() -> pd.DataFrame:
# cohs = read_multi_fastas(root_path+'cohs_specific_pos.fasta', suffix_to_remove='/')
# docs = read_multi_fastas(root_path+'docs_specific_pos.fasta', suffix_to_remove='/')
cohs, docs = retrive_relevant_poses()
data = binding_data()
colums = ['coh', 'doc'] + \
['core coh %i %s' % (i, aa) for i in [1, 2] for aa in aa2num.keys()] + \
['core doc %i %s' % (i, aa) for i in [1, 2, 3] for aa in aa2num.keys()] + \
['rim coh %i %s' % (i, t) for i in range(1, 19, 1) for t in types] + \
['rim doc %i %s' % (i, t) for i in range(1, 8, 1) for t in types] + ['binding']
df = pd.DataFrame(columns=colums)
i = 1
for coh, doc_dict in data.items():
coh_seq = cohs[coh].get_seq
for doc, res in doc_dict.items():
doc_seq = docs[doc].get_seq
df.loc[i] = [coh, doc] + seqs2row(coh_seq, doc_seq) + [1 if res else 0]
i += 1
return df
def parse_binding_data() -> pd.DataFrame:
# cohs = read_multi_fastas(root_path+'cohs_specific_pos.fasta', suffix_to_remove='/')
# docs = read_multi_fastas(root_path+'docs_specific_pos.fasta', suffix_to_remove='/')
cohs, docs = retrive_relevant_poses()
data = binding_data()
colums = ['coh', 'doc'] + \
['core coh %i %s' % (i, aa) for i in [1, 2] for aa in aa2num.keys()] + \
['core doc %i %s' % (i, aa) for i in [1, 2, 3] for aa in aa2num.keys()] + \
['rim coh %i %s' % (i, t) for i in range(1, 19, 1) for t in types] + \
['rim doc %i %s' % (i, t) for i in range(1, 8, 1) for t in types] + ['binding']
df = pd.DataFrame(columns=colums)
i = 1
for coh, doc_dict in data.items():
coh_seq = cohs[coh].get_seq
for doc, res in doc_dict.items():
doc_seq = docs[doc].get_seq
df.loc[i] = [coh, doc] + seqs2row(coh_seq,
doc_seq) + [1 if res else 0]
i += 1
return df
def show_predcition_matrix(prediction: dict) -> None:
obs_pre = {0: {0: 0, 1: 2}, 1: {0: 3, 1: 1}}
binding_ = binding_data()
df = pd.DataFrame(data=0,
index=list(binding_.values())[0].keys(),
columns=binding_.keys())
for coh, doc_ in binding_.items():
for doc, obs in doc_.items():
df[coh][doc] = obs_pre[binding_[coh][doc]][prediction[coh][doc]]
plt.figure()
axis = plt.gca()
cmap = colors.ListedColormap(
['white', 'cornflowerblue', 'red', 'darkorange'])
bounds = [-0.5, 0.5, 1.5, 2.5, 3.5]
norm = colors.BoundaryNorm(bounds, cmap.N)
heatmap = plt.pcolor(array(df),
cmap=cmap,
norm=norm,
edgecolors='k',
linewidth=2)
for y in range(array(df.shape)[0]):
for x in range(array(df.shape)[1]):
if array(df)[y, x] >= 0:
plt.text(x + 0.5,
y + 0.5,
array(df)[y, x],
horizontalalignment='center',
verticalalignment='center')
plt.yticks(arange(0.5, len(df.index), 1), df.index)
plt.xticks(arange(0.5, len(df.columns), 1), df.columns, rotation=70)
plt.xlabel('Cohesin name', style='oblique')
plt.ylabel('Dockerin name', style='oblique')
axis.set_aspect('equal')
plt.title('Cohesin dockerin cross binding')
plt.suptitle(
'0: obs no pred no, 1: obs yes, pred yes\n2: obs no pred yes, 3: obs yes pred no'
)
plt.show()
