plt.savefig("heatmap_3+_power_0.05+.svg")
plt.savefig("heatmap_3+_power_0.05+.eps")
plt.close()
"""
for sample_set_name in sample_set_names_list:
print("Handling %s" % sample_set_name)
os.chdir(workdir)
#os.system("mkdir -p %s" % sample_set_name)
os.chdir(sample_set_name)
#os.system("mkdir -p %s %s" % (clustering_dir, rainfall_dir))
#os.system("pwd")
clusters_dict[sample_set_name] = CollectionCCF(
from_file=True,
input_file="%s%s_adjusted_size_3+_power_0.1+.ccf" %
(clustering_dir, sample_set_name))
data[sample_set_name] = clusters_dict[sample_set_name].get_data_for_stat(additional_data=("Median", "Power")) \
if "HAP" in sample_set_name \
else clusters_dict[sample_set_name].get_data_for_stat(additional_data=("Median", "Power", "Homogeneity"))
data[sample_set_name] = [
data[sample_set_name][:, i]
for i in range(0, 4 if "HAP" in sample_set_name else 5)
]
os.chdir(workdir)
os.system("mkdir -p %s" % heatmap_dir)
os.chdir(heatmap_dir)
"_adjusted_size_3+_power_0.05+.ccf",
"_adjusted_size_3+_power_0.1+.ccf"]
homogeneity_dir = "homogeneity/"
y_name = "N of clusters"
os.system("mkdir -p %s" % homogeneity_dir)
data = {}
power_list = ["All", "Power >= 0.05", "Power >= 0.1"]
for sample in sample_set_names_list:
os.chdir(workdir + sample + "/" + clustering_dir)
data[sample] = dict([(key, 0) for key in power_list])
for suffix, name in zip(suffix_list, power_list):
print(sample + suffix)
clusters = CollectionCCF(from_file=True, input_file=sample + suffix)
data[sample][name] = clusters.get_data_for_stat(additional_data=["Homogeneity", "Median", "Power"])
parameters_dict = OrderedDict({"Length": 0,
"Size": 1,
"Homogeneity": 2,
"Median": 3,
"Power": 4})
index = 1
for parameter in parameters_dict:
os.chdir(workdir + homogeneity_dir)
plt.figure(index, dpi=150, figsize=(24, 18))
for j in range(0, len(sample_set_names_list)):
sample = sample_set_names_list[j]
for i in range(0, len(power_list)):
"HAP",
"PmCDA1_sub1_3d",
"PmCDA1_6d",
"HAP_sub1",
"PmCDA1_sub1_6d",
]
samples_dir = workdir + all_files_subdir
power_limits = [f / 100 for f in range(1, 11)]
size_limits = [i for i in range(3, 11)]
os.chdir(workdir)
data_dict = {}
for sample in sample_set_names_list:
data_dict[sample] = CollectionCCF(from_file=True, input_file=samples_dir + sample + all_files_suffix).get_data_for_stat(additional_data=["Power"])
figure = plt.figure(1, dpi=150, figsize=(18, 12))
for sample in sample_set_names_list:
size_data = data_dict[sample][:, 1]
max_size = max(size_data)
bins_size_data = np.linspace(3, max_size, max_size + 1)
power_data = data_dict[sample][:, 2]
max_power = max(power_data)
n_power_bins = int(max_power / 0.01) + 1
bins_power_data = np.linspace(0, n_power_bins * 0.01, n_power_bins + 1)
hist, xedges, yedges = np.histogram2d(size_data, power_data, bins=(bins_size_data, bins_power_data))
ax = figure.add_subplot(3, 2, sample_set_names_list.index(sample) + 1, projection="3d")
#print(index)
if index == 7 or index == 10:
index += 1
subplot_list.append(None)
continue
file_name = "%s_size_%s+_power_%s+_good.ccf" % (sample_set_name, size, power) if power != "all" \
else "%s_size_%s+_good.ccf" % (sample_set_name, size)
print("Handling %s" % file_name)
data[sample_set_name] = {}
if "HAP" in sample_set_name:
data_names = ["Size", "Power"]
else:
data_names = ["Size", "Power", "Homogeneity"]
tmp_data = CollectionCCF(
from_file=True, input_file=file_name).get_data_for_stat(
additional_data=data_names[1:])
for data_name in data_names:
data[sample_set_name][
data_name] = tmp_data[:, value_names_dict[data_name]]
if index == 1:
subplot_list.append(plt.subplot(4, 3, index))
elif index == 8:
subplot_list.append(
plt.subplot(4, 3, index, sharex=subplot_list[0]))
elif index >= 9:
subplot_list.append(
plt.subplot(4,
3,
index,
sharex=subplot_list[7],
def get_clusters(self,
extracting_method="inconsistent",
threshold=0.8,
cluster_distance='average',
dendrogramm_max_y=2000,
sample_name=None,
save_clustering=False,
clustering_dir="clustering",
split_by_regions=False,
dendrogramm_color_threshold=1000,
draw_dendrogramm=True,
return_collection=True,
write_inconsistent=True,
write_correlation=True):
from Parsers.CCF import RecordCCF, CollectionCCF, MetadataCCF, HeaderCCF
if self.linkage_dict:
linkage_dict = self.linkage_dict
else:
region_dict, linkage_dict = self.hierarchical_clustering(method=cluster_distance,
dendrogramm_max_y=dendrogramm_max_y,
sample_name=sample_name,
save=save_clustering,
clustering_dir=clustering_dir,
dendrogramm_color_threshold=dendrogramm_color_threshold,
draw_dendrogramm=draw_dendrogramm,
write_correlation=write_correlation,
write_inconsistent=write_inconsistent)
if split_by_regions:
mut_clusters_dict = OrderedDict({})
else:
mut_clusters_list = []
clusters = OrderedDict()
for region in linkage_dict:
clusters[region] = fcluster(linkage_dict[region], threshold, criterion=extracting_method)
if return_collection:
for region in region_dict:
# http://docs.scipy.org/doc/scipy/reference/generated/scipy.cluster.hierarchy.fcluster.html#scipy.cluster.hierarchy.fcluster
clusters_dict = OrderedDict({})
for i in range(0, len(clusters[region])):
if clusters[region][i] not in clusters_dict:
clusters_dict[clusters[region][i]] = [region_dict[region][i]]
else:
clusters_dict[clusters[region][i]].append(region_dict[region][i])
if split_by_regions:
mut_clusters_dict[region] = \
CollectionCCF(record_list=[RecordCCF(collection_vcf=CollectionVCF(record_list=clusters_dict[cluster], from_file=False),
from_records=True) for cluster in clusters_dict],
metadata=MetadataCCF(self.samples, vcf_metadata=self.metadata, vcf_header=self.header),
header=HeaderCCF("CLUSTER_ID\tCHROM\tSTART\tEND\tDESCRIPTION".split("\t")))
else:
mut_clusters_list += [RecordCCF(collection_vcf=CollectionVCF(record_list=clusters_dict[cluster], from_file=False), from_records=True)
for cluster in clusters_dict]
if split_by_regions:
return mut_clusters_dict
return CollectionCCF(record_list=mut_clusters_list, metadata=MetadataCCF(self.samples, vcf_metadata=self.metadata, vcf_header=self.header),
header=HeaderCCF("CLUSTER_ID\tCHROM\tSTART\tEND\tDESCRIPTION".split("\t")))
else:
return clusters
#"AID_6d"
]
power = "0.03"
size = "5"
bin_size = 50
workdir = "/media/mahajrod/d9e6e5ee-1bf7-4dba-934e-3f898d9611c8/Data/LAN2xx/combined_vcf/clusters/%s/%s/" % (
size, power)
suffix = "_size_%s+_power_%s+_good.ccf" % (size, power)
hist_subfolder = "length_distribution/"
os.chdir(workdir)
os.system("mkdir -p %s" % hist_subfolder)
for extension in ".svg", ".eps", ".png", ".pdf":
os.system("mkdir -p %s/%s" % (hist_subfolder, extension[1:]))
for sample_set in sample_set_names_list:
collection = CollectionCCF(from_file=True,
input_file=workdir + sample_set + suffix)
length_data = collection.get_data_for_stat(additional_data=None)[:, 0]
#print(stat_data)
print(sample_set)
#print(length_data)
total = len(length_data)
minimum = min(length_data)
maximum = max(length_data)
if len(length_data) == 0:
continue
plt.figure(1, figsize=(5, 5))
plt.subplot(1, 1, 1)
bins = np.linspace(1, 2500, 2500 / bin_size + 1)
plt.hist(length_data,
bins=bins,
label="Min L: %i bp\nMax L: %i bp" % (minimum, maximum))
#"PmCDA1_sub1_6d",
#"A1_3d",
#"A1_6d",
#"A3G_3d",
#"AID_3d",
#"AID_6d"
]
power_limits = [f / 100 for f in range(1, 11)]
size_limits = [i for i in range(3, 11)]
os.chdir(workdir)
for sample_set in sample_set_names_list:
stat_dict = TwoLvlDict(OrderedDict({}))
print("Handling %s" % sample_set)
all_clusters = CollectionCCF(from_file=True,
input_file=workdir + all_files_subdir +
sample_set + all_files_suffix)
if "HAP" not in sample_set:
all_clusters.check_strandness()
for min_size in size_limits:
stat_dict[min_size] = OrderedDict({})
os.system("mkdir -p %i %i/all " % (min_size, min_size))
above_size_clusters, below_size_clusters = all_clusters.filter_by_expression(
"record.size >= %i" % min_size)
above_size_clusters.write(
"%i/all/%s_size_%i+%s" %
(min_size, sample_set, min_size, all_files_suffix))
stat_dict[min_size][0.00] = len(above_size_clusters)
for min_power in power_limits:
os.system("mkdir -p %i/%.2f" % (min_size, min_power))
return 0
start_shift = start1 - start2
start_coef_shift = 0 if start_shift < 0 else 1
end_shift = end1 - end2
end_coef_shift = 0 if end_shift > 0 else 1
return (end2 - start2 + 1) - start_coef_shift * start_shift + end_coef_shift * end_shift
overlap_clusters_percent = TwoLvlDict({})
#size = 8
#power = 0.05
print([float(f) / float(100) for f in range(1, 11)])
for size in range(3, 11):
overlap_clusters_percent[size] = {}
for power in [float(f) / float(100) for f in range(1, 11)]:
PmCDA1_3d_clusters = CollectionCCF(from_file=True, input_file="/media/mahajrod/d9e6e5ee-1bf7-4dba-934e-3f898d9611c8/Data/LAN2xx/combined_vcf/clusters/%i/%.2f/PmCDA1_3d_size_%i+_power_%.2f+_good.ccf" % (size, power, size, power))
PmCDA1_3d_sub_clusters = CollectionCCF(from_file=True, input_file="/media/mahajrod/d9e6e5ee-1bf7-4dba-934e-3f898d9611c8/Data/LAN2xx/combined_vcf/clusters/%i/%.2f/PmCDA1_sub1_3d_size_%i+_power_%.2f+_good.ccf" % (size, power, size, power))
PmCDA1_3d_clusters.write_gff("/media/mahajrod/d9e6e5ee-1bf7-4dba-934e-3f898d9611c8/Data/LAN2xx/combined_vcf/clusters/%i/%.2f/PmCDA1_3d_size_%i+_power_%.2f+_good.gff" % (size, power, size, power))
PmCDA1_3d_sub_clusters.write_gff("/media/mahajrod/d9e6e5ee-1bf7-4dba-934e-3f898d9611c8/Data/LAN2xx/combined_vcf/clusters/%i/%.2f/PmCDA1_sub1_3d_size_%i+_power_%.2f+_good.gff" % (size, power, size, power))
#cluster_3d_dict = OrderedDict({})
cluster_3d_dict = TwoLvlDict({})
for cluster_3d in PmCDA1_3d_clusters:
cluster_3d_dict[cluster_3d.id] = OrderedDict({"length": cluster_3d.len,
"N of clusters": 0,
"length of clusters": [],
"intersection": [],
"intersection % of main cluster": [],
"interscection % of clusters": [],