dist_matrix = sp.spatial.distance.squareform(
sp.spatial.distance.pdist(sim_data_obj.pi.dot(sim_data_obj.MU),
'cosine'))
if len(dist_matrix[dist_matrix > 0]):
min_dist = np.min(dist_matrix[dist_matrix > 0])
max_dist = np.max(dist_matrix[dist_matrix > 0])
avg_dist = np.mean(dist_matrix[dist_matrix > 0])
else:
min_dist, max_dist, avg_dist = np.nan, np.nan, np.nan
avg_major_cn = np.mean(sim_data_obj.C_tumor_tot -
sim_data_obj.C_tumor_minor)
avg_tot_cn = np.mean(sim_data_obj.C_tumor_tot)
actual_perc_diploid = sum((sim_data_obj.C_tumor_tot == 2) & (
sim_data_obj.C_tumor_minor == 1)) / sim_data_obj.N
MU = get_MU()
subMU = get_MU(cancer_type=cancer_type)
id_list = list()
metrics_list = list()
mu_mat = {'all': MU, 'cancertype': subMU}
for setting in ('all', 'cancertype', 'prefit'):
mixture_file = pd.read_csv(
'{}/tracksig/tracksig_mixtures_{}.csv'.format(
folder_path, setting),
sep=',')
try:
changepoint_file = pd.read_csv(
'{}/tracksig/tracksig_changepoints_{}.txt'.format(
folder_path, setting),
header=None,
folder_path = sys.argv[1]
MIXTURE_THRESHOLD = 0.05
'''
folder_path = '20190623_simulations_clonesig_cn_cancer_type/type2-perc_diploid100-nb_clones4-nb_mut100'
'''
nb_pi = int(folder_path.split('pi')[1].split('-')[0])
nb_phi = int(folder_path.split('phi')[1].split('-')[0])
depth = int(folder_path.split('depth')[1].split('-')[0])
nb_mut = int(folder_path.split('nb_mut')[1].split('-')[0])
perc_diploid = float(folder_path.split('percdip')[1].split('-')[0])
nb_clones = 2
cancer_type = None
MU = get_MU()
data_df = pd.read_csv('{}/input_t.tsv'.format(folder_path), sep='\t')
with open('{}/purity.txt'.format(folder_path), 'r') as f:
purity = float(f.read())
# get metrics from simulated data
with open('{}/sim_data'.format(folder_path), 'rb') as sim_pickle_file:
sim_pickle = pickle.Unpickler(sim_pickle_file)
sim_data_obj = sim_pickle.load()
dist_matrix = sp.spatial.distance.squareform(
sp.spatial.distance.pdist(sim_data_obj.pi.dot(sim_data_obj.MU), 'cosine'))
if len(dist_matrix[dist_matrix > 0]):
min_dist = np.min(dist_matrix[dist_matrix > 0])
max_dist = np.max(dist_matrix[dist_matrix > 0])
avg_dist = np.mean(dist_matrix[dist_matrix > 0])