def add_CN(chrlen, cn_num, del_rate, min_cn_size, exp_theta, amp_p):
# for each branch, the copy number change happens on only one allele
CN_array = []
new_chrlen = [row[:] for row in chrlen]
CN_Tot = np.random.poisson(cn_num, 1)
for i in range(CN_Tot):
# allele
CN_Ale = np.random.binomial(1, 0.5)
# deletion versus amplification
CN_Del = np.random.binomial(1, del_rate)
CN_chromosome = get_chr(new_chrlen[CN_Ale])
#print CN_Ale, CN_chromosome
CN_p1, CN_p2 = get_range(new_chrlen[CN_Ale][CN_chromosome],
min_cn_size, exp_theta)
CN_amp_num = 0
#print new_chrlen
#print "before changing:"
#print chrlen, new_chrlen
if CN_Del == 0:
# get amplification copy number
# starting from 0
CN_amp_num = int(np.random.geometric(amp_p, 1) - 1)
#print CN_amp_num, CN_p2, CN_p1
new_chrlen[CN_Ale][CN_chromosome] = new_chrlen[CN_Ale][
CN_chromosome] + CN_amp_num * (CN_p2 - CN_p1)
#print new_chrlen
else:
new_chrlen[CN_Ale][CN_chromosome] = new_chrlen[CN_Ale][
CN_chromosome] - (CN_p2 - CN_p1)
#print new_chrlen
#print "after changing:"
#print chrlen, new_chrlen
CN_array.append(
CN(CN_Ale, CN_Del, CN_chromosome, CN_p1, CN_p2, CN_amp_num))
return CN_array, new_chrlen
def make_fa(ID, tree, ref, chr_name_array, fa_prefix):
fa_f_prefix = fa_prefix + str(ID) + "_"
# record all the nodes that this route visited (from leaf to root)
trace = [ID]
visit = ID
while visit != 0:
visit = tree[visit].parentID
trace.append(visit)
# now reverse the trace so that the CNV can be applied from root to leaf
CN = []
for i in range(len(trace)):
j = len(trace) - i - 1
# gather all CNs together
for cn in tree[trace[j]].cn:
CN.append(cn)
new_ref = gen_ref(ref, CN)
write_ref(new_ref, chr_name_array, fa_f_prefix)
def add_whole_amp(chrlen, whole_amp_rate, whole_amp_num, corres,
amp_num_geo_par):
new_chrlen = [row[:] for row in chrlen]
new_corres = [row[:] for row in corres]
# like CN
wholeamp = []
for i in range(len(chrlen)):
# allele
for j in range(len(chrlen[i])):
# chromosome
random_ = np.random.uniform(0.0, 1.0)
if random_ < whole_amp_rate:
# chosen to be amplified
# get the amplification number
# geometric distribution's mean is 1/p = 1.
amp_num = whole_amp_num * np.random.geometric(amp_num_geo_par)
while amp_num < 1:
amp_num = whole_amp_num * np.random.geometric(
amp_num_geo_par)
new_CN = CN(i, 0, j, 0, new_chrlen[i][j], amp_num, new_corres)
new_chrlen[i][j] = new_chrlen[i][j] * (amp_num + 1)
new_corres = get_new_corres(new_CN, new_corres)
wholeamp.append(new_CN)
return wholeamp, new_chrlen, new_corres
def add_CN(chrlen, cn_num, del_rate, min_cn_size, exp_theta, amp_p, corres,
CN_LIST_ID):
# for each branch, the copy number change happens on only one allele
CN_array = []
# a variable used only for fixing CN, for test
new_chrlen = [row[:] for row in chrlen]
new_corres = [row[:] for row in corres]
if random == 0:
CN_Tot = 9
else:
CN_Tot = int(np.random.poisson(cn_num, 1))
for i in range(CN_Tot):
# allele
if random == 0:
CN_Ale = allele_list[i]
else:
CN_Ale = np.random.binomial(1, 0.5)
# deletion versus amplification
if random == 0:
CN_Del = if_del_list[i]
else:
CN_Del = np.random.binomial(1, del_rate)
CN_chromosome = get_chr(new_chrlen[CN_Ale])
#print CN_Ale, CN_chromosome
CN_p1, CN_p2 = get_range(new_chrlen[CN_Ale][CN_chromosome],
min_cn_size, exp_theta, i)
# to protect from str to int, int to str, make it not dividable by 10
if CN_p1 % 10 == 0:
CN_p1 = CN_p1 + 1
if CN_p2 % 10 == 0:
CN_p2 = CN_p2 - 1
# think about how to get the actual coordinates given the previous ones.
CN_amp_num = 0
#print new_chrlen
#print "before changing:"
#print chrlen, new_chrlen
if CN_Del == 0:
# get amplification copy number
# starting from 0
while 1:
CN_amp_num_ = np.random.geometric(amp_p, 1)
CN_amp_num = int(CN_amp_num_[0])
if CN_amp_num >= 1:
break
if random == 0:
CN_amp_num = amp_num_list[i]
#CN_amp_num = int(np.random.geometric(amp_p, 1) - 1)
#print CN_amp_num, CN_p2, CN_p1
new_chrlen[CN_Ale][CN_chromosome] = new_chrlen[CN_Ale][
CN_chromosome] + CN_amp_num * (CN_p2 - CN_p1)
#print new_chrlen
else:
new_chrlen[CN_Ale][CN_chromosome] = new_chrlen[CN_Ale][
CN_chromosome] - (CN_p2 - CN_p1)
#print new_chrlen
#print "after changing:"
#print chrlen, new_chrlen
# corresp is the corresponding interval of ref and genome
new_CN = CN(CN_Ale, CN_Del, CN_chromosome, CN_p1, CN_p2, CN_amp_num,
new_corres)
new_corres = get_new_corres(new_CN, new_corres)
#print_corrs(new_corres)
CN_array.append(new_CN)
return CN_array, new_chrlen, new_corres