def ccc_test(LucN, hccN, MixN, mm2):
r_LucN, r_hccN, r_MixN = graph.num_read_cells(mm2)
time_point = len(r_LucN[0]) #8
num_experiments = len(r_LucN)
sim_tmp = len(LucN) / time_point #1.25
LucN_p = []
hccN_p = []
MixN_p = []
for t in range(time_point):
LucN_p.append(LucN[int(round(t * sim_tmp))])
hccN_p.append(hccN[int(round(t * sim_tmp))])
MixN_p.append(MixN[int(round(t * sim_tmp))])
ave_LucN_r = np.average(r_LucN, axis=0)
var_LucN_r = np.var(r_LucN, axis=0)
ccc = []
for t in range(len(r_LucN[0])): #timepoint
numerator = 0
for j in range(num_experiments): #experiment
numerator += (r_LucN[j][t] - ave_LucN_r[t]) * 1
denominator = var_LucN_r[t]**2 + 1 + (ave_LucN_r[t] - LucN_p[t])**2
tmp = 2 * numerator / (denominator * num_experiments)
ccc.append(tmp)
print ccc
def ccc_test(LucN, hccN, MixN, mm2):
r_LucN, r_hccN, r_MixN = graph.num_read_cells(mm2)
time_point = len(r_LucN[0])#8
num_experiments = len(r_LucN)
sim_tmp = len(LucN)/time_point #1.25
LucN_p = []
hccN_p = []
MixN_p = []
for t in range(time_point):
LucN_p.append(LucN[int(round(t*sim_tmp))])
hccN_p.append(hccN[int(round(t*sim_tmp))])
MixN_p.append(MixN[int(round(t*sim_tmp))])
ave_LucN_r = np.average(r_LucN, axis=0)
var_LucN_r = np.var(r_LucN, axis=0)
ccc = []
for t in range(len(r_LucN[0])):#timepoint
numerator = 0
for j in range(num_experiments):#experiment
numerator += (r_LucN[j][t]-ave_LucN_r[t])*1
denominator = var_LucN_r[t]**2 + 1 + (ave_LucN_r[t] - LucN_p[t])**2
tmp = 2*numerator/(denominator*num_experiments)
ccc.append(tmp)
print ccc
def num_cell_corrcoef(LucN, hccN, MixN, mm2):
#Raw Data
r_LucN, r_hccN, r_MixN = graph.num_read_cells(mm2)
#Adjustment of Timeseries
time_point = len(r_LucN[0]) # Experiments
sim_time_point = len(LucN) # Simulation time
sim_tmp = (len(LucN) * 0.66) / (time_point - 1) #1.25
#Simulation timepoint
LucN_p = []
hccN_p = []
MixN_p = []
for t in range(time_point):
if t != 0:
print int(sim_time_point * 0.3 + round(t * sim_tmp))
LucN_p.append(LucN[int(sim_time_point * 0.3 + round(t * sim_tmp))])
hccN_p.append(hccN[int(sim_time_point * 0.3 + round(t * sim_tmp))])
MixN_p.append(MixN[int(sim_time_point * 0.3 + round(t * sim_tmp))])
else:
LucN_p.append(LucN[int(round(t * sim_tmp))])
hccN_p.append(hccN[int(round(t * sim_tmp))])
MixN_p.append(MixN[int(round(t * sim_tmp))])
print 'Simulation vs Raw'
print 'Luc'
print check_number_of_cells(LucN_p, r_LucN, 'Luc')
print 'Mix'
print check_number_of_cells(MixN_p, r_MixN, 'Mix')
print 'hcc'
print check_number_of_cells(hccN_p, r_hccN, 'HCC')
draw.all_cells_fig(LucN_p, r_LucN, MixN_p, r_MixN, hccN_p, r_hccN)
corr_Luc = []
corr_hcc = []
corr_Mix = []
for i in range(len(r_LucN)): #times of experiments
tmp_Luc = np.corrcoef(r_LucN[i], LucN_p)
tmp_hcc = np.corrcoef(r_hccN[i], hccN_p)
tmp_Mix = np.corrcoef(r_MixN[i], MixN_p)
corr_Luc.append(tmp_Luc[0, 1])
corr_hcc.append(tmp_hcc[0, 1])
corr_Mix.append(tmp_Mix[0, 1])
print 'Average Correlation Luc = %s, HCC = %s, Mix ~ %s ' % (np.average(
np.array(corr_Luc)), np.average(
np.array(corr_hcc)), np.average(np.array(corr_Mix)))
return 0
def num_cell_corrcoef(LucN, hccN, MixN, mm2):
#Raw Data
r_LucN, r_hccN, r_MixN = graph.num_read_cells(mm2)
#Adjustment of Timeseries
time_point = len(r_LucN[0]) # Experiments
sim_time_point = len(LucN) # Simulation time
sim_tmp = (len(LucN)*0.66)/(time_point-1) #1.25
#Simulation timepoint
LucN_p = []
hccN_p = []
MixN_p = []
for t in range(time_point):
if t != 0:
print int(sim_time_point*0.3 + round(t*sim_tmp))
LucN_p.append(LucN[int(sim_time_point*0.3 + round(t*sim_tmp))])
hccN_p.append(hccN[int(sim_time_point*0.3 + round(t*sim_tmp))])
MixN_p.append(MixN[int(sim_time_point*0.3 + round(t*sim_tmp))])
else:
LucN_p.append(LucN[int(round(t*sim_tmp))])
hccN_p.append(hccN[int(round(t*sim_tmp))])
MixN_p.append(MixN[int(round(t*sim_tmp))])
print 'Simulation vs Raw'
print 'Luc'
print check_number_of_cells(LucN_p, r_LucN, 'Luc')
print 'Mix'
print check_number_of_cells(MixN_p, r_MixN, 'Mix')
print 'hcc'
print check_number_of_cells(hccN_p, r_hccN, 'HCC')
draw.all_cells_fig(LucN_p, r_LucN, MixN_p, r_MixN, hccN_p, r_hccN)
corr_Luc = []
corr_hcc = []
corr_Mix = []
for i in range(len(r_LucN)): #times of experiments
tmp_Luc = np.corrcoef(r_LucN[i], LucN_p)
tmp_hcc = np.corrcoef(r_hccN[i], hccN_p)
tmp_Mix = np.corrcoef(r_MixN[i], MixN_p)
corr_Luc.append(tmp_Luc[0,1])
corr_hcc.append(tmp_hcc[0,1])
corr_Mix.append(tmp_Mix[0,1])
print 'Average Correlation Luc = %s, HCC = %s, Mix ~ %s ' % (np.average(np.array(corr_Luc)), np.average(np.array(corr_hcc)), np.average(np.array(corr_Mix)))
return 0
def exp_test():
#parameter
luc_node = 100
time = 10
luc_gro = 10
#Generate Graph
lucG = nx.barabasi_albert_graph(luc_node, luc_gro)
#Time series cell volume
LucN = []
#Number of initial cell
LucN0 = 100
LucN_init = 100
mm2 = 43
for t in range(time):
LucN.append(calc.convert_volume(LucN0))
lucG = graph.update_graph(lucG, luc_gro)
LucN0 = LucN_init * math.exp(1 / 10 * (t + 1))
r_LucN, r_hccN, r_MixN = graph.num_read_cells(mm2)
time_point = len(r_LucN[0]) #8
sim_tmp = len(LucN) / time_point #1.25
LucN_p = []
for t in range(time_point):
LucN_p.append(LucN[int(round(t * sim_tmp))])
corr_Luc = []
for i in range(len(r_LucN)): #times of experiments
tmp_Luc = np.corrcoef(r_LucN[i], LucN_p)
corr_Luc.append(tmp_Luc[0, 1])
print np.average(np.array(corr_Luc))
return 0
def exp_test():
#parameter
luc_node = 100
time = 10
luc_gro = 10
#Generate Graph
lucG = nx.barabasi_albert_graph(luc_node, luc_gro)
#Time series cell volume
LucN = []
#Number of initial cell
LucN0 = 100
LucN_init = 100
mm2 = 43
for t in range(time):
LucN.append(calc.convert_volume(LucN0))
lucG = graph.update_graph(lucG, luc_gro)
LucN0 = LucN_init*math.exp(1/10*(t+1))
r_LucN, r_hccN, r_MixN = graph.num_read_cells(mm2)
time_point = len(r_LucN[0])#8
sim_tmp = len(LucN)/time_point #1.25
LucN_p = []
for t in range(time_point):
LucN_p.append(LucN[int(round(t*sim_tmp))])
corr_Luc = []
for i in range(len(r_LucN)): #times of experiments
tmp_Luc = np.corrcoef(r_LucN[i], LucN_p)
corr_Luc.append(tmp_Luc[0,1])
print np.average(np.array(corr_Luc))
return 0