def convert_to_bp(alelle, data, LIZ_500):
data_105 = list(data)
i = len(data)
indexes = findpeaks.findpeaks(data_105, spacing=50, limit=200)
ind = []
j = len(indexes) - 1
k = 0
while j >= 0:
if k == 15:
ind.append(indexes[j])
break
ind.append(indexes[j])
j -= 1
k += 1
for c in range(0, 16):
if alelle > ind[c] or alelle == ind[c]:
bp_pred = ((alelle - ind[c]) /
((ind[c - 1] - ind[c]) /
(LIZ_500[c - 1] - LIZ_500[c]))) + LIZ_500[c]
break
else:
bp_pred = 0
return bp_pred
def convert_to_bp(alelle, data, LIZ_500):
data_105 = list(data)
i = len(data)
indexes = findpeaks.findpeaks(data_105, spacing=35, limit=100)
ind = []
j = len(indexes) - 1
k = 0
# print(indexes)
while j >= 0:
if k == 15:
ind.append(indexes[j])
break
counter = 1
next_index = j - counter
del_ratio = (indexes[j] - indexes[next_index]) / (LIZ_500[k] -
LIZ_500[k + 1])
# print(del_ratio)
in_range = del_ratio > 8.5 and del_ratio < 13.5
if in_range:
ind.append(indexes[j])
k += 1
j -= 1
else:
while del_ratio < 8.5 or del_ratio > 13.5:
counter += 1
next_index = j - counter
if next_index < 0:
j -= 1
counter = 1
next_index = j - counter
del_ratio = (indexes[j] - indexes[next_index]) / (
LIZ_500[k] - LIZ_500[k + 1])
ind.append(indexes[j])
j = next_index
k += 1
# print(ind)
for c in range(0, 16):
if alelle > ind[c] or alelle == ind[c]:
bp_pred = ((alelle - ind[c]) /
((ind[c - 1] - ind[c]) /
(LIZ_500[c - 1] - LIZ_500[c]))) + LIZ_500[c]
break
else:
bp_pred = 0
return bp_pred
def find_lower(data, dye):
data_105 = list(data)
i = len(data)
indexes = findpeaks.findpeaks(data_105, spacing=50, limit=200)
ind = []
j = len(indexes) - 1
k = 0
while j >= 0:
if k == 15:
ind.append(indexes[j])
break
ind.append(indexes[j])
j -= 1
k += 1
return ind[len(ind) - 1]
def find_upper(data, LIZ_500):
data_105 = list(data)
i = len(data)
indexes = findpeaks.findpeaks(data_105, spacing=35, limit=100)
ind = []
j = len(indexes) - 1
k = 0
while j >= 0:
if k == 15:
ind.append(indexes[j])
break
counter = 1
next_index = j - counter
del_ratio = (indexes[j] - indexes[next_index]) / (LIZ_500[k] -
LIZ_500[k + 1])
# print(del_ratio)
in_range = del_ratio > 8.5 and del_ratio < 13.5
if in_range:
ind.append(indexes[j])
k += 1
j -= 1
else:
while del_ratio < 8.5 or del_ratio > 13.5:
counter += 1
next_index = j - counter
if next_index < 0:
j -= 1
counter = 1
next_index = j - counter
del_ratio = (indexes[j] - indexes[next_index]) / (
LIZ_500[k] - LIZ_500[k + 1])
ind.append(indexes[j])
j = next_index
k += 1
if ind[0] == ind[len(ind) - 1]:
ind.pop(len(ind) - 1)
return ind[0]
def convert_to_index(bp, data, LIZ_500):
data_105 = list(data)
i = len(data)
indexes = findpeaks.findpeaks(data_105, spacing=50, limit=200)
ind = []
j = len(indexes) - 1
k = 0
while j >= 0:
if k == 15:
ind.append(indexes[j])
break
ind.append(indexes[j])
j -= 1
k += 1
for c in range(0, 16):
if bp > LIZ_500[c]:
index_pred = ((bp - LIZ_500[c]) / ((LIZ_500[c - 1] - LIZ_500[c]) /
(ind[c - 1] - ind[c]))) + ind[c]
break
return index_pred
break alelle2_index = index_of_peaks[height.index(max(height))] if alelle2_index > alelle1_index: if alelle1_index+80>=alelle2_index: two_peaks = True else: two_peaks = False else: if alelle1_index-80<=alelle2_index: two_peaks = True else: two_peaks = False all_peaks = fp.findpeaks(data, spacing=25, limit=25) for i in range(len(all_peaks)-1): lower = all_peaks[i] - 80 upper = all_peaks[i] + 20 if lower < index_min or upper > index_max: continue if noise_count > 2320: if alelle1 != alelle2: if all_peaks[i] == alelle1_index or all_peaks[i] == alelle2_index: file_name.append(filename) area_of_peaks.append(np.trapz(data[lower:upper])) number_of_peaks.append(len(fp.findpeaks(data[lower:upper], spacing=5,limit=15))) length_in_bp.append(round(convert_to_bp(all_peaks[i], record.annotations['abif_raw'][e], dye))) height_of_peaks.append(data[all_peaks[i]])
dipFound = False while dipFound == False: slope = (channel[peak+1]-channel[peak])/((peak+1)-peak) if slope >= 0: dipFound = True plt.scatter(peak- a, channeldata[peak],color='red') peak -= 1 peak += 1 return(peak) currentfile = "A_BOH_12_12.fsa" my_dir="/home/bo/PGC/microsat/testdata/training/GetHeight/" record = SeqIO.read(my_dir+currentfile,"abi") channeldata = record.annotations['abif_raw']['DATA1'] smol_test = channeldata[a:b] all_peaks = fp.findpeaks(smol_test,spacing=5,limit=1000) print(all_peaks) print(leftwarddip(channeldata,all_peaks)) print(rightwarddip(channeldata,all_peaks)) plt.scatter(all_peaks,[channeldata[3350 + i] for i in all_peaks]) plt.plot(smol_test,'-o',markersize=4) plt.show()
def plotgraph(directory, filename, peakwindow, threshold=2000):
filename = filename
my_dir = directory
threshold = 500
#Load Data from FSA file
record = SeqIO.read(my_dir + filename, "abi")
channeldata = np.array(record.annotations['abif_raw']['DATA1'])
ladderdata = np.array(record.annotations['abif_raw']['DATA105'])
#What if the user wants to reset the scan with lower threshold?
#what if true peak is actually lower than the default 2000 threshold?
#This while block allows the user to repeat the scanning procedure
cond = False
while cond == False:
all_pk = fp.findpeaks(channeldata, spacing=25, limit=threshold)
all_pk = [
a for a in all_pk
if a > int(peakwindow[0]) and a < int(peakwindow[1])
]
peak_height = [channeldata[a] for a in all_pk]
#What if no peaks is detected?
#This block shows the user the graph and that nothing is detected
if len(all_pk) == 0:
print("No peaks found in channel with threshold=%s" % threshold)
plt.plot(channeldata)
plt.show()
print("Peaks detected from channel: %s \n%s\n" % (len(all_pk), all_pk))
seg_ranges = [[x - 80, x + 40] for x in all_pk] #Segment Ranges
seg_area = [np.trapz(channeldata[s[0]:s[1]]) for s in seg_ranges]
pk_area = [
np.trapz(channeldata[Peakboundaries(channeldata, i)[0]:
Peakboundaries(channeldata, i)[1]])
for i in all_pk
]
stu_area = [seg_area[i] - pk_area[i] for i in range(len(seg_area))]
print(f'peaks at {all_pk}')
print(f'peak height {peak_height}')
print(f'seg area {seg_area}')
print(f'peak area {pk_area}')
print(f'stutter area {stu_area}\n')
for i in range(len(seg_ranges)):
a, b = seg_ranges[i][0], seg_ranges[i][1]
segment = np.array(channeldata[a:b])
# the points
plt.style.use('seaborn')
plt.subplot(2, len(seg_ranges), i + 1)
plt.plot(list(range(a, b)), segment, color='blue') #peak+stutter
plt.title(f'{i+1}\nlikelihood: 5%')
plt.yticks([])
plt.ylim(0, max(peak_height) + 1000)
#bottom
plt.subplot(2, 1, 2)
plt.plot(ladderdata, color='black', alpha=0.3)
plt.plot(channeldata)
plt.plot(all_pk,
peak_height,
'o',
color='red',
markersize=5,
label="Suggested Peaks")
plt.ylim(0, max(peak_height) + 1000)
plt.xlim(500, 7000)
for i in range(len(seg_ranges)):
plt.axvspan(seg_ranges[i][0],
seg_ranges[i][1],
color='m',
alpha=0.5)
plt.axvspan(0,
peakwindow[0],
color='black',
alpha=0.3,
label="Excluded from peak search")
plt.axvspan(int(peakwindow[1]),
len(channeldata),
color='black',
alpha=0.3)
plt.axhline(y=0, color='k') #x axis line
plt.title(filename)
plt.legend() #show labels
plt.show(block=False) #print the graph
print("%s potential peaks detected. Please select peaks " %
len(seg_ranges),
end='')
print(
"separated by commas. \ne.g: 1,2 (heterozygous) or 1,1 (homozygous)\n"
)
print("[0] -> Repeat scan with lower threshold")
for i in range(len(seg_ranges)):
print("[{}] -> {}".format(i + 1, seg_ranges[i]))
#Asks for user input and select peaks to choose
#Idea: add option to repeat findpeaks or manual [0] if desired peak is not detected.
while True:
sel_peaks = str(input("\n>")).split(',')
sel_peaks = [int(p) - 1 for p in sel_peaks]
if len(sel_peaks) == 1 and sel_peaks[0] == -1:
#not all thres is detected, repeat scan
threshold -= 500
print("Repeating scan with lowered threshold (-500)")
break
elif len(sel_peaks) > len(seg_ranges) or min(
sel_peaks) + 1 < 0 or max(sel_peaks) + 1 > len(seg_ranges):
print("Invalid input")
else:
cond = True #condition is fulfulled
break
print("Please close graph to continue.....")
plt.show()
#figure out a way to move these to the top
# sel_peaks = [p-1 for p in sel_peaks]
# not_height = [height[i] for i in range(len(height)) if i not in sel_peaks]
# not_seg_area = [seg_area[i] for i in range(len(seg_area)) if i not in sel_peaks]
# not_sel_peaks = [all_pk[i] for i in range(len(all_pk)) if i not in sel_peaks]
# height = [height[i] for i in range(len(height)) if i in sel_peaks]
# seg_area = [seg_area[i] for i in range(len(seg_area)) if i in sel_peaks]
# sel_peaks = [all_pk[p] for p in sel_peaks]
# pk_area = [np.trapz(channeldata[Peakboundaries(channeldata,i)[0]:Peakboundaries(channeldata,i)[1]])
# for i in sel_peaks]
# pk_area_not = [np.trapz(channeldata[Peakboundaries(channeldata,i)[0]:Peakboundaries(channeldata,i)[1]])
# for i in not_sel_peaks]
# stu_area = [seg_area[i] - pk_area[i] for i in range(len(seg_area))]
# stu_area_not = [not_seg_area[i] - pk_area_not[i] for i in range(len(not_seg_area))]
"""
break
alelle2_index = index_of_peaks[height.index(max(height))]
if alelle2_index > alelle1_index:
if alelle1_index + 80 >= alelle2_index:
two_peaks = True
else:
two_peaks = False
else:
if alelle1_index - 80 <= alelle2_index:
two_peaks = True
else:
two_peaks = False
all_peaks = fp.findpeaks(data, spacing=25, limit=25)
for i in range(len(all_peaks) - 1):
lower = all_peaks[i] - 40
upper = all_peaks[i] + 40
if lower < index_min or upper > index_max:
continue
if noise_count >= 1268:
if alelle1 != alelle2:
if all_peaks[i] == alelle1_index or all_peaks[
i] == alelle2_index:
file_name.append(filename)
area_of_peaks.append(np.trapz(data[lower:upper]))
# number_of_peaks.append(len(fp.findpeaks(data[lower:upper], spacing=5,limit=15)))
length_in_bp.append(
file = input('File to analyze: ')
record = SeqIO.read(file + '.fsa', 'abi')
data = record.annotations['abif_raw'][a]
label = []
area_of_peaks = []
no_of_peaks = []
length = []
peaks = []
height = []
index_35 = find_lower(record.annotations['abif_raw']['DATA105'], dye)
index_500 = find_upper(record.annotations['abif_raw']['DATA105'], dye)
detected_peaks = fp.findpeaks(data, spacing=25, limit=25)
for i in range(len(detected_peaks) - 1):
lower_end_of_window = detected_peaks[i] - 80
upper_end_of_window = detected_peaks[i] + 20
if lower_end_of_window < index_35 or upper_end_of_window > index_500:
continue
peaks.append(detected_peaks[i])
area_of_peaks.append(
np.trapz(data[lower_end_of_window:upper_end_of_window]))
no_of_peaks.append(
len(
fp.findpeaks(data[lower_end_of_window:upper_end_of_window],
spacing=5,
limit=15)))
