def liveDead(row, f):
path, name = os.path.split(f)
img = ogre.site_to_array(f)
b = img[:, :, 0] # DAPI
g = img[:, :, 1] # FITC
r = img[:, :, 2] # TexasRed
bc = ogre.correct_illumination(b)
bcmask = ogre.adapt_thresh(bc)
blabs, bprops = ogre.watershed_label(bcmask, bc)
gc = ogre.correct_illumination(g)
gcmask = ogre.adapt_thresh(gc)
glabs, gprops = ogre.watershed_label(gcmask, gc)
rc = ogre.correct_illumination(r)
rcmask = ogre.adapt_thresh(rc)
rlabs, rprops = ogre.watershed_label(rcmask, rc)
# fig, ax = plt.subplots(3,3, figsize=(15,12))
# fig.suptitle(f, fontsize=15)
# ax[0,0].imshow(bc, cmap=plt.cm.gray)
# ax[0,1].imshow(gc, cmap=plt.cm.gray)
# ax[0,2].imshow(rc, cmap=plt.cm.gray)
# ax[1,0].imshow(bcmask, cmap=plt.cm.gray)
# ax[1,1].imshow(gcmask, cmap=plt.cm.gray)
# ax[1,2].imshow(rcmask, cmap=plt.cm.gray)
# ax[2,0].imshow(label2rgb(blabs))
# ax[2,1].imshow(label2rgb(glabs))
# ax[2,2].imshow(label2rgb(rlabs))
# sns.despine()
# fig.tight_layout()
# Raw cell counter seems to be doing a better job here...
raw_dapi = ogre.count_raw(bprops)
# bcount_filter = ogre.count_filter(bprops)
raw_fitc = ogre.count_raw(gprops)
# gcount_filter = ogre.count_filter(gprops)
raw_texas = ogre.count_raw(rprops)
# rcount_filter = ogre.count_filter(rprops)
glabmask = glabs != 0
rlabmask = rlabs != 0
live = 0
dead = 0
dying = 0
for i in range(1, blabs.max()):
nucleus_roi = blabs == i
if ((nucleus_roi * glabmask * rlabmask).any()):
dying += 1
elif ((nucleus_roi * glabmask).any()):
live += 1
elif ((nucleus_roi * rlabmask).any()):
dead += 1
return (row,
[path, name, raw_dapi, raw_fitc, raw_texas, live, dead, dying])
import seaborn as sns
import pandas as pd
import os
except ImportError:
raise ImportError("Please verify the module is in the working directory.")
#def ogre_show(axrow, x, y, z):
# axrow[0].imshow(x, cmap=plt.cm.gray)
# axrow[1].imshow(y, cmap=plt.cm.gray)
# axrow[2].imshow(z, cmap=plt.cm.gray)
# - MAIN - #
platepath = 'B:\\Projects\\Research\\2015-05-11 AlyssaScreen1'
output_filename = '2015-05-11 AlyssaScreen1.csv'
output_filepath = os.path.join(platepath, output_filename)
w1_files = ogre.iter_plate(platepath)
w1_files_length = len(w1_files)
columns = [
"path", "file", "raw_dapi", "raw_fitc", "raw_texas", "live", "dead", "all"
]
df = pd.DataFrame(index=range(0, w1_files_length), columns=columns)
for f in w1_files:
row = w1_files.index(f)
path, name = os.path.split(f)
print "Status Update:"
print "File: {0}".format(f)
print "Item: {0}/{1}".format(row, w1_files_length)
elif ((nucleus_roi * glabmask).any()):
live += 1
elif ((nucleus_roi * rlabmask).any()):
dead += 1
return (row,
[path, name, raw_dapi, raw_fitc, raw_texas, live, dead, dying])
if __name__ == '__main__':
# - MAIN - #
platepath = 'B:\\Projects\\Research\\2015-05-11 AlyssaScreen1\\20150511-AlyssaC2D-D3-LD_Plate_4714'
output_filename = '20150511-AlyssaC2D-D3-LD_Plate_4714.csv'
output_filepath = os.path.join(platepath, output_filename)
w1_files = ogre.iter_plate(platepath)
w1_files_length = len(w1_files)
columns = [
"path", "file", "raw_dapi", "raw_fitc", "raw_texas", "live", "dead",
"all"
]
df = pd.DataFrame(index=range(0, w1_files_length), columns=columns)
# Parallel for loop equivalent to "for f in w1_files:"
num_cores = multiprocessing.cpu_count()
results = Parallel(n_jobs=num_cores, verbose=11)(
delayed(liveDead)(row, f)
for row, f in zip(range(w1_files_length), w1_files))
for result in results:
import seaborn as sns
import pandas as pd
import os
except ImportError:
raise ImportError("Please verify the module is in the working directory.")
# def ogre_show(axrow, x, y, z):
# axrow[0].imshow(x, cmap=plt.cm.gray)
# axrow[1].imshow(y, cmap=plt.cm.gray)
# axrow[2].imshow(z, cmap=plt.cm.gray)
# - MAIN - #
platepath = "B:\\Projects\\Research\\2015-05-11 AlyssaScreen1"
output_filename = "2015-05-11 AlyssaScreen1.csv"
output_filepath = os.path.join(platepath, output_filename)
w1_files = ogre.iter_plate(platepath)
w1_files_length = len(w1_files)
columns = ["path", "file", "raw_dapi", "raw_fitc", "raw_texas", "live", "dead", "all"]
df = pd.DataFrame(index=range(0, w1_files_length), columns=columns)
for f in w1_files:
row = w1_files.index(f)
path, name = os.path.split(f)
print "Status Update:"
print "File: {0}".format(f)
print "Item: {0}/{1}".format(row, w1_files_length)
img = ogre.site_to_array(f)
#img = ogre.site_to_array(f)
#%%
#b = img[:, :, 0] # DAPI
#g = img[:, :, 1] # FITC
#r = img[:, :, 2] # TexasRed
b = plt.imread('20150421-BCP-WTC11-chip2-full_A03_s54_w1_TimePoint_1.TIF')
g = plt.imread('20150421-BCP-WTC11-chip2-full_A03_s54_w2_TimePoint_1.TIF')
r = np.zeros_like(b)
#%%
bc = ogre.correct_illumination(b)
bcmask = ogre.adapt_thresh(bc)
blabs, bprops = ogre.watershed_label(bcmask, bc)
gc = ogre.correct_illumination(g)
gcmask = ogre.adapt_thresh(gc)
glabs, gprops = ogre.watershed_label(gcmask, gc)
rc = ogre.correct_illumination(r)
rcmask = ogre.adapt_thresh(rc)
rlabs, rprops = ogre.watershed_label(rcmask, rc)
#%%
bmask = (blabs != 0)
gmask = (glabs != 0)
def liveDead(row, f):
path, name = os.path.split(f)
img = ogre.site_to_array(f)
b = img[:, :, 0] # DAPI
g = img[:, :, 1] # FITC
r = img[:, :, 2] # TexasRed
bc = ogre.correct_illumination(b)
bcmask = ogre.adapt_thresh(bc)
blabs, bprops = ogre.watershed_label(bcmask, bc)
gc = ogre.correct_illumination(g)
gcmask = ogre.adapt_thresh(gc)
glabs, gprops = ogre.watershed_label(gcmask, gc)
rc = ogre.correct_illumination(r)
rcmask = ogre.adapt_thresh(rc)
rlabs, rprops = ogre.watershed_label(rcmask, rc)
# fig, ax = plt.subplots(3,3, figsize=(15,12))
# fig.suptitle(f, fontsize=15)
# ax[0,0].imshow(bc, cmap=plt.cm.gray)
# ax[0,1].imshow(gc, cmap=plt.cm.gray)
# ax[0,2].imshow(rc, cmap=plt.cm.gray)
# ax[1,0].imshow(bcmask, cmap=plt.cm.gray)
# ax[1,1].imshow(gcmask, cmap=plt.cm.gray)
# ax[1,2].imshow(rcmask, cmap=plt.cm.gray)
# ax[2,0].imshow(label2rgb(blabs))
# ax[2,1].imshow(label2rgb(glabs))
# ax[2,2].imshow(label2rgb(rlabs))
# sns.despine()
# fig.tight_layout()
# Raw cell counter seems to be doing a better job here...
raw_dapi = ogre.count_raw(bprops)
# bcount_filter = ogre.count_filter(bprops)
raw_fitc = ogre.count_raw(gprops)
# gcount_filter = ogre.count_filter(gprops)
raw_texas = ogre.count_raw(rprops)
# rcount_filter = ogre.count_filter(rprops)
glabmask = glabs != 0
rlabmask = rlabs != 0
live = 0
dead = 0
dying = 0
for i in range(1, blabs.max()):
nucleus_roi = blabs == i
if((nucleus_roi * glabmask * rlabmask).any()):
dying += 1
elif((nucleus_roi*glabmask).any()):
live += 1
elif((nucleus_roi * rlabmask).any()):
dead += 1
return (row, [path, name, raw_dapi, raw_fitc, raw_texas, live, dead, dying])
live += 1
elif((nucleus_roi * rlabmask).any()):
dead += 1
return (row, [path, name, raw_dapi, raw_fitc, raw_texas, live, dead, dying])
if __name__ == '__main__':
# - MAIN - #
platepath = 'B:\\Projects\\Research\\2015-05-11 AlyssaScreen1\\20150511-AlyssaC2D-D3-LD_Plate_4714'
output_filename = '20150511-AlyssaC2D-D3-LD_Plate_4714.csv'
output_filepath = os.path.join(platepath, output_filename)
w1_files = ogre.iter_plate(platepath)
w1_files_length = len(w1_files)
columns = ["path", "file", "raw_dapi", "raw_fitc", "raw_texas", "live", "dead", "all"]
df = pd.DataFrame(index=range(0, w1_files_length), columns=columns)
# Parallel for loop equivalent to "for f in w1_files:"
num_cores = multiprocessing.cpu_count()
results = Parallel(n_jobs=num_cores, verbose=11)(delayed(liveDead)(row, f) for row, f in zip(range(w1_files_length), w1_files))
for result in results:
df.iloc[result[0]] = result[1]
df.to_csv(output_filepath)
