def main(args=None):
args = parse_arguments().parse_args(args)
corr = Correlation(
args.corData,
labels=args.labels,
)
args.plotFile.close()
corr.plot_pca(args.plotFile.name,
plot_title=args.plotTitle,
image_format=args.plotFileFormat)
if args.outFileNameData is not None:
import matplotlib
mlab_pca = matplotlib.mlab.PCA(corr.matrix)
n = len(corr.labels)
of = args.outFileNameData
of.write("Component\t{}\tEigenvalue\n".format("\t".join(corr.labels)))
for i in xrange(n):
of.write("{}".format(i + 1))
for v in mlab_pca.Wt[i, :]:
of.write("\t{}".format(v))
of.write("\t{}\n".format(mlab_pca.s[i]))
args.outFileNameData.close()
def main(args=None):
args = parse_arguments().parse_args(args)
corr = Correlation(args.corData,
labels=args.labels,)
args.plotFile.close()
corr.rowCenter = args.rowCenter
corr.plot_pca(args.plotFile.name,
plot_title=args.plotTitle,
image_format=args.plotFileFormat)
if args.outFileNameData is not None:
import matplotlib
mlab_pca = matplotlib.mlab.PCA(corr.matrix)
n = len(corr.labels)
of = args.outFileNameData
of.write("Component\t{}\tEigenvalue\n".format("\t".join(corr.labels)))
for i in range(n):
of.write("{}".format(i + 1))
for v in mlab_pca.Wt[i, :]:
of.write("\t{}".format(v))
of.write("\t{}\n".format(mlab_pca.s[i]))
args.outFileNameData.close()
def main(args=None):
args = parse_arguments().parse_args(args)
corr = Correlation(args.corData,
labels=args.labels,)
args.plotFile.close()
corr.plot_pca(args.plotFile.name,
plot_title=args.plotTitle,
image_format=args.plotFileFormat)
def main(args=None):
args = parse_arguments().parse_args(args)
if args.plotFile is None and args.outFileCorMatrix is None:
sys.exit("At least one of --plotFile and --outFileCorMatrix must be specified!\n")
corr = Correlation(args.corData,
args.corMethod,
labels=args.labels,
remove_outliers=args.removeOutliers,
skip_zeros=args.skipZeros)
if args.corMethod == 'pearson':
# test if there are outliers and write a message recommending the removal
if len(corr.get_outlier_indices(np.asarray(corr.matrix).flatten())) > 0:
if args.removeOutliers:
sys.stderr.write("\nOutliers were detected in the data. They "
"will be removed to avoid bias "
"in the pearson correlation.\n")
else:
sys.stderr.write("\nOutliers were detected in the data. Consider "
"using the --removeOutliers parameter to avoid a bias "
"in the pearson correlation.\n")
if args.colorMap:
try:
plt.get_cmap(args.colorMap)
except ValueError as error:
sys.stderr.write(
"A problem was found. Message: {}\n".format(error))
exit()
if args.plotFile is not None:
if args.whatToPlot == 'scatterplot':
corr.plot_scatter(args.plotFile,
plot_title=args.plotTitle,
image_format=args.plotFileFormat,
xRange=args.xRange,
yRange=args.yRange,
log1p=args.log1p)
else:
corr.plot_correlation(args.plotFile,
vmax=args.zMax,
vmin=args.zMin,
colormap=args.colorMap,
plot_title=args.plotTitle,
image_format=args.plotFileFormat,
plot_numbers=args.plotNumbers,
plotWidth=args.plotWidth,
plotHeight=args.plotHeight)
if args.outFileCorMatrix:
o = open(args.outFileCorMatrix, "w")
o.write("#plotCorrelation --outFileCorMatrix\n")
corr.save_corr_matrix(o)
o.close()
def main(args=None):
args = parse_arguments().parse_args(args)
if args.plotFile is None and args.outFileCorMatrix is None:
sys.exit("At least one of --plotFile and --outFileCorMatrix must be specified!\n")
corr = Correlation(args.corData,
args.corMethod,
labels=args.labels,
remove_outliers=args.removeOutliers,
skip_zeros=args.skipZeros)
if args.corMethod == 'pearson':
# test if there are outliers and write a message recommending the removal
if len(corr.get_outlier_indices(np.asarray(corr.matrix).flatten())) > 0:
if args.removeOutliers:
sys.stderr.write("\nOutliers were detected in the data. They "
"will be removed to avoid bias "
"in the pearson correlation.\n")
else:
sys.stderr.write("\nOutliers were detected in the data. Consider "
"using the --removeOutliers parameter to avoid a bias "
"in the pearson correlation.\n")
if args.colorMap:
try:
plt.get_cmap(args.colorMap)
except ValueError as error:
sys.stderr.write(
"A problem was found. Message: {}\n".format(error))
exit()
if args.plotFile is not None:
if args.whatToPlot == 'scatterplot':
corr.plot_scatter(args.plotFile,
plot_title=args.plotTitle,
image_format=args.plotFileFormat,
xRange=args.xRange,
yRange=args.yRange,
log1p=args.log1p)
else:
corr.plot_correlation(args.plotFile,
vmax=args.zMax,
vmin=args.zMin,
colormap=args.colorMap,
plot_title=args.plotTitle,
image_format=args.plotFileFormat,
plot_numbers=args.plotNumbers,
plotWidth=args.plotWidth,
plotHeight=args.plotHeight)
if args.outFileCorMatrix:
o = open(args.outFileCorMatrix, "w")
o.write("#plotCorrelation --outFileCorMatrix\n")
corr.save_corr_matrix(o)
o.close()
def main(args=None):
args = parse_arguments().parse_args(args)
if args.plotFile is None and args.outFileNameData is None:
sys.exit(
"At least one of --plotFile and --outFileNameData must be specified!\n"
)
if args.ntop < 0:
sys.exit("The value specified for --ntop must be >= 0!\n")
if args.PCs[0] == args.PCs[1]:
sys.exit("You must specify different principal components!\n")
if args.PCs[0] <= 0 or args.PCs[1] <= 0:
sys.exit("The specified principal components must be at least 1!\n")
corr = Correlation(
args.corData,
labels=args.labels,
)
corr.rowCenter = args.rowCenter
corr.transpose = args.transpose
corr.ntop = args.ntop
corr.log2 = args.log2
Wt, eigenvalues = corr.plot_pca(args.plotFile,
PCs=args.PCs,
plot_title=args.plotTitle,
image_format=args.plotFileFormat,
plotWidth=args.plotWidth,
plotHeight=args.plotHeight,
cols=args.colors,
marks=args.markers)
if args.outFileNameData is not None:
of = open(args.outFileNameData, "w")
of.write("#plotPCA --outFileNameData\n")
of.write("Component\t{}\tEigenvalue\n".format("\t".join(corr.labels)))
n = eigenvalues.shape[0]
for i in range(n):
of.write("{}\t{}\t{}\n".format(
i + 1, "\t".join(["{}".format(x) for x in Wt[i, :]]),
eigenvalues[i]))
of.close()
def main(args=None):
args = parse_arguments().parse_args(args)
corr = Correlation(
args.corData,
labels=args.labels,
)
if args.outFileCorMatrix:
corr.save_corr_matrix(args.outFileCorMatrix)
args.plotFile.close()
corr.plot_pca(args.plotFile.name,
plot_title=args.plotTitle,
image_format=args.plotFileFormat)
def main(args=None):
args = parse_arguments().parse_args(args)
if args.plotFile is None and args.outFileNameData is None:
sys.exit("At least one of --plotFile and --outFileNameData must be specified!\n")
if args.ntop < 0:
sys.exit("The value specified for --ntop must be >= 0!\n")
if args.PCs[0] == args.PCs[1]:
sys.exit("You must specify different principal components!\n")
if args.PCs[0] <= 0 or args.PCs[1] <= 0:
sys.exit("The specified principal components must be at least 1!\n")
corr = Correlation(args.corData,
labels=args.labels,)
corr.rowCenter = args.rowCenter
corr.transpose = args.transpose
corr.ntop = args.ntop
corr.log2 = args.log2
Wt, eigenvalues = corr.plot_pca(args.plotFile,
PCs=args.PCs,
plot_title=args.plotTitle,
image_format=args.plotFileFormat,
plotWidth=args.plotWidth,
plotHeight=args.plotHeight,
cols=args.colors,
marks=args.markers)
if args.outFileNameData is not None:
of = open(args.outFileNameData, "w")
of.write("#plotPCA --outFileNameData\n")
of.write("Component\t{}\tEigenvalue\n".format("\t".join(corr.labels)))
n = eigenvalues.shape[0]
for i in range(n):
of.write("{}\t{}\t{}\n".format(i + 1, "\t".join(["{}".format(x) for x in Wt[i, :]]), eigenvalues[i]))
of.close()
def get_labels_and_correlation(
bw_files,
# chrs_to_skip,
bin_size=10000,
method='pearson',
fileset_name='result',
blacklist=None,
labels=bw_labels,
output_dir=BASE_DIR):
my_listUnnested = []
# my_labels_list = []
assert method in ('pearson', 'spearman'), 'Invalid correlation method'
# Autogenerate labels from filename if not provided
if not labels:
labels = [
filename.split('/')[-1].split('.')[0] for filename in bw_files
]
# Generate a name for the unique combination
test_name = fileset_name + '_' + method
if blacklist:
blacklist_title = 'Blacklisted'
test_name += '_blacklisted'
else:
blacklist_title = ''
image_name = test_name + '.png'
# Bin the bigwig data in 10kb increments
num_reads_per_bin = score_bw.getScorePerBin(
bw_files,
bin_size,
# chrsToSkip=chrs_to_skip,
blackListFileName=blacklist)
# Write to npz file
print("right before npz")
os.system('pwd')
os.system('ls -lat')
print('svo')
print(output_dir)
print(test_name)
filename = output_dir + test_name + '.npz'
print(filename)
print('svo.. ')
with open(filename, "wb") as f:
np.savez_compressed(f, matrix=num_reads_per_bin, labels=labels)
# Compute the correlations
corr = Correlation(filename, method, labels=labels)
np_array = corr.compute_correlation()
print("ALERT CORR")
print(np_array)
listNested = np_array.tolist()
def removeNestings(listNest):
for i in listNest:
if type(i) == list:
print(str(i) + "is: " + str(type(i)))
removeNestings(i)
print(str(i) + "is after: " + str(type(i)))
else:
print(str(i) + "is finally " + str(type(i)))
my_listUnnested.append(i)
removeNestings(listNested)
print("FINAL CORRELATION VALUES")
print(my_listUnnested)
with open("corrScores.txt", "w") as f:
f.write(str(my_listUnnested))
plot_title = '{}{} Correlation of {}'.format(blacklist_title,
method.capitalize(),
fileset_name)
# Create a png file of correlation heatmap
image_path = output_dir + image_name
corr.plot_correlation(image_path, plot_title=plot_title)
# return np_ar
my_labels_list = labels
return image_path, my_labels_list, my_listUnnested
def get_labels_and_correlation(
bw_files,
# chrs_to_skip,
bin_size=10000,
method='pearson',
fileset_name='result',
blacklist=None,
labels=bw_labels,
output_dir='/Users/baditya02/Downloads/treatment-data/graphs/test/'
):
my_listUnnested = []
my_labels_list = []
assert method in ('pearson', 'spearman'), 'Invalid correlation method'
# Autogenerate labels from filename if not provided
if not labels:
labels = [filename.split( '/' )[-1].split( '.' )[0] for filename in bw_files]
# Generate a name for the unique combination
test_name = fileset_name + '_' + method
if blacklist:
blacklist_title = 'Blacklisted '
test_name += '_blacklisted'
else:
blacklist_title = ''
image_name = test_name + '.png'
# Bin the bigwig data in 10kb increments
num_reads_per_bin = score_bw.getScorePerBin(
bw_files,
bin_size,
# chrsToSkip=chrs_to_skip,
blackListFileName=blacklist
)
# Write to npz file
filename = output_dir + test_name + '.npz'
with open(filename, "wb") as f:
np.savez_compressed(f, matrix=num_reads_per_bin, labels=labels)
# Compute the correlations
corr = Correlation(filename, method, labels=labels)
np_array = corr.compute_correlation()
listNested = np_array.tolist()
def removeNestings(listNest):
for i in listNest:
if type(i) == list:
removeNestings(i)
else:
my_listUnnested.append(i)
removeNestings(listNested)
plot_title = '{}{} Correlation of {}'.format(
blacklist_title,
method.capitalize(),
fileset_name
)
# Create a png file of correlation heatmap
image_path = output_dir + image_name
corr.plot_correlation( image_path, plot_title=plot_title )
# return np_ar
my_labels_list = labels
return image_path, my_labels_list, my_listUnnested
