def main(args=None):
args = process_args(args)
if not args.plotFile and not args.outRawCounts and not args.outQualityMetrics:
sys.stderr.write("\nAt least one of --plotFile, --outRawCounts or --outQualityMetrics is required.\n")
sys.exit(1)
cr = sumR.SumCoveragePerBin(
args.bamfiles,
args.binSize,
args.numberOfSamples,
blackListFileName=args.blackListFileName,
numberOfProcessors=args.numberOfProcessors,
verbose=args.verbose,
region=args.region,
extendReads=args.extendReads,
minMappingQuality=args.minMappingQuality,
ignoreDuplicates=args.ignoreDuplicates,
center_read=args.centerReads,
samFlag_include=args.samFlagInclude,
samFlag_exclude=args.samFlagExclude,
minFragmentLength=args.minFragmentLength,
maxFragmentLength=args.maxFragmentLength)
num_reads_per_bin = cr.run()
if num_reads_per_bin.sum() == 0:
import sys
sys.stderr.write(
"\nNo reads were found in {} regions sampled. Check that the\n"
"min mapping quality is not overly high and that the \n"
"chromosome names between bam files are consistant.\n"
"\n".format(num_reads_per_bin.shape[0]))
exit(1)
if args.skipZeros:
num_reads_per_bin = countR.remove_row_of_zeros(num_reads_per_bin)
total = len(num_reads_per_bin[:, 0])
x = np.arange(total).astype('float') / total # normalize from 0 to 1
if args.plotFile is not None:
i = 0
# matplotlib won't iterate through line styles by itself
pyplot_line_styles = sum([7 * ["-"], 7 * ["--"], 7 * ["-."], 7 * [":"]], [])
plotly_colors = ["#d73027", "#fc8d59", "#f33090", "#e0f3f8", "#91bfdb", "#4575b4"]
plotly_line_styles = sum([6 * ["solid"], 6 * ["dot"], 6 * ["dash"], 6 * ["longdash"], 6 * ["dashdot"], 6 * ["longdashdot"]], [])
data = []
for i, reads in enumerate(num_reads_per_bin.T):
count = np.cumsum(np.sort(reads))
count = count / count[-1] # to normalize y from 0 to 1
if args.plotFileFormat == 'plotly':
trace = go.Scatter(x=x, y=count, mode='lines', name=args.labels[i])
trace['line'].update(dash=plotly_line_styles[i % 36], color=plotly_colors[i % 6])
data.append(trace)
else:
j = i % 35
plt.plot(x, count, label=args.labels[i], linestyle=pyplot_line_styles[j])
plt.xlabel('rank')
plt.ylabel('fraction w.r.t. bin with highest coverage')
# set the plotFileFormat explicitly to None to trigger the
# format from the file-extension
if not args.plotFileFormat:
args.plotFileFormat = None
if args.plotFileFormat == 'plotly':
fig = go.Figure()
fig['data'] = data
fig['layout'].update(title=args.plotTitle)
fig['layout']['xaxis1'].update(title="rank")
fig['layout']['yaxis1'].update(title="fraction w.r.t bin with highest coverage")
py.plot(fig, filename=args.plotFile, auto_open=False)
else:
plt.legend(loc='upper left')
plt.suptitle(args.plotTitle)
plt.savefig(args.plotFile, bbox_inches=0, format=args.plotFileFormat)
plt.close()
if args.outRawCounts is not None:
of = open(args.outRawCounts, "w")
of.write("#plotFingerprint --outRawCounts\n")
of.write("'" + "'\t'".join(args.labels) + "'\n")
fmt = "\t".join(np.repeat('%d', num_reads_per_bin.shape[1])) + "\n"
for row in num_reads_per_bin:
of.write(fmt % tuple(row))
of.close()
if args.outQualityMetrics is not None:
of = open(args.outQualityMetrics, "w")
of.write("Sample\tAUC\tSynthetic AUC\tX-intercept\tSynthetic X-intercept\tElbow Point\tSynthetic Elbow Point")
if args.JSDsample:
of.write("\tJS Distance\tSynthetic JS Distance\t% genome enriched\tdiff. enrichment\tCHANCE divergence")
else:
of.write("\tSynthetic JS Distance")
of.write("\n")
line = np.arange(num_reads_per_bin.shape[0]) / float(num_reads_per_bin.shape[0] - 1)
for idx, reads in enumerate(num_reads_per_bin.T):
counts = np.cumsum(np.sort(reads))
counts = counts / float(counts[-1])
AUC = np.sum(counts) / float(len(counts))
XInt = (np.argmax(counts > 0) + 1) / float(counts.shape[0])
elbow = (np.argmax(line - counts) + 1) / float(counts.shape[0])
expected = getExpected(np.mean(reads)) # A tuple of expected (AUC, XInt, elbow)
of.write("{0}\t{1}\t{2}\t{3}\t{4}\t{5}\t{6}".format(args.labels[idx], AUC, expected[0], XInt, expected[1], elbow, expected[2]))
if args.JSDsample:
JSD = getJSD(args, idx, num_reads_per_bin)
syntheticJSD = getSyntheticJSD(num_reads_per_bin[:, idx])
CHANCE = getCHANCE(args, idx, num_reads_per_bin)
of.write("\t{0}\t{1}\t{2}\t{3}\t{4}".format(JSD, syntheticJSD, CHANCE[0], CHANCE[1], CHANCE[2]))
else:
syntheticJSD = getSyntheticJSD(num_reads_per_bin[:, idx])
of.write("\t{0}".format(syntheticJSD))
of.write("\n")
of.close()
def main(args=None):
args = process_args(args)
if not args.plotFile and not args.outRawCounts and not args.outQualityMetrics:
sys.stderr.write(
"\nAt least one of --plotFile, --outRawCounts or --outQualityMetrics is required.\n"
)
sys.exit(1)
cr = sumR.SumCoveragePerBin(args.bamfiles,
args.binSize,
args.numberOfSamples,
blackListFileName=args.blackListFileName,
numberOfProcessors=args.numberOfProcessors,
verbose=args.verbose,
region=args.region,
extendReads=args.extendReads,
minMappingQuality=args.minMappingQuality,
ignoreDuplicates=args.ignoreDuplicates,
center_read=args.centerReads,
samFlag_include=args.samFlagInclude,
samFlag_exclude=args.samFlagExclude,
minFragmentLength=args.minFragmentLength,
maxFragmentLength=args.maxFragmentLength)
num_reads_per_bin = cr.run()
if num_reads_per_bin.sum() == 0:
import sys
sys.stderr.write(
"\nNo reads were found in {} regions sampled. Check that the\n"
"min mapping quality is not overly high and that the \n"
"chromosome names between bam files are consistant.\n"
"\n".format(num_reads_per_bin.shape[0]))
exit(1)
if args.skipZeros:
num_reads_per_bin = countR.remove_row_of_zeros(num_reads_per_bin)
total = len(num_reads_per_bin[:, 0])
x = np.arange(total).astype('float') / total # normalize from 0 to 1
if args.plotFile:
i = 0
# matplotlib won't iterate through line styles by itself
pyplot_line_styles = sum(
[7 * ["-"], 7 * ["--"], 7 * ["-."], 7 * [":"]], [])
for i, reads in enumerate(num_reads_per_bin.T):
count = np.cumsum(np.sort(reads))
count = count / count[-1] # to normalize y from 0 to 1
j = i % 35
plt.plot(x,
count,
label=args.labels[i],
linestyle=pyplot_line_styles[j])
plt.xlabel('rank')
plt.ylabel('fraction w.r.t. bin with highest coverage')
plt.legend(loc='upper left')
plt.suptitle(args.plotTitle)
# set the plotFileFormat explicitly to None to trigger the
# format from the file-extension
if not args.plotFileFormat:
args.plotFileFormat = None
plt.savefig(args.plotFile.name,
bbox_inches=0,
format=args.plotFileFormat)
plt.close()
if args.outRawCounts:
args.outRawCounts.write("'" + "'\t'".join(args.labels) + "'\n")
fmt = "\t".join(np.repeat('%d', num_reads_per_bin.shape[1])) + "\n"
for row in num_reads_per_bin:
args.outRawCounts.write(fmt % tuple(row))
args.outRawCounts.close()
if args.outQualityMetrics:
args.outQualityMetrics.write(
"Sample\tAUC\tSynthetic AUC\tX-intercept\tSynthetic X-intercept\tElbow Point\tSynthetic Elbow Point"
)
if args.JSDsample:
args.outQualityMetrics.write(
"\tJS Distance\tSynthetic JS Distance\t% genome enriched\tdiff. enrichment\tCHANCE divergence"
)
args.outQualityMetrics.write("\n")
line = np.arange(
num_reads_per_bin.shape[0]) / float(num_reads_per_bin.shape[0] - 1)
for idx, reads in enumerate(num_reads_per_bin.T):
counts = np.cumsum(np.sort(reads))
counts = counts / float(counts[-1])
AUC = np.sum(counts) / float(len(counts))
XInt = (np.argmax(counts > 0) + 1) / float(counts.shape[0])
elbow = (np.argmax(line - counts) + 1) / float(counts.shape[0])
expected = getExpected(
np.mean(reads)) # A tuple of expected (AUC, XInt, elbow)
args.outQualityMetrics.write(
"{0}\t{1}\t{2}\t{3}\t{4}\t{5}\t{6}".format(
args.labels[idx], AUC, expected[0], XInt, expected[1],
elbow, expected[2]))
if args.JSDsample:
JSD = getJSD(args, idx, num_reads_per_bin)
syntheticJSD = getSyntheticJSD(num_reads_per_bin[:, idx])
CHANCE = getCHANCE(args, idx, num_reads_per_bin)
args.outQualityMetrics.write(
"\t{0}\t{1}\t{2}\t{3}\t{4}".format(JSD, syntheticJSD,
CHANCE[0], CHANCE[1],
CHANCE[2]))
args.outQualityMetrics.write("\n")
args.outQualityMetrics.close()