def status_plot(bams):
'''
Brief: Generate histograms showing average length of MS region depending on MSI status
Args: lst
Return: none
'''
for locus in _MSI_LOCI:
msi_avgs = []
mss_avgs = []
for bam in bams:
bam_name = bam.split('/')[-1].replace('A.bam', '')
if bam_name in _ANNOTATIONS:
status = _ANNOTATIONS[bam_name]
if status != 'MSI' and status != 'MSS':
continue
average = avg_length(count_reads(bam, locus))
if average == 'Insufficient reads':
continue
if status == 'MSI':
msi_avgs.append(float(average))
elif status == 'MSS':
mss_avgs.append(float(average))
if len(msi_avgs) != 0 or len(mss_avgs) != 0:
plt.hist([msi_avgs, mss_avgs],
color=['red', 'blue'],
label=['MSI', 'MSS'])
plt.title('%s Distribution (subset)' % locus)
plt.legend(loc='best')
plt.xlabel = ('Average MS length (bp)')
plt.ylabel('Number of BAM files')
saveloc = '/home/upload/msi_project/status_corr_dist/subsetA/%s_dist.png' % locus
plt.savefig(saveloc)
plt.clf()
def print_mm_depth(bams, mismatch=2, length=7):
'''
Brief: used for optimizing flank length and mismatch parameters, prints edit distance and read depth to a file indicated below
Args: list, int, int
Returns: none
'''
outfile = '/home/upload/msi_project/mm_depth_analysis/subsetA-mismatch_depth-%d-%d.txt' % (
mismatch, length)
with open(outfile, 'w') as f:
f.write('#mismatches: %d, flank length: %d\n' % (mismatch, length))
f.write('\t')
for locus in _MSI_LOCI:
f.write(locus + '\t\t\t')
f.write('\n\t')
for i in range(len(_MSI_LOCI)):
f.write('f1 mm\tf2mm\t% accepted reads\t')
f.write('\n')
for bam in bams:
f.write(bam.split('/')[-1].replace('.bam', '') + '\t')
for locus in _MSI_LOCI:
accepted_reads, f1_mm, f2_mm, num_reads = count_reads(
bam,
locus,
flank_length=length,
flank_mismatch=mismatch,
return_mms=True)
if num_reads == 0:
percent_accepted = "no coverage"
else:
percent_accepted = float(
len(accepted_reads)) / float(num_reads) * 100
f.write(
str(f1_mm) + '\t' + str(f2_mm) + '\t' +
str(percent_accepted) + '\t')
f.write('\n')
def test_count_reads_count_reads(fn):
from count_reads import count_reads
res = count_reads(
fn,
(24, 44),
)
print(res.most_common(10))
def report_std_dev(bams, reporting_threshold=.9, mismatch=2, length=7):
'''
Brief: Reports to file the MSI status of a patient to compare with the known status, determining MSI status
based on standard deviation
Args: lst, int, int
Return: None
'''
outfile = '/home/upload/msi_project/diag_analysis/method_3/mss_training_set_statuses_stdev.txt'
with open(outfile, 'w') as f:
f.write('#mismatch: %s, flank length: %s, reporting_threshold: %s\n' %
(str(mismatch), str(length), str(reporting_threshold)))
f.write('locus\t')
for locus in _MSI_LOCI:
f.write(locus + '\t')
f.write('Average\tCall\tKnown status\n')
avg_stdevs = [] #average for each bamfile all loci
for bam in bams:
bam_name = bam.split('/')[-1].replace('A.bam', '')
f.write(bam_name + '\t')
locus_stdevs = []
for locus in _MSI_LOCI:
accepted_reads = (count_reads(bam,
locus,
flank_length=length,
flank_mismatch=mismatch))
if len(accepted_reads) == 0:
std_dev = 'n/a'
f.write('n/a\t')
else:
lengths = [len(e) for e in accepted_reads]
std_dev = np.std(lengths)
f.write(str(std_dev) + '\t')
locus_stdevs.append(std_dev)
bam_stdev = avg_value(locus_stdevs)
avg_stdevs.append(bam_stdev)
if len(locus_stdevs) == 0:
msi_status = 'Indeterminate'
else:
if bam_stdev < reporting_threshold:
msi_status = 'MSS'
else:
msi_status = 'MSI'
if bam_name in _ANNOTATIONS:
known_status = _ANNOTATIONS[bam_name]
else:
known_status = 'Not reported'
f.write(
str(bam_stdev) + '\t' + msi_status + '\t' + known_status +
'\n')
return avg_stdevs
def report_dist_mode(bams, mismatch=2, length=7):
'''
Brief: Reports to file the MSI status of a patient to compare with the known status, determining MSI status
based on absolute distance from the mode
Args: list, int, int
Returns: none
'''
outfile = '/home/upload/msi_project/diag_analysis/method_2/subsetA_statuses_mode.txt'
all_reads = []
modes = []
#Create 2D array of accepted read by locus and bam file
for bam in bams:
bam_reads = []
for locus in _MSI_LOCI:
accepted_reads = count_reads(bam,
locus,
flank_length=length,
flank_mismatch=mismatch)
bam_reads.append(accepted_reads)
all_reads.append(bam_reads)
#Generate a list of the mode length for each locus
for i in range(len(_MSI_LOCI)):
for j in range(len(all_reads)):
locus = []
locus.extend(all_reads[j][i])
mode = mode_length(locus)
modes.append(mode)
#find average distance from the mode for each bam each locus, average for all loci per bam, correlate with annotations
with open(outfile, 'w') as f:
f.write('BAM\n')
for i in range(len(all_reads)): #iterate over all bam files
bam_name = bams[i].split('/')[-1].replace('A.bam', '')
f.write(bam_name + '\t')
for j in range(len(modes)): #iterate over all loci
if modes[j] == 'error':
avg_distance = 'low loc covg'
elif len(all_reads[i][j]) == 0:
avg_distance = 'low bam covg'
else:
total_distance = 0
mode = modes[j]
for read in all_reads[i][j]:
total_distance += abs(float(mode) - len(read))
avg_distance = float(total_distance) / len(all_reads[i][j])
f.write(str(avg_distance) + '\t')
if bam_name in _ANNOTATIONS:
known_status = _ANNOTATIONS[bam_name]
else:
known_status = 'Not reported'
f.write(known_status + '\n')
def get_z_score(bamfile, locus, mismatch=2, length=7):
if float(_MSS_LOCUS_DATA[locus][1]) == 0:
return 'error'
accepted_reads = count_reads(bamfile,
locus,
flank_length=length,
flank_mismatch=mismatch)
if len(accepted_reads) == 0:
return 'error'
else:
lengths = [len(e) for e in accepted_reads]
std_dev = np.std(lengths)
z_score = ((float(_MSS_LOCUS_DATA[locus][0]) - float(std_dev)) /
float(_MSS_LOCUS_DATA[locus][1]))
z_score = abs(z_score)
return z_score
def report_num_lengths(bams, mismatch=2, length=7):
'''
Brief: Reports to file the MSI status of a patient to compare with the known status, determining MSI status
based on number of different lengths
Args: list, int, int
Returns: none
'''
outfile = '/home/upload/msi_project/diag_analysis/method_1/subsetA_statuses_length.txt'
with open(outfile, 'w') as f:
f.write('#mismatch: %s, flank length: %s\n' %
(str(mismatch), str(length)))
#f.write('BAM\tNUM DIF ELEMS\tSTATUS\tKNOWN STATUS\tAGREE?\n')
f.write('locus\t')
for locus in _MSI_LOCI:
f.write(locus + '\t')
f.write('\n')
for bam in bams:
status_marker = 0
bam_name = bam.split('/')[-1].replace('A.bam', '')
f.write(bam_name + '\t')
for locus in _MSI_LOCI:
accepted_reads = (count_reads(bam,
locus,
flank_length=length,
flank_mismatch=mismatch))
if len(accepted_reads) == 0:
f.write('n/a\t')
else:
f.write(str(len(set(accepted_reads))) + '\t')
status_marker += 1
msi_status = 'MSS'
if status_marker > 0:
msi_status = 'MSI'
if bam_name in _ANNOTATIONS:
known_status = _ANNOTATIONS[bam_name]
else:
known_status = 'Not reported'
agree = False
if msi_status == known_status:
agree == True
f.write(msi_status + '\t' + known_status + '\t' + str(agree) +
'\n')
def bw_plot(bams):
'''
Brief: Print a candlestick plot of the number of accepted reads at each locus
Args: lst, dict
Return: none
'''
data = []
label = []
for locus in _MSI_LOCI:
temp = []
label.append(locus)
for bam in bams:
runs, favg, bavg, num_reads = count_reads(bam,
locus,
return_mms=True)
temp.append(num_reads)
data.append(temp)
plt.boxplot(data, labels=label)
plt.xticks(rotation=90)
plt.title('Subset Read Depth')
plt.savefig('/home/upload/msi_project/subsetA_depth_plot.png')