def convert_to_sam(self, regions, output):
if (self.verbosity == "verbose"):
print " - Converting to SAM: " + output
if (output == "-"):
fh = sys.stdout
else:
fh = open(output, "w")
i = 0
# 1: write header
fh.write("@HD VN:1.0 SO:unsorted\n")
for region in regions:
if (self.input_format == 'bam'):
aligned_reads = BAMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
elif (self.input_format == 'sslm'):
aligned_reads = SSLMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
iterator = aligned_reads.parse_reads()
if (next(iterator, None)):
fh.write("@SQ SN:" + region[0] + " LN:" +
str(region[2] - region[1] + 1) + "\n")
del (iterator, aligned_reads)
fh.write("@PG ID:0 PN:manual_conversion_script VN:0.0\n")
# 2: write alignment
for region in regions:
if (self.verbosity == "verbose"):
print " - Masked region: " + region[0] + ":" + str(
region[1]) + "-" + str(region[2])
if (self.input_format == 'bam'):
aligned_reads = BAMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
elif (self.input_format == 'sslm'):
aligned_reads = SSLMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
for read in aligned_reads.parse_reads():
if (read.name):
fh.write(read.name)
else:
fh.write("unknown_read_" + str(i))
i += 1
strand = "60"
fh.write("\t0\t" + region[0] + "\t" + str(read.start + 1) +
"\t" + strand + "\t" + str(read.stop - read.start) +
"M\t*\t0\t0\t" + read.sequence + "\t*\tNH:i:1\n")
fh.close()
def run(self, regions, fasta_file):
if (self.verbosity == "verbose"):
print " - Running fragment detection"
self.fasta_file = fasta_file
for region in regions:
if (self.verbosity == "verbose"):
print " - Masked region: " + region[0] + ":" + str(
region[1]) + "-" + str(region[2])
print " * Acquiring statistics"
if (self.input_format == 'bam'):
aligned_reads = BAMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
elif (self.input_format == 'sslm'):
aligned_reads = SSLMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
aligned_reads.parse_stats()
if (self.verbosity == "verbose"):
print " * Detecting fragments"
predicted_fragments = FragmentFinder(region, aligned_reads)
self.add_fragments(predicted_fragments, self.fasta_file)
def test_01_a(self):
command = ['tar', '-xzf', '../share/small_RNA-seq_alignments/SRP028959/SRR954958.tar.gz']
subprocess.call(command)
args = CLI_sslm2sam(['-o', 'tmp/tests/test.sam', 'SRR954958'])
sslm2bed_converter = SSLMParser(args.sslm_directory)
sslm2bed_converter.convert_to_sam(args.output)
assertion = (os.stat("tmp/tests/test.sam").st_size == 46985661)
self.assertTrue(assertion, "Incorrect ../share/small_RNA-seq_alignments/SRP028959/test.sam") # Assume file size is sufficient :)
if assertion:
os.remove("tmp/tests/test.sam")
os.remove("SRR954958.bam")
os.remove("SRR954958.bam.bai")
shutil.rmtree("SRR954958")
def convert_to_bed(self, regions, output):
if (self.verbosity == "verbose"):
print " - Converting to BED: " + output
if (output == "-"):
fh = sys.stdout
else:
fh = open(output, "w")
i = 0
for region in regions:
if (self.verbosity == "verbose"):
print " - Masked region: " + region[0] + ":" + str(
region[1]) + "-" + str(region[2])
if (self.input_format == 'bam'):
aligned_reads = BAMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
elif (self.input_format == 'sslm'):
aligned_reads = SSLMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
for read_stacked in aligned_reads.parse_reads_stacked():
read = read_stacked[0]
numberofhits = read_stacked[1]
if (read.name):
fh.write(region[0] + "\t" + str(read.start) + "\t" +
str(read.stop) + "\t" + read.name + "\t" +
str(numberofhits) + "\t-\n")
else:
fh.write(region[0] + "\t" + str(read.start) + "\t" +
str(read.stop) + "\tunknown_read_" + str(i) +
"\t" + str(numberofhits) + "\t-\n")
i += 1
fh.close()
def count_error_with_intensity(self,
regions,
links,
masked_regions,
reference_offset=0):
"""
All sequences in our library of ncRNAs have been extended with 10 bases.
"""
out = []
if (self.verbosity == "verbose"):
print " - Running fragment detection"
for region in masked_regions:
ncRNA = region[0]
if (links.has_key(ncRNA)):
if (self.verbosity == "verbose"):
print " - Analysing: " + ncRNA
annotations = regions.index[links[ncRNA]]
if (self.input_format == 'bam'):
aligned_reads = BAMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
elif (self.input_format == 'sslm'):
aligned_reads = SSLMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
aligned_reads.parse_stats()
predicted_fragments_obj = FragmentFinder(ncRNA, aligned_reads)
predicted_fragments_obj.run()
predicted_fragments = predicted_fragments_obj.getResults()
aligned_reads.count_reads_per_region(
predicted_fragments_obj.getResults())
for mirna_annotation in annotations.fragments:
closest_fragment = self.find_closest_overlapping_fragment(
mirna_annotation, predicted_fragments,
reference_offset)
if (closest_fragment):
errors = self.find_errors(mirna_annotation, [
(closest_fragment.start - reference_offset),
(closest_fragment.stop - reference_offset)
]) #@todo ,reference_offset
err_5p = errors[0]
err_3p = errors[1]
#out.append({'5p':[closest_fragment[2],err_5p],'3p':[closest_fragment[3],err_3p]})
out.append({
'5p':
[closest_fragment.supporting_reads_start, err_5p],
'3p':
[closest_fragment.supporting_reads_stop, err_3p],
'coverage':
closest_fragment.supporting_reads
})
return out
def count_reads_per_region(self,
regions,
links,
masked_regions,
reference_offset=0):
"""
All sequences in our library of ncRNAs have been extended with 10 bases.
"""
stats_table = {}
stats_table['experimental'] = {
'error_5p': {
"<-5": 0,
-5: 0,
-4: 0,
-3: 0,
-2: 0,
-1: 0,
0: 0,
1: 0,
2: 0,
3: 0,
4: 0,
5: 0,
">5": 0
},
'error_3p': {
"<-5": 0,
-5: 0,
-4: 0,
-3: 0,
-2: 0,
-1: 0,
0: 0,
1: 0,
2: 0,
3: 0,
4: 0,
5: 0,
">5": 0
},
'predicted': 0,
'not_predicted_no_reads': 0,
'not_predicted_with_reads': 0
}
stats_table['not_experimental'] = {
'error_5p': {
"<-5": 0,
-5: 0,
-4: 0,
-3: 0,
-2: 0,
-1: 0,
0: 0,
1: 0,
2: 0,
3: 0,
4: 0,
5: 0,
">5": 0
},
'error_3p': {
"<-5": 0,
-5: 0,
-4: 0,
-3: 0,
-2: 0,
-1: 0,
0: 0,
1: 0,
2: 0,
3: 0,
4: 0,
5: 0,
">5": 0
},
'predicted': 0,
'not_predicted_no_reads': 0,
'not_predicted_with_reads': 0
}
if (self.verbosity == "verbose"):
print " - Running fragment detection"
i = 0
j = 0
#for ncRNA in self.alignment_directories_indexed.keys():
for region in masked_regions:
ncRNA = region[0]
if (links.has_key(ncRNA)):
if (self.verbosity == "verbose"):
print " - Analysing: " + ncRNA
annotations = regions.index[links[ncRNA]]
if (self.input_format == 'bam'):
aligned_reads = BAMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
elif (self.input_format == 'sslm'):
aligned_reads = SSLMParser(region[0], region[1], region[2],
self.alignments, self.verbosity)
aligned_reads.parse_stats()
predicted_fragments_obj = FragmentFinder(ncRNA, aligned_reads)
predicted_fragments_obj.run()
predicted_fragments = predicted_fragments_obj.results
i += 1
for annotation in annotations.fragments:
closest = self.find_closest_overlapping_fragment(
annotation, predicted_fragments, reference_offset)
j += 1
if (closest):
errors = self.find_errors(annotation, closest,
reference_offset)
err_5p = errors[0]
err_3p = errors[1]
if (err_5p > 5):
err_5p = ">5"
elif (err_5p < -5):
err_5p = "<-5"
if (err_3p > 5):
err_3p = ">5"
elif (err_3p < -5):
err_3p = "<-5"
if (annotation.evidence == "experimental"):
stats_table['experimental']["predicted"] += 1
stats_table['experimental']["error_5p"][
err_5p] += 1
stats_table['experimental']["error_3p"][
err_3p] += 1
else:
stats_table['not_experimental']["predicted"] += 1
stats_table['not_experimental']["error_5p"][
err_5p] += 1
stats_table['not_experimental']["error_3p"][
err_3p] += 1
else:
if (annotation.evidence == "experimental"):
if (annotation.get_supporting_reads() == 0):
stats_table['experimental'][
"not_predicted_no_reads"] += 1
else:
stats_table['experimental'][
"not_predicted_with_reads"] += 1
else:
if (annotation.get_supporting_reads() == 0):
stats_table['not_experimental'][
"not_predicted_no_reads"] += 1
else:
stats_table['not_experimental'][
"not_predicted_with_reads"] += 1
print i, "annotated pre-miRNAs"
print j, "annotated miRNAs"
return stats_table