def test_fragend_pairs(self):
''' Test identification of acceptable fragend pairs. '''
pairData = fragendPair.fragendPairs(pairIn=self.pairIn,
fasta=self.fastaIn,
resite='gATc',
maxDistance=20,
fragendOut=self.fragendOut)
output = []
with open(self.fragendOut) as fileIn:
for line in fileIn:
output.append(line.strip().split('\t'))
self.assertEqual(output, [['chr1', '10', '+', 'chr3', '32', '+'],
['chr2', '20', '-', 'chr2', '16', '-'],
['chr3', '32', '+', 'chr3', '3', '-']])
self.assertEqual(
pairData,
collections.defaultdict(
int, {
'total': 6,
'none': 2,
'distant': 1,
'intrachromosomal': 2,
'interchromosomal': 1,
'fragDist': [8, 8, 8, 21, 8, 11, 8, 8],
'ligDist': [4, 29]
}))
def test_fragend_pairs(self):
''' Test identification of acceptable fragend pairs. '''
pairData = fragendPair.fragendPairs(pairIn = self.pairIn,
fasta = self.fastaIn, resite = 'gATc', maxDistance = 20,
fragendOut = self.fragendOut)
output = []
with open(self.fragendOut) as fileIn:
for line in fileIn:
output.append(line.strip().split('\t'))
self.assertEqual(
output,
[['chr1','10','+','chr3','32','+'],
['chr2','20','-','chr2','16','-'],
['chr3','32','+','chr3','3','-']]
)
self.assertEqual(
pairData,
collections.defaultdict(int, {'total':6, 'none':2, 'distant':1,
'intrachromosomal':2, 'interchromosomal':1,
'fragDist':[8,8,8,21,8,11,8,8], 'ligDist':[4,29]})
)
'unique: %s' %(pairMetrics['unique']),
'discordant: %s' %(pairMetrics['discord']),
'unique discordant: %s' %(pairMetrics['discorduni']),
'duplication rate: %.4f' %(1 - (pairMetrics['unique'] /
float(pairMetrics['total']))),
'concordant rate: %.4f' %(1 - (pairMetrics['discorduni'] /
float(pairMetrics['unique'])))
)
##############################################################################
## Extract fragend pairs
##############################################################################
fragendMetrics = fragendPair.fragendPairs(
pairIn = args.outPairs,
fragendOut = args.outFrags,
fasta = args.bwaFasta,
maxDistance = args.maxDistance,
resite = args.cutSite
)
# Print fragend metrics
print '\nFragend Data:\n\t%s\n\t%s\n\t%s\n\t%s\n\t%s' %(
'total: %s' %(fragendMetrics['total']),
'no fragend: %s' %(fragendMetrics['none']),
'too distant: %s' %(fragendMetrics['distant']),
'interchromosomal: %s' %(fragendMetrics['interchromosomal']),
'intrachromosomal: %s' %(fragendMetrics['intrachromosomal'])
)
# Extract fragned and ligation metrics
fragendDist = fragendMetrics['fragDist']
ligationDist = fragendMetrics['ligDist']
fragendCount = len(fragendDist)