def testSequenceScoringWithStream(self):
Kahn2014Rgn = nc.loadGFF('tutorial/Kahn2014.GFF')
Kahn2014Seq = Kahn2014Rgn \
.recenter(3000) \
.extract(genome)
tneg = MC.generateSet(n=len(Kahn2014Seq), length=3000)
trainingSet = Kahn2014Seq.label(nc.positive) + tneg.label(nc.negative)
PyPREdictor = nc.motifs.Ringrose2003GTGT() \
.pairFreq(distCut = 219) \
.model(nc.logOdds(
labelPositive = nc.positive,
labelNegative = nc.negative)
) \
.sequenceModel(name = 'PyPREdictor (M2003+GTGT)',
windowSize = 500, windowStep = 250) \
.train(trainingSet)
PyPREdictor.batchsize = 100
MC.generateSet(n=1000,
length=3000).saveFASTA('./temp/test.Background.fasta')
seq1 = nc.streamFASTA('./temp/test.Background.fasta')
seq2 = nc.loadFASTA('./temp/test.Background.fasta')
score1 = PyPREdictor.getSequenceScores(seq1, nStreamFetch=1000)
score2 = PyPREdictor.getSequenceScores(seq2, nStreamFetch=1000)
diff = sum(abs(b - a) for a, b in zip(score1, score2))
assert (diff == 0.0)
assert (len(score1) == len(score2))
def testExtractRegionSequences(self):
testSeqs.saveFASTA('temp/test.fasta')
seqByName = {seq.name: seq.seq for seq in testSeqs}
# Ensure that saving and re-loading sequences yields identical sequences
seqs = rsA.extractSequences(nc.loadFASTA('temp/test.fasta'))
self.assertEqual([s.seq for s in seqs],
[seqByName[r.seq][r.start:r.end + 1] for r in rsA])
seqs = rsA.extractSequences(nc.streamFASTA('temp/test.fasta'))
self.assertEqual([s.seq for s in seqs],
[seqByName[r.seq][r.start:r.end + 1] for r in rsA])
# Ensure that streaming of short blocks yields identical final sequences
seqs = rsA.extractSequences(
nc.streamFASTA('temp/test.fasta', wantBlockSize=50))
self.assertEqual([s.seq for s in seqs],
[seqByName[r.seq][r.start:r.end + 1] for r in rsA])
def testSaveLoadFASTA(self):
testSeqs.saveFASTA('temp/test.fasta')
sA = '\n'.join('%s: %s' % (s.name, s.seq) for s in testSeqs)
sB = '\n'.join('%s: %s' % (s.name.split(' from FASTA file')[0], s.seq)
for s in nc.loadFASTA('temp/test.fasta'))
self.assertEqual(sA, sB)