def test_get_secondary_structure(self):
"""get_secondary_structure should give correct result."""
test_seq = 'AAACCCGGGUUU'
expected_struct = '((((....))))'
expected_energy = -0.80
obs_seq, obs_struct, obs_energy = \
get_secondary_structure(test_seq)
#Test get back correct seq and struct
self.assertEqual(obs_seq, test_seq)
self.assertEqual(obs_struct, expected_struct)
self.assertEqual(obs_energy, expected_energy)
def test_get_secondary_structure(self):
"""get_secondary_structure should give correct result."""
test_seq = 'AAACCCGGGUUU'
expected_struct = '((((....))))'
expected_energy = -0.80
obs_seq, obs_struct, obs_energy = \
get_secondary_structure(test_seq)
#Test get back correct seq and struct
self.assertEqual(obs_seq, test_seq)
self.assertEqual(obs_struct, expected_struct)
self.assertEqual(obs_energy, expected_energy)
def run_locarnap(seqsin, numkept, cpus=1,foldless=False):
'''Runs locarna-p on a set of sequences in MinimalFastaParser format
[(header, seq), (header, seq)] and returns alignment and structure'''
#make sure group has enough sequences before continuing
if len(seqsin) < numkept and not foldless:
return "", ""
if len(seqsin) == 1:
#raise ValueError("Need at least two sequences for locarna-p")
return LoadSeqs(data=seqsin, moltype=RNA), get_secondary_structure(seqsin[0][1])[1]
#headers come out in format Header_# so split to get # and sort by abundance
seqsin.sort(reverse=True, key=lambda count: int(count[0].split('_')[1]))
#cut to numkept most abundant sequences
if len(seqsin) > numkept:
seqsin = seqsin[:numkept]
aln, struct = create_locarnap_alignment(seqsin, RNA, struct=True, params={'--cpus': cpus})
struct = struct.replace('-', ".")
return aln, struct
def plot_rna_structure(seq, path='', subseqs=[], name='test'):
"""plot RNA structure using Vienna package"""
import cogent.app.vienna_package as vienna
colors = [" 1. 0. .2", " 0. .9 .5"]
seq, struct, e = vienna.get_secondary_structure(seq)
seqname = 'test'
rp = vienna.RNAplot()
i = 0
x = ''
if len(subseqs) > 0:
for s in subseqs:
ind = seq.find(s) + 1
e = ind + len(s)
x += format_cmark_values(range(ind, e), rgb=colors[i])
i += 1
rp.Parameters['--pre'].on('"%s"' % x)
rp(['>' + seqname, seq, struct])
filename = os.path.join(path, '%s.png' % name)
os.system('convert test_ss.ps %s' % filename)
return filename
def plot_rna_structure(seq, path='', subseqs=[], name='test'):
"""plot RNA structure using Vienna package"""
import cogent.app.vienna_package as vienna
colors = [" 1. 0. .2", " 0. .9 .5"]
seq,struct,e = vienna.get_secondary_structure(seq)
seqname='test'
rp = vienna.RNAplot()
i=0
x=''
if len(subseqs) > 0:
for s in subseqs:
ind = seq.find(s)+1
e = ind+len(s)
x += format_cmark_values(range(ind,e), rgb=colors[i])
i+=1
rp.Parameters['--pre'].on('"%s"' %x)
rp(['>'+seqname,seq,struct])
filename = os.path.join(path,'%s.png' %name)
os.system('convert test_ss.ps %s' %filename)
return filename
#Now need to iteratively refine the groups down
#check to make sure we need to first
if not skipiter:
startcount = 1
endcount = 0
iteration = 1
secs = time()
#try to match sequences in unfoldable clusters to other clusters that were folded
#write seq to orphanseqs file if can't be grouped
print len(nostruct), "of", count, "structureless"
orphanseqs = open(otufolder + "orphanseqs.txt", 'w')
orphanseqs.write("initial clustering\n")
for cluster in nostruct:
for seq in clusters[cluster]:
found = False
sequence, structure, energy = get_secondary_structure(seq[1])
for gstruct in structgroups:
if score_rnaforester(structure, gstruct) >= foresterscore:
structgroups[gstruct].append(seq)
found = True
break
if not found:
orphanseqs.write('>%s\n%s\n' % seq)
structgroups[structure] = [seq]
orphanseqs.close()
#wipe out clusters dict to save memory
clusters = 0
print "start: " + str(len(structgroups)) + " initial groups"
#initial clustering by structures generated in first folding
