def run_hmmtop(self):
ht = hmmtop.tools()
ht.add_library('res', self.master_seq_file)
ht.scan_libraries()
self.tms = ht.results
self.hmmtop = []
gap_seqs = {}
for i in SeqIO.parse('hmmtop.in', 'fasta'):
gap_seqs[i.id] = str(i.seq)
for i, j in ht.results['res'].items():
for k in j.values():
korrect = [
self.get_pos(k[0], gap_seqs[i]),
self.get_pos(k[1], gap_seqs[i])
]
point = range(korrect[0], korrect[1] + 1)
self.hmmtop.append(median(point))
graph = {}
for i in range(1, len(self.master_seq[0]) + 1):
graph[i] = 0
for tms in self.hmmtop:
graph[i] += 1
for key, hits in graph.items():
ratio = Decimal(hits) / Decimal(len(self.master_seq))
self.tms_ratio[key] = ratio
return True
def hmmtop(self):
db = self.indir + '/hmmtop.db'
if os.path.exists(db) and self.debug:
self.tms = pickle.load(open(db, 'r'))
return
ht = hmmtop.tools()
ht.add_library('queries', self.indir + "/myqueries.faa") # Genome
ht.add_library('tcdb', self.indir + "/mytcdb.faa")
ht.scan_libraries()
pickle.dump(ht.results, open(db, 'wb'))
self.tms = ht.results
return
def __call__(self):
if os.path.exists(self.outdir) is False:
os.mkdir(self.outdir)
self.tssearch()
# Swap subjects & targets if needed - subject is always shorter
if self.swap:
(self.subject_file,self.target_file) = (self.target_file,self.subject_file)
(self.subject_name,self.target_name) = (self.target_name,self.subject_name)
(self.srestrict,self.trestrict) = (self.trestrict,self.srestrict)
# Set subject & target names
if self.subject_name is None:
self.subject_name = self.subject_file.split('/')[-1]
if self.target_name is None:
self.target_name = self.target_file.split('/')[-1]
# Write selected subjects & targets, then run HMMTOP
subject_ids = self.tssdata.keys()
target_ids = []
for i in self.tssdata.values():
target_ids.extend([x['target_symbol'] for x in i])
target_ids = list(set(target_ids))
subjects = SeqIO.parse(self.subject_file,'fasta')
self.subjects_dict = SeqIO.to_dict(subjects)
targets = SeqIO.parse(self.target_file,'fasta')
self.targets_dict = SeqIO.to_dict(targets)
with open(self.outdir+'/subjects.faa','wb+') as sf:
for sid in subject_ids:
SeqIO.write(self.subjects_dict[sid],sf,'fasta')
with open(self.outdir+'/targets.faa','wb+') as tf:
for tid in target_ids:
SeqIO.write(self.targets_dict[tid],tf,'fasta')
hmt = hmmtop.tools()
hmt.add_library('subjects',self.outdir+'/subjects.faa')
hmt.add_library('targets',self.outdir+'/targets.faa')
hmt.scan_libraries()
self.tms = hmt.results
# Run GSAT on all of these
self.run_global_alignments()
# Sort global data
self.globaldata.sort(key=lambda x:float(x['gsat_score']),reverse=True)
for ii,i in enumerate(self.globaldata):
try:
z = int(i['gsat_score'])
z = 0 if z<0 else z
except:
z = 0
self.globaldata[ii]['gsat_score'] = z
self.globaldata.sort(key=lambda x:int(x['gsat_score']),reverse=True)
# Generate HTML report
self.generate_report()
def __call__(self, seq, gap):
self.annotate = []
self.gap = str(gap)
try:
self.seq = SeqIO.read(seq, 'fasta')
except:
self.seq = seq
try:
if self.hmmtop is False:
hmt = hmmtop.tools()
hmt.add_library("gap", seq)
hmt.scan_libraries()
self.global_tms_ranges = hmt.results['gap'][self.seq.id]
else:
self.global_tms_ranges = self.hmmtop[self.seq.id]
except:
return None
self.parse_gap()
self.build_label()
return self.annotation
def __call__(self):
# First, run HMMTOP of course
ht = hmmtop.tools()
ht.add_library('subject', self.subject_file)
ht.add_library('target', self.target_file)
ht.scan_libraries()
# Now, build the TMS sequence
self.subject_loops = self.build_flanks(self.subject_file,
ht.results['subject'])
self.target_loops = self.build_flanks(self.target_file,
ht.results['target'])
# Glue the loops together and write file
if os.path.exists(self.outdir) is False:
os.mkdir(self.outdir)
self.write_fastas()
# Give names to each unit
if self.subjectname is None:
self.subjectname = self.mysubject_file.name.split('/')[-1]
if self.targetname is None:
self.targetname = self.mytarget_file.name.split('/')[-1]
self.build_target_matrix()
self.cross_matrix()
self.construct_report()
def calculate(fasta, label, out):
#os.environ['BIOV_DEBUG'] = 'True'
ht = hmmtop.tools()
ht.add_library('c', fasta)
ht.scan_libraries()
tms = ht.results['c']
total = open(fasta, 'r').read().count('>')
zeros = total - len(tms.keys())
tabs = {}
for i in tms.values():
count = len(i)
tabs.setdefault(count, 0)
tabs[count] += 1
tabs.setdefault(0, zeros)
biggest = tabs.keys()
biggest.sort()
biggest = biggest[-1]
[tabs.setdefault(i, 0) for i in range(biggest)]
xlocations = na.array(range(len(tabs.keys())))
plt.xticks(xlocations + 0.3 / 2, xlocations)
plt.bar(xlocations, tabs.values(), width=0.3)
plt.title(label)
plt.savefig(out, dpi=80, format="png")
return tabs, tms
#fig.frameon = False
#fig.add_axes([0,0,0,0],frameon=False)
fig.set_size_inches(width, 5.5)
self.outfile = title_in
plt.savefig(self.outfile,
dpi=80,
format="png",
bbox_inches='tight',
pad_inches=0.003)
else:
plt.show()
if __name__ == '__main__':
ht = hmmtop.tools()
ht.add_library('HT', fastafile.name)
ht.scan_libraries()
try:
tms = ht.results['HT']['what'].values()
except:
tms = []
eh = what(seq, window, angle)
eh.graph(tms, 'Hydropathy & Amphipathicity', True)
print '<HTML><head><title>Hydropathy & Amphipathicty plot</title></head><body>'
print "<img src ='%s'><br><br>" % eh.outfile
print '<font color="Blue">Blue lines denote Hydropathy</font><br>'
print '<font color="Red">Red lines denote Amphipathicity</font><br>'
print '<font color="#C5860D">Orange bars mark transmembrane segments as predicted by HMMTOP</font>'
print '</body></HTML>'