def setTranslatedForName(**kwargs):
global k
global translation
libname = kwargs.get('libname')
o = getKMERsForName( **mem.sr(kwargs, libname = libname))
translated = zeros((len(o),k))
idxed_mers = dict([(i,k) for i,k in enumerate(o.keys())])
occurrences=array([ o[idxed_mers[i]] for i in range(len(translated))])
d = translation
for i in idxed_mers.keys():
translated[i] = [d.get(l,4) for l in idxed_mers[i]]
return idxed_mers,translated, occurrences
def run(alltweets, freebase_type = 'band', **kwargs):
'''
yield a list of statistically significant tweets.
cross check it against a set of aliases generated by freebase.
called with a freebase type, it will call the fetch_type routine
provided by freebase.py in order to grab a list of aliases to match
against discovered terms.
'''
counts = count(tokenize(alltweets))
long_keys = set([k for k in counts.keys() if len(k)>=5])
freebase_aliases = fbu.fetch_type(freebase_type, **mem.sr(kwargs))
matched = [ a.lower() for a in freebase_aliases if (a.lower() in long_keys)]
raise Exception()
def setAllGenes(**kwargs):
allPeaks = getPeaks()
all_results = {}
#if you were running for a larger dataset you might want to
#break this loop after a single iteration and just choose a chromosome
for num in range(1,20) + ['X']:
print 'Parsing Chromosome: chr{0}'.format(num)
genes_dict = {}
all_results['chr{0}'.format(num)] = genes_dict
#get the genes on a chromosome
chrgenes = getTrackChrGenes(**mem.sr(kwargs, num = num))
#get the peaks on a chromosome
peaks = allPeaks['chr{0}'.format(num)]
for i, g in enumerate(chrgenes):
name = g['name']
startpos = g['start'] if g['strand'] == 1 else g['end']
hits = []
#list features near this gene.
for p in peaks:
stranded_offset =array([ g['strand'] * (p['start'] - startpos),
g['strand'] * (p['end'] - startpos)])
if( np.min(abs(stranded_offset)) < 2000 \
or np.prod(stranded_offset) < 0):
stranded_offset.sort()
hits.append({'peak_info':p,
'peak_stranded_offset':stranded_offset})
#store some extra information in the dictionary that we'll output
hits = sorted(hits,key = lambda x: x['peak_stranded_offset'][0])
gene_object = {
'dnase_peaks':hits,
'name':name,
'gene_info':g,
'start':g['start'],
'end':g['end'],
'strand':g['strand']
}
genes_dict[name] = gene_object
if (mod(i,100) == 0):
print 'Gene {0}: {1}, {2} hits'.format(i, g['name'], len(hits))
return all_results;
