def memefiles2tamo(files, tamoname):
global probefile, PROBESET, fsafile
motifs = []
for filename in files:
print ">>>SDFSD>F ",filename
if re.search('\.ace$',filename):
mdobject = AlignAce.AlignAce(filename)
if not mdobject.fastafile: mdobject.fastafile=filename.replace('.ace','.fsa')
elif re.search('\.meme.*$',filename):
mdobject = Meme.Meme(filename)
if not mdobject.fastafile:
mdobject.fastafile=re.sub('\..\.meme','.meme',filename).replace('.meme','.fsa')
motifs.extend(mdobject.motifs)
#fsaname = find_fsa(mdobject.fastafile)
print mdobject.fastafile
if fsafile: fsaname = fsafile
else: fsaname = Fasta.find(mdobject.fastafile)
fsaD = Fasta.load(fsaname)
probes = fsaD.keys()
if not probefile:
PROBESET = MotifMetrics.ProbeSet('YEAST')
#PROBESET= pick_genome(fsaname)
for key,seq in fsaD.items():
PROBESET.probes[key] = seq
for motif in motifs:
if motif.pvalue == 1: motif.pvalue = PROBESET.p_value(motif,probes,'v')
if motif.church == 1: motif.church = PROBESET.church(motif,probes,'v')
#if motif.E_site == None: motif.E_site = PROBESET.E_sitef(motif,probes,3,'v')
#if motif.E_chi2 == None: motif.E_chi2 = PROBESET.E_chi2(motif,probes,None,'v')
#if motif.E_seq == None: motif.E_seq = PROBESET.E_seq(motif,probes,'v')
if motif.ROC_auc== None: motif.ROC_auc= PROBESET.ROC_AUC(motif,probes,'v')
#if motif.MNCP == None: motif.MNCP = PROBESET.MNCP(motif,probes,'v')
if motif.frac == None: motif.frac = PROBESET.frac(motif,probes,'v',0.7)
if re.search('\.meme$',filename):
motif.MAP = -math.log(motif.evalue)/math.log(10)
if 0 and (motif.CRA == None):
try:
pass
CRA, Cfrac = PROBESET.cons_ROC_AUC(motif,probes,'v',tuple='YES')
motif.CRA = CRA
motif.Cfrac = Cfrac
except: pass
if re.search('\.meme$',filename):
mdobject.motifs.sort(lambda x,y: cmp(x.pvalue, y.pvalue))
else:
mdobject.motifs.sort(lambda x,y: cmp(x.church, y.church))
MotifTools.save_motifs(motifs,tamoname)
def tamo2tamo(file, outname):
global probefile, PROBESET, fsafile
motifs = MotifTools.load(file)
if fsafile:
fsaname = fsafile
else:
fsaname = find_fsa(file)
print '# FSA ', fsaname
fsaD = MotifMetrics.fasta2seqs(fsaname, 'want_dict')
probes = fsaD.keys()
if not probefile:
PROBESET = MotifMetrics.ProbeSet('YEAST')
#PROBESET= pick_genome(fsaname)
#for key,seq in fsaD.items():
# PROBESET.probes[key] = seq
print "# %d motifs" % len(motifs)
for motif in motifs:
#motif.pvalue, motif.church = 1,1 #Comment this!
if motif.pvalue == 1:
motif.pvalue = PROBESET.p_value(motif, probes, 'v')
if motif.church == 1:
motif.church = PROBESET.church(motif, probes, 'v')
#if motif.E_site == None: motif.E_site = PROBESET.E_sitef(motif,probes,3,'v')
#if motif.E_chi2 == None: motif.E_chi2 = PROBESET.E_chi2(motif,probes,None,'v')
#if motif.E_seq == None: motif.E_seq = PROBESET.E_seq(motif,probes,'v')
if motif.ROC_auc == None:
motif.ROC_auc = PROBESET.ROC_AUC(motif, probes, 'v')
#if motif.MNCP == None: motif.MNCP = PROBESET.MNCP(motif,probes,'v')
if motif.frac == None:
motif.frac = PROBESET.frac(motif, probes, 'v', 0.7)
if motif.numbound == 0:
matching = PROBESET.matching_ids(motif, [], factor=0.7)
matchbound = [x for x in matching if x in probes]
motif.numbound = len(probes)
motif.nummotif = len(matching)
motif.numboundmotif = len(matchbound)
if 0 and motif.CRA == None:
try:
pass
CRA, Cfrac = PROBESET.cons_ROC_AUC(motif,
probes,
'v',
tuple='YES')
motif.CRA = CRA
motif.Cfrac = Cfrac
except:
pass
MotifTools.save_motifs(motifs, outname)
def tamo2tamo(file, outname):
global probefile, PROBESET, fsafile
motifs = MotifTools.load(file)
if fsafile:
fsaname = fsafile
else:
fsaname = find_fsa(file)
print '# FSA ',fsaname
fsaD = MotifMetrics.fasta2seqs(fsaname,'want_dict')
probes = fsaD.keys()
if not probefile:
PROBESET = MotifMetrics.ProbeSet('YEAST')
#PROBESET= pick_genome(fsaname)
#for key,seq in fsaD.items():
# PROBESET.probes[key] = seq
print "# %d motifs"%len(motifs)
for motif in motifs:
#motif.pvalue, motif.church = 1,1 #Comment this!
if motif.pvalue == 1: motif.pvalue = PROBESET.p_value(motif,probes,'v')
if motif.church == 1: motif.church = PROBESET.church(motif,probes,'v')
#if motif.E_site == None: motif.E_site = PROBESET.E_sitef(motif,probes,3,'v')
#if motif.E_chi2 == None: motif.E_chi2 = PROBESET.E_chi2(motif,probes,None,'v')
#if motif.E_seq == None: motif.E_seq = PROBESET.E_seq(motif,probes,'v')
if motif.ROC_auc== None: motif.ROC_auc= PROBESET.ROC_AUC(motif,probes,'v')
#if motif.MNCP == None: motif.MNCP = PROBESET.MNCP(motif,probes,'v')
if motif.frac == None: motif.frac = PROBESET.frac(motif,probes,'v',0.7)
if motif.numbound == 0:
matching = PROBESET.matching_ids(motif,[],factor=0.7)
matchbound = [x for x in matching if x in probes]
motif.numbound = len(probes)
motif.nummotif = len(matching)
motif.numboundmotif = len(matchbound)
if 0 and motif.CRA == None:
try:
pass
CRA, Cfrac = PROBESET.cons_ROC_AUC(motif,probes,'v',tuple='YES')
motif.CRA = CRA
motif.Cfrac = Cfrac
except: pass
MotifTools.save_motifs(motifs,outname)
def ace2tamo(filename, tamoname):
global probefile, PROBESET
if re.search('\.ace$',filename):
mdobject = AlignAce.AlignAce(filename)
elif re.search('\.meme$',filename):
mdobject = Meme.Meme(filename)
fsaname = find_fsa(mdobject.fastafile)
fsaD = MotifMetrics.fasta2seqs(fsaname,'want_dict')
probes = fsaD.keys()
if not probefile:
PROBESET = MotifMetrics.ProbeSet('HUMAN_250')
#PROBESET= pick_genome(fsaname)
for key,seq in fsaD.items():
PROBESET.probes[key] = seq
for motif in mdobject.motifs:
if motif.pvalue == 1: motif.pvalue = PROBESET.p_value(motif,probes,'v')
if motif.church == 1: motif.church = PROBESET.church(motif,probes,'v')
if motif.E_site == None: motif.E_site = PROBESET.E_sitef(motif,probes,3,'v')
#if motif.E_chi2 == None: motif.E_chi2 = PROBESET.E_chi2(motif,probes,None,'v')
if motif.E_seq == None: motif.E_seq = PROBESET.E_seq(motif,probes,'v')
if motif.ROC_auc== None: motif.ROC_auc= PROBESET.ROC_AUC(motif,probes,'v')
if motif.MNCP == None: motif.MNCP = PROBESET.MNCP(motif,probes,'v')
if re.search('\.meme$',filename):
motif.MAP = -math.log(motif.evalue)/math.log(10)
sys.stdout.flush()
i = 0
for motif in mdobject.motifs:
motif.seednum = i ; i=i+1
kmers = motif.bogus_kmers(100)
motif.maxscore = -100
scores = [motif.scan(kmer)[2][0] for kmer in kmers]
print Arith.avestd(scores)
if re.search('\.meme$',filename):
mdobject.motifs.sort(lambda x,y: cmp(x.pvalue, y.pvalue))
else:
mdobject.motifs.sort(lambda x,y: cmp(x.church, y.church))
MotifTools.save_motifs(mdobject.motifs,tamoname)
def ace2tamo(filename, tamoname):
global probefile, PROBESET
if re.search('\.ace$',filename):
mdobject = AlignAce.AlignAce(filename)
elif re.search('\.meme$',filename):
mdobject = Meme.Meme(filename)
fsaname = find_fsa(mdobject.fastafile)
fsaD = MotifMetrics.fasta2seqs(fsaname,'want_dict')
probes = fsaD.keys()
if not probefile:
PROBESET = MotifMetrics.ProbeSet('HUMAN_250')
#PROBESET= pick_genome(fsaname)
for key,seq in fsaD.items():
PROBESET.probes[key] = seq
for motif in mdobject.motifs:
if motif.pvalue == 1: motif.pvalue = PROBESET.p_value(motif,probes,'v')
if motif.church == 1: motif.church = PROBESET.church(motif,probes,'v')
if motif.E_site == None: motif.E_site = PROBESET.E_sitef(motif,probes,3,'v')
#if motif.E_chi2 == None: motif.E_chi2 = PROBESET.E_chi2(motif,probes,None,'v')
if motif.E_seq == None: motif.E_seq = PROBESET.E_seq(motif,probes,'v')
if motif.ROC_auc== None: motif.ROC_auc= PROBESET.ROC_AUC(motif,probes,'v')
if motif.MNCP == None: motif.MNCP = PROBESET.MNCP(motif,probes,'v')
if re.search('\.meme$',filename):
motif.MAP = -math.log(motif.evalue)/math.log(10)
sys.stdout.flush()
i = 0
for motif in mdobject.motifs:
motif.seednum = i ; i=i+1
kmers = motif.bogus_kmers(100)
motif.maxscore = -100
scores = [motif.scan(kmer)[2][0] for kmer in kmers]
print Arith.avestd(scores)
if re.search('\.meme$',filename):
mdobject.motifs.sort(lambda x,y: cmp(x.pvalue, y.pvalue))
else:
mdobject.motifs.sort(lambda x,y: cmp(x.church, y.church))
MotifTools.save_motifs(mdobject.motifs,tamoname)
def motifs2tamo(motifs, outname):
global probefile, PROBESET
fsaname = find_fsa(outname)
fsaD = MotifMetrics.fasta2seqs(fsaname,'want_dict')
probes = fsaD.keys()
if not probefile:
PROBESET = MotifMetrics.ProbeSet('YEAST')
#PROBESET= pick_genome(fsaname)
#for key,seq in fsaD.items():
# PROBESET.probes[key] = seq
print "# %d motifs"%len(motifs)
for motif in motifs:
if motif.pvalue == 1: motif.pvalue = PROBESET.p_value(motif,probes,'v')
if motif.church == 1: motif.church = PROBESET.church(motif,probes,'v')
if motif.E_site == None: motif.E_site = PROBESET.E_sitef(motif,probes,3,'v')
#if motif.E_chi2 == None: motif.E_chi2 = PROBESET.E_chi2(motif,probes,None,'v')
if motif.E_seq == None: motif.E_seq = PROBESET.E_seq(motif,probes,'v')
if motif.ROC_auc== None: motif.ROC_auc= PROBESET.ROC_AUC(motif,probes,'v')
if motif.MNCP == None: motif.MNCP = PROBESET.MNCP(motif,probes,'v')
MotifTools.save_motifs(motifs,outname)
filename = sys.argv[1]
motif_list = open(filename).read().split('\nMOTIF')[1:]
tamo_list = []
motif_counter = 1
nsites_pat = re.compile("(w= [0-9]+)")
for motif in motif_list:
m_info1, m_info2 = motif.split('letter-probability matrix: ')
m_mat = m_info2.split('--------------------------------------------------------------------------------', 1)[0]
m_mat_header, m_prob_mat = m_mat.split('\n', 1)
nsites = int(nsites_pat.findall(m_mat_header)[0].split('= ')[1])
count_pos = m_prob_mat.split('\n')[:-1]
count_mat = []
site_list = []
for count in count_pos:
sites = [float(i) for i in count.split()]
site_list.append(sites)
count_dict = {'A': int(sites[0] * nsites),
'C': int(sites[1] * nsites),
'G': int(sites[2] * nsites),
'T': int(sites[3] * nsites)}
count_mat.append(count_dict)
m = MotifTools.Motif_from_counts(count_mat)
m.source = "Motif%s | %s" % (motif_counter, m_mat_header)
tamo_list.append(m)
motif_counter += 1
MotifTools.save_motifs(tamo_list, "MEME_motifs_%s.tamo" % filename.split('.')[0])
# Sort on log'd pVals
testMotifs.sort(key=lambda x: x[1])
comboMotifs = []
for i in range(0,int(len(testMotifs)*0.2)):
simMotifs = getKmersWithOneMisMtch(testMotifs[i][0],testMotifs)
alndMotifs = alignSimilarMotifs([x[0] for x in simMotifs])
#for m in simMotifs:
#print m[0].oneletter
comboMotifs.append(MotifTools.sum(alndMotifs,[-x[1] for x in simMotifs])) # -x[1] to convert neg logs to pos weights
print len(comboMotifs)
t2 = time.time()
oFile = '/Users/biggus/Documents/James/Collaborations/Campbell/data/Results_HyperGeoScreen/masked/Results_gGEMS/CCupAt4Days.gte2x.5-16mers.shfSeq.3.gGEMS.tmo'
pFile = '/Users/biggus/Documents/James/Collaborations/Campbell/data/Results_HyperGeoScreen/masked/Results_gGEMS/CCupAt4Days.gte2x.5-16mers.shfSeq.3.gGEMS.pkl'
MotifTools.save_motifs(comboMotifs,oFile,kmer_count=60)
pFile = open(pFile, 'w')
cPickle.dump(comboMotifs,pFile)
t3 = time.time()
print 'Calculations took %.3f min.\nWriting/Pickling took %.3f min.' % ((float(t2)-t1)/60, (float(t3)-t2)/60)
None
# Create a general list with all the motifs from all algorithms
genlist = []
genlist.extend(seederlist)
genlist.extend(memelist)
genlist.extend(weederlist)
# Perform clustering on the general list of motifs
clusterinf = clusterinfo(genlist)
averages = clusteravg(genlist,clusterinf)
# Trim the final average list
# averages = trim(averages,0.5)
# print clusterinf
# Save new list of cluster averages
MotifTools.save_motifs(averages,tamooutput)
# WEBLOGO IMAGE GENERATION
# Generate giflogos of all average motifs
for index in range(len(averages)):
cluster = 'Cluster ' + str(index + 1)
clustergif = argv[1] + '/other/cluster' + str(index + 1)
averages[index].giflogo(clustergif,title=cluster,scale=2)
# SUMMARY REPORT
# Determine location of the markdown file for the summary report
reportout = open(str(argv[1] + '/final/' + listname + '_cluster_report.md'), 'w')
# Write the header of the report
rundate = date.today()
header = "# Summary report for `" + listname + "`\nThis analysis was run on: " + str(rundate) + \
def memefiles2tamo(files, tamoname):
global probefile, PROBESET
motifs = []
for filename in files:
print ">>>SDFSD>F ", filename
if re.search('\.ace$', filename):
mdobject = AlignAce.AlignAce(filename)
if not mdobject.fastafile:
mdobject.fastafile = filename.replace('.ace', '.fsa')
elif re.search('\.meme.*$', filename):
mdobject = Meme.Meme(filename)
if not mdobject.fastafile:
mdobject.fastafile = re.sub('\..\.meme', '.meme',
filename).replace('.meme', '.fsa')
motifs.extend(mdobject.motifs)
#fsaname = find_fsa(mdobject.fastafile)
print mdobject.fastafile
fsaname = Fasta.find(mdobject.fastafile)
fsaD = Fasta.load(fsaname)
probes = fsaD.keys()
if not probefile:
PROBESET = MotifMetrics.ProbeSet('YEAST')
#PROBESET= pick_genome(fsaname)
for key, seq in fsaD.items():
PROBESET.probes[key] = seq
for motif in motifs:
if motif.pvalue == 1:
motif.pvalue = PROBESET.p_value(motif, probes, 'v')
if motif.church == 1:
motif.church = PROBESET.church(motif, probes, 'v')
if motif.E_site == None:
motif.E_site = PROBESET.E_sitef(motif, probes, 3, 'v')
#if motif.E_chi2 == None: motif.E_chi2 = PROBESET.E_chi2(motif,probes,None,'v')
#if motif.E_seq == None: motif.E_seq = PROBESET.E_seq(motif,probes,'v')
if motif.ROC_auc == None:
motif.ROC_auc = PROBESET.ROC_AUC(motif, probes, 'v')
if motif.MNCP == None: motif.MNCP = PROBESET.MNCP(motif, probes, 'v')
if motif.frac == None:
motif.frac = PROBESET.frac(motif, probes, 'v', 0.7)
if re.search('\.meme$', filename):
motif.MAP = -math.log(motif.evalue) / math.log(10)
if 1 and (motif.CRA == None):
try:
pass
CRA, Cfrac = PROBESET.cons_ROC_AUC(motif,
probes,
'v',
tuple='YES')
motif.CRA = CRA
motif.Cfrac = Cfrac
except:
pass
if re.search('\.meme$', filename):
mdobject.motifs.sort(lambda x, y: cmp(x.pvalue, y.pvalue))
else:
mdobject.motifs.sort(lambda x, y: cmp(x.church, y.church))
MotifTools.save_motifs(motifs, tamoname)
from gusPyCode.MDAP_proj.MDAP_defs import alignAndCombineMotifs
from TAMO import MotifTools
Motif = MotifTools.Motif
outFile = '/Users/biggus/Documents/James/Collaborations/Campbell/data/Results_HyperGeoScreen/masked/Results_gGEMS/CCupAt4Days.6-8mers.gGEMS.top6PlusCombos.motifs.stdThresh.tmo'
m = MotifTools.load('/Users/biggus/Documents/James/Collaborations/Campbell/data/Results_HyperGeoScreen/masked/Results_gGEMS/CCupAt4Days.6-8mers.gGEMS.top6.motifs.stdThresh.tmo')
w = [5.8952,
5.6523,
5.0585,
4.9788,
4.9678,
4.7688]
toTmo = []
toTmo.append(alignAndCombineMotifs([m[0],m[1]],[w[0],w[1]]))
toTmo.append(alignAndCombineMotifs([m[0],m[4]],[w[0],w[4]]))
toTmo.append(alignAndCombineMotifs([m[1],m[4]],[w[1],w[4]]))
toTmo.append(alignAndCombineMotifs([m[2],m[3]],[w[2],w[3]]))
toTmo.append(alignAndCombineMotifs([m[2],m[5]],[w[2],w[5]]))
for e in toTmo:
print e.oneletter
MotifTools.save_motifs(m+toTmo,outFile)
None
from gusPyCode.MDAP_proj.MDAP_defs import loadMotifsFromOutFile
from TAMO import MotifTools
mdOutFiles = ['/Users/biggus/Documents/James/Data/ReClustering/PrelimData_Grant_Feb09/RandSplitFastas/AceResults/Clus2_247gene_0.8_Apr16_14-46-33.ace.1.txt',
'/Users/biggus/Documents/James/Data/ReClustering/PrelimData_Grant_Feb09/RandSplitFastas/AceResults/Clus2_247gene_0.8_Apr16_14-46-33.ace.2.txt',
'/Users/biggus/Documents/James/Data/ReClustering/PrelimData_Grant_Feb09/RandSplitFastas/AceResults/Clus2_247gene_0.8_Apr16_14-46-33.ace.3.txt']
for mdFile in mdOutFiles:
motifs = loadMotifsFromOutFile(mdFile,'list') # ['Meme', 'AlignAce', 'MDscan', 'Weeder','list']
MotifTools.save_motifs(motifs,mdFile+'.tmo')
print 'Done.'
# generate PWM for indicated motifs
#module load TAMO (on hpc)
import sys
from TAMO import MotifTools
#input file: motif list
input=open(sys.argv[1], 'r')
motif=[]
for line in input:
if line.startswith("#"):
pass
else:
line=line.strip("\n").strip("\r").strip()
motif.append(line)
input.close()
print (motif)
pw=[]
for i in range(0, len(motif)):
m=MotifTools.Motif_from_text(motif[i])
pw.append(m)
MotifTools.save_motifs(pw,sys.argv[1]+'.tamo')
def run_CV(self,models):
num_bg = len(self.all_probes.keys())
num_fg = len(self.fg_seqs.keys())
for key in self.fg_seqs.keys():
self.all_probes[key] = self.fg_seqs[key]
self.probes = self.all_probes.keys()
if (self.randomize): trials = 50
else: trials = 1
self.models = models
std = 0.0
mean = 0.0
sum_sq = 0.0
sum_mean = 0.0
for trial in range(trials):
if (self.randomize):
N = num_bg/num_fg
if (N>10): N = 10
fg = []
bg = []
total = len(self.probes)
while(len(fg)<num_fg):
sp = random.randint(0,(total-1))
if (not(fg.count(self.probes[sp])>0)):
fg.append(self.probes[sp])
while(len(bg)<(N*num_fg)):
sp = random.randint(0,(total-1))
if ( (not(bg.count(self.probes[sp])>0)) and (not(fg.count(self.probes[sp])>0)) ):
bg.append(self.probes[sp])
else:
#select a random under-sampled set of background sequences
N = num_bg/num_fg
(N, bg) = self.undersample(N)
num_bg = N*num_fg
fg = self.fg_seqs.keys()
fg_group_size = len(fg)/self.cv_level #determine foreground group size
bg_group_size = len(bg)/self.cv_level #determine background group size
#separate sequences into groups
fg_groups = {}
bg_groups = {}
temp_fg = []
temp_bg = []
for a in fg: temp_fg.append(a)
for b in bg: temp_bg.append(b)
for i in range(self.cv_level):
fg_groups[i] = []
bg_groups[i] = []
if (i==(self.cv_level-1)):
fg_groups[i] = temp_fg[0:]
bg_groups[i] = temp_bg[0:]
else:
for j in range(fg_group_size):
entry = temp_fg[random.randint(0,len(temp_fg)-1)]
temp_fg.remove(entry)
fg_groups[i].append(entry)
for j in range(bg_group_size):
entry = temp_bg[random.randint(0,len(temp_bg)-1)]
temp_bg.remove(entry)
bg_groups[i].append(entry)
####################################################################################
#for each seed we run EM to train a probability model, calculate the maximum LLR for
#each input sequence in a group, train a SVM classifier on the training data, and
#calculate classification error on the test set. We repeat this for each cv-group
#and determine the mean cross-validation error for each hypothesis.
####################################################################################
self.classification_errors = {}
self.refined_motifs = {}
classifier = {}
self.best_motif = []
num_models = len(self.models)
for k in range(num_models):
self.best_motif.append(None)
(self.refined_motifs[k], self.classification_errors[k]) = self.train_classifier(k, fg_groups, bg_groups, N, trial)
#print out some information
best_c = 0
best_beta = 0.0
min = 1.0
mean = 0.0
for beta in self.classification_errors[k].keys():
for j in range(len(self.classification_errors[k][beta][0])):
mean = 0.0
for i in range(self.cv_level):
mean = mean + self.classification_errors[k][beta][i][j]
mean = mean/self.cv_level
if (mean<min):
min = mean
best_c = j
best_beta = beta
if (self.randomize):
sum_sq = sum_sq + (min*min)
sum_mean = sum_mean + min
cv_mean = sum_mean/(trial+1)
if (trial>0):
try: std = math.sqrt((sum_sq - (sum_mean*sum_mean)/(trial+1))/(trial))
except: std = 0.0
stddev_err = 0.71*std/math.sqrt(trial+1)
else:
(self.best_motif[k], classifier[k], fn) = self.train_final(k, fg, bg, N, best_beta)
print "\r\r---------------New Hypothesis---------------"
print "Unrefined Hypothesis: %s"%(self.models[k])
print "Refined Hypothesis: %s Optimal Beta: %f"%(self.best_motif[k], best_beta/(best_beta+1.0))
print "Mean %i-fold cv error: %f"%(self.cv_level,min)
LLR_thresh = (classifier[k][2]*classifier[k][1]/classifier[k][0] - classifier[k][3])
print "LLR match threshold: %f True positives: %i False Negatives: %i"%(LLR_thresh/self.best_motif[k].maxscore, (len(fg)-fn), fn)
self.best_motif[k].source = self.models[k].source
if (self.family!=''): self.best_motif[k].family = self.family
self.best_motif[k].dataset = self.datafiles[0]
self.best_motif[k].bgfile = self.datafiles[1]
self.best_motif[k].beta = best_beta/(best_beta+1.0)
self.best_motif[k].match_thresh = LLR_thresh
self.best_motif[k].cverror = min
print "Motif matches in positive input set:"
best_pssm = MDsupport.Motif2c_PSSM(self.best_motif[k])
for seq in fg:
sites = []
matches = best_pssm.matchstarts(self.all_probes[seq],LLR_thresh)
if (matches):
line = seq + '------> '
for match in matches:
entry = str(match) + ': ' + self.all_probes[seq][match:match+self.best_motif[k].width] + ' '
line = line + entry
print line
if (self.randomize):
print "Mean: %f Std.Dev: %f Error: %f Percent error: %f"%(cv_mean,std,stddev_err,(stddev_err/std))
else:
MotifTools.save_motifs(self.best_motif, self.motif_file)
return((self.best_motif, min))
motiflist.pop(blankindex)
# Build matrix dictionaries and substitute the strings for the dictionaries
for num in range(len(motiflist)):
# Prepare emtpy list of dictionaries of the length of the motif
tempmotif = []
motiflength = len(motiflist[num][1].strip().split('\t')) - 1
for item in range(motiflength):
tempmotif.append({})
# Start filling in the dictionaries with the information in the matrix
for line in range(1, 5):
nucleotide = motiflist[num][line].strip().split('\t')[0]
problist = motiflist[num][line].strip().split('\t')[1:]
for position in range(motiflength):
tempmotif[position][nucleotide] = float(problist[position])
# Save the list of dictionaries in the general variable
motiflist[num] = tempmotif
#print motiflist
# TAMO CONVERSION
# TAMO formated list of files
tamomotifs = []
# Convert dictionaries found in sigmotifs and store them as TAMO motifs
for motif in motiflist:
tamomotif = MotifTools.Motif_from_counts(motif[:], beta=0.01, bg=bkgrddict)
tamomotifs.append(tamomotif)
MotifTools.save_motifs(tamomotifs, output)