def __call_peaks_wo_control (self):
"""To call peaks w/o control data.
"""
#self.lambda_bg = float(self.scan_window)*self.treat.total/self.gsize
#self.debug("#3 background lambda: %.2f " % (self.lambda_bg))
#self.min_tags = poisson_cdf_inv(1-pow(10,self.pvalue/-10),self.lambda_bg)+1
#self.debug("#3 min tags: %d" % (self.min_tags))
self.info("#3 shift treatment data")
self.__shift_trackI(self.treat)
self.info("#3 merge +/- strand of treatment data")
self.treat.merge_plus_minus_locations_naive ()
self.debug("#3 after shift and merging, tags: %d" % (self.treat.total))
if self.opt.store_bdg:
self.info("#3 save the shifted and merged tag counts into bedGraph file...")
zbdg_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.d,log=self.info,single=self.opt.single_profile)
elif self.opt.store_wig:
self.info("#3 save the shifted and merged tag counts into wiggle file...")
#if self.opt.wigextend:
# zwig_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.opt.wigextend,log=self.info,space=self.opt.space,single=self.opt.single_profile)
#else:
zwig_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.d,log=self.info,space=self.opt.space,single=self.opt.single_profile)
self.info("#3 call peak candidates")
peak_candidates = self.__call_peaks_from_trackI (self.treat)
self.info("#3 use self to calculate local lambda and filter peak candidates...")
self.final_peaks = self.__filter_w_control(peak_candidates,self.treat,self.treat,pass_sregion=True)
return self.final_peaks
def __call_peaks_wo_control(self):
"""To call peaks w/o control data.
"""
#self.lambda_bg = float(self.scan_window)*self.treat.total/self.gsize
#self.debug("#3 background lambda: %.2f " % (self.lambda_bg))
#self.min_tags = poisson_cdf_inv(1-pow(10,self.pvalue/-10),self.lambda_bg)+1
#self.debug("#3 min tags: %d" % (self.min_tags))
self.info("#3 shift treatment data")
self.__shift_trackI(self.treat)
self.info("#3 merge +/- strand of treatment data")
self.treat.merge_plus_minus_locations_naive()
self.debug("#3 after shift and merging, tags: %d" % (self.treat.total))
if self.opt.store_bdg:
self.info(
"#3 save the shifted and merged tag counts into bedGraph file..."
)
zbdg_write(self.treat,
self.opt.wig_dir_tr,
self.zwig_tr,
self.d,
log=self.info,
single=self.opt.single_profile)
elif self.opt.store_wig:
self.info(
"#3 save the shifted and merged tag counts into wiggle file..."
)
#if self.opt.wigextend:
# zwig_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.opt.wigextend,log=self.info,space=self.opt.space,single=self.opt.single_profile)
#else:
zwig_write(self.treat,
self.opt.wig_dir_tr,
self.zwig_tr,
self.d,
log=self.info,
space=self.opt.space,
single=self.opt.single_profile)
self.info("#3 call peak candidates")
peak_candidates = self.__call_peaks_from_trackI(self.treat)
self.info(
"#3 use self to calculate local lambda and filter peak candidates..."
)
self.final_peaks = self.__filter_w_control(peak_candidates,
self.treat,
self.treat,
pass_sregion=True)
return self.final_peaks
def __call_peaks_w_control(self):
"""To call peaks with control data.
A peak info type is a: dictionary
key value: chromosome
items: (peak start,peak end, peak length, peak summit, peak
height, number of tags in peak region, peak pvalue, peak
fold_enrichment) <-- tuple type
"""
#self.lambda_bg = float(self.scan_window)*self.treat.total/self.gsize
#self.debug("#3 background lambda: %.2f " % (self.lambda_bg))
#self.min_tags = poisson_cdf_inv(1-pow(10,self.pvalue/-10),self.lambda_bg)+1
#self.debug("#3 min tags: %d" % (self.min_tags))
self.ratio_treat2control = float(self.treat.total) / self.control.total
if self.ratio_treat2control > 2 or self.ratio_treat2control < 0.5:
self.warn(
"Treatment tags and Control tags are uneven! FDR may be wrong!"
)
self.info("#3 shift treatment data")
self.__shift_trackI(self.treat)
self.info("#3 merge +/- strand of treatment data")
self.treat.merge_plus_minus_locations_naive()
self.debug("#3 after shift and merging, tags: %d" % (self.treat.total))
if self.opt.store_bdg:
self.info(
"#3 save the shifted and merged tag counts into bedGraph file..."
)
zbdg_write(self.treat,
self.opt.wig_dir_tr,
self.zwig_tr,
self.d,
log=self.info,
single=self.opt.single_profile)
elif self.opt.store_wig:
self.info(
"#3 save the shifted and merged tag counts into wiggle file..."
)
#if self.opt.wigextend:
# zwig_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.opt.wigextend,log=self.info,space=self.opt.space,single=self.opt.single_profile)
#else:
zwig_write(self.treat,
self.opt.wig_dir_tr,
self.zwig_tr,
self.d,
log=self.info,
space=self.opt.space,
single=self.opt.single_profile)
self.info("#3 call peak candidates")
peak_candidates = self.__call_peaks_from_trackI(self.treat)
self.info("#3 shift control data")
self.info("#3 merge +/- strand of control data")
self.__shift_trackI(self.control)
self.control.merge_plus_minus_locations_naive()
self.debug("#3 after shift and merging, tags: %d" %
(self.control.total))
if self.opt.store_bdg:
self.info(
"#3 save the shifted and merged tag counts into bedGraph file..."
)
zbdg_write(self.control,
self.opt.wig_dir_ctl,
self.zwig_ctl,
self.d,
log=self.info,
single=self.opt.single_profile)
if self.opt.store_wig:
self.info(
"#3 save the shifted and merged tag counts into wiggle file..."
)
#if self.opt.wigextend:
# zwig_write(self.control,self.opt.wig_dir_ctl,self.zwig_ctl,self.opt.wigextend,log=self.info,space=self.opt.space,single=self.opt.single_profile)
#else:
zwig_write(self.control,
self.opt.wig_dir_ctl,
self.zwig_ctl,
self.d,
log=self.info,
space=self.opt.space,
single=self.opt.single_profile)
self.info("#3 call negative peak candidates")
negative_peak_candidates = self.__call_peaks_from_trackI(self.control)
# build score
#if self.save_score:
# self.info("#4 build scores")
# scoreswig = self.__build_score_wigtrackI(treatwig,controlbkI,self.d,space=self.opt.space,bglambda=self.lambda_bg)
# zwigfile = file(self.opt.name+".score.wig","w")
# self.info("#4.1 save scores to wiggle file")
# scoreswig.write_wig(zwigfile,"score")
# self.info("compress the wiggle file using gzip...")
# os.system("gzip "+self.opt.name+".score.wig")
self.info("#3 use control data to filter peak candidates...")
#self.final_peaks = self.__filter_w_control(peak_candidates,self.treat,self.control, self.ratio_treat2control,fake_when_missing=True)
self.final_peaks = self.__filter_w_control(
peak_candidates,
self.treat,
self.control,
fake_when_missing=True,
to_small_sample=self.opt.tosmall)
self.info("#3 find negative peaks by swapping treat and control")
#self.final_negative_peaks = self.__filter_w_control(negative_peak_candidates,self.control,self.treat, 1.0/self.ratio_treat2control,fake_when_missing=True)
self.final_negative_peaks = self.__filter_w_control(
negative_peak_candidates,
self.control,
self.treat,
fake_when_missing=True,
to_small_sample=self.opt.tosmall)
return self.__add_fdr(self.final_peaks, self.final_negative_peaks)
def __call_peaks_w_control (self):
"""To call peaks with control data.
A peak info type is a: dictionary
key value: chromosome
items: (peak start,peak end, peak length, peak summit, peak
height, number of tags in peak region, peak pvalue, peak
fold_enrichment) <-- tuple type
"""
#self.lambda_bg = float(self.scan_window)*self.treat.total/self.gsize
#self.debug("#3 background lambda: %.2f " % (self.lambda_bg))
#self.min_tags = poisson_cdf_inv(1-pow(10,self.pvalue/-10),self.lambda_bg)+1
#self.debug("#3 min tags: %d" % (self.min_tags))
self.ratio_treat2control = float(self.treat.total)/self.control.total
if self.ratio_treat2control > 2 or self.ratio_treat2control < 0.5:
self.warn("Treatment tags and Control tags are uneven! FDR may be wrong!")
self.info("#3 shift treatment data")
self.__shift_trackI(self.treat)
self.info("#3 merge +/- strand of treatment data")
self.treat.merge_plus_minus_locations_naive ()
self.debug("#3 after shift and merging, tags: %d" % (self.treat.total))
if self.opt.store_bdg:
self.info("#3 save the shifted and merged tag counts into bedGraph file...")
zbdg_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.d,log=self.info,single=self.opt.single_profile)
elif self.opt.store_wig:
self.info("#3 save the shifted and merged tag counts into wiggle file...")
#if self.opt.wigextend:
# zwig_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.opt.wigextend,log=self.info,space=self.opt.space,single=self.opt.single_profile)
#else:
zwig_write(self.treat,self.opt.wig_dir_tr,self.zwig_tr,self.d,log=self.info,space=self.opt.space,single=self.opt.single_profile)
self.info("#3 call peak candidates")
peak_candidates = self.__call_peaks_from_trackI (self.treat)
self.info("#3 shift control data")
self.info("#3 merge +/- strand of control data")
self.__shift_trackI(self.control)
self.control.merge_plus_minus_locations_naive ()
self.debug("#3 after shift and merging, tags: %d" % (self.control.total))
if self.opt.store_bdg:
self.info("#3 save the shifted and merged tag counts into bedGraph file...")
zbdg_write(self.control,self.opt.wig_dir_ctl,self.zwig_ctl,self.d,log=self.info,single=self.opt.single_profile)
if self.opt.store_wig:
self.info("#3 save the shifted and merged tag counts into wiggle file...")
#if self.opt.wigextend:
# zwig_write(self.control,self.opt.wig_dir_ctl,self.zwig_ctl,self.opt.wigextend,log=self.info,space=self.opt.space,single=self.opt.single_profile)
#else:
zwig_write(self.control,self.opt.wig_dir_ctl,self.zwig_ctl,self.d,log=self.info,space=self.opt.space,single=self.opt.single_profile)
self.info("#3 call negative peak candidates")
negative_peak_candidates = self.__call_peaks_from_trackI (self.control)
# build score
#if self.save_score:
# self.info("#4 build scores")
# scoreswig = self.__build_score_wigtrackI(treatwig,controlbkI,self.d,space=self.opt.space,bglambda=self.lambda_bg)
# zwigfile = file(self.opt.name+".score.wig","w")
# self.info("#4.1 save scores to wiggle file")
# scoreswig.write_wig(zwigfile,"score")
# self.info("compress the wiggle file using gzip...")
# os.system("gzip "+self.opt.name+".score.wig")
self.info("#3 use control data to filter peak candidates...")
#self.final_peaks = self.__filter_w_control(peak_candidates,self.treat,self.control, self.ratio_treat2control,fake_when_missing=True)
self.final_peaks = self.__filter_w_control(peak_candidates,self.treat,self.control, fake_when_missing=True, to_small_sample=self.opt.tosmall)
self.info("#3 find negative peaks by swapping treat and control")
#self.final_negative_peaks = self.__filter_w_control(negative_peak_candidates,self.control,self.treat, 1.0/self.ratio_treat2control,fake_when_missing=True)
self.final_negative_peaks = self.__filter_w_control(negative_peak_candidates,self.control,self.treat, fake_when_missing=True, to_small_sample=self.opt.tosmall)
return self.__add_fdr (self.final_peaks, self.final_negative_peaks)