def ConvertToBedViaR(cel_file, bpmapfile):
'''function for t calling R to get Bed-like values'''
# original script - minus redundant variables.
# library(rMAT)
# library(Biobase)
#
# expName <- "2012-07-12_SIR2_IP"; # cell file name
# bpmapFile <- "Sc03b_MR_v04.bpmap"; # the mapping!
# seqHeader <- ReadBPMAPAllSeqHeader(bpmapFile);
# arrayFile1 <- c("2012-07-12_SIR2_IP.CEL");
# ScSet <- BPMAPCelParser(bpmapFile, arrayFile1, verbose = FALSE, groupName = "Sc", seqName="chr");
# data <- list(chrNo = ScSet1@featureChromosome, probePos = ScSet1@featurePosition, MATScore = exprs(ScSet1)); # last parameter is the raw data.
# write.table(data, file = paste(expName,"_exp1_AllData.txt",sep=''), append = FALSE, row.names = FALSE, sep = "\t");
bedfile = StringUtils.rreplace(cel_file, 'CEL', 'BED', 2)
bedfile += "like"
print "input file: %s" % (cel_file)
print "Will be writing out to %s" % (bedfile)
print "importing rMat and Biobase libraries"
importr('rMAT')
importr('Biobase')
# print "creating seqHeader"
# robjects.r('seqHeader <- ReadBPMAPAllSeqHeader(\"' + bpmapfile + '\")')
print "creating scSet"
robjects.r('ScSet <- BPMAPCelParser(\"' +
bpmapfile + '\", c(\"' +
cel_file + '\"), verbose = FALSE, groupName = "Sc", seqName="chr")')
print "creating data"
robjects.r('data <- list(chrNo = ScSet@featureChromosome, probePos = ScSet@featurePosition, MATScore = exprs(ScSet))') # last parameter is the raw data.
print "writing table"
robjects.r('write.table(data, file = \"' + bedfile + '\", append = FALSE, quote = FALSE, row.names = FALSE, sep = "\t")')
files = os.listdir(args.BEDlikefiles)
starttime = time.time()
num_probes = 2635714
print "initializing array..."
probeset = [[0.0 for y in xrange(num_probes)] for x in xrange(0, 2)]
print "reading baseline from %s" % (args.Baselinefile)
print "probeset[%i][%i]" % (len(probeset), len(probeset[0]))
ReadBaseline(args.Baselinefile, probeset)
blaverage = FindAverageProbeIntensity(probeset)
print "baseline average probe intensity is:", blaverage
print "done."
# op = StringUtils.rreplace(sys.argv[1], 'BED', 'NORMAL', 1)
for i, f in enumerate(files):
filetime = time.time()
of = StringUtils.rreplace(f, '.BEDlike', '.normalized.BEDlike', 1)
bedprobes = [[0.0 for y in xrange(num_probes)] for x in xrange(0, 2)]
print "determining average probe intensity from %s%s" % (args.BEDlikefiles, f)
print "bedprobes[%i][%i]" % (len(bedprobes), len(bedprobes[0]))
ReadBaseline("%s%s" % (args.BEDlikefiles, f), bedprobes)
bedaverage = FindAverageProbeIntensity(bedprobes)
print "average probe intensity is:", bedaverage
scaling = float(bedaverage) / blaverage
if scaling < 1:
scaling = 1
print "baseline needs to be scaled by a factor of", scaling
ApplyBaseline(i + 1, "%s%s" % (args.BEDlikefiles, f), "%s%s" % (args.output_path, of), probeset, scaling) # first file is first file, use zero for chr/pos
print "File %i - %s processed in %f seconds" % (i + 1, f, time.time() - filetime)
# ProduceStats(len(files) + 1, probeset)
print 'Completed in %s seconds' % int((time.time() - starttime))
def run(PARAM, metadata_file, tohide):
'''reads metadata_file, and updates sample entries in database'''
while not os.path.isfile(metadata_file):
print "Unable to find file %s" % metadata_file
sys.exit()
print "Data being imported into database: ", PARAM.get('default_database')
print "opening connection(s) to MongoDB..."
mongo = Mongo_Connector.MongoConnector(PARAM.get("server"), PARAM.get("port"), PARAM.get('default_database'))
print "processing %s..." % metadata_file
f = open(metadata_file, 'r')
firstrow = True
collection_name = 'samples'
missing = 0
for line in f:
sample_update = {}
if not firstrow:
a = line.split("\t")
# # Find file name
file_name = StringUtils.rreplace(a[0], ".CEL", ".normalized.waves", 1)
print "Checking for file %s in database %s" % (file_name, PARAM.get('default_database'))
# check if entry exists in database
if mongo.find(collection_name, {"file_name":file_name}, {"_id":1}).count() > 0 :
print "Entry exists; updating metadata for ", file_name
# Update other parameters
sample_update["exp_date"] = a[1]
sample_update["strain_number"] = a[2]
sample_update["strain_background"] = a[3]
sample_update["mutations"] = a[4]
sample_update["researcher"] = a[5]
sample_update["type"] = a[6]
sample_update["antibody"] = a[7]
sample_update["catalog_number"] = a[8]
sample_update["antibody_volume"] = a[9]
sample_update["array_type"] = a[10]
sample_update["array_lot_number"] = a[11]
sample_update["protocol"] = a[12]
sample_update["crosslinking_time"] = a[13]
sample_update["pubmed_id"] = a[14]
sample_update["sample_id"] = a[15]
sample_update["comments"] = a[16]
if a[16] == "\n":
sample_update["comments"] = ""
if tohide:
sample_update['hide'] = True
else:
sample_update['hide'] = False # now that metadata is added, unhide sample
# update entries
mongo.update(collection_name, {"file_name":file_name}, {"$set": sample_update})
print "\tFinished updating metadata for ", file_name
else:
print "NOT FOUND: ", file_name
missing += 1
else:
firstrow = False
print "Done all updates from metadata file ", metadata_file
if missing > 0:
print "%s files were unable to be updated as no matching file_name was found in the waves table." % missing
mongo.close()
def find_waves_in_spliced_genes(splicefile, wavesfile, output, autothresh, heisig, thresh):
# body of function goes here
chromosomes = {1:"I", 2:"II", 3:"III", 4:"IV", 5:"V", 6:"VI", 7:"VII", 8:"VIII", 9:"IX", 10:"X", 11:"XI", 12:"XII", 13:"XIII", 14:"XIV", 15:"XV", 16:"XVI"}
waves = []
print "Reading files..."
f = open(wavesfile, 'r')
for line in f:
if line.startswith("#"):
continue
else:
a = line.split("\t")
wave = {}
wave["chr"] = "chr" + a[0].replace("chr", "").upper()
wave["pos"] = int(a[1])
wave["stddev"] = int(a[2])
wave["height"] = float(a[3])
wave["used"] = False
waves.append(wave)
f.close()
if(autothresh):
if(heisig):
# print "\nNow determining background levels for height of peaks"
bins = 70 # based on max peak height of 7
counts = [0] * bins
thresh = [0] * bins
for i in range(0, bins):
thresh[i] = (i + 1) * 0.1
for i in waves:
# check all heights to determine background levels
counts[int(i['height'] / 0.1)] += 1 # increment count where height1 is in bin of size 0.1
# print "counts are: ", counts
x = []
y = []
for i in range(10, 15): # (0 to 9 correspond to heights 0 to 0.9, do not have)
x.append(thresh[i])
y.append(counts[i])
# print "x is ", x
# print "y is ", y
slr = scipystats.linregress(x, y)
slope = slr[0]
intercept = slr[1]
# print "slope: %s and intercept: %s for background peak height" % (slope, intercept)
# find x-intercept, threshold for noise-signal
xint = abs(intercept / slope)
print "height threshold between noise and signal is ", xint
thresh = round(xint, 2)
else:
# determine threshold for sigma
bins = 300 # based on max sigma of 300
counts = [0] * bins
threshes = [0] * bins
for i in range(0, bins):
threshes[i] = (i + 1)
for i in waves:
# check all heights to determine background levels
counts[int(i['stddev'])] += 1 # increment count where height1 is in bin of size 0.1
# print "counts are: ", counts
highest = 0
ind = -1
for i in range(0, bins):
if counts[i] > highest:
highest = counts[i]
ind = i
thresh = threshes[ind]
print "sigma threshold between noise and signal is ", thresh
# remove waves that don't meet threshold
if(heisig):
waves[:] = [x for x in waves if x['height'] > thresh]
else:
waves[:] = [x for x in waves if x['stddev'] > thresh]
waves.sort(key = lambda x: (x['chr'], x['pos'])) # list sorted by position
splice = []
f = open(splicefile, 'r')
for line in f:
a = line.split("\t")
region = {}
region["gene"] = a[0]
region["chr"] = "chr" + chromosomes[int(a[1])]
if int(a[2]) == 1: # forward
# promoter
region["p1"] = int(a[3]) - 301 # p1 is left boundary of promoter
region["p2"] = int(a[3]) - 1 # p2 is right boundary of promoter
# exon1
region["e11"] = int(a[3]) # e11 is left boundary of exon1
region["e12"] = int(a[4]) # e12 us right boundary of exon1
# intron
region["i1"] = int(a[5])
region["i2"] = int(a[6])
# exon2
region["e21"] = int(a[7])
region["e22"] = int(a[8])
# 3' UTR
region["u1"] = int(a[8]) + 1 # 3'UTR
region["u2"] = int(a[8]) + 301
else: # reverse
# promoter
region["p1"] = int(a[4]) - 301
region["p2"] = int(a[4]) - 1
# exon1
region["e11"] = int(a[4])
region["e12"] = int(a[3])
# intron
region["i1"] = int(a[6])
region["i2"] = int(a[5])
# exon2
region["e21"] = int(a[8])
region["e22"] = int(a[7])
# 3' UTR
region["u1"] = int(a[7]) + 1
region["u2"] = int(a[7]) + 301
# counts in each region
region['p'] = 0
region['e1'] = 0
region['i'] = 0
region['e2'] = 0
region['u'] = 0
# splice site regions (+/- 100 bp)
region['pe11'] = region['e11'] - 99 # promoter to exon1 left boundary
region['pe12'] = region['e11'] + 101 # promoter to exon1 right boundary
region['e1i1'] = region['i1'] - 99
region['e1i2'] = region['i1'] + 101
region['ie21'] = region['e21'] - 99
region['ie22'] = region['e21'] + 101
region['e2u1'] = region['u1'] - 99
region['e2u2'] = region['u1'] + 101
# splice site counts
region['pe1'] = 0
region['e1i'] = 0
region['ie2'] = 0
region['e2u'] = 0
splice.append(region)
splice.sort(key = lambda x: (x['chr'], x['p1']))
print_queue = multiprocessing.Queue()
# launch thread to read and process the print queue
print_thread = PrintThread.StringWriter(print_queue, output, StringUtils.rreplace(os.path.basename(wavesfile), ".normalized.waves", "_" + str(thresh) + "_splice_summary.txt", 1), True, True)
# print header
print_queue.put("gene\tchr\tp1\tp2\te11\te12\ti1\ti2\te21\te22\tu1\tu2\tp\tpe1\te1\te1i\ti\tie2\te2\te2u\tu")
print "Now finding peaks in each region..."
w = 0
b = 0
while b < len(splice) and w < len(waves):
if(waves[w]['chr'] == splice[b]['chr'] and (waves[w]['pos'] + 1 * waves[w]['stddev']) >= splice[b]['p1'] and (waves[w]['pos'] - 1 * waves[w]['stddev']) <= splice[b]['u2']):
# print "found a wave in gene ", splice["gene"]
# find which segment of gene it is present in
# only use standard deviation threshold when at beginning of promoter and end of UTR
if (waves[w]['pos'] + 1 * waves[w]['stddev']) >= splice[b]['p1'] and waves[w]['pos'] <= splice[b]['p2']:
splice[b]['p'] += 1.0 / (splice[b]['p2'] - splice[b]['p1'])
elif waves[w]['pos'] >= splice[b]['e11'] and waves[w]['pos'] <= splice[b]['e12']:
splice[b]['e1'] += 1.0 / (splice[b]['e12'] - splice[b]['e11'])
elif waves[w]['pos'] >= splice[b]['i1'] and waves[w]['pos'] <= splice[b]['i2']:
splice[b]['i'] += 1.0 / (splice[b]['i2'] - splice[b]['i1'])
elif waves[w]['pos'] >= splice[b]['e21'] and waves[w]['pos'] <= splice[b]['e22']:
splice[b]['e2'] += 1.0 / (splice[b]['e22'] - splice[b]['e21'])
elif waves[w]['pos'] >= splice[b]['u1'] and (waves[w]['pos'] - 1 * waves[w]['stddev']) <= splice[b]['u2']:
splice[b]['u'] += 1.0 / (splice[b]['u2'] - splice[b]['u1'])
if waves[w]['pos'] >= splice[b]['pe11'] and waves[w]['pos'] <= splice[b]['pe12']:
splice[b]['pe1'] += 1.0 / (splice[b]['pe12'] - splice[b]['pe11'])
elif waves[w]['pos'] >= splice[b]['e1i1'] and waves[w]['pos'] <= splice[b]['e1i2']:
splice[b]['e1i'] += 1.0 / (splice[b]['e1i2'] - splice[b]['e1i1'])
elif waves[w]['pos'] >= splice[b]['ie21'] and waves[w]['pos'] <= splice[b]['ie22']:
splice[b]['ie2'] += 1.0 / (splice[b]['ie22'] - splice[b]['ie21'])
elif waves[w]['pos'] >= splice[b]['e2u1'] and waves[w]['pos'] <= splice[b]['e2u2']:
splice[b]['e2u'] += 1.0 / (splice[b]['e2u2'] - splice[b]['e2u1'])
waves[w]['used'] = True
w += 1
elif waves[w]['chr'] > splice[b]['chr'] or ((waves[w]['pos'] - 1 * waves[w]['stddev']) > splice[b]['u2'] and waves[w]['chr'] == splice[b]['chr']):
# wave is past bin, move to next bin
# write to file
# print "b: %i, w: %i, wave is AFTER, start: %s, end: %s, pos: %s, %s %s" % (b, w, splice[b]['start'], splice[b]['end'], waves[w]['pos'], splice[b]['chr'], waves[w]['chr'])
print_queue.put(splice[b]['gene'] + "\t" + splice[b]['chr'] + "\t" + \
str(splice[b]['p1']) + "\t" + str(splice[b]['p2']) + "\t" + \
str(splice[b]['e11']) + "\t" + str(splice[b]['e12']) + "\t" + \
str(splice[b]['i1']) + "\t" + str(splice[b]['i2']) + "\t" + \
str(splice[b]['e21']) + "\t" + str(splice[b]['e22']) + "\t" + \
str(splice[b]['u1']) + "\t" + str(splice[b]['u2']) + "\t" + \
str(splice[b]['p']) + "\t" + str(splice[b]['pe1']) + "\t" + \
str(splice[b]['e1']) + "\t" + str(splice[b]['e1i']) + "\t" + \
str(splice[b]['i']) + "\t" + str(splice[b]['ie2']) + "\t" + \
str(splice[b]['e2']) + "\t" + str(splice[b]['e2u']) + "\t" + \
str(splice[b]['u']))
b += 1
else:
# wave is before bin
# print "b: %i, w: %i, wave is BEFORE, start: %s, end: %s, pos: %s, %s %s" % (b, w, splice[b]['start'], splice[b]['end'], waves[w]['pos'], splice[b]['chr'], waves[w]['chr'])
w += 1
while b < len(splice):
# if ended because got to end of waves
# need to write statistics for remaining bins
# print "filling in remaining bins with 0s"
print_queue.put(splice[b]['gene'] + "\t" + splice[b]['chr'] + "\t" + \
str(splice[b]['p1']) + "\t" + str(splice[b]['p2']) + "\t" + \
str(splice[b]['e11']) + "\t" + str(splice[b]['e12']) + "\t" + \
str(splice[b]['i1']) + "\t" + str(splice[b]['i2']) + "\t" + \
str(splice[b]['e21']) + "\t" + str(splice[b]['e22']) + "\t" + \
str(splice[b]['u1']) + "\t" + str(splice[b]['u2']) + "\t" + \
str(splice[b]['p']) + "\t" + str(splice[b]['pe1']) + "\t" + \
str(splice[b]['e1']) + "\t" + str(splice[b]['e1i']) + "\t" + \
str(splice[b]['i']) + "\t" + str(splice[b]['ie2']) + "\t" + \
str(splice[b]['e2']) + "\t" + str(splice[b]['e2u']) + "\t" + \
str(splice[b]['u']))
b += 1
# end printing
if print_thread is None or not print_thread.is_alive():
pass
else:
while print_queue.qsize() > 0:
print "waiting on print_queue to empty", print_queue.qsize()
time.sleep(1)
print_thread.END_PROCESSES = True
print_thread.f.close()
# while not print_thread.is_closed():
# print "waiting for print_thread to close ", print_queue.qsize(), " ", print_thread.is_closed(), " ", print_thread.END_PROCESSES
# time.sleep(1)
# quick summary statistics
unassigned = 0
for i in waves:
if not i["used"]:
unassigned += 1
if(heisig):
print "Height threshold applied to waves:", thresh
else:
print "Sigma threshold applied to waves:", thresh
ln = 0 # total number of genes
ctp = 0 # count of peaks in promoters
ctpe1 = 0 # count of peaks in promoter/exon1 region
cte1 = 0
cte1i = 0
cti = 0
ctie2 = 0
cte2 = 0
cte2u = 0
ctu = 0
emptygenes = 0
for line in splice:
ln += 1
ctp += line['p']
ctpe1 += line['pe1']
cte1 += line['e1']
cte1i += line['e1i']
cti += line['i']
ctie2 += line['ie2']
cte2 += line['e2']
cte2u += line['e2u']
ctu += line['u']
if line['p'] == 0 and line['e1'] == 0 and line['i'] == 0 and line['e2'] == 0 and line['u'] == 0:
# no waves in this gene
emptygenes += 1
ln -= 1 # dont use in calculating the average
avgp = float(ctp) / ln
avgpe1 = float(ctpe1) / ln
avge1 = float(cte1) / ln
avge1i = float(cte1i) / ln
avgi = float(cti) / ln
avgie2 = float(ctie2) / ln
avge2 = float(cte2) / ln
avge2u = float(cte2u) / ln
avgu = float(ctu) / ln
print "%s waves were not part of a spliced gene" % unassigned
print "%s genes had no waves" % emptygenes
print "Region\tAverage_num_waves"
print "Prom\t%f" % avgp
print "Exon_1\t%f" % avge1
print "Intron\t%f" % avgi
print "Exon_2\t%f" % avge2
print "3'_UTR\t%f" % avgu
print "\n"
print "prom_exon1\t%f" % avgpe1
print "exon1_intron\t%f" % avge1i
print "intron_exon2\t%f" % avgie2
print "exon2_utr\t%f" % avge2u
def FindBaseline(file_name, normalize = False):
'''Find the baseline - TODO: break this into smaller functions'''
f = open(file_name, 'r') # open file
print "processing data file (" + file_name + ")..."
first_line = True
headers = []
data = []
for line in f:
if (first_line):
headers = line.split("\t")
for h in range(len(headers)):
headers[h] = headers[h].lower()
if headers[h].find("\n") != -1:
headers[h] = headers[h].replace("\n", "")
first_line = False
else:
a = line.split("\t")
r = row(a[0], int(a[1]), float(a[2]))
data.append(r)
# find most common number
f.close()
if normalize:
v_sum = 0
hist = {}
for x in range(len(data)):
v_sum += data[x].value
if hist.has_key(str(data[x].value)):
hist[str(data[x].value)] += 1
else:
hist[str(data[x].value)] = 1
v_avg = v_sum / len(data)
print "v average = %f" % (v_avg)
point = 0
largest = 0
for g, y in hist.iteritems():
if y > largest:
largest = y
point = g
# print "g, y (%s, %i)" % (g, y)
# print "point, largest (%s, %i)" % (point, largest)
point = float(point)
for g in range(len(data)):
v = data[g].value - point
if v < 0:
v = 0
data[g].setv(v)
# create wig file
f_w_name = StringUtils.rreplace(file_name, '.BEDlike', '', 1)
# f_w_name = StringUtils.rreplace(f_w_name, 'BED', 'WIG', 1) # #should these two lines just replace .BEDlike with .WIG?
trackname = os.path.basename(f_w_name)
# print "Writing to %s" % (f_w_name)
# f = open(f_w_name, 'w') # open file
current_chr = data[0].chromosome
print "New Chromosome %s (%s)" % (current_chr, chr_yeast[current_chr])
last_bp = 0
last_ht = 0
wigfile = WigFileThread.WigFileWriter(None)
wigfile.start_wig_writer(os.path.dirname(f_w_name), os.path.basename(f_w_name), trackname)
# for x in range (0, 5020):
# print "x=%i position=%i, value=%f" % (x, data[x].position + 1, data[x].value)
x = 0
a = 0
# gc.disable()
coverage_map = []
l = len(data)
x = 1
block_left = 0
while x < l:
# print "x of len: %i/%i" % (x, l)
if data[x].chromosome != current_chr:
if len(coverage_map) > 0:
wigfile.add_map(coverage_map, chr_yeast[current_chr], block_left) # for yeast chromosome nomenclature (roman numeral)
# wigfile.add_map(coverage_map, current_chr, block_left)
coverage_map = []
# TODO: taper off chromosome
# TODO: taper "on" new chromosome
print "New Chromosome %s (%s)" % (data[x].chromosome, chr_yeast[data[x].chromosome])
current_chr = data[x].chromosome # switch chromosomes,
if data[x - 1].chromosome == data[x].chromosome:
block_left = data[x - 1].position + 2 # shift by 1, and block starts after the zero.
last_bp = data[x - 1].position + 1
else:
block_left = data[x].position + 2
last_bp = data[x].position + 1
# print "x=%i data=%i, value=%f" % (x, data[x].position + 1, data[x].value)
while x < l and data[x].value >= 0 and data[x].chromosome == current_chr:
diff = (data[x].position + 1) - last_bp
# if diff > 6:
# diff = 6
if (diff > 1):
slope = float(data[x].value - last_ht) / (diff) # slope between the two
for y in range(1, diff):
coverage_map.append(round(last_ht + (slope * y), 2))
coverage_map.append(round(data[x].value, 2))
else:
coverage_map.append(round(data[x].value, 2))
a += 1
last_bp = data[x].position + 1
last_ht = data[x].value
if x < l - 1 and (data[x + 1].value <= 0 or data[x + 1].chromosome != current_chr):
diff = (data[x + 1].position + 1) - (data[x].position + 1)
slope = float(0 - data[x].value) / (diff)
for y in range(1, diff):
coverage_map.append(round(data[x].value + (slope * y), 2))
last_ht = 0
x += 1
while x < l and data[x].value <= 0:
x += 1
last_ht = 0
if len(coverage_map) > 0:
wigfile.add_map(coverage_map, chr_yeast[current_chr], block_left) # for yeast chromosome nomenclature (roman numeral)
# print "writing map"
# wigfile.add_map(coverage_map, current_chr, block_left)
coverage_map = []
# gc.enable()
print "Closing Wigwriter. This may take some time."
wigfile.close_wig_writer()
print "Wigwriter closed."
def find_waves_in_promoter(orffile, wavesfile, output, autothresh, heisig):
# body of function goes here
chromosomes = {1:"I", 2:"II", 3:"III", 4:"IV", 5:"V", 6:"VI", 7:"VII", 8:"VIII", 9:"IX", 10:"X", 11:"XI", 12:"XII", 13:"XIII", 14:"XIV", 15:"XV", 16:"XVI"}
waves = []
print "Reading files..."
f = open(wavesfile, 'r')
for line in f:
if line.startswith("#"):
continue
else:
a = line.split("\t")
wave = {}
wave["chr"] = "chr" + a[0].replace("chr", "").upper()
wave["pos"] = int(a[1])
wave["stddev"] = int(a[2])
wave["height"] = float(a[3])
wave["used"] = False
waves.append(wave)
f.close()
# automatically determine noise
if(autothresh):
if(heisig):
# print "\nNow determining background levels for height of peaks"
bins = 70 # based on max peak height of 7
counts = [0] * bins
thresh = [0] * bins
for i in range(0, bins):
thresh[i] = (i + 1) * 0.1
for i in waves:
# check all heights to determine background levels
counts[int(i['height'] / 0.1)] += 1 # increment count where height1 is in bin of size 0.1
# print "counts are: ", counts
x = []
y = []
for i in range(10, 15): # (0 to 9 correspond to heights 0 to 0.9, do not have)
x.append(thresh[i])
y.append(counts[i])
# print "x is ", x
# print "y is ", y
slr = scipystats.linregress(x, y)
slope = slr[0]
intercept = slr[1]
# print "slope: %s and intercept: %s for background peak height" % (slope, intercept)
# find x-intercept, threshold for noise-signal
xint = abs(intercept / slope)
print "height threshold between noise and signal is ", xint
thresh = round(xint, 2)
else:
# determine threshold for sigma
bins = 300 # based on max sigma of 300
counts = [0] * bins
threshes = [0] * bins
for i in range(0, bins):
threshes[i] = (i + 1)
for i in waves:
# check all heights to determine background levels
counts[int(i['stddev'])] += 1 # increment count where height1 is in bin of size 0.1
# print "counts are: ", counts
highest = 0
ind = -1
for i in range(0, bins):
if counts[i] > highest:
highest = counts[i]
ind = i
thresh = threshes[ind]
print "sigma threshold between noise and signal is ", thresh
else:
if(heisig):
usr_in = raw_input("What would you like to use as the minimum wave height? ")
thresh = float(usr_in)
else:
usr_in = raw_input("What would you like to use as the minimum wave sigma? ")
thresh = int(usr_in)
# remove waves that don't meet threshold
if(heisig):
waves[:] = [x for x in waves if x['height'] > thresh]
else:
waves[:] = [x for x in waves if x['stddev'] > thresh]
waves.sort(key = lambda x: (x['chr'], x['pos'])) # list sorted by position
usr_in = raw_input("Distance upstream of TSS to check for waves: ")
prom = int(usr_in)
bed = []
f = open(orffile, 'r')
next(f)
for line in f:
a = line.split("\t")
if int(a[1]) == 17: # assumed mitochondrial chromosome
pass
else:
region = {}
region['gene'] = a[0]
# print a[1]
region["chr"] = "chr" + chromosomes[int(a[1])]
# print "chromosome is: ", region["chr"]
region["start"] = int(a[2])
# if int(a[4]) == 1:
# region["start"] = int(a[2]) - prom
# region["end"] = int(a[2])
# else:
# region["start"] = int(a[3])
# region["end"] = int(a[3]) + prom
if int(a[4]) == 1:
region["start"] = int(a[2])
region["end"] = int(a[2]) + prom
else:
region["start"] = int(a[3]) - prom
region["end"] = int(a[3])
bed.append(region)
bed.sort(key = lambda x: (x['chr'], x['start']))
maxperbin = 0
print_queue = multiprocessing.Queue()
# launch thread to read and process the print queue
# print "printing to: ", output + StringUtils.rreplace(os.path.basename(wavesfile), ".normalized.waves", "_" + str(thresh) + "_summary.txt", 1)
print_thread = PrintThread.StringWriter(print_queue, output, StringUtils.rreplace(os.path.basename(wavesfile), ".waves", "_" + str(thresh) + "_promoter_summary.txt", 1), True, True)
print "Now finding peaks in each promoter..."
w = 0
b = 0
count = 0
height = 0
while b < len(bed) and w < len(waves):
if(waves[w]['chr'] == bed[b]['chr'] and (waves[w]['pos'] + 1 * waves[w]['stddev']) >= bed[b]['start'] and (waves[w]['pos'] - 1 * waves[w]['stddev']) <= bed[b]['end']):
# print "found a wave in bin ", b
count += 1
if count > maxperbin:
maxperbin = count
height += waves[w]['height']
waves[w]['used'] = True
w += 1
elif waves[w]['chr'] > bed[b]['chr'] or ((waves[w]['pos'] + 1 * waves[w]['stddev']) > bed[b]['end'] and waves[w]['chr'] == bed[b]['chr']):
# wave is past bin, move to next bin
# write to file
# print "b: %i, w: %i, wave is AFTER, start: %s, end: %s, pos: %s, %s %s" % (b, w, bed[b]['start'], bed[b]['end'], waves[w]['pos'], bed[b]['chr'], waves[w]['chr'])
print_queue.put(bed[b]['gene'] + "\t" + bed[b]['chr'] + "\t" + str(bed[b]['start']) + "\t" + str(bed[b]['end']) + "\t" + str(count) + "\t" + str(height))
b += 1
count = 0
height = 0
else:
# wave is before bin
# print "b: %i, w: %i, wave is BEFORE, start: %s, end: %s, pos: %s, %s %s" % (b, w, bed[b]['start'], bed[b]['end'], waves[w]['pos'], bed[b]['chr'], waves[w]['chr'])
w += 1
while b < len(bed):
# if ended because got to end of waves
# need to write statistics for remaining bins
# print "filling in remaining bins with 0s"
start = bed[b]["start"]
end = bed[b]["end"]
chrom = bed[b]["chr"]
count = 0
height = 0
print_queue.put(bed[b]['gene'] + "\t" + str(chrom) + "\t" + str(start) + "\t" + str(end) + "\t" + str(count) + "\t" + str(height))
b += 1
# end printing
if print_thread is None or not print_thread.is_alive():
pass
else:
while print_queue.qsize() > 0:
print "waiting on print_queue to empty", print_queue.qsize()
time.sleep(1)
print_thread.END_PROCESSES = True
print_thread.f.close()
# while not print_thread.is_closed():
# print "waiting for print_thread to close ", print_queue.qsize(), " ", print_thread.is_closed(), " ", print_thread.END_PROCESSES
# time.sleep(1)
print_queue = multiprocessing.Queue()
# launch thread to read and process the print queue
print_thread = PrintThread.StringWriter(print_queue, output, StringUtils.rreplace(os.path.basename(wavesfile), ".waves", "_" + str(thresh) + "_promoter_counts.txt", 1), True, True)
# quick summary statistics
print_queue.put("nwaves\tnbins\tavgBinSize")
unassigned = 0
un_height = 0
for i in waves:
if not i["used"]:
unassigned += 1
un_height += i["height"]
# print i['height']
# for each bin of height 0.5, determine % of waves not mapped
total = [0] * 13
unmap = [0] * 13
for i in waves:
total[int(i['height'] / 0.5) - 2] += 1 # TODO: this is almost certainly wrong!
if not i['used']:
unmap[int(i['height'] / 0.5) - 2] += 1 # TODO: most certainly wrong.
# TODO: Print these values in some meaningful way.
prop = [0.0] * 13
for i in range(0, len(prop)):
if total[i] != 0:
prop[i] = float(unmap[i]) / float(total[i])
else:
prop[i] = 0
print "Proportions of unused waves in each bin:"
print "[1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0]"
print total
print unmap
print prop
# print "%s waves were unassigned to a BED bin." % unassigned
print "Max per bin is:", maxperbin
print "Height threshold applied to waves:", thresh
counts = [0] * (maxperbin + 1)
# sizes = [[]] * (maxperbin + 1)
sizes = [[] for x in range(0, maxperbin + 1)]
f = open(output + StringUtils.rreplace(os.path.basename(wavesfile), ".waves", "_" + str(thresh) + "_promoter_summary.txt", 1), 'r', 0)
ln = 0
for line in f:
ln += 1
a = line.split("\t")
# print "on line ", ln, " \n line is: ", line, " \nand value is: ", a[4]
counts[int(a[4])] += 1
sizes[int(a[4])].append(int(a[3]) - int(a[2]))
f.close()
print "%s waves of %s were unassigned to a BED bin (%f%%)." % (unassigned, len(waves), round(float(unassigned) / len(waves) * 100, 2))
print "Average height of waves not part of a bin:", (un_height / unassigned)
print_queue.put(str(unassigned) + "\t0\t0")
print "Total number of bins:", len(bed)
for i in range(0, maxperbin + 1):
tot = 0
# print len(sizes[i])
for j in range(0, len(sizes[i])):
tot += sizes[i][j]
# print "size is: ", sizes[i][j]
if len(sizes[i]) != 0:
avg = float(tot) / len(sizes[i])
else:
avg = 0
if i != 0:
print_queue.put(str(i) + "\t" + str(counts[i]) + "\t" + str(avg))
else:
print "%i bins have no waves mapped to them (%s%%)" % (counts[i], float(counts[i]) / len(bed) * 100)
print "%i bins have at least 1 wave mapped to them." % (len(bed) - counts[i])
print_queue.put(str(i) + "\t" + str(counts[i]) + "\t" + str(avg))
# end printing
if print_thread is None or not print_thread.is_alive():
pass
else:
while print_queue.qsize() > 0:
print "waiting on print_queue to empty", print_queue.qsize()
time.sleep(1)
print_thread.END_PROCESSES = True
print_thread.f.close()
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("parameter_file", help = ".input parameterfile", type = str)
parser.add_argument("-data_file", help = "override the source file in the parameter input file, containing data upon which waves will be called", type = str)
parser.add_argument("-output_path", help = "override the path for output files", type = str)
parser.add_argument("-noise_compensation", "-nc", help = "increase the amount of noise acceptable for wave calling - set to 16 for chip-chip, 1 for chip_seq.", type = int, default = 1)
args = parser.parse_args()
p = Parameters.parameter()
param = create_param_obj(args.parameter_file)
param.set("noise_compensation", args.noise_compensation)
# override parameter file with cmdline args
if args.data_file :
param.set("input_file", args.data_file) # override input_file
# set file_name in param to be based on input file.
ofile = StringUtils.rreplace(os.path.basename(args.data_file), '.wig', '', 1)
param.set("file_name", ofile) # override output file_name (.waves gets added later)
if args.output_path:
param.set("output_path", args.output_path) # override output_path
print "param file: ", args.parameter_file
print "input_file: ", param.get("input_file")
print "output_path: ", param.get("output_path")
print "output_file_name: ", param.get("file_name")
main(param)
