def genebody_percentile(refbed, mRNA_len_cut=100):
'''
return percentile points of gene body
mRNA length < mRNA_len_cut will be skipped
'''
import numpy
if refbed is None:
print >> sys.stderr, "You must specify a bed file representing gene model\n"
exit(0)
g_percentiles = {}
transcript_count = 0
for line in open(refbed, 'r'):
try:
if line.startswith(('#', 'track', 'browser')): continue
# Parse fields from gene tabls
fields = line.split()
chrom = fields[0]
tx_start = int(fields[1])
tx_end = int(fields[2])
geneName = fields[3]
strand = fields[5]
geneID = '_'.join(
[str(j) for j in (chrom, tx_start, tx_end, geneName, strand)])
exon_starts = map(int, fields[11].rstrip(',\n').split(','))
exon_starts = map((lambda x: x + tx_start), exon_starts)
exon_ends = map(int, fields[10].rstrip(',\n').split(','))
exon_ends = map((lambda x, y: x + y), exon_starts, exon_ends)
transcript_count += 1
except:
print >> sys.stderr, "[NOTE:input bed must be 12-column] skipped this line: " + line,
continue
gene_all_base = []
mRNA_len = 0
flag = 0
for st, end in zip(exon_starts, exon_ends):
gene_all_base.extend(range(st + 1, end +
1)) #1-based coordinates on genome
if len(gene_all_base) < mRNA_len_cut:
continue
#get 100 points
pos_select = []
g_percentiles[geneID] = (
chrom, strand,
mystat.percentile_list(gene_all_base[-mRNA_len_cut:])
) #get 100 points from each gene's coordinates
if len(gene_all_base) <= mRNA_len_cut:
continue
#get 100 points
pos_select = []
g_percentiles[geneID] = (
chrom, strand, mystat.percentile_list(gene_all_base[mRNA_len_cut:])
) #get 100 points from each gene's coordinates
printlog("Total " + str(transcript_count) + ' transcripts loaded')
return g_percentiles
def genebody_percentile(anno, gene_filter, mRNA_len_cut = 100):
'''
return percentile points of gene body
mRNA length < mRNA_len_cut will be skipped
'''
g_percentiles = {}
g_filter = []
if gene_filter:
with open(gene_filter) as f:
for line in f:
line = line.rstrip()
word = line.split("\t")
g_filter.append(word[0])
for line in open(anno,'r'):
if line.startswith('Ensembl'):continue
# Parse fields from gene tabls
fields = line.split()
if fields[1] == "MT": chrom = "chrM"
elif fields[1] == "X": chrom = "chrX"
elif fields[1] == "Y": chrom = "chrY"
elif fields[1].isdigit(): chrom = "chr" + fields[1]
else:
continue
tx_start = int( fields[2] )
tx_end = int( fields[3] )
geneName = fields[0]
if fields[4] == "1":
strand = "+"
else:
strand = "-"
geneID = '_'.join([str(j) for j in (chrom, tx_start, tx_end, geneName, strand)])
gene_all_base=[]
if g_filter:
if geneName in g_filter:
gene_all_base.extend(range(tx_start+1,tx_end+1)) #1-based coordinates on genome
if len(gene_all_base) < mRNA_len_cut:
continue
g_percentiles[geneID] = (chrom, strand, mystat.percentile_list (gene_all_base)) #get 100 points from each gene's coordinates
else:
gene_all_base.extend(range(tx_start+1,tx_end+1)) #1-based coordinates on genome
if len(gene_all_base) < mRNA_len_cut:
continue
g_percentiles[geneID] = (chrom, strand, mystat.percentile_list (gene_all_base)) #get 100 points from each gene's coordinates
return g_percentiles
def Rcode_write(dataset, file_prefix, format='pdf', length=100):
'''generate R script for visualization'''
ROUT = open(file_prefix + '.r', 'w')
names = []
datas = []
for name, data, tmp in dataset:
names.append(name)
datas.append(data)
print >> ROUT, name + ' <- c(' + ','.join([str(i) for i in data]) + ')'
x = mystat.percentile_list(range(1, length + 1))
print >> ROUT, '\n'
print >> ROUT, '%s(\"%s.%s\")' % (format.lower(), file_prefix + ".curves",
format.lower())
print >> ROUT, "x %s" % (' <- c(' + ','.join([str(i) for i in x]) + ')')
print >> ROUT, 'icolor = colorRampPalette(c("#7fc97f","#beaed4","#fdc086","#ffff99","#386cb0","#f0027f"))(%d)' % (
len(names))
if len(names) == 1:
print >> ROUT, "plot(x,%s,type='l',xlab=\"Gene body, bp (5\'->3\')\", ylab=\"Starts coverage\",lwd=0.8,col=icolor[1])" % (
names[0])
elif len(names) >= 2 and len(names) <= 6:
print >> ROUT, "plot(x,%s,type='l',xlab=\"Gene body, bp (5\'->3\')\", ylab=\"Starts coverage\",lwd=0.8,col=icolor[1])" % (
names[0])
for i in range(1, len(names)):
print >> ROUT, "lines(x,%s,type='l',col=icolor[%d])" % (names[i],
i + 1)
print >> ROUT, "legend(0,1,fill=icolor[%d:%d], legend=c(%s))" % (
1, len(names), ','.join(["'" + str(n) + "'" for n in names]))
elif len(names) > 6:
print >> ROUT, 'layout(matrix(c(1,1,1,2,1,1,1,2,1,1,1,2), 4, 4, byrow = TRUE))'
print >> ROUT, "plot(x,%s,type='l',xlab=\"Gene body, bp (5\'->3\')\", ylab=\"Starts coverage\",lwd=0.8,col=icolor[1])" % (
names[0])
for i in range(1, len(names)):
print >> ROUT, "lines(x,%s,type='l',col=icolor[%d])" % (names[i],
i + 1)
print >> ROUT, 'par(mar=c(1,0,2,1))'
print >> ROUT, 'plot.new()'
print >> ROUT, "legend(0,1,fill=icolor[%d:%d],legend=c(%s))" % (
1, len(names), ','.join(["'" + str(n) + "'" for n in names]))
print >> ROUT, 'dev.off()'
ROUT.close()
def genebody_percentile(refbed, mRNA_len_cut = 100):
'''
return percentile points of gene body
mRNA length < mRNA_len_cut will be skipped
'''
if refbed is None:
print >>sys.stderr,"You must specify a bed file representing gene model\n"
exit(0)
g_percentiles = {}
transcript_count = 0
for line in open(refbed,'r'):
try:
if line.startswith(('#','track','browser')):continue
# Parse fields from gene tabls
fields = line.split()
chrom = fields[0]
tx_start = int( fields[1] )
tx_end = int( fields[2] )
geneName = fields[3]
strand = fields[5]
geneID = '_'.join([str(j) for j in (chrom, tx_start, tx_end, geneName, strand)])
exon_starts = map( int, fields[11].rstrip( ',\n' ).split( ',' ) )
exon_starts = map((lambda x: x + tx_start ), exon_starts)
exon_ends = map( int, fields[10].rstrip( ',\n' ).split( ',' ) )
exon_ends = map((lambda x, y: x + y ), exon_starts, exon_ends)
transcript_count += 1
except:
print >>sys.stderr,"[NOTE:input bed must be 12-column] skipped this line: " + line,
continue
gene_all_base=[]
mRNA_len =0
flag=0
for st,end in zip(exon_starts,exon_ends):
gene_all_base.extend(range(st+1,end+1)) #1-based coordinates on genome
if len(gene_all_base) < mRNA_len_cut:
continue
g_percentiles[geneID] = (chrom, strand, mystat.percentile_list (gene_all_base)) #get 100 points from each gene's coordinates
printlog("Total " + str(transcript_count) + ' transcripts loaded')
return g_percentiles
def coverageGeneBody_bigwig(bigFile, refbed, outfile, gtype="png"):
'''Calculate reads coverage over gene body, from 5'to 3'. each gene will be equally divided
into 100 regsions. bigFile is bigwig format file'''
if refbed is None:
print >> sys.stderr, "You must specify a bed file representing gene model\n"
exit(0)
OUT1 = open(outfile + ".geneBodyCoverage_plot.r", 'w')
OUT2 = open(outfile + ".geneBodyCoverage.txt", 'w')
bw = BigWigFile(file=open(bigFile))
print >> sys.stderr, "calculating coverage over gene body ..."
coverage = collections.defaultdict(int)
flag = 0
gene_count = 0
for line in open(refbed, 'r'):
try:
if line.startswith(('#', 'track', 'browser')): continue
gene_count += 1
# Parse fields from gene tabls
fields = line.split()
chrom = fields[0]
tx_start = int(fields[1])
tx_end = int(fields[2])
geneName = fields[3]
strand = fields[5]
exon_starts = map(int, fields[11].rstrip(',\n').split(','))
exon_starts = map((lambda x: x + tx_start), exon_starts)
exon_ends = map(int, fields[10].rstrip(',\n').split(','))
exon_ends = map((lambda x, y: x + y), exon_starts, exon_ends)
except:
print >> sys.stderr, "[NOTE:input bed must be 12-column] skipped this line: " + line,
continue
gene_all_base = []
percentile_base = []
mRNA_len = 0
flag = 0
for st, end in zip(exon_starts, exon_ends):
gene_all_base.extend(range(st + 1, end +
1)) #0-based coordinates on genome
mRNA_len = len(gene_all_base)
if mRNA_len < 100:
flag = 1
break
if flag == 1: continue
if strand == '-':
gene_all_base.sort(reverse=True) #deal with gene on minus stand
else:
gene_all_base.sort(reverse=False)
percentile_base = mystat.percentile_list(
gene_all_base) #get 101 points from each gene's coordinates
for i in range(0, len(percentile_base)):
#try:
sig = bw.get_as_array(chrom, percentile_base[i] - 1,
percentile_base[i])
if sig is None: continue
coverage[i] += np.nan_to_num(sig[0])
#except:
# continue
print >> sys.stderr, " %d genes finished\r" % gene_count,
x_coord = []
y_coord = []
print >> OUT2, "percentile\tcount"
for i in coverage:
x_coord.append(str(i))
y_coord.append(str(coverage[i]))
print >> OUT2, str(i) + '\t' + str(coverage[i])
print >> OUT1, "%s(\'%s\')" % (gtype,
outfile + ".geneBodyCoverage." + gtype)
print >> OUT1, "x=0:100"
print >> OUT1, "y=c(" + ','.join(y_coord) + ')'
print >> OUT1, "plot(x,y/%s,xlab=\"percentile of gene body (5'->3')\",ylab='average wigsum',type='s')" % gene_count
print >> OUT1, "dev.off()"
def coverageGeneBody_bigwig(bigFile, refbed, outfile, gtype='png'):
'''Calculate reads coverage over gene body, from 5'to 3'. each gene will be equally divided
into 100 regsions. bigFile is bigwig format file'''
if refbed is None:
print("You must specify a bed file representing gene model",
file=sys.stderr)
exit(0)
bw = openBigWig(bigFile, 'r')
# Get chromosomes present in the bigwig file
chroms = bw.chroms().keys()
print("calculating coverage over gene body ...", file=sys.stderr)
coverage = defaultdict(int)
flag = 0
gene_count = 0
with open(refbed, 'r') as handle:
# Loop through the genes in the BED file
for line in handle:
try:
if line.startswith(('#', 'track', 'browser')):
continue
# Parse fields from gene tabls
fields = line.split()
chrom = fields[0]
tx_start = int(fields[1])
tx_end = int(fields[2])
geneName = fields[3]
strand = fields[5]
# Skip chromosomes present in the bed file but not present in the bigwig file.
# This could happen with PATCHES or Unplaced chromosomes.
if chrom not in chroms:
continue
exon_starts = list(
map(int, fields[11].rstrip(',\n').split(',')))
exon_starts = list(map((lambda x: x + tx_start), exon_starts))
exon_ends = list(map(int, fields[10].rstrip(',\n').split(',')))
exon_ends = list(
map((lambda x, y: x + y), exon_starts, exon_ends))
except:
print(
"[NOTE: input bed must be 12-column] skipped this line: " +
line,
end='\n',
file=sys.stderr)
continue
# Count gene if it was properly read
gene_count += 1
gene_all_base = []
mRNA_len = 0
flag = 0
for st, end in zip(exon_starts, exon_ends):
gene_all_base.extend(range(st + 1, end +
1)) # 0-based coordinates on genome
mRNA_len = len(gene_all_base)
if mRNA_len < 100:
flag = 1
break
if flag == 1:
continue
# Sort coordinates according to the strand
if strand == '-':
gene_all_base.sort(reverse=True)
else:
gene_all_base.sort(reverse=False)
# Get 100 points from each gene's coordinates
percentile_base = []
percentile_base = mystat.percentile_list(gene_all_base)
for i in range(0, len(percentile_base)):
sig = bw.values(chrom, percentile_base[i] - 1,
percentile_base[i])
coverage[i] += nan_to_num(sig[0])
print(" \t%d genes finished\r" % gene_count,
end=' ',
file=sys.stderr)
# Close bigwig file
bw.close()
print("\n", file=sys.stderr)
x_coord = []
y_coord = []
with open(outfile + ".geneBodyCoverage.txt", 'w') as handle:
handle.write("percentile\tcount\n")
for i in coverage:
x_coord.append(str(i))
y_coord.append(str(coverage[i]))
handle.write("%i\t%i\n" % (i, coverage[i]))
with open(outfile + ".geneBodyCoverage_plot.r", 'w') as handle:
handle.write("%s(\'%s\')\n" %
(gtype, outfile + ".geneBodyCoverage." + gtype))
handle.write("x=1:100\n")
handle.write("y=c(%s)\n" % ','.join(y_coord))
handle.write(
"plot(x, y/%s, xlab=\"percentile of gene body (5'->3')\", ylab='average wigsum', type='s')\n"
% gene_count)
handle.write("dev.off()\n")
def coverageGeneBody_bigwig(bigFile,refbed,outfile,gtype="png"):
'''Calculate reads coverage over gene body, from 5'to 3'. each gene will be equally divided
into 100 regsions. bigFile is bigwig format file'''
if refbed is None:
print >>sys.stderr,"You must specify a bed file representing gene model\n"
exit(0)
OUT1 = open(outfile + ".geneBodyCoverage_plot.r",'w')
OUT2 = open(outfile + ".geneBodyCoverage.txt",'w')
bw = BigWigFile( file = open(bigFile) )
print >>sys.stderr, "calculating coverage over gene body ..."
coverage=collections.defaultdict(int)
flag=0
gene_count = 0
for line in open(refbed,'r'):
try:
if line.startswith(('#','track','browser')):continue
gene_count += 1
# Parse fields from gene tabls
fields = line.split()
chrom = fields[0]
tx_start = int( fields[1] )
tx_end = int( fields[2] )
geneName = fields[3]
strand = fields[5]
exon_starts = map( int, fields[11].rstrip( ',\n' ).split( ',' ) )
exon_starts = map((lambda x: x + tx_start ), exon_starts)
exon_ends = map( int, fields[10].rstrip( ',\n' ).split( ',' ) )
exon_ends = map((lambda x, y: x + y ), exon_starts, exon_ends);
except:
print >>sys.stderr,"[NOTE:input bed must be 12-column] skipped this line: " + line,
continue
gene_all_base=[]
percentile_base=[]
mRNA_len =0
flag=0
for st,end in zip(exon_starts,exon_ends):
gene_all_base.extend(range(st+1,end+1)) #0-based coordinates on genome
mRNA_len = len(gene_all_base)
if mRNA_len <100:
flag=1
break
if flag==1: continue
if strand == '-':
gene_all_base.sort(reverse=True) #deal with gene on minus stand
else:
gene_all_base.sort(reverse=False)
percentile_base = mystat.percentile_list (gene_all_base) #get 101 points from each gene's coordinates
for i in range(0,len(percentile_base)):
#try:
sig = bw.get_as_array(chrom,percentile_base[i]-1,percentile_base[i])
if sig is None:continue
coverage[i] += np.nan_to_num(sig[0])
#except:
# continue
print >>sys.stderr, " %d genes finished\r" % gene_count,
x_coord=[]
y_coord=[]
print >>OUT2, "percentile\tcount"
for i in coverage:
x_coord.append(str(i))
y_coord.append(str(coverage[i]))
print >>OUT2, str(i) + '\t' + str(coverage[i])
print >>OUT1, "%s(\'%s\')" % (gtype, outfile + ".geneBodyCoverage." + gtype)
print >>OUT1, "x=0:100"
print >>OUT1, "y=c(" + ','.join(y_coord) + ')'
print >>OUT1, "plot(x,y/%s,xlab=\"percentile of gene body (5'->3')\",ylab='average wigsum',type='s')" % gene_count
print >>OUT1, "dev.off()"
