def getSeeds(fafile,sense=True,start_at=2,stop_at=8):
if sense:
for fa in IO.BioReader(fafile,'fasta'):
yield fa.id,fa.seq.seq[start_at-1:stop_at-1]
else:
for fa in IO.BioReader(fafile,'fasta'):
yield fa.id,Utils.rc(fa.seq.seq)[start_at-1:stop_at-1]
def calPvalues(siRNASeeds,seedfile,outfile,method='RRA',N=10000):
# read headers
print >>sys.stderr, Utils.touchtime(),"reading header"
fh = gzip.open(seedfile,'rb')
genes,lengths = SeedUtils.parseHeader(fh)
num_genes = len(genes)
# aRRA
print >>sys.stderr, Utils.touchtime(),"start analyzing data."
method = {'RRA':SeedUtils.aRRA}.get(method,SeedUtils.aRRA)
escores = numpy.zeros(num_genes)
for i in range(0,num_genes,N):
n = N if i+N < num_genes else num_genes%N
if n:
geneSeeds = numpy.sum(numpy.frombuffer(fh.read(2**17*n),dtype=numpy.uint16,count=2**16*n).reshape((n,4,2**14)),axis=1)
escores[i:i+n] = method(siRNASeeds,geneSeeds)
print Utils.touchtime(), "analyzed", i+n
fh.close()
# Writting result to file
print >>sys.stderr, Utils.touchtime(), 'writing result to file.'
df = pandas.DataFrame({'gid':genes,'length':lengths, 'escore':escores})
rv = SeedUtils.evdplot(df,outfile+".pdf")
df.loc[:,'pvalue'] = rv.pdf(df.escore)
df = df.sort_values(by='pvalue',ascending=False)
df.to_csv(outfile,index=False,sep='\t',columns=['gid','length','escore','pvalue'])
print >>sys.stderr, Utils.touchtime(), "finished."
return df
def parseGeneSeeds(gfile,dumpfile,N=10):
'''
Parse seeds on gene sequences.
Parameters:
gfile: string
Gene sequence file in Fasta format.
N: int
Parse N sequence at a time instead all to reduce memory usage. [Default=10000]
N = 10000, memory usage for the two huge matrix is up to 1.83G.
'''
# create seed tables
SeedUtils.createTables()
# read genes
ofh = gzip.open(dumpfile,'wb')
genes = [fa for fa in IO.BioReader(gfile,'fasta')]
# write headers
gids = ';'.join(["{0}:{1}".format(fa.id,len(fa)) for fa in genes])
header = numpy.array([0x19840405,len(genes),len(gids)],dtype=numpy.uint64)
ofh.write(header.data)
ofh.write(gids)
# parse genes
cnt = 0
mat = numpy.zeros((N,2**15),dtype=numpy.uint16)
seeds = numpy.zeros((N,4,2**14),dtype=numpy.uint16) # N x type x seeds
for fa in genes:
SeedUtils.findSeeds(fa.seq.seq.upper(),mat[cnt%N])
cnt += 1
if cnt%(N)==0:
# calculate seeds for the N genes
SeedUtils.parseSeeds(mat,seeds)
ofh.write(seeds.data)
print >>sys.stderr, touchtime(), "Parsed {0} genes .".format(cnt)
seeds.fill(0)
mat.fill(0)
# calculate seeds for the rest of genes
rest = cnt%N
if rest:
SeedUtils.parseSeeds(mat[:rest],seeds[:rest])
ofh.write(seeds[:rest].data)
print >>sys.stderr, Utils.touchtime(), "Parsed {0} genes. ".format(cnt)
ofh.close()
p._optionals.title = "Options"
p.add_argument("-i","--input",dest='ifname',type=str,metavar="input.bed",required=True,help="Input file. Can be stdin.")
p.add_argument("-f","--format",dest="ftype",type=str,metavar="bed",default="bed",help="Format of input file. Default is 'bed'. Can be 'bed3', 'bedgraph', 'bed','peak','wig', 'sam2bed' or 'genepred'.")
p.add_argument("-g","--genome",dest='genome',type=str,metavar="Genome",default=None,help="Genome version (hg19, mm10 .etc) or genome size file with chrom and size in each line.")
p.add_argument("-u","--up",dest="up",type=int,metavar="upstream",default=0,help="bps extended to upstream. If minus, trim the 5' end.")
p.add_argument("-d","--down",dest="down",type=int,metavar="downstream",default=0,help="bps extended to downstream. If minus, trim the 3' end.")
p.add_argument("-o","--output",dest="ofname",type=str,metavar="output.bed",default="stdout",help="Output file. Default is stdout.")
if len(sys.argv)==1:
sys.exit(p.print_help())
args = p.parse_args()
return args
# ------------------------------------
# Classes
# ------------------------------------
# ------------------------------------
# Main
# ------------------------------------
if __name__=="__main__":
# Get parameters
args=argParser()
fh = IO.mopen(args.ofname, 'w')
if args.genome:
genome=Utils.genomeSize(args.genome)
for item in IO.BioReader(args.ifname,args.ftype):
tbed=item.extend(args.up,args.down, args.genome)
print >> fh, tbed
IO.mclose(fh)
if __name__=="__main__":
if len(sys.argv)==1:
print "Usage: "+sys.argv[0]+" annotation.tab/bed genomesize *.bed"
print " Find the nearest annotation for given bed."
else:
# check file
if '.tab' in sys.argv[1]:
ftype='gene'
else:
ftype='bed'
# initiation annotations.
annos={}
#for chrom in IO.genomeSize('hg19'):
for chrom in Utils.genomeSize(sys.argv[2]):
if ftype=='bed':
annos[chrom]=BedList()
else:
annos[chrom]=GeneBedList()
# read annotations.
for anno in IO.BioReader(sys.argv[1],ftype=ftype):
if annos.has_key(anno.chrom):
annos[anno.chrom].append(anno)
# sort
for chrom in annos:
annos[chrom].sort()
# Find nearest annoations
def findSeeds(seq,sary):
sary.fill(0)
seq = Utils.rc(seq)
for i in range(len(seq)-7):
sary[twoBytesTable[seq[i:i+8]]] += 1
"--output",
dest="ofname",
type=str,
metavar="output.bed",
default="stdout",
help="Output file. Default is stdout.")
if len(sys.argv) == 1:
sys.exit(p.print_help())
args = p.parse_args()
return args
# ------------------------------------
# Classes
# ------------------------------------
# ------------------------------------
# Main
# ------------------------------------
if __name__ == "__main__":
# Get parameters
args = argParser()
fh = IO.mopen(args.ofname, 'w')
if args.genome:
genome = Utils.genomeSize(args.genome)
for item in IO.BioReader(args.ifname, args.ftype):
tbed = item.extend(args.up, args.down, args.genome)
print >> fh, tbed
IO.mclose(fh)
