052f3c799ed1d1d2468ebedc6d2e70cb261b52df max Wed Sep 9 01:59:39 2020 -0700 allowing comments, fixing bug and removing dead code from chromToUcsc, refs #26053 diff --git src/utils/chromToUcsc/chromToUcsc src/utils/chromToUcsc/chromToUcsc index 8d95d5f..7b55228 100755 --- src/utils/chromToUcsc/chromToUcsc +++ src/utils/chromToUcsc/chromToUcsc @@ -1,165 +1,156 @@ #!/usr/bin/env python import logging, optparse, gzip from sys import stdin, stdout, stderr, exit, modules from os.path import basename try: from urllib.request import urlopen # py2 except ImportError: from urllib2 import urlopen # py3 try: from cStringIO import StringIO # py2 except ImportError: from io import BytesIO # py3 # ==== functions ===== def parseArgs(): " setup logging, parse command line arguments and options. -h shows auto-generated help page " parser = optparse.OptionParser("""usage: %prog [options] filename - change NCBI or Ensembl chromosome names to UCSC names using the chromAlias table of the genome browser. Requires a <genome>.chromAlias.tsv file which can be downloaded like this: %prog --get hg19 # download the file hg19.chromAlias.tsv into current directory If you do not want to use the --get option to retrieve the mapping tables, you can also download the alias mapping files yourself, e.g. for mm10 with 'wget https://hgdownload.soe.ucsc.edu/goldenPath/mm10/database/chromAlias.txt.gz' Then the script can be run like this: %prog -i in.bed -o out.bed -a hg19.chromAlias.tsv %prog -i in.bed -o out.bed -a https://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/chromAlias.txt.gz Or in pipes, like this: cat test.bed | %prog -a mm10.chromAlias.tsv > test.ucsc.bed """) - parser.add_option("", "--get", dest="doDownload", action="store", help="download a chrom alias table from UCSC for the genomeDb into the current directory and exit") + parser.add_option("", "--get", dest="downloadDb", action="store", help="download a chrom alias table from UCSC for the genomeDb into the current directory and exit") parser.add_option("-a", "--chromAlias", dest="aliasFname", action="store", help="a UCSC chromAlias file in tab-sep format. or a URL to one") parser.add_option("-i", "--in", dest="inFname", action="store", help="input filename, default: /dev/stdin") parser.add_option("-o", "--out", dest="outFname", action="store", help="output filename, default: /dev/stdout") parser.add_option("-d", "--debug", dest="debug", action="store_true", help="show debug messages") parser.add_option("-k", "--field", dest="fieldNo", action="store", type="int", \ help="index of field to convert, default is %default (1-based), for most formats, e.g. BED. For genePred, the chromosome is field 2, for PSL it is 10 or 14.", default=1) (options, args) = parser.parse_args() - if options.doDownload and not options.db: - print("If you use --get you need to provide a genome assembly code like 'mm10' with the -g option") - exit(1) - - if options.doDownload is None and options.aliasFname is None: + if options.downloadDb is None and options.aliasFname is None: parser.print_help() exit(1) if options.debug: logging.basicConfig(level=logging.DEBUG) else: logging.basicConfig(level=logging.INFO) return args, options # ----------- main -------------- def splitLines(ifh): " yield (chromName, restOfLine) for all lines of ifh " sep = -1 #if (sys.version_info > (3, 0)): lineNo = 0 for line in ifh: if sep==-1: if "\t" in line: sep = "\t" else: sep = None # = split on any whitespace, consec. whitespc counts as one #chrom, rest = line.split(sep, 1) row = line.rstrip("\n\t").split(sep) lineNo += 1 yield lineNo, sep, row def parseAlias(fname): " parse tsv file with at least two columns, orig chrom name and new chrom name " toUcsc = {} if fname.startswith("http://") or fname.startswith("https://"): ifh = urlopen(fname) if fname.endswith(".gz"): data = gzip.GzipFile(fileobj=ifh).read().decode() ifh = data.splitlines() elif fname.endswith(".gz"): ifh = gzip.open(fname) else: ifh = open(fname) for line in ifh: if line.startswith("alias"): continue row = line.rstrip("\n").split("\t") toUcsc[row[0]] = row[1] return toUcsc def chromToUcsc(aliasFname, fieldIdx, ifh, ofh): - " convert the first column to UCSC-style chrom names " + " convert column number fieldIdx to UCSC-style chrom names " toUcsc = parseAlias(aliasFname) ucscChroms = set(toUcsc.values()) - mtSkipCount = 0 - for lineNo, sep, row in splitLines(ifh): # just pass through any UCSC chrom names chrom = row[fieldIdx] - if chrom in ucscChroms: + if row[0].startswith("#") or chrom in ucscChroms: ucscChrom = chrom else: ucscChrom = toUcsc.get(chrom) if ucscChrom is None: logging.error("line %d: chrom name %s is not in chromAlias table" % (lineNo, repr(chrom))) exit(1) continue row[fieldIdx] = ucscChrom ofh.write(sep.join(row)) ofh.write("\n") - if mtSkipCount!=0: - stderr.write("%d features were skipped because they were located on the M chromosome. hg19 includes an older version of the mitochondrial genome and these features cannot be mapped yet.\n" % mtSkipCount) - def download(db): url = "http://hgdownload.soe.ucsc.edu/goldenPath/%s/database/chromAlias.txt.gz" % db gzData = urlopen(url).read() if 'cStringIO' in modules: data = StringIO(gzData) else: data = BytesIO(gzData) data = gzip.GzipFile(fileobj=data).read().decode() outFname = db+".chromAlias.tsv" open(outFname, "w").write(data) print("Wrote %s to %s" % (url, outFname)) print("You can now convert a file with 'chromToUcsc -a %s -i infile.bed -o outfile.bed'" % outFname) exit(0) def main(): args, options = parseArgs() aliasFname = options.aliasFname inFname = options.inFname outFname = options.outFname - if options.doDownload: - download(db) + if options.downloadDb: + download(options.downloadDb) if aliasFname is None: logging.error("You need to provide an alias table with the -a option or use --get to download one.") exit(1) if inFname is None: ifh = stdin else: ifh = open(inFname) if outFname is None: ofh = stdout else: ofh = open(outFname, "w") fieldIdx = options.fieldNo-1 chromToUcsc(aliasFname, fieldIdx, ifh, ofh) main()