def main(argv=None):
if argv is None:
argv = sys.argv
parser = E.ArgumentParser(description=__doc__)
parser.add_argument("--version", action='version', version="1.0")
parser.add_argument(
"-m",
"--method",
dest="method",
type=str,
choices=("add-flank", "add-upstream-flank", "add-downstream-flank",
"crop", "crop-unique", "complement-groups", "combine-groups",
"filter-range", "join-features", "merge-features", "sanitize",
"to-forward-coordinates", "to-forward-strand", "rename-chr"),
help="method to apply ")
parser.add_argument("--ignore-strand",
dest="ignore_strand",
help="ignore strand information.",
action="store_true")
parser.add_argument("--is-gtf",
dest="is_gtf",
action="store_true",
help="input will be treated as gtf.")
parser.add_argument("-c",
"--contigs-tsv-file",
dest="input_filename_contigs",
type=str,
help="filename with contig lengths.")
parser.add_argument(
"--agp-file",
dest="input_filename_agp",
type=str,
help="agp file to map coordinates from contigs to scaffolds.")
parser.add_argument("-g",
"--genome-file",
dest="genome_file",
type=str,
help="filename with genome.")
parser.add_argument("--crop-gff-file",
dest="filename_crop_gff",
type=str,
help="GFF/GTF file to crop against.")
parser.add_argument(
"--group-field",
dest="group_field",
type=str,
help="""gff field/attribute to group by such as gene_id, "
"transcript_id, ... .""")
parser.add_argument(
"--filter-range",
dest="filter_range",
type=str,
help="extract all elements overlapping a range. A range is "
"specified by eithor 'contig:from..to', 'contig:+:from..to', "
"or 'from,to' .")
parser.add_argument("--sanitize-method",
dest="sanitize_method",
type=str,
choices=("ucsc", "ensembl", "genome"),
help="method to use for sanitizing chromosome names. "
".")
parser.add_argument(
"--flank-method",
dest="flank_method",
type=str,
choices=("add", "extend"),
help="method to use for adding flanks. ``extend`` will "
"extend existing features, while ``add`` will add new features. "
".")
parser.add_argument("--skip-missing",
dest="skip_missing",
action="store_true",
help="skip entries on missing contigs. Otherwise an "
"exception is raised .")
parser.add_argument(
"--contig-pattern",
dest="contig_pattern",
type=str,
help="a comma separated list of regular expressions specifying "
"contigs to be removed when running method sanitize .")
parser.add_argument(
"--assembly-report",
dest="assembly_report",
type=str,
help="path to assembly report file which allows mapping of "
"ensembl to ucsc contigs when running method sanitize .")
parser.add_argument(
"--assembly-report-hasids",
dest="assembly_report_hasIDs",
type=int,
help="path to assembly report file which allows mapping of "
"ensembl to ucsc contigs when running method sanitize .")
parser.add_argument(
"--assembly-report-ucsccol",
dest="assembly_report_ucsccol",
type=int,
help="column in the assembly report containing ucsc contig ids"
".")
parser.add_argument(
"--assembly-report-ensemblcol",
dest="assembly_report_ensemblcol",
type=int,
help="column in the assembly report containing ensembl contig ids")
parser.add_argument(
"--assembly-extras",
dest="assembly_extras",
type=str,
help="additional mismatches between gtf and fasta to fix when"
"sanitizing the genome .")
parser.add_argument("--extension-upstream",
dest="extension_upstream",
type=float,
help="extension for upstream end .")
parser.add_argument("--extension-downstream",
dest="extension_downstream",
type=float,
help="extension for downstream end .")
parser.add_argument("--min-distance",
dest="min_distance",
type=int,
help="minimum distance of features to merge/join .")
parser.add_argument("--max-distance",
dest="max_distance",
type=int,
help="maximum distance of features to merge/join .")
parser.add_argument("--min-features",
dest="min_features",
type=int,
help="minimum number of features to merge/join .")
parser.add_argument("--max-features",
dest="max_features",
type=int,
help="maximum number of features to merge/join .")
parser.add_argument(
"--rename-chr-file",
dest="rename_chr_file",
type=str,
help="mapping table between old and new chromosome names."
"TAB separated 2-column file.")
parser.set_defaults(input_filename_contigs=False,
filename_crop_gff=None,
input_filename_agp=False,
genome_file=None,
rename_chr_file=None,
add_up_flank=None,
add_down_flank=None,
complement_groups=False,
crop=None,
crop_unique=False,
ignore_strand=False,
filter_range=None,
min_distance=0,
max_distance=0,
min_features=1,
max_features=0,
extension_upstream=1000,
extension_downstream=1000,
sanitize_method="ucsc",
flank_method="add",
output_format="%06i",
skip_missing=False,
is_gtf=False,
group_field=None,
contig_pattern=None,
assembly_report=None,
assembly_report_hasIDs=1,
assembly_report_ensemblcol=4,
assembly_report_ucsccol=9,
assembly_extras=None)
(args) = E.start(parser, argv=argv)
contigs = None
genome_fasta = None
chr_map = None
if args.input_filename_contigs:
contigs = Genomics.readContigSizes(
iotools.open_file(args.input_filename_contigs, "r"))
if args.genome_file:
genome_fasta = IndexedFasta.IndexedFasta(args.genome_file)
contigs = genome_fasta.getContigSizes()
if args.rename_chr_file:
chr_map = {}
with open(args.rename_chr_file, 'r') as filein:
reader = csv.reader(filein, delimiter='\t')
for row in reader:
if len(row) != 2:
raise ValueError(
"Mapping table must have exactly two columns")
chr_map[row[0]] = row[1]
if not len(chr_map.keys()) > 0:
raise ValueError("Empty mapping dictionnary")
if args.assembly_report:
df = pd.read_csv(args.assembly_report,
comment="#",
header=None,
sep="\t")
# fixes naming inconsistency in assembly report: ensembl chromosome
# contigs found in columnn 0, ensembl unassigned contigs found in
# column 4.
if args.assembly_report_hasIDs == 1:
ucsccol = args.assembly_report_ucsccol
ensemblcol = args.assembly_report_ensemblcol
df.loc[df[1] == "assembled-molecule",
ensemblcol] = df.loc[df[1] == "assembled-molecule", 0]
if args.sanitize_method == "ucsc":
assembly_dict = df.set_index(ensemblcol)[ucsccol].to_dict()
elif args.sanitize_method == "ensembl":
assembly_dict = df.set_index(ucsccol)[ensemblcol].to_dict()
else:
raise ValueError(''' When using assembly report,
please specify sanitize method as either
"ucsc" or "ensembl" to specify direction of conversion
''')
else:
assembly_dict = {}
if args.assembly_extras is not None:
assembly_extras = args.assembly_extras.split(",")
for item in assembly_extras:
item = item.split("-")
assembly_dict[item[0]] = item[1]
if args.method in ("forward_coordinates", "forward_strand",
"add-flank", "add-upstream-flank",
"add-downstream-flank") \
and not contigs:
raise ValueError("inverting coordinates requires genome file")
if args.input_filename_agp:
agp = AGP.AGP()
agp.readFromFile(iotools.open_file(args.input_filename_agp, "r"))
else:
agp = None
gffs = GTF.iterator(args.stdin)
if args.method in ("add-upstream-flank", "add-downstream-flank",
"add-flank"):
add_upstream_flank = "add-upstream-flank" == args.method
add_downstream_flank = "add-downstream-flank" == args.method
if args.method == "add-flank":
add_upstream_flank = add_downstream_flank = True
upstream_flank = int(args.extension_upstream)
downstream_flank = int(args.extension_downstream)
extend_flank = args.flank_method == "extend"
if args.is_gtf:
iterator = GTF.flat_gene_iterator(gffs)
else:
iterator = GTF.joined_iterator(gffs, args.group_field)
for chunk in iterator:
is_positive = Genomics.IsPositiveStrand(chunk[0].strand)
chunk.sort(key=lambda x: (x.contig, x.start))
lcontig = contigs[chunk[0].contig]
if extend_flank:
if add_upstream_flank:
if is_positive:
chunk[0].start = max(0,
chunk[0].start - upstream_flank)
else:
chunk[-1].end = min(lcontig,
chunk[-1].end + upstream_flank)
if add_downstream_flank:
if is_positive:
chunk[-1].end = min(lcontig,
chunk[-1].end + downstream_flank)
else:
chunk[0].start = max(0,
chunk[0].start - downstream_flank)
else:
if add_upstream_flank:
gff = GTF.Entry()
if is_positive:
gff.copy(chunk[0])
gff.end = gff.start
gff.start = max(0, gff.start - upstream_flank)
chunk.insert(0, gff)
else:
gff.copy(chunk[-1])
gff.start = gff.end
gff.end = min(lcontig, gff.end + upstream_flank)
chunk.append(gff)
gff.feature = "5-Flank"
gff.mMethod = "gff2gff"
if add_downstream_flank:
gff = GTF.Entry()
if is_positive:
gff.copy(chunk[-1])
gff.start = gff.end
gff.end = min(lcontig, gff.end + downstream_flank)
chunk.append(gff)
else:
gff.copy(chunk[0])
gff.end = gff.start
gff.start = max(0, gff.start - downstream_flank)
chunk.insert(0, gff)
gff.feature = "3-Flank"
gff.mMethod = "gff2gff"
if not is_positive:
chunk.reverse()
for gff in chunk:
args.stdout.write(str(gff) + "\n")
elif args.method == "complement-groups":
iterator = GTF.joined_iterator(gffs, group_field=args.group_field)
for chunk in iterator:
if args.is_gtf:
chunk = [x for x in chunk if x.feature == "exon"]
if len(chunk) == 0:
continue
chunk.sort(key=lambda x: (x.contig, x.start))
x = GTF.Entry()
x.copy(chunk[0])
x.start = x.end
x.feature = "intron"
for c in chunk[1:]:
x.end = c.start
args.stdout.write(str(x) + "\n")
x.start = c.end
elif args.method == "combine-groups":
iterator = GTF.joined_iterator(gffs, group_field=args.group_field)
for chunk in iterator:
chunk.sort(key=lambda x: (x.contig, x.start))
x = GTF.Entry()
x.copy(chunk[0])
x.end = chunk[-1].end
x.feature = "segment"
args.stdout.write(str(x) + "\n")
elif args.method == "join-features":
for gff in combineGFF(gffs,
min_distance=args.min_distance,
max_distance=args.max_distance,
min_features=args.min_features,
max_features=args.max_features,
merge=False,
output_format=args.output_format):
args.stdout.write(str(gff) + "\n")
elif args.method == "merge-features":
for gff in combineGFF(gffs,
min_distance=args.min_distance,
max_distance=args.max_distance,
min_features=args.min_features,
max_features=args.max_features,
merge=True,
output_format=args.output_format):
args.stdout.write(str(gff) + "\n")
elif args.method == "crop":
for gff in cropGFF(gffs, args.filename_crop_gff):
args.stdout.write(str(gff) + "\n")
elif args.method == "crop-unique":
for gff in cropGFFUnique(gffs):
args.stdout.write(str(gff) + "\n")
elif args.method == "filter-range":
contig, strand, interval = None, None, None
try:
contig, strand, start, sep, end = re.match(
"(\S+):(\S+):(\d+)(\.\.|-)(\d+)", args.filter_range).groups()
except AttributeError:
pass
if not contig:
try:
contig, start, sep, end = re.match("(\S+):(\d+)(\.\.|-)(\d+)",
args.filter_range).groups()
strand = None
except AttributeError:
pass
if not contig:
try:
start, end = re.match("(\d+)(\.\.|\,|\-)(\d+)",
args.filter_range).groups()
except AttributeError:
raise "can not parse range %s" % args.filter_range
contig = None
strand = None
if start:
interval = (int(start), int(end))
else:
interval = None
E.debug("filter: contig=%s, strand=%s, interval=%s" %
(str(contig), str(strand), str(interval)))
for gff in GTF.iterator_filtered(gffs,
contig=contig,
strand=strand,
interval=interval):
args.stdout.write(str(gff) + "\n")
elif args.method == "sanitize":
def assemblyReport(id):
if id in assembly_dict.keys():
id = assembly_dict[id]
# if not in dict, the contig name is forced
# into the desired convention, this is helpful user
# modified gff files that contain additional contigs
elif args.sanitize_method == "ucsc":
if not id.startswith("contig") and not id.startswith("chr"):
id = "chr%s" % id
elif args.sanitize_method == "ensembl":
if id.startswith("contig"):
return id[len("contig"):]
elif id.startswith("chr"):
return id[len("chr"):]
return id
if args.sanitize_method == "genome":
if genome_fasta is None:
raise ValueError("please specify --genome-file= when using "
"--sanitize-method=genome")
f = genome_fasta.getToken
else:
if args.assembly_report is None:
raise ValueError(
"please specify --assembly-report= when using "
"--sanitize-method=ucsc or ensembl")
f = assemblyReport
skipped_contigs = collections.defaultdict(int)
outofrange_contigs = collections.defaultdict(int)
filtered_contigs = collections.defaultdict(int)
for gff in gffs:
try:
gff.contig = f(gff.contig)
except KeyError:
if args.skip_missing:
skipped_contigs[gff.contig] += 1
continue
else:
raise
if genome_fasta:
lcontig = genome_fasta.getLength(gff.contig)
if lcontig < gff.end:
outofrange_contigs[gff.contig] += 1
continue
if args.contig_pattern:
to_remove = [
re.compile(x) for x in args.contig_pattern.split(",")
]
if any([x.search(gff.contig) for x in to_remove]):
filtered_contigs[gff.contig] += 1
continue
args.stdout.write(str(gff) + "\n")
if skipped_contigs:
E.info("skipped %i entries on %i contigs: %s" %
(sum(skipped_contigs.values()),
len(list(skipped_contigs.keys())), str(skipped_contigs)))
if outofrange_contigs:
E.warn(
"skipped %i entries on %i contigs because they are out of range: %s"
% (sum(outofrange_contigs.values()),
len(list(
outofrange_contigs.keys())), str(outofrange_contigs)))
if filtered_contigs:
E.info("filtered out %i entries on %i contigs: %s" %
(sum(filtered_contigs.values()),
len(list(filtered_contigs.keys())), str(filtered_contigs)))
elif args.method == "rename-chr":
if not chr_map:
raise ValueError("please supply mapping file")
for gff in renameChromosomes(gffs, chr_map):
args.stdout.write(str(gff) + "\n")
else:
for gff in gffs:
if args.method == "forward_coordinates":
gff.invert(contigs[gff.contig])
if args.method == "forward_strand":
gff.invert(contigs[gff.contig])
gff.strand = "+"
if agp:
# note: this works only with forward coordinates
gff.contig, gff.start, gff.end = agp.mapLocation(
gff.contig, gff.start, gff.end)
args.stdout.write(str(gff) + "\n")
E.stop()
def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if not argv:
argv = sys.argv
# setup command line parser
parser = E.OptionParser(version="%prog version: $Id$",
usage=globals()["__doc__"])
parser.add_option("-m",
"--method",
dest="methods",
type="choice",
action="append",
choices=(
"geneprofile",
"tssprofile",
"utrprofile",
"intervalprofile",
"midpointprofile",
"geneprofilewithintrons",
"geneprofileabsolutedistancefromthreeprimeend",
"separateexonprofile",
"separateexonprofilewithintrons",
),
help='counters to use. Counters describe the '
'meta-gene structure to use. '
'Note using geneprofilewithintrons, or '
'geneprofileabsolutedistancefromthreeprimeend will '
'automatically turn on the --use-base-accuracy option'
'[%default].')
parser.add_option("-b",
"--bam-file",
"--bedfile",
"--bigwigfile",
dest="infiles",
metavar="BAM",
type="string",
action="append",
help="BAM/bed/bigwig files to use. Do not mix "
"different types [%default]")
parser.add_option("-c",
"--control-bam-file",
dest="controlfiles",
metavar="BAM",
type="string",
action="append",
help="control/input to use. Should be of the same "
"type as the bam/bed/bigwig file"
" [%default]")
parser.add_option("-g",
"--gtf-file",
dest="gtffile",
type="string",
metavar="GTF",
help="GTF file to use. "
"[%default]")
parser.add_option("--normalize-transcript",
dest="transcript_normalization",
type="choice",
choices=("none", "max", "sum", "total-max", "total-sum"),
help="normalization to apply on each transcript "
"profile before adding to meta-gene profile. "
"[%default]")
parser.add_option("--normalize-profile",
dest="profile_normalizations",
type="choice",
action="append",
choices=("all", "none", "area", "counts", "background"),
help="normalization to apply on meta-gene "
"profile normalization. "
"[%default]")
parser.add_option(
"-r",
"--reporter",
dest="reporter",
type="choice",
choices=("gene", "transcript"),
help="report results for genes or transcripts."
" When 'genes` is chosen, exons across all transcripts for"
" a gene are merged. When 'transcript' is chosen, counts are"
" computed for each transcript separately with each transcript"
" contributing equally to the meta-gene profile."
" [%default]")
parser.add_option("-i",
"--shift-size",
dest="shifts",
type="int",
action="append",
help="shift reads in :term:`bam` formatted file "
"before computing densities (ChIP-Seq). "
"[%default]")
parser.add_option("-a",
"--merge-pairs",
dest="merge_pairs",
action="store_true",
help="merge pairs in :term:`bam` formatted "
"file before computing "
"densities (ChIP-Seq). "
"[%default]")
parser.add_option("-u",
"--use-base-accuracy",
dest="base_accuracy",
action="store_true",
help="compute densities with base accuracy. The default "
"is to only use the start and end of the aligned region "
"(RNA-Seq) "
"[%default]")
parser.add_option("-e",
"--extend",
dest="extends",
type="int",
action="append",
help="extend reads in :term:`bam` formatted file "
"(ChIP-Seq). "
"[%default]")
parser.add_option("--resolution-upstream",
dest="resolution_upstream",
type="int",
help="resolution of upstream region in bp "
"[%default]")
parser.add_option("--resolution-downstream",
dest="resolution_downstream",
type="int",
help="resolution of downstream region in bp "
"[%default]")
parser.add_option("--resolution-upstream-utr",
dest="resolution_upstream_utr",
type="int",
help="resolution of upstream UTR region in bp "
"[%default]")
parser.add_option("--resolution-downstream-utr",
dest="resolution_downstream_utr",
type="int",
help="resolution of downstream UTR region in bp "
"[%default]")
parser.add_option("--resolution-cds",
dest="resolution_cds",
type="int",
help="resolution of cds region in bp "
"[%default]")
parser.add_option("--resolution-first-exon",
dest="resolution_first",
type="int",
help="resolution of first exon in gene, in bp"
"[%default]")
parser.add_option("--resolution-last-exon",
dest="resolution_last",
type="int",
help="resolution of last exon in gene, in bp"
"[%default]")
parser.add_option("--resolution-introns",
dest="resolution_introns",
type="int",
help="resolution of introns region in bp "
"[%default]")
parser.add_option("--resolution-exons-absolute-distance-topolya",
dest="resolution_exons_absolute_distance_topolya",
type="int",
help="resolution of exons absolute distance "
"topolya in bp "
"[%default]")
parser.add_option("--resolution-introns-absolute-distance-topolya",
dest="resolution_introns_absolute_distance_topolya",
type="int",
help="resolution of introns absolute distance "
"topolya in bp "
"[%default]")
parser.add_option("--extension-exons-absolute-distance-topolya",
dest="extension_exons_absolute_distance_topolya",
type="int",
help="extension for exons from the absolute "
"distance from the topolya in bp "
"[%default]")
parser.add_option(
"--extension-introns-absolute-distance-topolya",
dest="extension_introns_absolute_distance_topolya",
type="int",
help="extension for introns from the absolute distance from "
"the topolya in bp [%default]")
parser.add_option("--extension-upstream",
dest="extension_upstream",
type="int",
help="extension upstream from the first exon in bp"
"[%default]")
parser.add_option("--extension-downstream",
dest="extension_downstream",
type="int",
help="extension downstream from the last exon in bp"
"[%default]")
parser.add_option("--extension-inward",
dest="extension_inward",
type="int",
help="extension inward from a TSS start site in bp"
"[%default]")
parser.add_option("--extension-outward",
dest="extension_outward",
type="int",
help="extension outward from a TSS start site in bp"
"[%default]")
parser.add_option("--scale-flank-length",
dest="scale_flanks",
type="int",
help="scale flanks to (integer multiples of) gene length"
"[%default]")
parser.add_option(
"--control-factor",
dest="control_factor",
type="float",
help="factor for normalizing control and foreground data. "
"Computed from data if not set. "
"[%default]")
parser.add_option("--output-all-profiles",
dest="output_all_profiles",
action="store_true",
help="keep individual profiles for each "
"transcript and output. "
"[%default]")
parser.add_option("--counts-tsv-file",
dest="input_filename_counts",
type="string",
help="filename with count data for each transcript. "
"Use this instead "
"of recomputing the profile. Useful for plotting the "
"meta-gene profile "
"from previously computed counts "
"[%default]")
parser.add_option(
"--background-region-bins",
dest="background_region_bins",
type="int",
help="number of bins on either end of the profile "
"to be considered for background meta-gene normalization "
"[%default]")
parser.add_option(
"--output-res",
dest="resolution_images",
type="int",
help="the output dpi for the figure plot - will default to "
"[%default]")
parser.add_option(
"--image-format",
dest="image_format",
type="string",
help="The output format for the figure plot - defaults to "
"[%default]")
parser.set_defaults(
remove_rna=False,
ignore_pairs=False,
force_output=False,
bin_size=10,
extends=[],
shifts=[],
sort=[],
reporter="transcript",
resolution_cds=1000,
resolution_introns=1000,
# 3kb is a good balance of seeing long enough 3 prime bias and not omit
# too many genes. Tim 31th Aug 2013
resolution_exons_absolute_distance_topolya=3000,
# introns is only for assess the noise level, thus do ont need a long
# region, a long region has the side effect of omit more genes. Tim
# 31th Aug 2013
resolution_introns_absolute_distance_topolya=500,
# extension can simply just be the same as resolution
extension_exons_absolute_distance_topolya=3000,
extension_introns_absolute_distance_topolya=500,
resolution_upstream_utr=1000,
resolution_downstream_utr=1000,
resolution_upstream=1000,
resolution_downstream=1000,
resolution_first=1000,
resolution_last=1000,
# mean length of transcripts: about 2.5 kb
extension_upstream=2500,
extension_downstream=2500,
extension_inward=3000,
extension_outward=3000,
plot=True,
methods=[],
infiles=[],
controlfiles=[],
gtffile=None,
profile_normalizations=[],
transcript_normalization=None,
scale_flanks=0,
merge_pairs=False,
min_insert_size=0,
max_insert_size=1000,
base_accuracy=False,
matrix_format="single",
control_factor=None,
output_all_profiles=False,
background_region_bins=10,
input_filename_counts=None,
resolution_images=None,
image_format="png",
)
# add common options (-h/--help, ...) and parse command line
(options, args) = E.start(parser, argv=argv, add_output_options=True)
# Keep for backwards compatability
if len(args) == 2:
infile, gtf = args
options.infiles.append(infile)
options.gtffile = gtf
if not options.gtffile:
raise ValueError("no GTF file specified")
if options.gtffile == "-":
options.gtffile = options.stdin
else:
options.gtffile = iotools.open_file(options.gtffile)
if len(options.infiles) == 0:
raise ValueError("no bam/wig/bed files specified")
for methodsRequiresBaseAccuracy in [
"geneprofilewithintrons",
"geneprofileabsolutedistancefromthreeprimeend",
]:
# If you implemented any methods that you do not want the
# spliced out introns or exons appear to be covered by
# non-existent reads, it is better you let those methods imply
# --base-accurarcy by add them here.
if methodsRequiresBaseAccuracy in options.methods:
options.base_accuracy = True
if options.reporter == "gene":
gtf_iterator = GTF.flat_gene_iterator(GTF.iterator(options.gtffile))
elif options.reporter == "transcript":
gtf_iterator = GTF.transcript_iterator(GTF.iterator(options.gtffile))
# Select rangecounter based on file type
if len(options.infiles) > 0:
if options.infiles[0].endswith(".bam"):
bamfiles = [pysam.AlignmentFile(x, "rb") for x in options.infiles]
if options.controlfiles:
controlfiles = [
pysam.AlignmentFile(x, "rb") for x in options.controlfiles
]
else:
controlfiles = None
format = "bam"
if options.merge_pairs:
range_counter = geneprofile.RangeCounterBAM(
bamfiles,
shifts=options.shifts,
extends=options.extends,
merge_pairs=options.merge_pairs,
min_insert_size=options.min_insert_size,
max_insert_size=options.max_insert_size,
controfiles=controlfiles,
control_factor=options.control_factor)
elif options.shifts or options.extends:
range_counter = geneprofile.RangeCounterBAM(
bamfiles,
shifts=options.shifts,
extends=options.extends,
controlfiles=controlfiles,
control_factor=options.control_factor)
elif options.base_accuracy:
range_counter = geneprofile.RangeCounterBAMBaseAccuracy(
bamfiles,
controlfiles=controlfiles,
control_factor=options.control_factor)
else:
range_counter = geneprofile.RangeCounterBAM(
bamfiles,
controlfiles=controlfiles,
control_factor=options.control_factor)
elif options.infiles[0].endswith(".bed.gz"):
bedfiles = [pysam.Tabixfile(x) for x in options.infiles]
if options.controlfiles:
controlfiles = [
pysam.Tabixfile(x) for x in options.controlfiles
]
else:
controlfiles = None
range_counter = geneprofile.RangeCounterBed(
bedfiles,
controlfiles=controlfiles,
control_factor=options.control_factor)
elif options.infiles[0].endswith(".bw"):
wigfiles = [BigWigFile(file=open(x)) for x in options.infiles]
range_counter = geneprofile.RangeCounterBigWig(wigfiles)
else:
raise NotImplementedError("can't determine file type for %s" %
str(options.infiles))
counters = []
for method in options.methods:
if method == "utrprofile":
counters.append(
geneprofile.UTRCounter(
range_counter,
options.resolution_upstream,
options.resolution_upstream_utr,
options.resolution_cds,
options.resolution_downstream_utr,
options.resolution_downstream,
options.extension_upstream,
options.extension_downstream,
))
elif method == "geneprofile":
counters.append(
geneprofile.GeneCounter(
range_counter, options.resolution_upstream,
options.resolution_cds, options.resolution_downstream,
options.extension_upstream, options.extension_downstream,
options.scale_flanks))
elif method == "geneprofilewithintrons":
counters.append(
geneprofile.GeneCounterWithIntrons(
range_counter, options.resolution_upstream,
options.resolution_cds, options.resolution_introns,
options.resolution_downstream, options.extension_upstream,
options.extension_downstream, options.scale_flanks))
elif method == "geneprofileabsolutedistancefromthreeprimeend":
# options.extension_exons_absolute_distance_tostartsite,
# options.extension_introns_absolute_distance_tostartsite,
# Tim 31th Aug 2013: a possible feature for future, if five prime
# bias is of your interest.
# (you need to create another class). It is not very difficult to
# derive from this class, but is not implemented yet
# This future feature is slightly different the TSS profile
# already implemented, because in this future feature introns are
# skipped,
counters.append(
geneprofile.GeneCounterAbsoluteDistanceFromThreePrimeEnd(
range_counter, options.resolution_upstream,
options.resolution_downstream,
options.resolution_exons_absolute_distance_topolya,
options.resolution_introns_absolute_distance_topolya,
options.extension_upstream, options.extension_downstream,
options.extension_exons_absolute_distance_topolya,
options.extension_introns_absolute_distance_topolya,
options.scale_flanks))
elif method == "tssprofile":
counters.append(
geneprofile.TSSCounter(range_counter,
options.extension_outward,
options.extension_inward))
elif method == "intervalprofile":
counters.append(
geneprofile.RegionCounter(range_counter,
options.resolution_upstream,
options.resolution_cds,
options.resolution_downstream,
options.extension_upstream,
options.extension_downstream))
elif method == "midpointprofile":
counters.append(
geneprofile.MidpointCounter(range_counter,
options.resolution_upstream,
options.resolution_downstream,
options.extension_upstream,
options.extension_downstream))
# add new method to split 1st and last exons out
# requires a representative transcript for reach gene
# gtf should be sorted gene-position
elif method == "separateexonprofile":
counters.append(
geneprofile.SeparateExonCounter(
range_counter, options.resolution_upstream,
options.resolution_first, options.resolution_last,
options.resolution_cds, options.resolution_downstream,
options.extension_upstream, options.extension_downstream))
elif method == "separateexonprofilewithintrons":
counters.append(
geneprofile.SeparateExonWithIntronCounter(
range_counter, options.resolution_upstream,
options.resolution_first, options.resolution_last,
options.resolution_cds, options.resolution_introns,
options.resolution_downstream, options.extension_upstream,
options.extension_downstream))
# set normalization
for c in counters:
c.setNormalization(options.transcript_normalization)
if options.output_all_profiles:
c.setOutputProfiles(
iotools.open_file(
E.get_output_file(c.name) + ".profiles.tsv.gz", "w"))
if options.input_filename_counts:
# read counts from file
E.info("reading counts from %s" % options.input_filename_counts)
all_counts = pandas.read_csv(iotools.open_file(
options.input_filename_counts),
sep='\t',
header=0,
index_col=0)
if len(counters) != 1:
raise NotImplementedError(
'counting from matrix only implemented for 1 counter.')
# build counter based on reference counter
counter = geneprofile.UnsegmentedCounter(counters[0])
counters = [counter]
geneprofile.countFromCounts(counters, all_counts)
else:
E.info("starting counting with %i counters" % len(counters))
feature_names = geneprofile.countFromGTF(counters, gtf_iterator)
# output matrices
if not options.profile_normalizations:
options.profile_normalizations.append("none")
elif "all" in options.profile_normalizations:
options.profile_normalizations = [
"none", "area", "counts", "background"
]
for method, counter in zip(options.methods, counters):
profiles = []
for norm in options.profile_normalizations:
# build matrix, apply normalization
profile = counter.getProfile(
normalize=norm,
background_region_bins=options.background_region_bins)
profiles.append(profile)
for x in range(1, len(profiles)):
assert profiles[0].shape == profiles[x].shape
# build a single matrix of all profiles for output
matrix = numpy.concatenate(profiles)
matrix.shape = len(profiles), len(profiles[0])
matrix = matrix.transpose()
with iotools.open_file(
E.get_output_file(counter.name) + ".matrix.tsv.gz",
"w") as outfile:
outfile.write("bin\tregion\tregion_bin\t%s\n" %
"\t".join(options.profile_normalizations))
fields = []
bins = []
for field, nbins in zip(counter.fields, counter.nbins):
fields.extend([field] * nbins)
bins.extend(list(range(nbins)))
for row, cols in enumerate(zip(fields, bins, matrix)):
outfile.write("%i\t%s\t" %
(row, "\t".join([str(x) for x in cols[:-1]])))
outfile.write("%s\n" % ("\t".join([str(x) for x in cols[-1]])))
with iotools.open_file(
E.get_output_file(counter.name) + ".lengths.tsv.gz",
"w") as outfile:
counter.writeLengthStats(outfile)
if options.output_all_profiles:
counter.closeOutputProfiles()
if options.plot:
import matplotlib
# avoid Tk or any X
matplotlib.use("Agg")
import matplotlib.pyplot as plt
for method, counter in zip(options.methods, counters):
if method in ("geneprofile", "geneprofilewithintrons",
"geneprofileabsolutedistancefromthreeprimeend",
"utrprofile", "intervalprofile",
"separateexonprofile",
"separateexonprofilewithintrons"):
plt.figure()
plt.subplots_adjust(wspace=0.05)
max_scale = max([max(x) for x in counter.aggregate_counts])
for x, counts in enumerate(counter.aggregate_counts):
plt.subplot(6, 1, x + 1)
plt.plot(list(range(len(counts))), counts)
plt.title(counter.fields[x])
plt.ylim(0, max_scale)
figname = counter.name + ".full"
fn = E.get_output_file(figname) + "." + options.image_format
plt.savefig(os.path.expanduser(fn),
format=options.image_format,
dpi=options.resolution_images)
plt.figure()
points = []
cuts = []
for x, counts in enumerate(counter.aggregate_counts):
points.extend(counts)
cuts.append(len(counts))
plt.plot(list(range(len(points))), points)
xx, xxx = 0, []
for x in cuts:
xxx.append(xx + x // 2)
xx += x
plt.axvline(xx, color="r", ls="--")
plt.xticks(xxx, counter.fields)
figname = counter.name + ".detail"
fn = E.get_output_file(figname) + "." + options.image_format
plt.savefig(os.path.expanduser(fn),
format=options.image_format,
dpi=options.resolution_images)
elif method == "tssprofile":
plt.figure()
plt.subplot(1, 3, 1)
plt.plot(
list(
range(-options.extension_outward,
options.extension_inward)),
counter.aggregate_counts[0])
plt.title(counter.fields[0])
plt.subplot(1, 3, 2)
plt.plot(
list(
range(-options.extension_inward,
options.extension_outward)),
counter.aggregate_counts[1])
plt.title(counter.fields[1])
plt.subplot(1, 3, 3)
plt.title("combined")
plt.plot(
list(
range(-options.extension_outward,
options.extension_inward)),
counter.aggregate_counts[0])
plt.plot(
list(
range(-options.extension_inward,
options.extension_outward)),
counter.aggregate_counts[1])
plt.legend(counter.fields[:2])
fn = E.get_output_file(
counter.name) + "." + options.image_format
plt.savefig(os.path.expanduser(fn),
format=options.image_format,
dpi=options.resolution_images)
elif method == "midpointprofile":
plt.figure()
plt.plot(numpy.arange(-options.resolution_upstream, 0),
counter.aggregate_counts[0])
plt.plot(numpy.arange(0, options.resolution_downstream),
counter.aggregate_counts[1])
fn = E.get_output_file(
counter.name) + "." + options.image_format
plt.savefig(os.path.expanduser(fn),
format=options.image_format,
dpi=options.resolution_images)
# write footer and output benchmark information.
E.stop()
