def classifyLncRNA(infiles, outfile):
'''
Classify lncRNA realtive to protein coding loci
Classify lincRNA in terms of their relationship to
protein coding genes - creates indices for intervals on the
fly - mayb should be creating additional annotations:
antisense
transcript overlapping protein coding exons on opposite strand
antisense_upstream
transcript < 2kb from tss on opposite strand
antisense_downstream
transcript < 2kb from gene end on opposite strand
sense_upstream
transcript < 2kb from tss on same strand
sense_downstream
transcript < 2kb from gene end on same strand
intergenic
transcript >2kb from any protein coding gene
intronic
overlaps protein coding gene intron on same strand
antisense_intronic
overlaps protein coding intron on opposite strand
'''
PipelineLncRNA.classifyLncRNAGenes(infiles[0],
infiles[1],
outfile,
dist=PARAMS["lncrna_dist"])
def flagExonStatus(infile, outfile):
'''
Adds two attributes to the gtf entry:
exon_status_locus - specifies whether the gene model is multi- or
single-exon
exon_status - specifies whether the transcript is mult- or single exon
'''
PipelineLncRNA.flagExonStatus(infile, outfile)
def classifyFilteredLncRNA(infiles, outfile):
'''
classifies all lincRNA before cpc filtering to define any classes that
are represented in the coding set that are filtered
NOTE: This task is not included when running the full pipeline
'''
PipelineLncRNA.classifyLncRNAGenes(infiles[0],
infiles[1],
outfile,
dist=PARAMS["lncrna_dist"])
def splitLncRNAFasta(infile, outfiles):
out_dir = "./phyloCSF/lncrna_fasta"
name_dict = {}
for mapping in PARAMS["phyloCSF_map_species_names"].split(","):
pair = mapping.split(":")
key = ">" + pair[0]
value = ">" + pair[1]
name_dict[key] = value
E.info("Name mapping: %s" % name_dict)
PipelineLncRNA.splitAlignedFasta(infile, out_dir, name_dict)
def buildRefnoncodingGeneSet(infile, outfile):
'''
filter the refnoncoding geneset for things that are described in ensembl
as being:
Ambiguous_orf
Retained_intron
Sense_intronic
antisense
Sense_overlapping
Processed transcript
'''
PipelineLncRNA.buildRefnoncodingGeneSet(infile, outfile)
def buildFilteredLncRNAGeneSet(infiles, outfile):
'''
Creates a filtered lncRNA geneset. That contains previously identified
gene models supplied in contig file.
'''
assert PARAMS["filtering_remove_single_exon"] in [
"loci", "transcripts", None
]
PipelineLncRNA.buildFilteredLncRNAGeneSet(
infiles[0],
outfile,
infiles[1:len(infiles)],
filter_se=PARAMS["filtering_remove_single_exon"])
def buildLncRNAGeneSetStats(infile, outfile):
'''
counts:
no. of transcripts
no. genes
average number of exons per transcript
average number of exons per gene
no. multi-exon transcripts
no. single exon transcripts
no. multi-exon genes
no. single exon genes
in the coding and lncRNA genesets
'''
outf = open(outfile, "w")
outf.write("\t".join([
"no_transcripts", "no_genes", "no_exons_per_transcript",
"no_exons_per_gene", "no_single_exon_transcripts",
"no_multi_exon_transcripts", "no_single_exon_genes",
"no_multi_exon_genes"
]) + "\n")
# For pep8 purposes
x = list(
map(str, [
PipelineLncRNA.CounterTranscripts(infile).count(),
PipelineLncRNA.CounterGenes(infile).count(),
PipelineLncRNA.CounterExonsPerTranscript(infile).count(),
PipelineLncRNA.CounterExonsPerGene(infile).count(),
PipelineLncRNA.CounterSingleExonTranscripts(infile).count(),
PipelineLncRNA.CounterMultiExonTranscripts(infile).count(),
PipelineLncRNA.CounterSingleExonGenes(infile).count(),
PipelineLncRNA.CounterMultiExonGenes(infile).count()
]))
outf.write("\t".join(x))
def buildFinalLncRNAGeneSet(infile, outfile):
'''
the final lncRNA gene set consists of transcripts that pass
the initial filtering stage i.e. are;
multi-exonic/previously seen single exon transcripts
display low evidence for coding potential
'''
# filter based on coding potential and rename
PipelineLncRNA.buildFinalLncRNAGeneSet(infile,
"lncrna_filtered_cpc_result",
outfile, PARAMS["filtering_cpc"],
PARAMS["filtering_cpc_threshold"],
PARAMS["final_geneset_rename"])
def buildLncRNAGeneSet(infiles, outfile):
'''
build lncRNA gene set.
This is a set of transcripts in the abinitio set that
do not overlap at any protein coding or pseudogene transcripts
or additional biotypes from ensembl that are unwanted
(exons) in a reference gene set.
Transcripts need to have a length of at least 200 bp.
'''
PipelineLncRNA.buildLncRNAGeneSet(infiles[0], infiles[1], infiles[2],
infiles[3], infiles[4], outfile,
PARAMS["lncrna_min_length"])
def extractLncRNAFastaAlignments(infiles, outfile):
"""
Recieves a MAF file containing pairwise alignments and a gtf12 file
containing intervals. Outputs a single fasta file containing aligned
sequence for each interval.
"""
bed_file, maf_file = infiles
maf_tmp = P.getTempFilename("./phyloCSF")
to_cluster = False
statement = ("gunzip -c %(maf_file)s > %(maf_tmp)s")
P.run()
target_genome = PARAMS["genome"]
query_genome = PARAMS["phyloCSF_query_genome"]
genome_file = os.path.join(PARAMS["genomedir"], PARAMS["genome"])
gene_models = PipelineLncRNA.extractMAFGeneBlocks(bed_file,
maf_tmp,
genome_file,
outfile,
target_genome,
query_genome,
keep_gaps=False)
E.info("%i gene_models extracted" % gene_models)
os.unlink(maf_tmp)
def createMAFAlignment(infiles, outfile):
"""
Takes all .axt files in the input directory, filters them to remove
files based on supplied regular expressions, converts to a single maf file
using axtToMaf, filters maf alignments under a specified length.
"""
outfile = P.snip(outfile, ".gz")
axt_dir = PARAMS["phyloCSF_location_axt"]
to_ignore = re.compile(PARAMS["phyloCSF_ignore"])
axt_files = []
for axt_file in os.listdir(axt_dir):
if axt_file.endswith("net.axt.gz") and not to_ignore.search(axt_file):
axt_files.append(os.path.join(axt_dir, axt_file))
axt_files = (" ").join(sorted(axt_files))
E.info("axt files from which MAF alignment will be created: %s" %
axt_files)
target_genome = PARAMS["phyloCSF_target_genome"]
target_contigs = os.path.join(PARAMS["annotations_dir"],
PARAMS["annotations_interface_contigs"])
query_genome = PARAMS["phyloCSF_query_genome"]
query_contigs = os.path.join(PARAMS["phyloCSF_query_assembly"],
PARAMS["annotations_interface_contigs"])
tmpf1 = P.getTempFilename("./phyloCSF")
tmpf2 = P.getTempFilename("./phyloCSF")
to_cluster = False
# concatenate axt files, then remove headers
statement = ("zcat %(axt_files)s"
" > %(tmpf1)s;"
" axtToMaf "
" -tPrefix=%(target_genome)s."
" -qPrefix=%(query_genome)s."
" %(tmpf1)s"
" %(target_contigs)s"
" %(query_contigs)s"
" %(tmpf2)s")
P.run()
E.info("Temporary axt file created %s" % os.path.abspath(tmpf1))
E.info("Temporary maf file created %s" % os.path.abspath(tmpf2))
removed = P.snip(outfile, ".maf") + "_removed.maf"
to_cluster = False
filtered = PipelineLncRNA.filterMAF(tmpf2, outfile, removed,
PARAMS["phyloCSF_filter_alignments"])
E.info("%s blocks were ignored in MAF alignment"
" because length of target alignment was too short" % filtered[0])
E.info("%s blocks were output to filtered MAF alignment" % filtered[1])
os.unlink(tmpf1)
os.unlink(tmpf2)
to_cluster = False
statement = ("gzip %(outfile)s;" " gzip %(removed)s")
P.run()
def buildCodingGeneSet(infiles, outfile):
'''
takes the output from cuffcompare of a transcript
assembly and filters for annotated protein coding
genes.
NB "pruned" refers to nomenclature in the transcript
building pipeline - transcripts that appear in at least
two samples.
Because an abinitio assembly will often contain
fragments of known transcripts and describe them as
novel, the default behaviour is to produce a set that
is composed of 'complete' or 'contained' transcripts
i.e. nothing novel. This may underestimate the number
of transcripts that are actually expressed
'''
PipelineLncRNA.buildCodingGeneSet(infiles[0], infiles[1], outfile)
def buildRefcodingGeneSetStats(infile, outfile):
'''
counts:
no. of transcripts
no. genes
average number of exons per transcript
average number of exons per gene
no. multi-exon transcripts
no. single exon transcripts
no. multi-exon genes
no. single exon genes
in the coding and lncRNA genesets
'''
# calculate exon status for refcoding genes.
tmpf = P.getTempFilename(".") + ".gz"
PipelineLncRNA.flagExonStatus(infile, tmpf)
outf = iotools.openFile(outfile, "w")
outf.write("\t".join([
"no_transcripts", "no_genes", "no_exons_per_transcript",
"no_exons_per_gene", "no_single_exon_transcripts",
"no_multi_exon_transcripts", "no_single_exon_genes",
"no_multi_exon_genes"
]) + "\n")
outf.write("\t".join(
map(str, [
PipelineLncRNA.CounterTranscripts(tmpf).count(),
PipelineLncRNA.CounterGenes(tmpf).count(),
PipelineLncRNA.CounterExonsPerTranscript(tmpf).count(),
PipelineLncRNA.CounterExonsPerGene(tmpf).count(),
PipelineLncRNA.CounterSingleExonTranscripts(tmpf).count(),
PipelineLncRNA.CounterMultiExonTranscripts(tmpf).count(),
PipelineLncRNA.CounterSingleExonGenes(tmpf).count(),
PipelineLncRNA.CounterMultiExonGenes(tmpf).count()
])))
os.unlink(tmpf)
os.unlink(tmpf + ".log")
os.unlink(P.snip(tmpf, ".gz"))
def extractControllLncRNAFastaAlignments(infiles, outfile):
bed_file, maf_file = infiles
maf_tmp = P.getTempFilename("/ifs/scratch")
to_cluster = False
statement = ("gunzip -c %(maf_file)s > %(maf_tmp)s")
P.run()
target_genome = PARAMS["genome"]
query_genome = PARAMS["phyloCSF_query_genome"]
genome_file = os.path.join(PARAMS["genomedir"], PARAMS["genome"])
gene_models = PipelineLncRNA.extractMAFGeneBlocks(bed_file,
maf_tmp,
genome_file,
outfile,
target_genome,
query_genome,
keep_gaps=False)
E.info("%i gene_models extracted" % gene_models)
os.unlink(maf_tmp)
def convertControlGTFToBed12(infile, outfile):
"""
Convert either ensembl lincRNA, or control gtf to bed12 format
"""
PipelineLncRNA.gtfToBed12(infile, outfile, "transcript")
def buildRefcodingGeneSet(infiles, outfile):
'''
builds a refcoding geneset based on the genes that are present in
the abinitio assembly
'''
PipelineLncRNA.buildRefcodingGeneSet(infiles[1], infiles[0], outfile)
def loadLncRNAPhyloCSF(infile, outfile):
tmpf = P.getTempFilename("/ifs/scratch")
PipelineLncRNA.parsePhyloCSF(infile, tmpf)
P.load(tmpf, outfile, options="--add-index=gene_id")
def convertGTFToBed12(infile, outfile):
"""
Transform the lncrna_final.gtf.gz into lncrna_final.bed
"""
PipelineLncRNA.gtfToBed12(infile, outfile, "transcript")