def getLevelMapper(level, arguments):
if level in koSyns:
return lambda h: h
if arguments.heirarchyType == 'cazy':
return getCazyGroup
lookupLevel = level if level not in level3Syns else '3'
if arguments.heirarchyType == 'kegg':
# Ideally, we'd be able to parse the heirachy once, but the current
# KEGG code just retuns simple mappings
logging.info("Reading KEGG level %s assignments from %s" %
(level, arguments.heirarchyFile))
geneTranslation = kegg.parse_KEGG_file(arguments.heirarchyFile,
lookupLevel)
else:
# SEED or COG/KOG
if arguments.heirarchyType == 'seed':
logging.info("Reading SEED subsystem assignments from %s" %
(arguments.heirarchyFile))
seedTree = kegg.readSEEDTree(arguments.heirarchyFile)
elif arguments.heirarchyType == 'cog':
logging.info("Reading COG subsystem assignments from %s" %
(arguments.heirarchyFile))
seedTree = kegg.readCogTree(arguments.heirarchyFile)
geneTranslation = seedTree[lookupLevel]
return lambda gene: geneTranslation.get(gene, gene)
def getLevelMapper(level, arguments):
if level in koSyns:
return lambda h: h
if arguments.heirarchyType == 'cazy':
return getCazyGroup
lookupLevel = level if level not in level3Syns else '3'
if arguments.heirarchyType == 'kegg':
# Ideally, we'd be able to parse the heirachy once, but the current
# KEGG code just retuns simple mappings
logging.info(
"Reading KEGG level %s assignments from %s" %
(level, arguments.heirarchyFile))
geneTranslation = kegg.readKEGGFile(
arguments.heirarchyFile, lookupLevel)
else:
# SEED or COG/KOG
if arguments.heirarchyType == 'seed':
logging.info(
"Reading SEED subsystem assignments from %s" %
(arguments.heirarchyFile))
seedTree = kegg.readSEEDTree(arguments.heirarchyFile)
elif arguments.heirarchyType == 'cog':
logging.info(
"Reading COG subsystem assignments from %s" %
(arguments.heirarchyFile))
seedTree = kegg.readCogTree(arguments.heirarchyFile)
geneTranslation = seedTree[lookupLevel]
return lambda gene: geneTranslation.get(gene, gene)
def printCountTablesByLevel(fileCounts, totals, fileNames, options):
"""
Create a new file for each level with a tab separated table of counts
"""
cutoff = options.cutoff
if options.heirarchyType == 'seed':
logging.info(
"Reading SEED subsystem assignments from %s",
options.heirarchyFile)
seedTree = kegg.readSEEDTree(options.heirarchyFile)
elif options.heirarchyType == 'cog':
logging.info(
"Reading COG subsystem assignments from %s",
options.heirarchyFile)
seedTree = kegg.readCogTree(options.heirarchyFile)
# create an output table for each requested level
for level in options.levels:
logging.debug("Processing level %s", level)
translateToPaths = level not in koSyns
descString = None
if translateToPaths:
if options.heirarchyType == 'cazy':
geneTranslator = getCazyGroup
else:
lookupLevel = level if level not in level3Syns else '3'
if options.heirarchyType == 'kegg':
# Ideally, we'd be able to parse the heirachy once, but the
# current KEGG code just retuns simple mappings
logging.info(
"Reading KEGG level %s assignments from %s",
level, options.heirarchyFile)
geneTranslation = kegg.readKEGGFile(
options.heirarchyFile, lookupLevel)
else:
# SEED or COG/KOG
geneTranslation = seedTree[lookupLevel]
geneTranslator = dict_lookup_default_to_query(geneTranslation)
elif level is not None and options.heirarchyType == 'kegg':
# return descriptions if level explicitly set to ko (or syn.)
descString = "Description"
logging.info(
"Reading KO descriptions from %s",
options.heirarchyFile)
geneTranslation = kegg.readKEGGFile(
options.heirarchyFile, "DESCRIPTION")
geneTranslator = lambda_get_gene_and_translation(geneTranslation)
elif level is not None and options.heirarchyType == 'cog':
# return descriptions if level explicitly set to ko (or syn.)
descString = "Description\tCategories"
geneTranslator = lambda_get_seed_translations(seedTree)
else:
# just return gene if no level set or not KEGG/COG/KOG
geneTranslator = passThrough
# For each level, try to force all counts to be at that level
fileLevelTotals = {}
levelCounts = {}
levelPaths = {}
thresholds = {}
for (filename, counts) in fileCounts.items():
fileLevelTotals[filename] = 0
thresholds[filename] = totals[filename] * cutoff
fileLevelCounts = levelCounts.setdefault(filename, {})
fileTotal = 0
for gene in sorted(counts.keys(),
key=lambda s: "" if s is None else s):
# get the counts from this node
geneCount = counts[gene]
fileTotal += geneCount
# translate gene to pathway (or not depending on above code)
pathway = geneTranslator(gene)
# update counts
# Some KOs will map to multiple pathways,
# so... allow for multiple translated values
if not(
isinstance(
pathway,
list) or isinstance(
pathway,
tuple)):
pathway = [pathway, ]
for indPathway in pathway:
fileLevelCounts[indPathway] = fileLevelCounts.get(
indPathway, 0) + geneCount
levelPaths[indPathway] = True
logging.debug(
"File %s has %d hits (had %d)",
filename, fileTotal, totals[filename])
# logging.debug(repr(levelPaths))
# logging.debug(repr(levelCounts))
if logging.getLogger().level <= logging.DEBUG:
for (filename, counts) in levelCounts.items():
logging.debug("File %s has %d counts",
filename, sum(counts.values()))
# apply cutoff
for pathway in list(levelPaths.keys()):
# check to see if pathway is over cutoff in any file
over = False
for (filename, fileLevelCount) in levelCounts.items():
flPathCount = fileLevelCount.get(pathway, 0)
fileLevelTotals[filename] += flPathCount
if flPathCount > thresholds[filename]:
over = True
if not over:
# this pathway is not over the cutoff for any file
levelPaths.pop(pathway)
other = 'Other'
levelPaths[other] = True
for (filename, fileLevelCount) in levelCounts.items():
fileLevelCount[other] = fileLevelCount.get(
other, 0) + fileLevelCount.pop(pathway, 0)
if logging.getLogger().level <= logging.DEBUG:
for (filename, counts) in levelCounts.items():
logging.debug("File %s has %d counts",
filename, sum(counts.values()))
missed = False
for path in counts.keys():
if path not in levelPaths:
missed = True
logging.debug(
"Missing pathway %s has %d counts for %s",
path, counts[path], filename)
if not missed:
logging.debug(
"There are no missing pathways from %s",
filename)
logging.debug("Final file counts: %r", fileLevelTotals)
# output file
if options.output_file is None:
outs = sys.stdout
else:
if len(options.levels) > 1:
outfile = "%s.%s" % (options.output_file, level)
else:
outfile = options.output_file
outs = open(outfile, 'w')
# write to file(s?)
# header
if level in koSyns:
# Header for when level is the gene
if descString is not None:
outs.write("Gene\t%s\t%s\n" % (descString,
'\t'.join(fileNames)))
else:
outs.write("Gene\t%s\n" % ('\t'.join(fileNames)))
else:
# Header for when level is a pathway or group
outs.write("Pathway\t%s\n" % ('\t'.join(fileNames)))
for pathway in sorted(levelPaths.keys(),
key=lambda s: "" if s is None else s):
outs.write(str(pathway))
for filename in fileNames:
outs.write("\t")
outs.write(str(levelCounts[filename].get(pathway, 0)))
outs.write("\n")
# close out stream
if options.output_file is not None:
outs.close()
def printCountTablesByLevel(fileCounts, totals, fileNames, options):
"""
Create a new file for each level with a tab separated table of counts
"""
cutoff = options.cutoff
if options.heirarchyType == 'seed':
logging.info("Reading SEED subsystem assignments from %s",
options.heirarchyFile)
seedTree = kegg.readSEEDTree(options.heirarchyFile)
elif options.heirarchyType == 'cog':
logging.info("Reading COG subsystem assignments from %s",
options.heirarchyFile)
seedTree = kegg.readCogTree(options.heirarchyFile)
# create an output table for each requested level
for level in options.levels:
logging.debug("Processing level %s", level)
translateToPaths = level not in koSyns
descString = None
if translateToPaths:
if options.heirarchyType == 'cazy':
geneTranslator = getCazyGroup
else:
lookupLevel = level if level not in level3Syns else '3'
if options.heirarchyType == 'kegg':
# Ideally, we'd be able to parse the heirachy once, but the
# current KEGG code just retuns simple mappings
logging.info("Reading KEGG level %s assignments from %s",
level, options.heirarchyFile)
geneTranslation = kegg.readKEGGFile(
options.heirarchyFile, lookupLevel)
else:
# SEED or COG/KOG
geneTranslation = seedTree[lookupLevel]
geneTranslator = dict_lookup_default_to_query(geneTranslation)
elif level is not None and options.heirarchyType == 'kegg':
# return descriptions if level explicitly set to ko (or syn.)
descString = "Description"
logging.info("Reading KO descriptions from %s",
options.heirarchyFile)
geneTranslation = kegg.readKEGGFile(options.heirarchyFile,
"DESCRIPTION")
geneTranslator = lambda_get_gene_and_translation(geneTranslation)
elif level is not None and options.heirarchyType == 'cog':
# return descriptions if level explicitly set to ko (or syn.)
descString = "Description\tCategories"
geneTranslator = lambda_get_seed_translations(seedTree)
else:
# just return gene if no level set or not KEGG/COG/KOG
geneTranslator = passThrough
# For each level, try to force all counts to be at that level
fileLevelTotals = {}
levelCounts = {}
levelPaths = {}
thresholds = {}
for (filename, counts) in fileCounts.items():
fileLevelTotals[filename] = 0
thresholds[filename] = totals[filename] * cutoff
fileLevelCounts = levelCounts.setdefault(filename, {})
fileTotal = 0
for gene in sorted(counts.keys(),
key=lambda s: "" if s is None else s):
# get the counts from this node
geneCount = counts[gene]
fileTotal += geneCount
# translate gene to pathway (or not depending on above code)
pathway = geneTranslator(gene)
# update counts
# Some KOs will map to multiple pathways,
# so... allow for multiple translated values
if not (isinstance(pathway, list)
or isinstance(pathway, tuple)):
pathway = [
pathway,
]
for indPathway in pathway:
fileLevelCounts[indPathway] = fileLevelCounts.get(
indPathway, 0) + geneCount
levelPaths[indPathway] = True
logging.debug("File %s has %d hits (had %d)", filename, fileTotal,
totals[filename])
# logging.debug(repr(levelPaths))
# logging.debug(repr(levelCounts))
if logging.getLogger().level <= logging.DEBUG:
for (filename, counts) in levelCounts.items():
logging.debug("File %s has %d counts", filename,
sum(counts.values()))
# apply cutoff
for pathway in list(levelPaths.keys()):
# check to see if pathway is over cutoff in any file
over = False
for (filename, fileLevelCount) in levelCounts.items():
flPathCount = fileLevelCount.get(pathway, 0)
fileLevelTotals[filename] += flPathCount
if flPathCount > thresholds[filename]:
over = True
if not over:
# this pathway is not over the cutoff for any file
levelPaths.pop(pathway)
other = 'Other'
levelPaths[other] = True
for (filename, fileLevelCount) in levelCounts.items():
fileLevelCount[other] = fileLevelCount.get(
other, 0) + fileLevelCount.pop(pathway, 0)
if logging.getLogger().level <= logging.DEBUG:
for (filename, counts) in levelCounts.items():
logging.debug("File %s has %d counts", filename,
sum(counts.values()))
missed = False
for path in counts.keys():
if path not in levelPaths:
missed = True
logging.debug(
"Missing pathway %s has %d counts for %s", path,
counts[path], filename)
if not missed:
logging.debug("There are no missing pathways from %s",
filename)
logging.debug("Final file counts: %r", fileLevelTotals)
# output file
if options.output_file is None:
outs = sys.stdout
else:
if len(options.levels) > 1:
outfile = "%s.%s" % (options.output_file, level)
else:
outfile = options.output_file
outs = open(outfile, 'w')
# write to file(s?)
# header
if level in koSyns:
# Header for when level is the gene
if descString is not None:
outs.write("Gene\t%s\t%s\n" %
(descString, '\t'.join(fileNames)))
else:
outs.write("Gene\t%s\n" % ('\t'.join(fileNames)))
else:
# Header for when level is a pathway or group
outs.write("Pathway\t%s\n" % ('\t'.join(fileNames)))
for pathway in sorted(levelPaths.keys(),
key=lambda s: "" if s is None else s):
outs.write(str(pathway))
for filename in fileNames:
outs.write("\t")
outs.write(str(levelCounts[filename].get(pathway, 0)))
outs.write("\n")
# close out stream
if options.output_file is not None:
outs.close()