def dump_diamond_matches(fasta_file, seed_orthologs_file, args):
cpu = args.cpu
score_thr = args.seed_ortholog_score
evalue_thr = args.seed_ortholog_evalue
excluded_taxa = args.excluded_taxa if args.excluded_taxa else None
if args.translate:
tool = 'blastx'
else:
tool = 'blastp'
if not DIAMOND:
raise ValueError("diamond not found in path")
tempdir = mkdtemp(prefix='emappertmp_dmdn_', dir=args.temp_dir)
raw_output_file = pjoin(tempdir, uuid.uuid4().hex)
if excluded_taxa:
cmd = '%s %s -d %s -q %s --more-sensitive --threads %s -e %f -o %s --max-target-seqs 25' %\
(DIAMOND, tool, EGGNOG_DMND_DB, fasta_file, cpu, evalue_thr, raw_output_file)
else:
cmd = '%s %s -d %s -q %s --more-sensitive --threads %s -e %f -o %s --top 3' %\
(DIAMOND, tool, EGGNOG_DMND_DB, fasta_file, cpu, evalue_thr, raw_output_file)
print colorify(' '+cmd, 'yellow')
status = subprocess.call(cmd, shell=True,
stdout=subprocess.PIPE, stderr=subprocess.PIPE)
if status == 0:
OUT = open('%s' %seed_orthologs_file, 'w')
if not args.no_file_comments:
print >>OUT, get_call_info()
print >>OUT, '#', cmd
visited = set()
for line in open(raw_output_file):
if not line.strip() or line.startswith('#'):
continue
fields = map(str.strip, line.split('\t'))
query = fields[0]
hit = fields[1]
evalue = float(fields[10])
score = float(fields[11])
if query in visited:
continue
if evalue > evalue_thr or score < score_thr:
continue
if excluded_taxa and hit.startswith("%s." % excluded_taxa):
continue
visited.add(query)
print >>OUT, '\t'.join(map(str, [query, hit, evalue, score]))
OUT.close()
else:
print cmd
raise ValueError('Error running diamond')
shutil.rmtree(tempdir)
def refine_matches(fasta_file, refine_file, hits_file, args):
refine_header = map(
str.strip, '''#query_name, best_hit_eggNOG_ortholog,
best_hit_evalue, best_hit_score'''.split(','))
print colorify("Hit refinement starts now", 'green')
start_time = time.time()
og2level = dict([
tuple(map(str.strip, line.split('\t')))
for line in gopen(OGLEVELS_FILE)
])
OUT = open(refine_file, "w")
if not args.no_file_comments:
print >> OUT, get_call_info()
print >> OUT, '\t'.join(refine_header)
qn = 0
for qn, r in enumerate(
process_nog_hits_file(hits_file,
fasta_file,
og2level,
translate=args.translate,
cpu=args.cpu,
excluded_taxa=args.excluded_taxa,
base_tempdir=args.temp_dir)):
if qn and (qn % 25 == 0):
total_time = time.time() - start_time
print >>sys.stderr, qn + \
1, total_time, "%0.2f q/s (refinement)" % (
(float(qn + 1) / total_time))
sys.stderr.flush()
query_name = r[0]
best_hit_name = r[1]
if best_hit_name == '-' or best_hit_name == 'ERROR':
continue
best_hit_evalue = float(r[2])
best_hit_score = float(r[3])
print >> OUT, '\t'.join(
map(str,
(query_name, best_hit_name, best_hit_evalue, best_hit_score)))
#OUT.flush()
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >> OUT, '# %d queries scanned' % (qn + 1)
print >> OUT, '# Total time (seconds):', elapsed_time
print >> OUT, '# Rate:', "%0.2f q/s" % ((float(qn + 1) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn+1, elapsed_time, "%0.2f q/s" % ((float(qn+1) / elapsed_time))), 'lblue')
def dump_diamond_matches(fasta_file, seed_orthologs_file, args):
cpu = args.cpu
score_thr = args.seed_ortholog_score
evalue_thr = args.seed_ortholog_evalue
excluded_taxa = args.excluded_taxa if args.excluded_taxa else None
if not DIAMOND:
raise ValueError("diamond not found in path")
tempdir = mkdtemp(prefix='emappertmp_dmdn_', dir=TEMPDIR)
raw_output_file = pjoin(tempdir, uuid.uuid4().hex)
cmd = '%s blastp -d %s -q %s --more-sensitive --threads %s -e %f -o %s --top 3' %\
(DIAMOND, EGGNOG_DMND_DB, fasta_file, cpu, evalue_thr, raw_output_file)
print colorify(' '+cmd, 'yellow')
status = subprocess.call(cmd, shell=True,
stdout=subprocess.PIPE, stderr=subprocess.PIPE)
if status == 0:
OUT = open('%s' %seed_orthologs_file, 'w')
if not args.no_file_comments:
print >>OUT, get_call_info()
print >>OUT, '#', cmd
visited = set()
for line in open(raw_output_file):
if not line.strip() or line.startswith('#'):
continue
fields = map(str.strip, line.split('\t'))
query = fields[0]
hit = fields[1]
evalue = float(fields[10])
score = float(fields[11])
if query in visited:
continue
if evalue > evalue_thr or score < score_thr:
continue
if excluded_taxa or hit.startswith("%s." % excluded_taxa):
continue
visited.add(query)
print >>OUT, '\t'.join(map(str, [query, hit, evalue, score]))
OUT.close()
else:
print cmd
raise ValueError('Error running diamond')
shutil.rmtree(tempdir)
def annotate_hmm_matches(hits_file, hits_annot_file, args):
hits_annot_header = map(
str.strip, '''#query_name, hit, level, evalue,
sum_score, query_length, hmmfrom, hmmto, seqfrom, seqto, query_coverage,
members_in_og, og_description, og_COG_categories'''.
split(','))
annota.connect()
print colorify("Functional annotation of hits starts now", 'green')
start_time = time.time()
if pexists(hits_file):
OUT = open(hits_annot_file, "w")
if not args.no_file_comments:
print >> OUT, get_call_info()
print >> OUT, '\t'.join(hits_annot_header)
qn = 0
t1 = time.time()
for line in open(hits_file):
if not line.strip() or line.startswith('#'):
continue
qn += 1
if qn and (qn % 10000 == 0):
total_time = time.time() - start_time
print >>sys.stderr, qn, total_time, "%0.2f q/s (refinement)" %\
((float(qn) / total_time))
sys.stderr.flush()
(query, hit, evalue, sum_score, query_length, hmmfrom, hmmto,
seqfrom, seqto, q_coverage) = map(str.strip, line.split('\t'))
if hit not in ['ERROR', '-']:
hitname = cleanup_og_name(hit)
level, nm, desc, cats = annota.get_og_annotations(hitname)
print >> OUT, '\t'.join(
map(str, [
query, hitname, level, evalue, sum_score, query_length,
hmmfrom, hmmto, seqfrom, seqto, q_coverage, nm, desc,
cats
]))
else:
print >> OUT, '\t'.join([query] + [hit] *
(len(hits_annot_header) - 1))
elapsed_time = time.time() - t1
if not args.no_file_comments:
print >> OUT, '# %d queries scanned' % (qn)
print >> OUT, '# Total time (seconds):', elapsed_time
print >> OUT, '# Rate:', "%0.2f q/s" % ((float(qn) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn, elapsed_time, "%0.2f q/s" % ((float(qn) / elapsed_time))), 'lblue')
def annotate_hmm_matches(hits_file, hits_annot_file, args):
hits_annot_header = map(str.strip, '''#query_name, hit, level, evalue,
sum_score, query_length, hmmfrom, hmmto, seqfrom, seqto, query_coverage,
members_in_og, og_description, og_COG_categories'''.split(','))
annota.connect()
print colorify("Functional annotation of hits starts now", 'green')
start_time = time.time()
if pexists(hits_file):
OUT = open(hits_annot_file, "w")
if not args.no_file_comments:
print >>OUT, get_call_info()
print >>OUT, '\t'.join(hits_annot_header)
qn = 0
t1 = time.time()
for line in open(hits_file):
if not line.strip() or line.startswith('#'):
continue
if qn and (qn % 10000 == 0):
total_time = time.time() - start_time
print >>sys.stderr, qn+1, total_time, "%0.2f q/s (refinement)" %\
((float(qn + 1) / total_time))
sys.stderr.flush()
qn += 1
(query, hit, evalue, sum_score, query_length, hmmfrom, hmmto,
seqfrom, seqto, q_coverage) = map(str.strip, line.split('\t'))
if hit not in ['ERROR', '-']:
hitname = cleanup_og_name(hit)
level, nm, desc, cats = annota.get_og_annotations(hitname)
print >>OUT, '\t'.join(map( str, [query, hitname, level, evalue,
sum_score, query_length,
hmmfrom, hmmto, seqfrom,
seqto, q_coverage, nm, desc,
cats]))
else:
print >>OUT, '\t'.join(
[query] + [hit] * (len(hits_annot_header) - 1))
elapsed_time = time.time() - t1
if not args.no_file_comments:
print >>OUT, '# %d queries scanned' % (qn + 1)
print >>OUT, '# Total time (seconds):', elapsed_time
print >>OUT, '# Rate:', "%0.2f q/s" % ((float(qn + 1) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn+1, elapsed_time, "%0.2f q/s" % ((float(qn+1) / elapsed_time))), 'lblue')
def refine_matches(fasta_file, refine_file, hits_file, args):
refine_header = map(str.strip, '''#query_name, best_hit_eggNOG_ortholog,
best_hit_evalue, best_hit_score'''.split(','))
print colorify("Hit refinement starts now", 'green')
start_time = time.time()
og2level = dict([tuple(map(str.strip, line.split('\t')))
for line in gopen(OGLEVELS_FILE)])
OUT = open(refine_file, "w")
if not args.no_file_comments:
print >>OUT, get_call_info()
print >>OUT, '\t'.join(refine_header)
qn = 0
for qn, r in enumerate(process_nog_hits_file(hits_file, fasta_file, og2level,
translate=args.translate,
cpu=args.cpu,
excluded_taxa=args.excluded_taxa)):
if qn and (qn % 25 == 0):
total_time = time.time() - start_time
print >>sys.stderr, qn + \
1, total_time, "%0.2f q/s (refinement)" % (
(float(qn + 1) / total_time))
sys.stderr.flush()
query_name = r[0]
best_hit_name = r[1]
if best_hit_name == '-' or best_hit_name == 'ERROR':
continue
best_hit_evalue = float(r[2])
best_hit_score = float(r[3])
print >>OUT, '\t'.join(map(str, (query_name, best_hit_name,
best_hit_evalue, best_hit_score)))
#OUT.flush()
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >>OUT, '# %d queries scanned' % (qn + 1)
print >>OUT, '# Total time (seconds):', elapsed_time
print >>OUT, '# Rate:', "%0.2f q/s" % ((float(qn + 1) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn+1, elapsed_time, "%0.2f q/s" % ((float(qn+1) / elapsed_time))), 'lblue')
def get_seq_hmm_matches(hits_file):
annota.connect()
print colorify("Reading HMM matches", 'green')
seq2oginfo = {}
start_time = time.time()
hitnames = set()
if pexists(hits_file):
for line in open(hits_file):
if not line.strip() or line.startswith('#'):
continue
(query, hit, evalue, sum_score, query_length, hmmfrom, hmmto,
seqfrom, seqto, q_coverage) = map(str.strip, line.split('\t'))
if query not in seq2oginfo and hit not in ['ERROR', '-']:
hitname = cleanup_og_name(hit)
seq2oginfo[query] = [hitname, evalue, sum_score, query_length,
hmmfrom, hmmto, seqfrom, seqto,
q_coverage]
return seq2oginfo
def get_seq_hmm_matches(hits_file):
annota.connect()
print colorify("Reading HMM matches", 'green')
seq2oginfo = {}
start_time = time.time()
hitnames = set()
if pexists(hits_file):
for line in open(hits_file):
if not line.strip() or line.startswith('#'):
continue
(query, hit, evalue, sum_score, query_length, hmmfrom, hmmto,
seqfrom, seqto, q_coverage) = map(str.strip, line.split('\t'))
if query not in seq2oginfo and hit not in ['ERROR', '-']:
hitname = cleanup_og_name(hit)
seq2oginfo[query] = [hitname, evalue, sum_score, query_length,
hmmfrom, hmmto, seqfrom, seqto,
q_coverage]
return seq2oginfo
action="store_true",
dest='quiet',
help='quiet_mode')
parser.add_argument("--data_dir",
metavar='',
type=existing_dir,
help='Directory to use for DATA_PATH.')
args = parser.parse_args()
if args.data_dir:
set_data_path(args.data_dir)
if args.force or not pexists(pjoin(get_data_path(), 'og2level.tsv.gz')):
print colorify('Downloading "og2level.tsv.gz" at %s' % get_data_path(),
'green')
download_og2level()
if 'all' in args.dbs:
args.dbs = EGGNOG_DATABASES
if args.force or not pexists(pjoin(get_data_path(), 'eggnog.db')):
if args.allyes or ask("Download main annotation database?") == 'y':
print colorify(
'Downloading "eggnog.db" at %s...' % get_data_path(), 'green')
download_annotations()
else:
print 'Skipping'
else:
if not args.quiet:
def parse_args(parser):
args = parser.parse_args()
if args.version:
print get_version()
sys.exit(0)
if not args.no_annot and not pexists(EGGNOGDB_FILE):
print colorify(
'Annotation database data/eggnog.db not present. Use download_eggnog_database.py to fetch it',
'red')
raise emapperException()
if args.mode == 'diamond' and not pexists(EGGNOG_DMND_DB):
print colorify(
'DIAMOND database data/eggnog_proteins.dmnd not present. Use download_eggnog_database.py to fetch it',
'red')
raise emapperException()
if args.cpu == 0:
args.cpu = multiprocessing.cpu_count()
# No --servermode available for diamond
if args.mode == 'diamond' and args.servermode:
parser.error(
'--mode [diamond] and --servermode are mutually exclusive')
# Output file required unless running in servermode
if not args.servermode and not args.output:
parser.error('An output project name is required (-o)')
# Servermode implies using mem-based databases
if args.servermode:
args.usemem = True
# Direct annotation implies no searches
if args.annotate_hits_table:
args.no_search = True
args.no_annot = False
# Check inputs for running sequence searches
if not args.no_search and not args.servermode:
if not args.input:
parser.error('An input fasta file is required (-i)')
# HMM
if args.mode == 'hmmer':
if not args.db and not args.guessdb:
parser.error(
'HMMER mode requires specifying a target database (i.e. -d, --guessdb ))'
)
if args.db and args.guessdb:
parser.error('-d and --guessdb options are mutually exclusive')
if args.guessdb:
from ete3 import NCBITaxa
ncbi = NCBITaxa()
lineage = ncbi.get_lineage(args.guessdb)
for tid in reversed(lineage):
if tid in TAXID2LEVEL:
print tid, TAXID2LEVEL[tid]
args.db = TAXID2LEVEL[tid]
break
# DIAMOND
elif args.mode == 'diamond':
#if args.db or args.guessdb:
# parser.error('diamond mode does not require -d or --guessdb options')
pass
return args
help='assume "yes" to all questions')
parser.add_argument(
'-s',
action="store_true",
dest='simulate',
help='simulate and print commands. Nothing is downloaded')
##
args = parser.parse_args()
if args.dbname is None or args.dbname == "":
print(
colorify(
f'A prefix name for the DB to be created is required. Use the --dbname option.',
'red'))
sys.exit(1)
if (args.taxids is None or args.taxids == "") and (args.taxa is None
or args.taxa == ""):
print(
colorify(f'Either --taxids or --taxa parameter is required',
'red'))
sys.exit(1)
if (args.taxids is not None and
args.taxids != "") and (args.taxa is not None and args.taxa != ""):
print(colorify(f'Use either --taxids or --taxa, not both', 'red'))
sys.exit(1)
def parse_args(parser):
args = parser.parse_args()
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
if args.version:
version = ""
try:
version = get_full_version_info()
except Exception:
version = get_version()
print(version)
sys.exit(0)
args.call_info = get_call_info()
if args.list_taxa:
from eggnogmapper.vars import LEVEL_DEPTH, LEVEL_DICT, LEVEL_NAMES, LEVEL_PARENTS
print("tax_name\ttax_id\tdepth\tparents\tparents_names")
for tax_name, tax_id in LEVEL_DICT.items():
depth = LEVEL_DEPTH.get(tax_id, "-")
parents = LEVEL_PARENTS.get(tax_id, "-")
parents_names = [LEVEL_NAMES.get(x, "-") for x in parents]
print(f"{tax_name}\t{tax_id}\t{depth}\t{','.join(parents)}\t{','.join(parents_names)}")
sys.exit(0)
if args.cpu == 0:
args.cpu = multiprocessing.cpu_count()
# translate
if args.itype in [ITYPE_GENOME, ITYPE_META, ITYPE_PROTS] and args.translate == True:
parser.error('"--translate" only can be used with "--itype CDS"')
# Gene prediction
if args.training_genome is not None and args.training_file is None:
parser.error('"--training_genome requires --training_file"')
if args.training_genome is None and args.training_file is not None:
if not os.path.isfile(args.training_file):
parser.error('"--training_file must point to an existing file, if no --training_genome is provided."')
# Search modes
if args.mode == SEARCH_MODE_DIAMOND:
dmnd_db = args.dmnd_db if args.dmnd_db else get_eggnog_dmnd_db()
if not pexists(dmnd_db):
print(colorify('DIAMOND database %s not present. Use download_eggnog_database.py to fetch it' % dmnd_db, 'red'))
raise EmapperException()
if args.input is not None:
if args.annotate_hits_table is not None:
print(colorify(f"--annotate_hits_table will be ignored, due to -m {SEARCH_MODE_DIAMOND}", 'blue'))
args.annotate_hits_table = None
else:
# the default -m is diamond, but we will consider -m no_search as default when
# --annotate_hits_table has been provided and -i has not been provided
if args.annotate_hits_table is not None:
print(colorify(f"Assuming -m {SEARCH_MODE_NO_SEARCH}", 'blue'))
args.mode = SEARCH_MODE_NO_SEARCH
else:
parser.error('An input fasta file is required (-i)')
# Output file required
if not args.output:
parser.error('An output project name is required (-o)')
if args.resume == True:
print(colorify("Diamond jobs cannot be resumed. --resume will be ignored.", 'blue'))
args.resume = False
elif args.mode == SEARCH_MODE_MMSEQS2:
mmseqs_db = args.mmseqs_db if args.mmseqs_db else get_eggnog_mmseqs_db()
if not pexists(mmseqs_db):
print(colorify('MMseqs2 database %s not present. Use download_eggnog_database.py to fetch it' % mmseqs_db, 'red'))
raise EmapperException()
if not args.input:
parser.error('An input fasta file is required (-i)')
# Output file required
if not args.output:
parser.error('An output project name is required (-o)')
if args.resume == True:
print(colorify("MMseqs2 jobs cannot be resumed. --resume will be ignored.", 'blue'))
args.resume = False
if args.annotate_hits_table is not None:
print(colorify(f"--annotate_hits_table will be ignored, due to -m {SEARCH_MODE_MMSEQS2}", 'blue'))
args.annotate_hits_table = None
elif args.mode == SEARCH_MODE_HMMER:
# if args.usemem == True:
# total_workers = args.num_workers * args.num_servers
# if args.cpu < total_workers:
# parser.error(f"Less cpus ({args.cpu}) than total workers ({total_workers}) were specified.")
# if args.cpu % total_workers != 0:
# parser.error(f"Number of cpus ({args.cpu}) must be a multiple of total workers ({total_workers}).")
# args.cpus_per_worker = int(args.cpu / total_workers)
# sys.stderr.write(f"CPUs per worker: {args.cpus_per_worker}\n")
# else:
# args.cpus_per_worker = args.cpu
if not args.input:
parser.error('An input file is required (-i)')
# Output file required
if not args.output:
parser.error('An output project name is required (-o)')
# Hmmer database
# NOTE: hmmer database format, name and checking if exists is done within hmmer module
if not args.db:
parser.error('HMMER mode requires a target database (-d, --database).')
if args.itype == ITYPE_CDS:
args.translate = True
if (args.itype == ITYPE_GENOME or args.itype == ITYPE_META) and args.genepred == GENEPRED_MODE_SEARCH:
parser.error('HMMER mode is not compatible with "--genepred search" option.')
if args.annotate_hits_table is not None:
print(colorify(f"--annotate_hits_table will be ignored, due to -m {SEARCH_MODE_HMMER}", 'blue'))
args.annotate_hits_table = None
if args.clean_overlaps is not None:
if args.clean_overlaps == "none":
args.clean_overlaps = None
elif args.mode == SEARCH_MODE_CACHE:
if args.cache_file is None:
parser.error('A file with annotations and md5 of queries is required (-c FILE)')
if args.decorate_gff != DECORATE_GFF_NONE:
print(colorify("WARNING: no GFF will be created for cache-based annotations. It is not implemented yet, sorry.", 'red'))
if args.no_annot == True:
parser.error(f'Cache mode (-m {SEARCH_MODE_CACHE}) should be used to annotate.')
elif args.mode == SEARCH_MODE_NO_SEARCH:
if args.no_annot == False and not args.annotate_hits_table:
parser.error(f'No search mode (-m {SEARCH_MODE_NO_SEARCH}) requires a hits table to annotate (--annotate_hits_table FILE.seed_orthologs)')
if args.md5 == True and args.input is None:
parser.error(f'--md5 requires an input FASTA file (-i FASTA).')
# if args.no_annot == True and args.report_orthologs == False:
# parser.error(f'Nothing to do if running in no search mode (-m {SEARCH_MODE_NO_SEARCH}), with --no_annot and without --report_orthologs.')
else:
parser.error(f'unrecognized search mode (-m {args.mode})')
# Search thresholds
args.dmnd_evalue = args.mmseqs_evalue = args.hmm_evalue = args.evalue
args.dmnd_score = args.mmseqs_score = args_hmm_score = args.score
args.qcov = args.query_cover
# Annotation options
if args.no_annot == False or args.report_orthologs == True:
if not pexists(get_eggnogdb_file()):
print(colorify('Annotation database data/eggnog.db not present. Use download_eggnog_database.py to fetch it', 'red'))
raise EmapperException()
args.tax_scope_mode, args.tax_scope_id = __parse_tax_scope(args.tax_scope)
if args.target_taxa is not None:
args.target_taxa = args.target_taxa.split(",")
if args.excluded_taxa is not None:
args.excluded_taxa = args.excluded_taxa.split(",")
# Sets GO evidence bases
if args.go_evidence == 'experimental':
args.go_evidence = set(["EXP","IDA","IPI","IMP","IGI","IEP"])
args.go_excluded = set(["ND", "IEA"])
elif args.go_evidence == 'non-electronic':
args.go_evidence = None
args.go_excluded = set(["ND", "IEA"])
elif args.go_evidence == 'all':
args.go_evidence = None
args.go_excluded = None
else:
raise ValueError('Invalid --go_evidence value')
# PFAM annotation options
if args.pfam_transfer in [PFAM_TRANSFER_BEST_OG, PFAM_TRANSFER_NARROWEST_OG, PFAM_TRANSFER_SEED_ORTHOLOG]:
pass
else:
raise ValueError(f'Invalid --pfam_transfer option {args.pfam_transfer}')
if args.pfam_realign == PFAM_REALIGN_NONE:
pass
elif args.pfam_realign == PFAM_REALIGN_REALIGN or args.pfam_realign == PFAM_REALIGN_DENOVO:
if not args.input:
parser.error(f'An input fasta file is required (-i) for --pfam_realign {args.pfam_realign}')
else:
raise ValueError(f'Invalid --pfam_realign option {args.pfam_realign}')
total_workers = args.num_workers * args.num_servers
if args.cpu < total_workers:
parser.error(f"Less cpus ({args.cpu}) than total workers ({total_workers}) were specified.")
if args.cpu % total_workers != 0:
parser.error(f"Number of cpus ({args.cpu}) must be a multiple of total workers ({total_workers}).")
args.cpus_per_worker = int(args.cpu / total_workers)
return args
def setup_hmm_search(args):
host = 'localhost'
idmap = None
if args.usemem:
scantype = 'mem'
else:
scantype = 'disk'
connecting_to_server = False
# If searching against a predefined database name
if args.db in EGGNOG_DATABASES:
dbpath, port = get_db_info(args.db)
print dbpath
db_present = [
pexists(dbpath + "." + ext)
for ext in 'h3f h3i h3m h3p idmap'.split()
]
if False in db_present:
print db_present
print colorify(
'Database %s not present. Use download_eggnog_database.py to fetch it'
% (args.db), 'red')
raise ValueError('Database not found')
if not args.no_refine:
if not pexists(pjoin(get_data_path(), 'OG_fasta')):
print colorify(
'Database data/OG_fasta/ not present. Use download_eggnog_database.py to fetch it',
'red')
raise ValueError('Database not found')
if scantype == 'mem':
idmap_file = dbpath + '.idmap'
end_port = 53200
# If searching against a custom hmm database
elif os.path.isfile(args.db + '.h3f'):
dbpath = args.db
if scantype == 'mem':
idmap_file = args.db + ".idmap"
if not pexists(idmap_file):
if generate_idmap(args.db):
idmap_file = args.db + ".idmap"
print >> sys.stderr, "idmap succesfully created!"
else:
raise ValueError("idmap could not be created!")
port = 53000
end_port = 53200
else:
idmap_file = None
port = None
# If searching against a emapper hmm server
elif ":" in args.db:
dbname, host, port = map(str.strip, args.db.split(":"))
scantype = 'mem'
port = int(port)
if dbname in EGGNOG_DATABASES:
dbfile, port = get_db_info(dbname)
args.db = dbname
else:
dbfile = dbname
idmap_file = dbfile + '.idmap'
if not pexists(idmap_file):
raise ValueError("idmap file not found: %s" % idmap_file)
dbpath = host
if not server_functional(host, port, args.dbtype):
print colorify(
"eggnog-mapper server not found at %s:%s" % (host, port),
'red')
exit(1)
connecting_to_server = True
else:
raise ValueError('Invalid database name/server')
# If memory based searches requested, start server
if scantype == "mem" and not connecting_to_server:
master_db, worker_db = None, None
for try_port in range(port, end_port, 2):
print colorify(
"Loading server at localhost, port %s-%s" %
(try_port, try_port + 1), 'lblue')
dbpath, master_db, worker_db = load_server(dbpath, try_port,
try_port + 1, args.cpu)
port = try_port
ready = False
for _ in xrange(TIMEOUT_LOAD_SERVER):
print "Waiting for server to become ready...", host, try_port
time.sleep(1)
if not master_db.is_alive() or not worker_db.is_alive():
master_db.terminate()
master_db.join()
worker_db.terminate()
worker_db.join()
break
elif server_functional(host, port, args.dbtype):
ready = True
break
if ready:
dbpath = host
break
elif scantype == "mem":
print colorify("DB Server already running or not needed!", 'yellow')
dbpath = host
# Preload seqid map to translate hits from hmmpgmd
if scantype == "mem":
print colorify("Reading idmap %s" % idmap_file, color='lblue')
idmap = {}
for _lnum, _line in enumerate(open(idmap_file)):
if not _line.strip():
continue
try:
_seqid, _seqname = map(str, _line.strip().split(' '))
except ValueError:
if _lnum == 0:
# idmap generated by esl_reformat has info line at beginning
continue
else:
raise
_seqid = int(_seqid)
idmap[_seqid] = [_seqname]
print len(idmap), "names loaded"
# If server mode, just listen for connections and exit when interrupted
if args.servermode:
while True:
print colorify(
"Server ready listening at %s:%s and using %d CPU cores" %
(host, port, args.cpu), 'green')
print colorify(
"Use `emapper.py -d %s:%s:%s (...)` to search against this server"
% (args.db, host, port), 'lblue')
time.sleep(10)
raise emapperException()
else:
return host, port, dbpath, scantype, idmap
args = parser.parse_args()
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
data_path = get_data_path()
##
# Annotation DB
if args.force or not pexists(get_eggnogdb_file()):
if args.allyes or ask("Download main annotation database?") == 'y':
print(colorify(f'Downloading "eggnog.db" at {data_path}...', 'green'))
download_annotations(data_path)
else:
print('Skipping')
else:
if not args.quiet:
print(colorify('Skipping eggnog.db database (already present). Use -f to force download', 'lblue'))
##
# NCBI taxa
if args.force or not pexists(get_ncbitaxadb_file()):
if args.allyes or ask("Download taxa database?") == 'y':
print(colorify(f'Downloading "eggnog.taxa.db" at {data_path}...', 'green'))
download_taxa(data_path)
else:
dbpath, host, idmap_file = setup_custom_db(args.db,
scantype=SCANTYPE_MEM,
dbtype=args.dbtype)
host = 'localhost'
port = args.port
end_port = args.end_port
wport = args.wport
dbpath, host, port, servers = create_servers(args.dbtype, dbpath, host,
port, end_port,
args.num_servers,
args.num_workers,
args.cpus_per_worker)
print(colorify("All servers ready and listening", 'green'))
if args.output_servers_list is not None:
print(f"Creating servers list file: {args.output_servers_list}")
with open(args.output_servers_list, 'w') as outfn:
for server in servers:
print(f"{server[0]}:{server[1]}", file=outfn)
print(f"File {args.output_servers_list} created successfully.")
print(
colorify(
f"Use `emapper.py (-d db:host:port or --servers_list {args.output_servers_list}) to search against these servers",
'lblue'))
else:
print(
colorify(
"Use `emapper.py (-d db:host:port or --servers_list FILE) to search against these servers",
def run(cmd):
print colorify(cmd, 'cyan')
if not args.simulate:
os.system(cmd)
def annotate_hits_file(seed_orthologs_file, annot_file, hmm_hits_file, args):
annot_header = (
"#query_name",
"seed_eggNOG_ortholog",
"seed_ortholog_evalue",
"seed_ortholog_score",
"predicted_gene_name",
"GO_terms",
"KEGG_KOs",
"BiGG_reactions",
"Annotation_tax_scope",
"OGs",
"bestOG|evalue|score",
"COG cat",
"eggNOG annot",
)
start_time = time.time()
seq2bestOG = {}
if pexists(hmm_hits_file):
seq2bestOG = get_seq_hmm_matches(hmm_hits_file)
seq2annotOG = annota.get_ogs_annotations(
set([v[0] for v in seq2bestOG.itervalues()]))
print colorify("Functional annotation of refined hits starts now", 'green')
OUT = open(annot_file, "w")
if args.report_orthologs:
ORTHOLOGS = open(annot_file + ".orthologs", "w")
if not args.no_file_comments:
print >> OUT, '# emapper version:', get_version(
), 'emapper DB:', get_db_version()
print >> OUT, '# command: ./emapper.py ', ' '.join(sys.argv[1:])
print >> OUT, '# time: ' + time.ctime()
print >> OUT, '\t'.join(annot_header)
qn = 0
pool = multiprocessing.Pool(args.cpu)
for result in pool.imap(annotate_hit_line,
iter_hit_lines(seed_orthologs_file, args)):
qn += 1
if qn and (qn % 500 == 0):
total_time = time.time() - start_time
print >> sys.stderr, qn, total_time, "%0.2f q/s (refinement)" % (
(float(qn) / total_time))
sys.stderr.flush()
if result:
(query_name, best_hit_name, best_hit_evalue, best_hit_score,
best_name, gos, kegg, bigg, annot_level_max, match_nogs,
orthologs) = result
if query_name in seq2bestOG:
(hitname, evalue, score, qlength, hmmfrom, hmmto, seqfrom,
seqto, q_coverage) = seq2bestOG[query_name]
bestOG = '%s|%s|%s' % (hitname, evalue, score)
og_cat, og_desc = seq2annotOG.get(hitname, ['', ''])
else:
bestOG = 'NA|NA|NA'
og_cat, og_desc = annota.get_best_og_description(match_nogs)
if args.report_orthologs:
print >> ORTHOLOGS, '\t'.join(
map(str, (query_name, ','.join(orthologs))))
print >> OUT, '\t'.join(
map(str, (
query_name,
best_hit_name,
best_hit_evalue,
best_hit_score,
best_name,
','.join(sorted(gos)),
','.join(sorted(kegg)),
','.join(sorted(bigg)),
annot_level_max,
','.join(match_nogs),
bestOG,
og_cat.replace('\n', ''),
og_desc.replace('\n', ' '),
)))
OUT.flush()
pool.terminate()
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >> OUT, '# %d queries scanned' % (qn)
print >> OUT, '# Total time (seconds):', elapsed_time
print >> OUT, '# Rate:', "%0.2f q/s" % ((float(qn) / elapsed_time))
OUT.close()
if args.report_orthologs:
ORTHOLOGS.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn, elapsed_time, "%0.2f q/s" % ((float(qn) / elapsed_time))), 'lblue')
def setup_hmm_search(args):
host = 'localhost'
idmap = None
if args.usemem:
scantype = 'mem'
else:
scantype = 'disk'
connecting_to_server = False
# If searching against a predefined database name
if args.db in EGGNOG_DATABASES:
dbpath, port = get_db_info(args.db)
db_present = [pexists(dbpath + "." + ext)
for ext in 'h3f h3i h3m h3p idmap'.split()]
if False in db_present:
print db_present
print colorify('Database %s not present. Use download_eggnog_database.py to fetch it' % (args.db), 'red')
raise ValueError('Database not found')
if not args.no_refine:
if not pexists(pjoin(DATA_PATH, 'OG_fasta')):
print colorify('Database data/OG_fasta/ not present. Use download_eggnog_database.py to fetch it', 'red')
raise ValueError('Database not found')
if scantype == 'mem':
idmap_file = dbpath + '.idmap'
end_port = 53200
# If searching against a custom hmm database
elif os.path.isfile(args.db + '.h3f'):
dbpath = args.db
if scantype == 'mem':
idmap_file = args.db + ".idmap"
if not pexists(idmap_file):
if generate_idmap(args.db):
idmap_file = args.db + ".idmap"
print >>sys.stderr, "idmap succesfully created!"
else:
raise ValueError("idmap could not be created!")
port = 53000
end_port = 53200
else:
idmap_file = None
port = None
# If searching against a emapper hmm server
elif ":" in args.db:
dbname, host, port = map(str.strip, args.db.split(":"))
scantype = 'mem'
port = int(port)
if dbname in EGGNOG_DATABASES:
dbfile, port = get_db_info(dbname)
args.db = dbname
else:
dbfile = dbname
idmap_file = dbfile + '.idmap'
if not pexists(idmap_file):
raise ValueError("idmap file not found: %s" % idmap_file)
dbpath = host
if not server_functional(host, port, args.dbtype):
print colorify("eggnog-mapper server not found at %s:%s" % (host, port), 'red')
exit(1)
connecting_to_server = True
else:
raise ValueError('Invalid database name/server')
# If memory based searches requested, start server
if scantype == "mem" and not connecting_to_server:
master_db, worker_db = None, None
for try_port in range(port, end_port, 2):
print colorify("Loading server at localhost, port %s-%s" %
(try_port, try_port + 1), 'lblue')
dbpath, master_db, worker_db = load_server(
dbpath, try_port, try_port + 1, args.cpu)
port = try_port
ready = False
for _ in xrange(TIMEOUT_LOAD_SERVER):
print "Waiting for server to become ready...", host, try_port
time.sleep(1)
if not master_db.is_alive() or not worker_db.is_alive():
master_db.terminate()
master_db.join()
worker_db.terminate()
worker_db.join()
break
elif server_functional(host, port, args.dbtype):
ready = True
break
if ready:
dbpath = host
break
elif scantype == "mem":
print colorify("DB Server already running or not needed!", 'yellow')
dbpath = host
# Preload seqid map to translate hits from hmmpgmd
if scantype == "mem":
print colorify("Reading idmap %s" % idmap_file, color='lblue')
idmap = {}
for _lnum, _line in enumerate(open(idmap_file)):
if not _line.strip():
continue
try:
_seqid, _seqname = map(str, _line.strip().split(' '))
except ValueError:
if _lnum == 0:
# idmap generated by esl_reformat has info line at beginning
continue
else:
raise
_seqid = int(_seqid)
idmap[_seqid] = [_seqname]
print len(idmap), "names loaded"
# If server mode, just listen for connections and exit when interrupted
if args.servermode:
while True:
print colorify("Server ready listening at %s:%s and using %d CPU cores" % (host, port, args.cpu), 'green')
print colorify("Use `emapper.py -d %s:%s:%s (...)` to search against this server" % (args.db, host, port), 'lblue')
time.sleep(10)
raise emapperException()
else:
return host, port, dbpath, scantype, idmap
return __author__+" "+__license__+" : "+__description__
if __name__ == "__main__":
parser = create_arg_parser()
args = parse_args(parser)
_total_time = time.time()
try:
print('# ', get_version())
print('# hmm_worker.py ', ' '.join(sys.argv[1:]))
worker_db = None
print(colorify(f"Loading worker at localhost, port {args.port}, connecting to {args.host}", 'green'))
worker_db = load_worker(args.host, args.port, args.cpu)
ready = False
for _ in range(TIMEOUT_LOAD_SERVER):
print(f"Waiting for worker to become ready at localhost:{args.port} ...")
time.sleep(1)
if worker_db.is_alive():
break
else:
worker_db.terminate()
worker_db.join()
print(colorify("worker not alive"), 'red')
break
print(colorify("Worker of master %s ready listening at localhost:%s and using %d CPU cores" % (args.host, args.port, args.cpu), 'lblue'))
def dump_hmm_matches(fasta_file, hits_file, dbpath, port, scantype, idmap, args):
hits_header = ("#query_name", "hit", "evalue", "sum_score", "query_length",
"hmmfrom", "hmmto", "seqfrom", "seqto", "query_coverage")
# Cache previous results if resuming is enabled
VISITED = set()
if args.resume and pexists(hits_file):
print colorify("Resuming previous run. Reading computed output from %s" % hits_file, 'yellow')
VISITED = set([line.split('\t')[0].strip()
for line in open(hits_file) if not line.startswith('#')])
print len(VISITED), 'queries skipped'
OUT = open(hits_file, 'a')
else:
OUT = open(hits_file, 'w')
print colorify("Sequence mapping starts now!", 'green')
if not args.no_file_comments:
print >>OUT, get_call_info()
print >>OUT, '# ' + '\t'.join(hits_header)
total_time = 0
last_time = time.time()
start_time = time.time()
qn = 0
for qn, (name, elapsed, hits, querylen, seq) in enumerate(search.iter_hits(
fasta_file,
args.translate,
args.qtype,
args.dbtype,
scantype,
dbpath,
port,
evalue_thr=args.evalue,
score_thr=args.score,
qcov_thr=args.qcov,
fixed_Z=args.Z,
max_hits=args.maxhits,
skip=VISITED,
maxseqlen=args.maxseqlen,
cpus=args.cpu)):
if elapsed == -1:
# error occurred
print >>OUT, '\t'.join(
[name] + ['ERROR'] * (len(hits_header) - 1))
elif not hits:
print >>OUT, '\t'.join([name] + ['-'] * (len(hits_header) - 1))
else:
for hitindex, (hid, heval, hscore, hmmfrom, hmmto, sqfrom, sqto, domscore) in enumerate(hits):
hitname = hid
if idmap:
hitname = idmap[hid][0]
print >>OUT, '\t'.join(map(str, [name, hitname, heval, hscore,
int(querylen), int(hmmfrom),
int(hmmto), int(sqfrom),
int(sqto),
float(sqto - sqfrom) / querylen]))
OUT.flush()
# monitoring
total_time += time.time() - last_time
last_time = time.time()
if qn and (qn % 25 == 0):
print >>sys.stderr, qn + \
1, total_time, "%0.2f q/s" % ((float(qn + 1) / total_time))
sys.stderr.flush()
# Writes final stats
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >>OUT, '# %d queries scanned' % (qn + 1)
print >>OUT, '# Total time (seconds):', elapsed_time
print >>OUT, '# Rate:', "%0.2f q/s" % ((float(qn + 1) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn+1, elapsed_time, "%0.2f q/s" % ((float(qn+1) / elapsed_time))), 'lblue')
def annotate_hits_file_sequential(seed_orthologs_file, annot_file,
hmm_hits_file, args):
annot_header = (
"#query_name",
"seed_eggNOG_ortholog",
"seed_ortholog_evalue",
"seed_ortholog_score",
"predicted_gene_name",
"GO_terms",
"KEGG_pathways",
"Annotation_tax_scope",
"OGs",
"bestOG|evalue|score",
"COG cat",
"eggNOG annot",
)
start_time = time.time()
seq2bestOG = {}
if pexists(hmm_hits_file):
seq2bestOG = get_seq_hmm_matches(hmm_hits_file)
seq2annotOG = annota.get_ogs_annotations(
set([v[0] for v in seq2bestOG.itervalues()]))
print colorify("Functional annotation of refined hits starts now", 'green')
OUT = open(annot_file, "w")
if not args.no_file_comments:
print >> OUT, '# ' + time.ctime()
print >> OUT, '# ' + ' '.join(sys.argv)
print >> OUT, '\t'.join(annot_header)
qn = 0
for line in open(seed_orthologs_file):
if not line.strip() or line.startswith('#'):
continue
if qn and (qn % 500 == 0):
total_time = time.time() - start_time
print >> sys.stderr, qn + 1, total_time, "%0.2f q/s (refinement)" % (
(float(qn + 1) / total_time))
sys.stderr.flush()
qn += 1
r = map(str.strip, line.split('\t'))
query_name = r[0]
best_hit_name = r[1]
if best_hit_name == '-' or best_hit_name == 'ERROR':
continue
best_hit_evalue = float(r[2])
best_hit_score = float(r[3])
if best_hit_score < args.seed_ortholog_score or best_hit_evalue > args.seed_ortholog_evalue:
continue
match_nogs = annota.get_member_ogs(best_hit_name)
if not match_nogs:
continue
match_levels = set([nog.split("@")[1] for nog in match_nogs])
if args.tax_scope == "auto":
for level in TAXONOMIC_RESOLUTION:
if level in match_levels:
annot_levels = set(LEVEL_CONTENT.get(level, [level]))
annot_levels.add(level)
annot_level_max = "%s[%d]" % (level, len(annot_levels))
break
else:
annot_levels = set(
LEVEL_CONTENT.get(args.tax_scope, [args.tax_scope]))
annot_levels.add(args.tax_scope)
annot_level_max = "%s[%d]" % (args.tax_scope, len(annot_levels))
all_orthologies = annota.get_member_orthologs(
best_hit_name, target_levels=annot_levels)
orthologs = sorted(all_orthologies[args.target_orthologs])
if args.excluded_taxa:
orthologs = [
o for o in orthologs
if not o.startswith("%s." % args.excluded_taxa)
]
if orthologs:
pname, gos, keggs = annota.get_member_annotations(
orthologs,
target_go_ev=args.go_evidence,
excluded_go_ev=args.go_excluded)
best_name = ''
if pname:
name_candidate, freq = pname.most_common(1)[0]
if freq >= 2:
best_name = name_candidate
else:
pname = []
best_name = ''
gos = set()
keggs = set()
if query_name in seq2bestOG:
(hitname, evalue, score, qlength, hmmfrom, hmmto, seqfrom, seqto,
q_coverage) = seq2bestOG[query_name]
bestOG = '%s|%s|%s' % (hitname, evalue, score)
og_cat, og_desc = seq2annotOG.get(hitname, ['', ''])
else:
bestOG = 'NA|NA|NA'
og_cat, og_desc = '', ''
print >> OUT, '\t'.join(
map(str, (
query_name,
best_hit_name,
best_hit_evalue,
best_hit_score,
best_name,
','.join(sorted(gos)),
','.join(sorted(map(lambda x: "map%05d" % x, map(int,
keggs)))),
annot_level_max,
','.join(match_nogs),
bestOG,
og_cat.replace('\n', ''),
og_desc.replace('\n', ' '),
)))
OUT.flush()
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >> OUT, '# %d queries scanned' % (qn + 1)
print >> OUT, '# Total time (seconds):', elapsed_time
print >> OUT, '# Rate:', "%0.2f q/s" % ((float(qn + 1) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn+1, elapsed_time, "%0.2f q/s" % ((float(qn+1) / elapsed_time))), 'lblue')
def main(args):
# Output and intermediate files
hmm_hits_file = "%s.emapper.hmm_hits" % args.output
seed_orthologs_file = "%s.emapper.seed_orthologs" % args.output
annot_file = "%s.emapper.annotations" % args.output
orthologs_file = "%s.emapper.predict_orthologs" % args.output
if args.no_search:
output_files = [annot_file]
elif args.no_annot:
output_files = [hmm_hits_file, seed_orthologs_file]
else:
output_files = [hmm_hits_file, seed_orthologs_file, annot_file]
# convert to absolute path before changing directory
if args.annotate_hits_table:
args.annotate_hits_table = os.path.abspath(args.annotate_hits_table)
# force user to decide what to do with existing files
os.chdir(args.output_dir)
files_present = set([pexists(fname) for fname in output_files])
if True in files_present and not args.resume and not args.override:
print "Output files detected in disk. Use --resume or --override to continue"
raise emapperException()
if args.override:
for outf in output_files:
silent_rm(outf)
print '# ', get_version()
print '# ./emapper.py ', ' '.join(sys.argv[1:])
if args.scratch_dir:
# If resuming in and using --scratch_dir, transfer existing files.
if args.resume and args.scratch_dir:
for f in output_files:
if pexists(f):
print " Copying input file %s to scratch dir %s" % (
f, args.scratch_dir)
shutil.copy(f, args.scratch_dir)
# Change working dir
os.chdir(args.scratch_dir)
# Step 1. Sequence search
if not args.no_search:
if args.mode == 'diamond' and not args.no_search:
dump_diamond_matches(args.input, seed_orthologs_file, args)
elif args.mode == 'hmmer' and not args.no_search:
host, port, dbpath, scantype, idmap = setup_hmm_search(args)
# Start HMM SCANNING sequences (if requested)
if not pexists(hmm_hits_file) or args.override:
dump_hmm_matches(args.input, hmm_hits_file, dbpath, port,
scantype, idmap, args)
if not args.no_refine and (not pexists(seed_orthologs_file)
or args.override):
if args.db == 'viruses':
print 'Skipping seed ortholog detection in "viruses" database'
elif args.db in EGGNOG_DATABASES:
refine_matches(args.input, seed_orthologs_file,
hmm_hits_file, args)
else:
print 'refined hits not available for custom hmm databases.'
# Step 2. Annotation
if not args.no_annot:
annota.connect()
if args.annotate_hits_table:
if not os.path.exists(args.annotate_hits_table):
raise IOError(errno.ENOENT, os.strerror(errno.ENOENT),
args.annotate_hits_table)
annotate_hits_file(args.annotate_hits_table, annot_file,
hmm_hits_file, args)
elif args.db == 'viruses':
annotate_hmm_matches(hmm_hits_file, hmm_hits_file + '.annotations',
args)
OUT = open(annot_file, 'w')
for line in open(hmm_hits_file + '.annotations'):
if line.startswith('#') or not line.strip():
continue
(query, hitname, level, evalue, sum_score, query_length,
hmmfrom, hmmto, seqfrom, seqto, q_coverage, nm, desc,
cats) = line.split("\t")
if hitname != '-' and hitname != 'ERROR':
print >> OUT, '\t'.join(
map(str,
(query, hitname, evalue, sum_score, '', '', '',
'viruses', hitname + "@viruses", "%s|%s|%s" %
(hitname, evalue, sum_score),
cats.replace('\n', ''), desc.replace('\n', ' '))))
OUT.close()
else:
annotate_hits_file(seed_orthologs_file, annot_file, hmm_hits_file,
args)
if args.predict_ortho:
orthology.connect()
dump_orthologs(seed_orthologs_file, orthologs_file, args)
# If running in scratch, move files to real output dir and clean up
if args.scratch_dir:
for fname in output_files:
if pexists(fname):
print " Copying result file %s from scratch to %s" % (
fname, args.output_dir)
shutil.copy(annot_file, args.output_dir)
print " Cleaning result file %s from scratch dir" % (fname)
# Finalize and exit
print colorify('Done', 'green')
for f in output_files:
colorify('Result files:', 'yellow')
if pexists(f):
print " %s" % (f)
print 'Total time: %g secs' % (time.time() - _total_time)
if args.mode == 'hmmer':
print get_citation(['hmmer'])
elif args.mode == 'diamond':
print get_citation(['diamond'])
shutdown_server()
def dump_hmm_matches(fasta_file, hits_file, dbpath, port, scantype, idmap,
args):
hits_header = ("#query_name", "hit", "evalue", "sum_score", "query_length",
"hmmfrom", "hmmto", "seqfrom", "seqto", "query_coverage")
# Cache previous results if resuming is enabled
VISITED = set()
if args.resume and pexists(hits_file):
print colorify(
"Resuming previous run. Reading computed output from %s" %
hits_file, 'yellow')
VISITED = set([
line.split('\t')[0].strip() for line in open(hits_file)
if not line.startswith('#')
])
print len(VISITED), 'queries skipped'
OUT = open(hits_file, 'a')
else:
OUT = open(hits_file, 'w')
print colorify("Sequence mapping starts now!", 'green')
if not args.no_file_comments:
print >> OUT, get_call_info()
print >> OUT, '# ' + '\t'.join(hits_header)
total_time = 0
last_time = time.time()
start_time = time.time()
qn = 0 # in case nothing to loop bellow
for qn, (name, elapsed, hits, querylen, seq) in enumerate(
search.iter_hits(fasta_file,
args.translate,
args.qtype,
args.dbtype,
scantype,
dbpath,
port,
evalue_thr=args.evalue,
score_thr=args.score,
qcov_thr=args.qcov,
fixed_Z=args.Z,
max_hits=args.maxhits,
skip=VISITED,
maxseqlen=args.maxseqlen,
cpus=args.cpu,
base_tempdir=args.temp_dir)):
if elapsed == -1:
# error occurred
print >> OUT, '\t'.join([name] + ['ERROR'] *
(len(hits_header) - 1))
elif not hits:
print >> OUT, '\t'.join([name] + ['-'] * (len(hits_header) - 1))
else:
for hitindex, (hid, heval, hscore, hmmfrom, hmmto, sqfrom, sqto,
domscore) in enumerate(hits):
hitname = hid
if idmap:
hitname = idmap[hid][0]
print >> OUT, '\t'.join(
map(str, [
name, hitname, heval, hscore,
int(querylen),
int(hmmfrom),
int(hmmto),
int(sqfrom),
int(sqto),
float(sqto - sqfrom) / querylen
]))
OUT.flush()
# monitoring
total_time += time.time() - last_time
last_time = time.time()
if qn and (qn % 25 == 0):
print >>sys.stderr, qn + \
1, total_time, "%0.2f q/s" % ((float(qn + 1) / total_time))
sys.stderr.flush()
# Writes final stats
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >> OUT, '# %d queries scanned' % (qn + 1)
print >> OUT, '# Total time (seconds):', elapsed_time
print >> OUT, '# Rate:', "%0.2f q/s" % ((float(qn + 1) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn+1, elapsed_time, "%0.2f q/s" % ((float(qn+1) / elapsed_time))), 'lblue')
def dump_diamond_matches(fasta_file, seed_orthologs_file, args):
cpu = args.cpu
score_thr = args.seed_ortholog_score
evalue_thr = args.seed_ortholog_evalue
excluded_taxa = args.excluded_taxa if args.excluded_taxa else None
if args.translate:
tool = 'blastx'
else:
tool = 'blastp'
dmnd_db = args.dmnd_db if args.dmnd_db else get_eggnog_dmnd_db()
query_cov = args.query_cover
subject_cov = args.subject_cover
dmnd_opts = ''
if args.matrix is not None:
dmnd_opts += ' --matrix %s' % args.matrix
if args.gapopen is not None:
dmnd_opts += ' --gapopen %d' % args.gapopen
if args.gapextend is not None:
dmnd_opts += ' --gapextend %d' % args.gapextend
if not DIAMOND:
raise ValueError("diamond not found in path")
tempdir = mkdtemp(prefix='emappertmp_dmdn_', dir=args.temp_dir)
raw_output_file = pjoin(tempdir, uuid.uuid4().hex)
if excluded_taxa:
cmd = '%s %s -d %s -q %s --more-sensitive --threads %s -e %f -o %s --max-target-seqs 25 --query-cover %s --subject-cover %s' %\
(DIAMOND, tool, dmnd_db, fasta_file, cpu, evalue_thr, raw_output_file, query_cov, subject_cov)
else:
cmd = '%s %s -d %s -q %s --more-sensitive --threads %s -e %f -o %s --top 3 --query-cover %s --subject-cover %s' %\
(DIAMOND, tool, dmnd_db, fasta_file, cpu, evalue_thr, raw_output_file, query_cov, subject_cov)
#diamond blastp --threads "${GALAXY_SLOTS:-12}" --db ./database --query '/panfs/roc/galaxy/GALAXYP/files/000/164/dataset_164640.dat' --query-gencode '1' --outfmt '6' qseqid sseqid sallseqid qlen slen pident length nident mismatch positive gapopen gaps ppos qstart qend sstart send qseq sseq evalue bitscore score qframe stitle salltitles qcovhsp --out '/panfs/roc/galaxy/GALAXYP/files/000/164/dataset_164759.dat' --compress '0' --gapopen '10' --gapextend '1' --matrix 'PAM30' --seg 'yes' --max-target-seqs '25' --evalue '0.001' --id '0' --query-cover '0' --block-size '2.0'
print colorify(' ' + cmd, 'yellow')
try:
subprocess.check_call(cmd, shell=True, stdout=subprocess.PIPE)
OUT = open('%s' % seed_orthologs_file, 'w')
if not args.no_file_comments:
print >> OUT, get_call_info()
print >> OUT, '#', cmd
visited = set()
for line in open(raw_output_file):
if not line.strip() or line.startswith('#'):
continue
fields = map(str.strip, line.split('\t'))
query = fields[0]
hit = fields[1]
evalue = float(fields[10])
score = float(fields[11])
if query in visited:
continue
if evalue > evalue_thr or score < score_thr:
continue
if excluded_taxa and hit.startswith("%s." % excluded_taxa):
continue
visited.add(query)
print >> OUT, '\t'.join(map(str, [query, hit, evalue, score]))
OUT.close()
except subprocess.CalledProcessError as e:
raise e
finally:
shutil.rmtree(tempdir)
def main(args):
# Output and intermediate files
hmm_hits_file = "%s.emapper.hmm_hits" % args.output
seed_orthologs_file = "%s.emapper.seed_orthologs" % args.output
annot_file = "%s.emapper.annotations" % args.output
if args.no_search:
output_files = [annot_file]
elif args.no_annot:
output_files = [hmm_hits_file, seed_orthologs_file]
else:
output_files = [hmm_hits_file, seed_orthologs_file, annot_file]
# force user to decide what to do with existing files
os.chdir(args.output_dir)
files_present = set([pexists(fname) for fname in output_files])
if True in files_present and not args.resume and not args.override:
print "Output files detected in disk. Use --resume or --override to continue"
raise emapperException()
if args.override:
for outf in output_files:
silent_rm(outf)
print '# ', get_version()
print '# ./emapper.py ', ' '.join(sys.argv[1:])
if args.scratch_dir:
# If resuming in and using --scratch_dir, transfer existing files.
if args.resume and args.scratch_dir:
for f in output_files:
if pexists(f):
print " Copying input file %s to scratch dir %s" % (f, args.scratch_dir)
shutil.copy(f, args.scratch_dir)
# Change working dir
os.chdir(args.scratch_dir)
# Step 1. Sequence search
if not args.no_search:
if args.mode == 'diamond' and not args.no_search:
dump_diamond_matches(args.input, seed_orthologs_file, args)
elif args.mode == 'hmmer' and not args.no_search:
host, port, dbpath, scantype, idmap = setup_hmm_search(args)
# Start HMM SCANNING sequences (if requested)
if not pexists(hmm_hits_file) or args.override:
dump_hmm_matches(args.input, hmm_hits_file, dbpath, port, scantype, idmap, args)
if not args.no_refine and (not pexists(seed_orthologs_file) or args.override):
if args.db == 'viruses':
print 'Skipping seed ortholog detection in "viruses" database'
elif args.db in EGGNOG_DATABASES:
refine_matches(args.input, seed_orthologs_file, hmm_hits_file, args)
else:
print 'refined hits not available for custom hmm databases.'
# Step 2. Annotation
if not args.no_annot:
annota.connect()
if args.annotate_hits_table:
annotate_hits_file(args.annotate_hits_table, annot_file, hmm_hits_file, args)
elif args.db == 'viruses':
annotate_hmm_matches(hmm_hits_file, hmm_hits_file+'.annotations', args)
OUT = open(annot_file, 'w')
for line in open(hmm_hits_file+'.annotations'):
if line.startswith('#') or not line.strip():
continue
(query, hitname, level, evalue, sum_score, query_length,
hmmfrom, hmmto, seqfrom, seqto, q_coverage, nm, desc, cats) = line.split("\t")
if hitname != '-' and hitname != 'ERROR':
print >>OUT, '\t'.join(map(str, (query,
hitname,
evalue,
sum_score,
'',
'',
'',
'viruses',
hitname+"@viruses",
"%s|%s|%s" %(hitname, evalue, sum_score),
cats.replace('\n', ''),
desc.replace('\n', ' '))))
OUT.close()
else:
annotate_hits_file(seed_orthologs_file, annot_file, hmm_hits_file, args)
# If running in scratch, move files to real output dir and clean up
if args.scratch_dir:
for fname in output_files:
if pexists(fname):
print " Copying result file %s from scratch to %s" % (fname, args.output_dir)
shutil.copy(annot_file, args.output_dir)
print " Cleaning result file %s from scratch dir" %(fname)
# Finalize and exit
print colorify('Done', 'green')
for f in output_files:
colorify('Result files:', 'yellow')
if pexists(f):
print " %s" % (f)
print 'Total time: %g secs' % (time.time()-_total_time)
if args.mode == 'hmmer':
print get_citation(['hmmer'])
elif args.mode == 'diamond':
print get_citation(['diamond'])
shutdown_server()
help='assume "yes" to all questions')
parser.add_argument('-f', action="store_true", dest='force',
help='forces download even if the files exist')
parser.add_argument('-s', action="store_true", dest='simulate',
help='simulate and print commands. Nothing is downloaded')
args = parser.parse_args()
if 'all' in args.dbs:
args.dbs = EGGNOG_DATABASES
if args.force or not pexists(pjoin(DATA_PATH, 'eggnog.db')):
if args.allyes or ask("Download main annotation database?") == 'y':
print colorify('Downloading "eggnog.db" at %s...' %DATA_PATH, 'green')
download_annotations()
else:
print 'Skipping'
else:
print colorify('Skipping eggnog.db database (already present). Use -f to force download', 'lblue')
if args.force or not pexists(pjoin(DATA_PATH, 'OG_fasta')):
if args.allyes or ask("Download OG fasta files for annotation refinement (~20GB after decompression)?") == 'y':
print colorify('Downloading fasta files " at %s/OG_fasta...' %DATA_PATH, 'green')
download_groups()
else:
print 'Skipping'
else:
def annotate_hits_file(seed_orthologs_file, annot_file, hmm_hits_file, args):
annot_header = ("#query_name",
"seed_eggNOG_ortholog",
"seed_ortholog_evalue",
"seed_ortholog_score",
"predicted_gene_name",
"GO_terms",
"KEGG_pathways",
"Annotation_tax_scope",
"OGs",
"bestOG|evalue|score",
"COG cat",
"eggNOG annot",
)
start_time = time.time()
seq2bestOG = {}
if pexists(hmm_hits_file):
seq2bestOG = get_seq_hmm_matches(hmm_hits_file)
seq2annotOG = annota.get_ogs_annotations(set([v[0] for v in seq2bestOG.itervalues()]))
print colorify("Functional annotation of refined hits starts now", 'green')
OUT = open(annot_file, "w")
if not args.no_file_comments:
print >>OUT, '# ' + time.ctime()
print >>OUT, '# ' + ' '.join(sys.argv)
print >>OUT, '\t'.join(annot_header)
qn = 0
for line in open(seed_orthologs_file):
if not line.strip() or line.startswith('#'):
continue
if qn and (qn % 500 == 0):
total_time = time.time() - start_time
print >>sys.stderr, qn+1, total_time, "%0.2f q/s (refinement)" % (
(float(qn + 1) / total_time))
sys.stderr.flush()
qn += 1
r = map(str.strip, line.split('\t'))
query_name = r[0]
best_hit_name = r[1]
if best_hit_name == '-' or best_hit_name == 'ERROR':
continue
best_hit_evalue = float(r[2])
best_hit_score = float(r[3])
if best_hit_score < args.seed_ortholog_score or best_hit_evalue > args.seed_ortholog_evalue:
continue
match_nogs = annota.get_member_ogs(best_hit_name)
if not match_nogs:
continue
match_levels = set([nog.split("@")[1] for nog in match_nogs])
if args.tax_scope == "auto":
for level in TAXONOMIC_RESOLUTION:
if level in match_levels:
annot_levels = set(LEVEL_CONTENT.get(level, [level]))
annot_levels.add(level)
annot_level_max = "%s[%d]" %(level, len(annot_levels))
break
else:
annot_levels = set(LEVEL_CONTENT.get(args.tax_scope, [args.tax_scope]))
annot_levels.add(args.tax_scope)
annot_level_max = "%s[%d]" %(args.tax_scope, len(annot_levels))
all_orthologies = annota.get_member_orthologs(best_hit_name, target_levels=annot_levels)
orthologs = sorted(all_orthologies[args.target_orthologs])
if args.excluded_taxa:
orthologs = [o for o in orthologs if not o.startswith("%s." %args.excluded_taxa)]
if orthologs:
pname, gos, keggs = annota.get_member_annotations(orthologs,
excluded_gos=set(["IEA", "ND"]))
best_name = ''
if pname:
name_candidate, freq = pname.most_common(1)[0]
if freq >= 2:
best_name = name_candidate
else:
pname = []
best_name = ''
gos = set()
keggs = set()
if query_name in seq2bestOG:
(hitname, evalue, score, qlength, hmmfrom, hmmto, seqfrom,
seqto, q_coverage) = seq2bestOG[query_name]
bestOG = '%s|%s|%s' %(hitname, evalue, score)
og_cat, og_desc = seq2annotOG.get(hitname, ['', ''])
else:
bestOG = 'NA|NA|NA'
og_cat, og_desc = '', ''
print >>OUT, '\t'.join(map(str, (query_name,
best_hit_name,
best_hit_evalue,
best_hit_score,
best_name,
','.join(sorted(gos)),
','.join(sorted(map(lambda x: "map%05d"%x, map(int, keggs)))),
annot_level_max,
','.join(match_nogs),
bestOG,
og_cat.replace('\n', ''),
og_desc.replace('\n', ' '),
)))
OUT.flush()
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >>OUT, '# %d queries scanned' % (qn + 1)
print >>OUT, '# Total time (seconds):', elapsed_time
print >>OUT, '# Rate:', "%0.2f q/s" % ((float(qn + 1) / elapsed_time))
OUT.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn+1, elapsed_time, "%0.2f q/s" % ((float(qn+1) / elapsed_time))), 'lblue')
def parse_args(parser):
args = parser.parse_args()
if args.version:
print get_version()
sys.exit(0)
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
if not args.no_annot and not pexists(get_eggnogdb_file()):
print colorify(
'Annotation database data/eggnog.db not present. Use download_eggnog_database.py to fetch it',
'red')
raise emapperException()
if args.mode == 'diamond':
dmnd_db = args.dmnd_db if args.dmnd_db else get_eggnog_dmnd_db()
if not pexists(dmnd_db):
print colorify(
'DIAMOND database %s not present. Use download_eggnog_database.py to fetch it'
% dmnd_db, 'red')
raise emapperException()
if args.cpu == 0:
args.cpu = multiprocessing.cpu_count()
# No --servermode available for diamond
if args.mode == 'diamond' and args.servermode:
parser.error(
'--mode [diamond] and --servermode are mutually exclusive')
# Output file required unless running in servermode
if not args.servermode and not args.output:
parser.error('An output project name is required (-o)')
# Servermode implies using mem-based databases
if args.servermode:
args.usemem = True
# Direct annotation implies no searches
if args.annotate_hits_table:
args.no_search = True
args.no_annot = False
# Sets GO evidence bases
if args.go_evidence == 'experimental':
args.go_evidence = set(["EXP", "IDA", "IPI", "IMP", "IGI", "IEP"])
args.go_excluded = set(["ND", "IEA"])
elif args.go_evidence == 'non-electronic':
args.go_evidence = None
args.go_excluded = set(["ND", "IEA"])
else:
raise ValueError('Invalid --go_evidence value')
# Check inputs for running sequence searches
if not args.no_search and not args.servermode:
if not args.input:
parser.error('An input fasta file is required (-i)')
# HMM
if args.mode == 'hmmer':
if not args.db and not args.guessdb:
parser.error(
'HMMER mode requires specifying a target database (i.e. -d, --guessdb ))'
)
if args.db and args.guessdb:
parser.error('-d and --guessdb options are mutually exclusive')
if args.guessdb:
from ete3 import NCBITaxa
ncbi = NCBITaxa()
lineage = ncbi.get_lineage(args.guessdb)
for tid in reversed(lineage):
if tid in TAXID2LEVEL:
print tid, TAXID2LEVEL[tid]
args.db = TAXID2LEVEL[tid]
break
# DIAMOND
elif args.mode == 'diamond':
#if args.db or args.guessdb:
# parser.error('diamond mode does not require -d or --guessdb options')
pass
return args
def parse_args(parser):
args = parser.parse_args()
if args.version:
print get_version()
sys.exit(0)
if not args.no_annot and not pexists(EGGNOGDB_FILE):
print colorify('Annotation database data/eggnog.db not present. Use download_eggnog_database.py to fetch it', 'red')
raise emapperException()
if args.mode == 'diamond' and not pexists(EGGNOG_DMND_DB):
print colorify('DIAMOND database data/eggnog_proteins.dmnd not present. Use download_eggnog_database.py to fetch it', 'red')
raise emapperException()
if args.cpu == 0:
args.cpu = multiprocessing.cpu_count()
# No --servermode available for diamond
if args.mode == 'diamond' and args.servermode:
parser.error('--mode [diamond] and --servermode are mutually exclusive')
# Output file required unless running in servermode
if not args.servermode and not args.output:
parser.error('An output project name is required (-o)')
# Servermode implies using mem-based databases
if args.servermode:
args.usemem = True
# Direct annotation implies no searches
if args.annotate_hits_table:
args.no_search = True
args.no_annot = False
# Check inputs for running sequence searches
if not args.no_search and not args.servermode:
if not args.input:
parser.error('An input fasta file is required (-i)')
# HMM
if args.mode == 'hmmer':
if not args.db and not args.guessdb:
parser.error('HMMER mode requires specifying a target database (i.e. -d, --guessdb ))')
if args.db and args.guessdb:
parser.error('-d and --guessdb options are mutually exclusive')
if args.guessdb:
from ete3 import NCBITaxa
ncbi = NCBITaxa()
lineage = ncbi.get_lineage(args.guessdb)
for tid in reversed(lineage):
if tid in TAXID2LEVEL:
print tid, TAXID2LEVEL[tid]
args.db = TAXID2LEVEL[tid]
break
# DIAMOND
elif args.mode == 'diamond':
#if args.db or args.guessdb:
# parser.error('diamond mode does not require -d or --guessdb options')
pass
return args
def dump_diamond_matches(fasta_file, seed_orthologs_file, args):
cpu = args.cpu
score_thr = args.seed_ortholog_score
evalue_thr = args.seed_ortholog_evalue
excluded_taxa = args.excluded_taxa if args.excluded_taxa else None
if args.translate:
tool = 'blastx'
else:
tool = 'blastp'
dmnd_db = args.dmnd_db if args.dmnd_db else get_eggnog_dmnd_db()
query_cov = args.query_cover
subject_cov = args.subject_cover
dmnd_opts = ''
if args.matrix is not None:
dmnd_opts += ' --matrix %s' % args.matrix
if args.gapopen is not None:
dmnd_opts += ' --gapopen %d' % args.gapopen
if args.gapextend is not None:
dmnd_opts += ' --gapextend %d' % args.gapextend
if not DIAMOND:
raise ValueError("diamond not found in path")
tempdir = mkdtemp(prefix='emappertmp_dmdn_', dir=args.temp_dir)
raw_output_file = pjoin(tempdir, uuid.uuid4().hex)
if excluded_taxa:
cmd = '%s %s -d %s -q %s --more-sensitive --threads %s -e %f -o %s --max-target-seqs 25 --query-cover %s --subject-cover %s' %\
(DIAMOND, tool, dmnd_db, fasta_file, cpu, evalue_thr, raw_output_file, query_cov, subject_cov)
else:
cmd = '%s %s -d %s -q %s --more-sensitive --threads %s -e %f -o %s --top 3 --query-cover %s --subject-cover %s' %\
(DIAMOND, tool, dmnd_db, fasta_file, cpu, evalue_thr, raw_output_file, query_cov, subject_cov)
print colorify(' ' + cmd, 'yellow')
try:
with open(raw_output_file + '.stdout', 'w') as STDOUT:
subprocess.check_call(cmd, shell=True, stdout=STDOUT)
OUT = open('%s' % seed_orthologs_file, 'w')
if not args.no_file_comments:
print >> OUT, get_call_info()
print >> OUT, '#', cmd
visited = set()
for line in open(raw_output_file):
if not line.strip() or line.startswith('#'):
continue
fields = map(str.strip, line.split('\t'))
query = fields[0]
hit = fields[1]
evalue = float(fields[10])
score = float(fields[11])
if query in visited:
continue
if evalue > evalue_thr or score < score_thr:
continue
if excluded_taxa and hit.startswith("%s." % excluded_taxa):
continue
visited.add(query)
print >> OUT, '\t'.join(map(str, [query, hit, evalue, score]))
OUT.close()
except subprocess.CalledProcessError as e:
raise e
finally:
shutil.rmtree(tempdir)
help='assume "yes" to all questions')
parser.add_argument('-f', action="store_true", dest='force',
help='forces download even if the files exist')
parser.add_argument('-s', action="store_true", dest='simulate',
help='simulate and print commands. Nothing is downloaded')
args = parser.parse_args()
if 'all' in args.dbs:
args.dbs = EGGNOG_DATABASES
if args.force or not pexists(pjoin(DATA_PATH, 'eggnog.db')):
if args.allyes or ask("Download main annotation database?") == 'y':
print colorify('Downloading "eggnog.db" at %s...' %DATA_PATH, 'green')
download_annotations()
else:
print 'Skipping'
else:
print colorify('Skipping eggnog.db database (already present). Use -f to force download', 'lblue')
if args.force or not pexists(pjoin(DATA_PATH, 'OG_fasta')):
if args.allyes or ask("Download OG fasta files for annotation refinement (~20GB after decompression)?") == 'y':
print colorify('Downloading fasta files " at %s/OG_fasta...' %DATA_PATH, 'green')
download_groups()
else:
print 'Skipping'
else:
def run(cmd):
print colorify(cmd, 'cyan')
if not args.simulate:
os.system(cmd)
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
# if args.force or not pexists(pjoin(get_data_path(), 'og2level.tsv.gz')):
# print colorify('Downloading "og2level.tsv.gz" at %s' %get_data_path(), 'green')
# download_og2level()
# if 'all' in args.dbs:
# args.dbs = EGGNOG_DATABASES
if args.force or not pexists(pjoin(get_data_path(), 'eggnog.db')):
if args.allyes or ask("Download main annotation database?") == 'y':
print colorify(
'Downloading "eggnog.db" at %s...' % get_data_path(), 'green')
download_annotations()
else:
print 'Skipping'
else:
if not args.quiet:
print colorify(
'Skipping eggnog.db database (already present). Use -f to force download',
'lblue')
# if args.force or not pexists(pjoin(get_data_path(), 'OG_fasta')):
# if args.allyes or ask("Download OG fasta files for annotation refinement (~20GB after decompression)?") == 'y':
# print colorify('Downloading fasta files " at %s/OG_fasta...' %get_data_path(), 'green')
# download_groups()
# else:
def annotate_hits_file(seed_orthologs_file, annot_file, hmm_hits_file, args):
HIT_HEADER = [
"#query_name",
"seed_eggNOG_ortholog",
"seed_ortholog_evalue",
"seed_ortholog_score",
"best_tax_level",
]
HIT_OG_HEADER = [
"taxonomic scope", "eggNOG OGs", "best eggNOG OG",
"COG Functional cat.", "eggNOG free text desc."
]
start_time = time.time()
seq2bestOG = {}
if pexists(hmm_hits_file):
seq2bestOG = get_seq_hmm_matches(hmm_hits_file)
seq2annotOG = annota.get_ogs_annotations(
set([v[0] for v in seq2bestOG.itervalues()]))
print colorify("Functional annotation of refined hits starts now", 'green')
OUT = open(annot_file, "w")
if args.report_orthologs:
ORTHOLOGS = open(annot_file + ".orthologs", "w")
if not args.no_file_comments:
print >> OUT, '# emapper version:', get_version(
), 'emapper DB:', get_db_version()
print >> OUT, '# command: ./emapper.py ', ' '.join(sys.argv[1:])
print >> OUT, '# time: ' + time.ctime()
print >> OUT, '\t'.join(HIT_HEADER + ANNOTATIONS_HEADER +
HIT_OG_HEADER)
qn = 0
pool = multiprocessing.Pool(args.cpu)
for result in pool.imap(annotate_hit_line,
iter_hit_lines(seed_orthologs_file, args)):
qn += 1
if qn and (qn % 500 == 0):
total_time = time.time() - start_time
print >> sys.stderr, qn, total_time, "%0.2f q/s (func. annotation)" % (
(float(qn) / total_time))
sys.stderr.flush()
if result:
(query_name, best_hit_name, best_hit_evalue, best_hit_score,
annotations, annot_level_max, swallowest_level, match_nogs,
orthologs) = result
if query_name in seq2bestOG:
(hitname, evalue, score, qlength, hmmfrom, hmmto, seqfrom,
seqto, q_coverage) = seq2bestOG[query_name]
bestOG = '%s|%s|%s' % (hitname, evalue, score)
og_cat, og_desc = seq2annotOG.get(hitname, ['', ''])
else:
bestOG = 'NA|NA|NA'
og_cat, og_desc = annota.get_best_og_description(match_nogs)
if args.report_orthologs:
print >> ORTHOLOGS, '\t'.join(
map(str, (query_name, ','.join(orthologs))))
# prepare annotations for printing
annot_columns = [
query_name, best_hit_name,
str(best_hit_evalue),
str(best_hit_score), LEVEL_NAMES[swallowest_level]
]
for h in ANNOTATIONS_HEADER:
if h in annotations:
annot_columns.append(','.join(sorted(annotations[h])))
else:
annot_columns.append('')
annot_columns.extend([
annot_level_max, ','.join(match_nogs), bestOG,
og_cat.replace('\n', ''),
og_desc.replace('\n', ' ')
])
print >> OUT, '\t'.join(annot_columns)
#OUT.flush()
pool.terminate()
elapsed_time = time.time() - start_time
if not args.no_file_comments:
print >> OUT, '# %d queries scanned' % (qn)
print >> OUT, '# Total time (seconds):', elapsed_time
print >> OUT, '# Rate:', "%0.2f q/s" % ((float(qn) / elapsed_time))
OUT.close()
if args.report_orthologs:
ORTHOLOGS.close()
print colorify(" Processed queries:%s total_time:%s rate:%s" %\
(qn, elapsed_time, "%0.2f q/s" % ((float(qn) / elapsed_time))), 'lblue')
def parse_args(parser):
args = parser.parse_args()
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
if args.version:
version = ""
try:
version = get_full_version_info()
except Exception:
version = get_version()
print(version)
sys.exit(0)
args.call_info = get_call_info()
if args.list_taxa:
print_taxa()
sys.exit(0)
if args.cpu == 0:
args.cpu = multiprocessing.cpu_count()
multiprocessing.set_start_method(args.mp_start_method)
if args.resume == True and args.override == True:
parser.error('Only one of --resume or --override is allowed.')
# Gene prediction
if args.training_genome is not None and args.training_file is None:
parser.error('"--training_genome requires --training_file"')
if args.training_genome is None and args.training_file is not None:
if not os.path.isfile(args.training_file):
parser.error('"--training_file must point to an existing file, if no --training_genome is provided."')
# Search modes
if args.mode == SEARCH_MODE_DIAMOND:
dmnd_db = args.dmnd_db if args.dmnd_db else get_eggnog_dmnd_db()
if not pexists(dmnd_db):
print(colorify('DIAMOND database %s not present. Use download_eggnog_database.py to fetch it' % dmnd_db, 'red'))
raise EmapperException()
if args.input is not None:
if args.annotate_hits_table is not None:
print(colorify(f"--annotate_hits_table will be ignored, due to -m {SEARCH_MODE_DIAMOND}", 'blue'))
args.annotate_hits_table = None
else:
# the default -m is diamond, but we will consider -m no_search as default when
# --annotate_hits_table has been provided and -i has not been provided
if args.annotate_hits_table is not None:
print(colorify(f"Assuming -m {SEARCH_MODE_NO_SEARCH}", 'blue'))
args.mode = SEARCH_MODE_NO_SEARCH
else:
parser.error('An input fasta file is required (-i)')
# Output file required
if not args.output:
parser.error('An output project name is required (-o)')
elif args.mode == SEARCH_MODE_MMSEQS2:
mmseqs_db = args.mmseqs_db if args.mmseqs_db else get_eggnog_mmseqs_db()
if not pexists(mmseqs_db):
print(colorify('MMseqs2 database %s not present. Use download_eggnog_database.py to fetch it' % mmseqs_db, 'red'))
raise EmapperException()
if not args.input:
parser.error('An input fasta file is required (-i)')
# Output file required
if not args.output:
parser.error('An output project name is required (-o)')
if args.annotate_hits_table is not None:
print(colorify(f"--annotate_hits_table will be ignored, due to -m {SEARCH_MODE_MMSEQS2}", 'blue'))
args.annotate_hits_table = None
elif args.mode == SEARCH_MODE_HMMER:
if not args.input:
parser.error('An input file is required (-i)')
# Output file required
if not args.output:
parser.error('An output project name is required (-o)')
# Hmmer database
# NOTE: hmmer database format, name and checking if exists is done within hmmer module
if not args.db:
parser.error('HMMER mode requires a target database (-d, --database).')
if args.itype == ITYPE_CDS:
args.translate = True
if (args.itype == ITYPE_GENOME or args.itype == ITYPE_META) and args.genepred == GENEPRED_MODE_SEARCH:
parser.error('HMMER mode is not compatible with "--genepred search" option.')
if args.annotate_hits_table is not None:
print(colorify(f"--annotate_hits_table will be ignored, due to -m {SEARCH_MODE_HMMER}", 'blue'))
args.annotate_hits_table = None
if args.clean_overlaps is not None:
if args.clean_overlaps == "none":
args.clean_overlaps = None
elif args.mode == SEARCH_MODE_CACHE:
if args.cache_file is None:
parser.error('A file with annotations and md5 of queries is required (-c FILE)')
if args.decorate_gff != DECORATE_GFF_NONE:
print(colorify("WARNING: no GFF will be created for cache-based annotations. It is not implemented yet, sorry.", 'red'))
if args.no_annot == True:
parser.error(f'Cache mode (-m {SEARCH_MODE_CACHE}) should be used to annotate.')
elif args.mode == SEARCH_MODE_NO_SEARCH:
if args.no_annot == False and not args.annotate_hits_table:
parser.error(f'No search mode (-m {SEARCH_MODE_NO_SEARCH}) requires a hits table to annotate (--annotate_hits_table FILE.seed_orthologs)')
if args.md5 == True and args.input is None:
parser.error(f'--md5 requires an input FASTA file (-i FASTA).')
else:
parser.error(f'unrecognized search mode (-m {args.mode})')
# Search thresholds
args.dmnd_evalue = args.mmseqs_evalue = args.hmm_evalue = args.evalue
args.dmnd_score = args.mmseqs_score = args_hmm_score = args.score
args.qcov = args.query_cover
# Annotation options
if args.no_annot == False or args.report_orthologs == True:
if not pexists(get_eggnogdb_file()):
print(colorify('Annotation database data/eggnog.db not present. Use download_eggnog_database.py to fetch it', 'red'))
raise EmapperException()
args.tax_scope_ids = parse_tax_scope(args.tax_scope)
if args.target_taxa is not None:
args.target_taxa = args.target_taxa.split(",")
if args.excluded_taxa is not None:
args.excluded_taxa = args.excluded_taxa.split(",")
# Sets GO evidence bases
if args.go_evidence == 'experimental':
args.go_evidence = set(["EXP","IDA","IPI","IMP","IGI","IEP"])
args.go_excluded = set(["ND", "IEA"])
elif args.go_evidence == 'non-electronic':
args.go_evidence = None
args.go_excluded = set(["ND", "IEA"])
elif args.go_evidence == 'all':
args.go_evidence = None
args.go_excluded = None
else:
raise ValueError('Invalid --go_evidence value')
# PFAM annotation options
if args.pfam_realign == PFAM_REALIGN_NONE:
pass
elif args.pfam_realign == PFAM_REALIGN_REALIGN or args.pfam_realign == PFAM_REALIGN_DENOVO:
if not args.input:
parser.error(f'An input fasta file is required (-i) for --pfam_realign {args.pfam_realign}')
else:
raise ValueError(f'Invalid --pfam_realign option {args.pfam_realign}')
total_workers = args.num_workers * args.num_servers
if args.cpu < total_workers:
parser.error(f"Less cpus ({args.cpu}) than total workers ({total_workers}) were specified.")
if args.cpu % total_workers != 0:
parser.error(f"Number of cpus ({args.cpu}) must be a multiple of total workers ({total_workers}).")
args.cpus_per_worker = int(args.cpu / total_workers)
return args
