def dump_orthologs(seed_orthologs_file, orthologs_file, args):
#Copy from predict_orthologs.py
OUT = open(orthologs_file, "w")
if args.predict_output_format == "per_query":
ortholog_header = ("#Query", "Orthologs")
elif args.predict_output_format == "per_species":
ortholog_header = ("#Query", "Species", "Orthologs")
print >> OUT, "\t".join(ortholog_header)
if args.target_taxa != 'all':
args._expanded_target_taxa = orthology.normalize_target_taxa(
args.target_taxa)
else:
# report orthologs from any species by default
args._expanded_target_taxa = None
pool = multiprocessing.Pool(args.cpu)
for result in pool.imap(find_orthologs_per_hit,
iter_hit_lines(seed_orthologs_file, args)):
if result:
write_orthologs_in_file(result, OUT, args)
pool.terminate()
def _annotate_hit_line(arguments):
annota.connect()
line, args = arguments
if not line.strip() or line.startswith('#'):
return None
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':
return None
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:
return None
match_nogs = annota.get_member_ogs(best_hit_name)
if not match_nogs:
return None
match_levels = set()
for nog in match_nogs:
match_levels.update(LEVEL_PARENTS[nog.split("@")[1]])
swallowest_level = sorted(match_levels & set(LEVEL_DEPTH.keys()),
key=lambda x: LEVEL_DEPTH[x],
reverse=True)[0]
annot_levels = set()
if args.tax_scope == "auto":
for level in TAXONOMIC_RESOLUTION:
if level in match_levels:
annot_levels.add(level)
annot_level_max = LEVEL_NAMES.get(level, level)
break
else:
annot_levels.add(args.tax_scope)
annot_level_max = LEVEL_NAMES.get(args.tax_scope, args.tax_scope)
if args.target_taxa != 'all':
target_taxa = orthology.normalize_target_taxa(args.target_taxa)
else:
target_taxa = None
try:
all_orthologies = annota.get_member_orthologs(
best_hit_name, target_taxa=target_taxa, target_levels=annot_levels)
except Exception:
orthologs = None
status = 'Error'
else:
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)
]
status = 'OK'
if orthologs:
annotations = annota.summarize_annotations(
orthologs,
target_go_ev=args.go_evidence,
excluded_go_ev=args.go_excluded)
else:
annotations = {}
return (query_name, best_hit_name, best_hit_evalue, best_hit_score,
annotations, annot_level_max, swallowest_level, match_nogs,
orthologs)
def annotate_hit_line(arguments):
annota.connect()
line, args = arguments
if not line.strip() or line.startswith('#'):
return None
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':
return None
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:
return None
match_nogs = annota.get_member_ogs(best_hit_name)
if not match_nogs:
return None
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))
if args.target_taxa != 'all':
target_taxa = orthology.normalize_target_taxa(args.target_taxa)
else:
target_taxa = None
all_orthologies = annota.get_member_orthologs(best_hit_name,
target_taxa=target_taxa,
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, kegg, bigg = annota.summarize_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()
kegg = set()
bigg = set()
return (query_name, best_hit_name, best_hit_evalue, best_hit_score,
best_name, gos, kegg, bigg, annot_level_max, match_nogs, orthologs)