def annotate_gene(self, reaction, gene_name):
try:
proteins = Protein.objects(organism=self.organism, gene=gene_name)
if not len(proteins):
proteins = Protein.objects(organism=self.organism,
alias__iexact=gene_name)
if len(proteins):
for protein in proteins:
if not hasattr(protein, "reactions"):
protein.reactions = [
] # self.add_pathways_info_to_gene(reaction, protein)
self.add_reaction_info_to_gene(reaction, protein)
self.add_properties_to_gene(protein, gene_name)
protein.save()
else:
_log.warn("%s not found" % gene_name)
self.unmapped_genes.append(gene_name)
except DoesNotExist:
self.unmapped_genes.append(gene_name)
_log.warn("%s not found" % gene_name)
except MultipleObjectsReturned:
self.unmapped_genes.append(gene_name)
_log.warn(gene_name)
def update_proteins(annotation_dir,
proteome,
seq_col_name,
tax_id,
identity=0.9,
cpus=multiprocessing.cpu_count(),
db_init=None):
# if db_init:
# from SNDG.Sequence.ProteinAnnotator import PABase
# PABase.sqldb.initialize(db_init)
# mkdir(annotation_dir)
# out = annotation_dir + "/species_blast.tbl"
#
# tax = Tax.select().where(Tax.ncbi_taxon_id == tax_id).get()
# species_tax = None
# for tax in Tax.parents(tax):
# if tax.node_rank == "genus":
# species_tax = tax
# break
# tax_data = "/data/xomeq/tax/"
# species_fasta = tax_data + str(int(species_tax.ncbi_taxon_id)) + ".fasta"
if not os.path.exists(out):
if not os.path.exists(species_fasta):
Uniprot.download_proteome_from_tax(str(species_tax.ncbi_taxon_id),
tax_data)
cmd = "blastp -query %s -db %s -evalue 0.00001 -outfmt 6 -max_hsps 1 -qcov_hsp_perc 0.9 -num_threads %i -out %s"
execute(cmd % (proteome, species_fasta, cpus, out))
species_desc = {
x.id.split("|")[1]: " ".join(x.description.split()[1:])
for x in bpio.parse(species_fasta, "fasta")
}
total = Protein.objects(organism=seq_col_name).count()
with tqdm(bpsio.parse(out, "blast-tab"), total=total) as pbar:
for query in pbar:
pbar.set_description(query.id)
if query[0][0].ident_pct > identity:
unip = query[0].id.split(
"|")[1] if "|" in query[0].id else query[0].id
dbxrefs = [
x.db + "||" + x.value
for x in Mapping.select().where(Mapping.uniprot == unip)
]
p = Protein.objects(gene=query.id,
organism=seq_col_name).no_cache().get()
if not p.description and unip in species_desc:
p.description = species_desc[unip].split(
"OS=")[0] + " | homology with: " + unip
p.save()
if dbxrefs:
p = SearchLoader.update_protein_with_dbxref(
query.id, dbxrefs, seq_col_name)
p.save()
def process(
self,
data_csv="/data/projects/23staphylo/raw/metadata/saureus_resist_snps.csv",
genome="SaureusN315",
user="******"):
Protein.objects(organism=genome).update(
__raw__={"$pull": {
"features": {
"type": "Aanensen2016"
}
}})
df = self.create_df(data_csv)
for i, r in df.iterrows():
gene = r["Core gene"]
gene = gene.strip()
assert gene
# regex = re.compile(gene + '.*')
has_prot = False
for prot in Protein.objects(organism=genome, gene__iexact=gene):
has_prot = True
# print len([x.gene[1] if len(x.gene) > 1 else x.gene[0] for x in prot])
self._process_prot(prot, r, i)
if not has_prot:
if "RNA" not in r["Core gene"]:
if r["Core gene"]:
print "Core gene not found %s" % gene
else:
print "%s not found" % gene
def _process_proteins(self, genome):
total = Protein.objects(organism=genome.name, reactions__0__exists=True).count()
iterprot = tqdm(Protein.objects(organism=genome.name, reactions__0__exists=True).no_cache().timeout(
False),total=total)
for protein in iterprot: # @UndefinedVariable
keywords1 = self._protein_keywords(protein)
keywords2 = self._add_drugability_props_to_protein(protein)
protein.keywords = list(
set([x.strip().lower() for x in keywords1 + keywords2 + protein.keywords if x.strip()]))
protein.save()
self.update_ont_org_idx()
def common_annotations(collection_name, tmp_dir, cpu=1, remove_tmp=False):
process_pdb = Protein.objects(
__raw__={
"organism": collection_name,
"features.type": SO_TERMS["polypeptide_structural_motif"]
}).count()
process_hmm = not (Protein.objects(
__raw__={
"organism": collection_name,
"features.type": SO_TERMS["polypeptide_domain"]
}).count())
_common_annotations(collection_name, tmp_dir, cpu, remove_tmp, process_pdb,
process_hmm)
def update_protein_with_dbxref(protein_gene, annotations, organism):
p = Protein.objects(gene=protein_gene,
organism=organism).no_cache().get()
for ann in annotations:
word = ann.split("||")[1]
if ann.lower().startswith("ec") or ann.lower().startswith("go"):
p.ontologies.append(word)
p.keywords.append(word)
for attr in """EcoGene
Ensembl_PRO
Ensembl_TRS
WormBase_PRO
WormBase_TRS
UniGene
GeneDB
EuPathDB""".split():
if ann.startswith(attr):
p.keywords.append(word)
p.alias.append(word)
for g in ["Gene_Name", "Gene_OrderedLocusName", "Gene_ORFName"]:
if ann.startswith(g):
p.keywords.append(word)
p.alias.append(word)
p.gene.append(word)
#if "Uncharacterized" in p.description:
# p.description =
p.ontologies = list(set([x.lower() for x in p.ontologies]))
p.keywords = list(set([x.lower() for x in p.keywords]))
p.alias = list(set(p.alias))
p.gene = [protein_gene] + [x for x in set(p.gene) if x != protein_gene]
return p
def clean_structures(self, organism):
proteins = list(
Protein.objects(organism=organism).no_cache().timeout(False))
for p in proteins:
experimentals = p.cristals()
models = p.models()
if experimentals and models:
for model in models:
for exp in experimentals:
hit = [
f for f in p.features
if f.identifier.startswith(exp.name)
][0]
model_query = model.templates[0].aln_query
model_range = set(
range(model_query.start, model_query.end))
exp_range = set(
range(hit.location.start, hit.location.end))
if ((len(model_range & exp_range) * 1.0 /
len(model_range)) > 0.8):
model.delete()
if len(models) > 1:
for i, j in combinations(range(len(models), 2)):
if ((models[i].templates[0].aln_query.start
== models[j].templates[0].aln_query.start)
and (models[i].templates[0].aln_query.end
== models[j].templates[0].aln_query.end)):
models[i].delete()
def common_annotations(collection_name,
tmp_dir,
cpu=1,
remove_tmp=False,
pfam_db="/data/databases/xfam/Pfam-A.hmm",
pdbs_path="/data/databases/pdb/pdb_seqres.txt"):
process_pdb = not Protein.objects(
__raw__={
"organism": collection_name,
"features.type": SO_TERMS["polypeptide_structural_motif"]
}).count()
process_hmm = not (Protein.objects(
__raw__={
"organism": collection_name,
"features.type": SO_TERMS["polypeptide_domain"]
}).count())
_common_annotations(collection_name, tmp_dir, cpu, remove_tmp, process_pdb,
process_hmm, pfam_db, pdbs_path)
def build_index(self, query={}, error_output="/tmp/index_struct.log"):
self.update_collection_params()
if not query:
proteins = Protein.objects(
organism=self.collection.name).no_cache().timeout(False)
prot_count = Protein.objects(organism=self.collection.name).count()
else:
query["organism"] = self.collection.name
proteins = Protein.objects(__raw__=query).no_cache().timeout(False)
prot_count = Protein.objects(__raw__=query).count()
with tqdm(proteins, total=prot_count) as pbar:
for protein in pbar:
pbar.set_description(protein.name)
try:
if not protein.search:
protein.search = ProteinDruggabilitySearch()
self.initDrugabilitySearch(protein.search)
cristals, models = self.get_protein_structures(protein)
if len(cristals + models) > 0:
self.annotate_protein_with_structures(
cristals, models, protein)
else:
protein.keywords = [
x for x in protein.keywords if x not in [
"poorly_druggable", "druggable",
"highly_druggable", "has_structure"
]
]
protein.keywords.append("non_druggable")
for x in StructuromeIndexer.search_params:
protein.search[x[0]] = None
protein.save()
except Exception as ex:
error_line = protein.name + "," + protein.organism + "," + str(
protein.id) + "," + str(ex)
_log.warn(error_line)
with open(error_output, "a") as h:
h.write(error_line)
def create_proteome(tmp_dir, collection_name):
protein_fasta = tmp_dir + "/proteins.fasta"
if not os.path.exists(protein_fasta) or (
not os.path.getsize(protein_fasta)):
with open(protein_fasta, "w") as h:
for p in Protein.objects(organism=collection_name).no_cache():
bpio.write(
SeqRecord(id=p.gene[0], description="", seq=Seq(p.seq)), h,
"fasta")
return protein_fasta
def properties_from_feature(self,
organism,
feature_type,
value_fn,
property_name=None,
url_fn=None):
proteins = Protein.objects(__raw__={
"organism": organism,
"features.type": feature_type
}).no_cache().timeout(False)
if property_name == "human_offtarget":
res = self.db.proteins.update(
{
"organism": organism,
"search.human_offtarget": {
"$exists": True
}
}, {"$unset": {
"search.human_offtarget": ""
}},
multi=True)
_log.info(res)
for p in proteins:
f = [x for x in p.features if x.type == feature_type]
if f:
f = f[0]
bp = BioProperty(
_type=feature_type,
property=property_name if property_name else feature_type,
value=value_fn(f),
)
if url_fn:
bp.url = url_fn(f)
props = [
x for x in p.properties
if not (x._type == feature_type and (
(not property_name) or (x.property == property_name)))
]
p.properties = props
p.properties.append(bp)
if property_name == "human_offtarget":
p.search.human_offtarget = 1 - f.aln.identity
p.save()
if property_name == "human_offtarget":
res = self.db.proteins.update(
{
"organism": organism,
"search.human_offtarget": {
"$exists": False
}
}, {"$set": {
"search.human_offtarget": 1
}},
multi=True)
_log.info(res)
def load_pdb_domains(organism,
blast_file,
feature_type="SO:0001079",
min_identity=0.9,
min_query_coverage=0.9,
min_hit_coverage=0.9):
queries = list(bpsio.parse(blast_file, 'blast-xml'))
features_added = 0
with tqdm(queries) as pbar:
for query in queries:
pbar.set_description(query.id)
pfam, dnstart, dnend = query.id.split("_")[-3:]
dnstart, dnend = int(dnstart), int(dnend)
gene = "_".join(query.id.split("_")[:-3])
proteins = Protein.objects(organism=organism,
gene=gene).no_cache().timeout(False)
change = False
for protein in proteins:
for hit in query:
hsp = hit[0]
dn = [
x for x in protein.domains()
if (abs(x.location.start - dnstart) < 10) and (
abs(x.location.end - dnend) < 10)
and x.identifier.split(".")[0] == pfam.split(".")[0]
]
if dn:
dn = dn[0]
ident_fn = lambda fident: "_".join(
fident.split("_")[0:1] + fident.split("_")[-2:])
pdb = [
x for x in protein.features
if x.type == feature_type
and ident_fn(x.identifier) == ident_fn(hit.id)
]
if not pdb:
posSet = set(
range(dn.location.start, dn.location.end))
dncover = 1.0 * len(
posSet & set(range(dnstart, dnend))) / (
dn.location.end - dn.location.start)
if (min_identity <= identity(hsp)) and (
dncover >= min_query_coverage):
hsp_feature = BioDocFactory.feature_from_hsp(
hsp, feature_type)
hsp_feature.location.start += dnstart
hsp_feature.location.end += dnstart
features_added = features_added + 1
change = True
protein.features.append(hsp_feature)
if change:
protein.save()
_log.info("Features added: " + str(features_added))
def update_genome_props(self):
if self.user == "demo":
user2 = ""
else:
user2 = self.user + "."
search_params = [("resistance", "Associated with resistance",
"variant-db", SeqColDruggabilityParamTypes.value,
["true", "false"], "true", "equal", "avg")]
search_params = search_params + [
(x.lower(), "Associated with " + x + " resistance", "variant-db",
SeqColDruggabilityParamTypes.value, ["true", "false"
], "true", "equal", "avg")
for x in Saureus.drugs
]
SeqCollection.objects(name=self.organism).update(
__raw__={
"$pull": {
"druggabilityParams": {
"target": "variant-db",
"uploader": self.user
}
}
})
collection = SeqCollection.objects(name=self.organism).get()
for name, description, target, _type, options, defaultValue, defaultOperation, defaultGroupOperation in search_params:
Protein.objects(organism=self.organism).update(
__raw__={"$set": {
"search." + user2 + name: False
}})
if not collection.has_druggability_param(name):
dp = SeqColDruggabilityParam(name=name,
description=description,
target=target,
type=_type,
uploader=self.user)
dp.options = options
dp.defaultValue = defaultValue
dp.defaultOperation = defaultOperation
dp.defaultGroupOperation = defaultGroupOperation
collection.druggabilityParams.append(dp)
collection.save()
def load_priam_hits(self, seq_collection_name, path_genomeEnzymes):
for line in open(path_genomeEnzymes):
arr_line = line.split("\t")
ec = "ec:" + arr_line[0]
prot_id = arr_line[1].split(" ")[0]
for prot in Protein.objects(organism=seq_collection_name, alias=prot_id):
prot.ontologies.append(ec)
prot.keywords.append(ec)
prot.save()
def load_from_emapper(self, organism, emapperv2_file):
from SNDG.Annotation.EMapper import EMapper
em = EMapper()
em.read_file(emapperv2_file)
for locus_tag, record in em.data.items():
prot = Protein.objects(organism=organism, gene=locus_tag).get()
for ec in record["EC"].split(","):
prot.ontologies.append("ec:" + ec)
for go in record["GOs"].split(","):
prot.ontologies.append(go.lower())
prot.save()
def load_hmm(organism,
hmm_file,
transform_query_regexp=None,
transform_hit_regexp=None):
assert os.path.exists(hmm_file)
for query in tqdm(bpsio.parse(hmm_file, 'hmmer3-text')):
for hit in query:
for hsp in hit:
gene = query.id
if transform_query_regexp:
gene = re.search(transform_query_regexp, query.id,
re.IGNORECASE).group(1)
hit_name = hit.id
if transform_query_regexp:
hit_name = re.search(transform_hit_regexp, hit_name,
re.IGNORECASE).group(1)
proteins = Protein.objects(
organism=organism, alias=gene).no_cache().timeout(False)
for protein in proteins:
dn = [
d for d in protein.domains()
if (d.identifier == hit_name) and (
d.location.start == hsp.query_start) and (
d.location.end == hsp.query_end)
]
if dn:
protein.features.remove(dn[0])
hsp_feature = Feature(
_id=ObjectId(),
location=Location(start=hsp.query_start,
end=hsp.query_end),
aln=SimpleAlignment(
evalue=hsp.evalue,
aln_query=AlnLine(name=hsp.query_id,
seq=str(hsp.aln[0].seq),
start=hsp.query_start,
end=hsp.query_end),
aln_hit=AlnLine(name=hsp.hit.id,
seq=str(hsp.aln[1].seq),
start=hsp.hit_start,
end=hsp.hit_end),
aln_cd=hsp.aln_annotation["CS"]
if "CS" in hsp.aln_annotation else "",
aln_pp=hsp.aln_annotation["PP"]
if "PP" in hsp.aln_annotation else ""),
identifier=hsp.hit.id,
type=SO_TERMS["polypeptide_domain"])
protein.features.append(hsp_feature)
protein.save()
def load_blast_features(organism,
blast_file,
feature_type,
min_identity=0,
min_query_coverage=0,
min_hit_coverage=0):
queries = list(bpsio.parse(blast_file, 'blast-xml'))
def check_overlap(features, new_feature, max_aa_overlap):
for f in features:
if (1.0 * len(new_feature & f) / len(f)) > 0.8:
return True
return False
features_added = 0
for query in tqdm(queries):
gene = query.id
proteins = Protein.objects(organism=organism,
gene=gene).no_cache().timeout(False)
change = False
for protein in proteins:
for hit in query:
hsp = hit[0]
if ((identity(hsp) >= min_identity)
and (coverage(query, hsp) >= min_query_coverage)
and (hit_coverage(hit, hsp) >= min_hit_coverage)):
hsp_feature = BioDocFactory.feature_from_hsp(
hsp, feature_type)
features_added = features_added + 1
change = True
protein.features.append(hsp_feature)
elif (identity(hsp) >= min_identity) and (hit_coverage(
hit, hsp) >= min_hit_coverage):
for dn in protein.domains():
posSet = set(range(dn.location.start, dn.location.end))
dncover = 1.0 * len(posSet & set(
range(hsp.query_start, hsp.query_end))) / (
dn.location.end - dn.location.start)
if dncover >= min_query_coverage:
hsp_feature = BioDocFactory.feature_from_hsp(
hsp, feature_type)
features_added = features_added + 1
change = True
protein.features.append(hsp_feature)
if change:
protein.save()
_log.info("Features added: " + str(features_added))
def main(argv=None): # IGNORE:C0111
program_version = "v%s" % __version__
program_build_date = str(__updated__)
program_version_message = '%%(prog)s %s (%s)' % (program_version, program_build_date)
program_shortdesc = __import__('__main__').__doc__.split("\n")[1]
program_license = '''%s
Created by user_name on %s.
Copyright 2015 BIA. All rights reserved.
Licensed under the Apache License 2.0
http://www.apache.org/licenses/LICENSE-2.0
Distributed on an "AS IS" basis without warranties
or conditions of any kind, either express or implied.
USAGE
''' % (program_shortdesc, str(__date__))
parser = ArgumentParser(description=program_license, formatter_class=RawDescriptionHelpFormatter)
parser.add_argument("-v", "--verbose", dest="verbose", action="count",
help="set verbosity level [default: %(default)s]")
parser.add_argument("-n", "--name", required=True)
parser.add_argument("-dir", "--structs_dir", required=True)
parser.add_argument("-db_structure", "--db_structure",help="Mongo structure db", default='pdb')
parser.add_argument("-db_genome", "--db_genome",help="Mongo proteins db", default='xomeq')
parser.add_argument("-host", "--db_host", default='127.0.0.1')
parser.add_argument('-V', '--version', action='version', version=program_version_message)
args = parser.parse_args()
BioMongoDB(args.db_genome)
db = pymongo.MongoClient(args.db_host)[args.db_structure]
sa = StructureAnotator(args.structs_dir + "/")
total = sa.total(db, args.name, {})
with tqdm(sa.iterator(db, args.name, {}), total=total) as pbar:
for model in pbar:
pbar.set_description(model.name)
template = model.templates[0]
try:
protein = Protein.objects(organism=args.name, alias=template.aln_query.name).get()
except DoesNotExist:
_log.warn(template.aln_query.name + " does not exists")
sa.annotate_model(model, protein.domains())
model.save()
def load_from_interpro(self, organism, interprot_gff):
for l in tqdm(open(interprot_gff)):
if l.startswith(">"):
break
if l.startswith("##"):
continue
l = l.replace("EC=", "EC ")
locus_tag, source, feature, start, end, score, strand, frame = l.split(
"\t")[:8]
attributes = " ".join(l.split("\t")[8:])
if feature == "polypeptide":
continue
start, end = int(start), int(end)
if "signature_desc=" in attributes:
repl = attributes.split("signature_desc=")[1].split(
";Name=")[0]
attributes = attributes.replace(
repl,
repl.replace("=", "%3D").replace(";", "%3B"))
attributes = {
x.split("=")[0]: x.split("=")[1]
for x in attributes.split(";")
}
# [seq,source,feature,start,end,score,strand,frame,attributes ])
feature = Feature(_id=ObjectId(),
location=Location(start=start, end=end),
identifier=attributes["Name"],
type=source)
prot = Protein.objects(organism=organism, gene=locus_tag).get()
if "signature_desc" in attributes:
feature.qualifiers = {
"description": attributes["signature_desc"]
}
if "Ontology_term" in attributes:
for ont in attributes["Ontology_term"].split(","):
ont = ont.replace('"', "").strip()
prot.ontologies.append(ont.lower())
if "Dbxref" in attributes:
for ont in attributes["Dbxref"].split(","):
ont = ont.replace('"', "").strip()
prot.ontologies.append(ont.lower())
prot.features.append(feature)
prot.save()
def props_from_dbxref(self, name):
i = 0
for p in Protein.objects(organism=name).no_cache().timeout(False):
if p.dbxrefs:
i += 1
prop = BioProperty(_type="dbxref",
property="links",
value=p.dbxrefs)
p.alias += [
x.split(":")[1] for x in p.dbxrefs
if x.lower().startswith("uniprot")
]
p.alias = list(set(p.alias))
p.properties.append(prop)
p.save()
def organism_iterator(self, organism, seq_map=None):
for dbcontig in Contig.objects(organism=organism).no_cache():
if seq_map:
seq = str(seq_map[dbcontig.name].seq)
else:
seq = dbcontig.seq
contig = SeqRecord(id=dbcontig.name, seq=Seq(seq))
for dbfeature in dbcontig.features:
qualifiers = {"locus_tag": [dbfeature.locus_tag]}
p = list(
Protein.objects(organism=organism,
gene=dbfeature.identifier))
if p:
p = p[0]
qualifiers["description"] = [p.description]
qualifiers["gene_symbol"] = p.gene
qualifiers["Note"] = [p.description]
ecs = [
x.upper() for x in p.ontologies if x.startswith("ec:")
]
gos = [
x.upper() for x in p.ontologies if x.startswith("go:")
]
if ecs:
qualifiers["EC"] = ecs
if gos:
qualifiers["GO"] = gos
feature = SeqFeature(id=dbfeature.identifier,
type=dbfeature.type,
qualifiers=qualifiers,
location=FeatureLocation(
start=dbfeature.location.start,
end=dbfeature.location.end,
strand=dbfeature.location.strand))
contig.features.append(feature)
yield contig
def correct_chokes(self, name):
metabolites_in = defaultdict(list)
metabolites_out = defaultdict(list)
for p in self.db.proteins.find({
"organism": name,
"reactions.0": {
"$exists": True
}
}):
for r in p["reactions"]:
for m in r["products"]:
metabolites_out[m["name"]].append(r["name"])
for m in r["substrates"]:
metabolites_in[m["name"]].append(r["name"])
for m, r in metabolites_in.items():
if (len(set(r)) > 1):
# or (self.db.proteins.count({"organism":name,"reactions.name": r[0]}) > 1)):
del metabolites_in[m]
for m, r in metabolites_out.items():
if (len(set(r)) > 1):
# or (self.db.proteins.count({"organism":name,"reactions.name": r[0]}) > 1)):
del metabolites_out[m]
choke_reactions_in = []
for rs in metabolites_in.values():
choke_reactions_in += rs
choke_reactions_out = []
for rs in metabolites_out.values():
choke_reactions_out += rs
reaction_metabolites = defaultdict(lambda: [])
for m, rs in metabolites_in.items():
for r in rs:
reaction_metabolites[r].append(m)
for m, rs in metabolites_out.items():
for r in rs:
reaction_metabolites[r].append(m)
for p in Protein.objects(
organism=name,
reactions__0__exists=True).no_cache().timeout(False):
cout = bool(
[r.name for r in p.reactions if r.name in choke_reactions_out])
cin = bool(
[r.name for r in p.reactions if r.name in choke_reactions_in])
p.search.chokepoint = cout | cin
if p.search.chokepoint:
p.search.chokepoint_type = "double" if (cout & cin) else (
"production" if cout else "consuming")
prop = [x for x in p.properties if x.property == "chokepoint"]
if prop:
prop = prop[0]
prop.metabolites = []
for x in p.reactions:
if x.name in reaction_metabolites:
prop.metabolites += reaction_metabolites[x.name]
else:
prop = BioProperty(_type="pathways",
property="chokepoint",
metabolites=[],
type=p.search.chokepoint_type)
for x in p.reactions:
if x.name in reaction_metabolites:
prop.metabolites += reaction_metabolites[x.name]
p.properties.append(prop)
else:
del p.search.chokepoint_type
p.properties = [
x for x in p.properties if x.property != "chokepoint"
]
p.save()
def create_protein(cls, seqrecord, feature, exons=[]):
alias = cls.alias(feature)
locus_tag = feature.qualifiers["locus_tag"][0]
protein_name = locus_tag
seq = str(seqrecord.seq)
# if "translation" in feature.qualifiers:
# seq = feature.qualifiers["translation"][0]
# else:
# if exons:
# seq = str(reduce(Seq.__add__, [exon.extract(seqrecord.seq) for exon in exons]).translate())
# else:
# seq = str(feature.extract(Seq(str(seqrecord.seq))).translate())
if "gene_symbol_source" in feature.qualifiers:
try:
int(feature.qualifiers["gene_symbol_source"][0])
protein_name = ""
except:
protein_name = feature.qualifiers['gene_symbol'][0]
# elif "product" in mrna_feature.qualifiers:
# protein_name = mrna_feature.qualifiers['product'][0]
p = Protein(seq=seq, name=protein_name)
if "description" in feature.qualifiers:
protein_description = feature.qualifiers['description'][0]
elif "Note" in feature.qualifiers:
protein_description = feature.qualifiers['Note'][0]
elif "product" in feature.qualifiers:
protein_description = feature.qualifiers['product'][0]
else:
protein_description = ""
bp = BioProperty(
_type="annotation",
description="homolog proteins and sources used for the annotation")
if "protein_id" in feature.qualifiers:
bp.ncbi_protein_id = feature.qualifiers["protein_id"][0]
if "db_xref" in feature.qualifiers:
bp.ncbi_db_xref = feature.qualifiers["db_xref"][0]
if "Dbxref" in feature.qualifiers:
bp.ncbi_db_xref = feature.qualifiers["Dbxref"][0]
if "top_cog_hit" in feature.qualifiers:
bp.cog = feature.qualifiers["top_cog_hit"][0]
if "gene_symbol_source" in feature.qualifiers:
bp.source = feature.qualifiers["gene_symbol_source"][0]
if "gene_product_name_source" in feature.qualifiers:
bp.source = feature.qualifiers["gene_product_name_source"][0]
ecs = []
if "EC" in feature.qualifiers:
ecs = ecs + [
x.lower() if x.lower().startswith("ec") else "ec:" + x.lower()
for x in feature.qualifiers["EC"] if "." in x
]
if "EC_number" in feature.qualifiers:
ecs = ecs + [
x.lower() if x.lower().startswith("ec") else "ec:" + x.lower()
for x in feature.qualifiers["EC_number"] if "." in x
]
if "Dbxref" in feature.qualifiers:
ecs = ecs + [
x.lower() if x.lower().startswith("ec") else "ec:" + x.lower()
for x in feature.qualifiers["Dbxref"] if "." in x
]
gos = []
if "db_xref" in feature.qualifiers:
gos = gos + [
x.lower()
for x in feature.qualifiers["db_xref"] if "GO:" in x and
(x not in ["GO:0008150", "GO:0003674", "GO:0005575"])
]
if "GO" in feature.qualifiers:
gos = gos + [
x.lower() for x in feature.qualifiers["GO"] if "GO:" in x and
(x not in ["GO:0008150", "GO:0003674", "GO:0005575"])
]
if "Ontology_term" in feature.qualifiers:
gos = [
x.lower() for x in feature.qualifiers["Ontology_term"]
if "GO:" in x and (
x not in ["GO:0008150", "GO:0003674", "GO:0005575"])
]
ontologies = list(set(ecs + gos))
p.gene = list([locus_tag, protein_name
]) if locus_tag != protein_name else [locus_tag]
p.name = protein_name
p.description = protein_description
p.ontologies = ontologies
p.properties = [bp]
p.alias = alias
return p
model_filename = model_file.split("/")[-1]
model_name = model_filename.split(".pdb")[0]
hunt_pockets(model_file)
build_assessments(os.path.dirname(model_file),
os.path.basename(model_filename))
seq_name = "_".join(model_name.split("_")[:-4])
org_model_name = model_name
template_name = "_".join(model_name.split("_")[-4:])
if ModeledStructure.objects(organism=organism,
name=model_name).count():
continue
prot = list(Protein.objects(organism=organism, gene=seq_name))
if len(prot) == 0:
_log.warn("Not found: " + seq_name)
continue
aln = [
hit[0] for hit in list(
bpsio.read(
basepath + "/" + seq_name +
"/profile_search.xml", "blast-xml"))
if hit.id == template_name
][0]
with open(model_file + ".json") as h:
assessments = json.load(h)
pockets = []
def from_TriTrypDB(name, gff, fasta, tax, tmp_dir=None):
genome = {x.id: x for x in sp(fasta)}
from BCBio import GFF
import re
annotation = list(GFF.parse(gff, base_dict=genome))
contig = annotation[0]
seqCol = BioDocFactory.create_genome(name, contig, tax, Tax)
seqCol.save()
if not tmp_dir:
tmp_dir = "/tmp/" + name + "/"
mkdir(tmp_dir)
gene_ids = {}
with tqdm(annotation) as pbar:
for contig in pbar:
pbar.set_description(contig.id)
if len(contig.seq) > 15000000:
contig.seq = ""
contigDoc, gene_ids2 = BioDocFactory.create_contig(
contig,
seqCol,
type_map={
"rRNA": "rRNA",
"ncRNA": "ncRNA",
NCBI.f_mRNA: "gene",
"exon": "exon",
"gene": "gene",
NCBI.f_CDS: NCBI.f_CDS,
"rRNA": "rRNA",
"tRNA": "tRNA",
"tmRNA": "tmRNA",
"snoRNA": "snoRNA",
"three_prime_UTR": "three_prime_UTR",
"five_prime_UTR": "five_prime_UTR"
})
gene_ids.update(gene_ids2)
contigDoc.save()
prots = []
with tqdm(tritryp_protein_iter(annotation)) as pbar:
for (protein, cds_f) in pbar:
protDoc = Protein(seq=str(protein.seq), name=protein.id)
if "description" in cds_f.qualifiers:
protein_description = cds_f.qualifiers['description'][0]
elif "Note" in cds_f.qualifiers:
protein_description = cds_f.qualifiers['Note'][0]
elif "product" in cds_f.qualifiers:
protein_description = cds_f.qualifiers['product'][0]
else:
protein_description = ""
protDoc.description = protein_description
gos = []
if "Ontology_term" in cds_f.qualifiers:
gos = [
x.lower() for x in cds_f.qualifiers["Ontology_term"]
if "GO:" in x and (
x not in ["GO:0008150", "GO:0003674", "GO:0005575"])
]
note = cds_f.qualifiers["Note"][0].split(
" ")[0] if "Note" in cds_f.qualifiers else ""
ecs = ["ec:" + note] if re.match(
'^[0-9]+\.[0-9\-]+\.[0-9\-]+\.[0-9\-]$', note) else []
ontologies = list(set(ecs + gos))
protDoc.gene = [protein.id]
protDoc.ontologies = ontologies
protDoc.alias = [protein.id]
if len(protDoc.seq) > 30000:
raise Exception("No existen proteinas tan largas...")
protDoc.gene_id = gene_ids[protein.id]
protDoc.organism = name
protDoc.auth = str(BioMongoDB.demo_id)
protDoc.seq_collection_id = seqCol
prots.append(protDoc)
if pbar.n and ((pbar.n % 1000) == 0):
Protein.objects.insert(prots)
prots = []
if prots:
Protein.objects.insert(prots)
_common_annotations(name, tmp_dir)
# pockets_json = model_file + ".pocket.json"
# if os.path.exists(pockets_json):
# rss = StructureAnotator.pocket_residue_set(pockets_json, model.get_atoms())
# strdoc.pockets = rss
# strdoc.save()
# except Exception as ex:
# _log.error(ex)
#
# print ("OK!")
sa = StructureAnotator(basepath,
struct_path=lambda wd, modeldoc: glob("/".join([
wd, modeldoc.templates[0].aln_query.name, modeldoc.
templates[0].aln_query.name, modeldoc.name, "*.pdb"
]))[0])
total = sa.total(db, organism, {})
with tqdm(sa.iterator(db, organism, {}), total=total) as pbar:
for model in pbar:
pbar.set_description(model.name)
template = model.templates[0]
try:
protein = Protein.objects(organism=organism,
alias=template.aln_query.name).get()
except:
print template.aln_query.name + " does not exists"
continue
sa.annotate_model(model, protein.domains())
model.save()
def annotate_variants_with_prots(organism_name, dbs, drugs, force=False):
"""
drugs: list of strings, example TBDream.drugs or Saureus.drugs
"""
for idx, p in enumerate(
Protein.objects(
__raw__={
"organism": organism_name,
"features.qualifiers.strain": {
"$exists": 1
}
}).no_cache()):
print idx
pvariants = list(
Variant.objects(organism=organism_name, gene__in=p.gene))
for vd in pvariants:
if ((vd.search == None) or force):
psearch = p.search
del psearch.structures
vd.search = psearch
vd.ontologies = p.ontologies
for r in p.reactions:
for pw in r.pathways:
vd.ontologies.append(pw)
vd.ontologies = list(set(vd.ontologies))
for drug in drugs:
vd.search[drug] = False
vd.search["resistance"] = False
for db in dbs:
for sample_allele in vd.sample_alleles:
aa_pos = sample_allele.aa_pos
feature = [
f for f in p.features
if (f.type == db) and (f.location.start == aa_pos)
]
if feature:
feature = feature[0]
if feature._data["qualifiers"]:
vd.search["resistance"] = True
feature = [
f for f in p.features
if (f.type == db) and (f.location.start == aa_pos)
and ("mut" in f._data["qualifiers"]) and (
(f._data["qualifiers"]["mut"] ==
sample_allele.aa_alt) or (
(f._data["qualifiers"]["change"].lower()
== "frameshift" and sample_allele.
variant_type == "frameshift_variant")))
]
if feature:
sample_allele.feature = feature[0]
if sample_allele.feature._data["qualifiers"]:
vd.search[sample_allele.feature._data[
"qualifiers"]["drug"]] = True #
vd.save()
def annotate_variants(organism_name, strain_name, database, parse_change):
"""
parse_change: function that transforms dbvar.qualifiers["change"] into aa_ref, aa_alt
"""
collection = SeqCollection.objects(name=organism_name).get()
prop = strain_name + "_" + database
Protein.objects(organism=organism_name).update(
__raw__={"$set": {
"search." + prop: False
}})
if not collection.has_druggability_param(prop):
dp = SeqColDruggabilityParam(name=prop,
description="Variant in strain " +
strain_name + " is reported in " +
database,
target="variant-strain",
type=SeqColDruggabilityParamTypes.value,
uploader="demo")
dp.options = ["true", "false"]
dp.defaultValue = "true"
dp.defaultOperation = "equal"
dp.defaultGroupOperation = "avg"
collection.druggabilityParams.append(dp)
prop = strain_name + "_" + database + "_pos"
Protein.objects(organism=organism_name).update(
__raw__={"$set": {
"search." + prop: False
}})
if not collection.has_druggability_param(prop):
dp = SeqColDruggabilityParam(
name=prop,
description="The position of the variant the strain " +
strain_name + " is reported in " + database,
target="variant-strain",
type=SeqColDruggabilityParamTypes.value,
uploader="demo")
dp.options = ["true", "false"]
dp.defaultValue = "true"
dp.defaultOperation = "equal"
dp.defaultGroupOperation = "avg"
collection.druggabilityParams.append(dp)
collection.save()
for p in Protein.objects(__raw__={
"organism": organism_name,
"features.qualifiers.strain": strain_name
}).no_cache():
dbvars = [f for f in p.features if f.type == database]
dirty = False
if dbvars:
strainvars = [
f for f in p.features if (f.type == "strain_variant") and (
f._data["qualifiers"]["strain"] == strain_name)
]
for dbvar in dbvars:
dirty = True
for strainvar in strainvars:
strainvar._data["qualifiers"]["ref_pos"] = False
if dbvar.location.start == strainvar.location.start:
p.search[strain_name + "_" + database + "_pos"] = True
strainvar._data["qualifiers"]["ref_pos"] = dbvar._id
try:
dref, dalt = parse_change(
dbvar._data["qualifiers"]["change"])
sref, salt = strainvar._data["qualifiers"][
"change"].strip().split("/")
sref = sref.strip()
salt = salt.strip()
if (dref == sref) and (dalt == salt):
p.search[strain_name + "_" + database] = True
strainvar._data["qualifiers"][
"ref"] = dbvar._id
except Exception as ex:
_log.warn(ex)
if (("frameshift"
in dbvar._data["qualifiers"]["change"].lower())
and
("frameshift" in strainvar._data["qualifiers"]
["change"].lower())):
p.search[strain_name + "_" + database] = True
if dirty:
p.save()
def load_metadata(self, organism_name, datafile, uploader=demo):
import pandas as pd
from tqdm import tqdm
seqCollection = list(SeqCollection.objects(name=organism_name))
seqCollection = seqCollection[0]
errors = []
upload = DataUpload(uploader=uploader, errors=errors)
df = pd.read_table(datafile, comment="#", index_col=False)
headerProperties = [
c for c in df.columns if c != BioMongoDB.GENE_FIELD_IMPORT
]
prots = Protein.objects(organism=organism_name)
for hp in headerProperties:
prots.update(
__raw__={
"$pull": {
"properties": {
"property": hp,
"_type": uploader
}
},
"$unset": {
"search." + hp: ""
}
})
upload.properties = headerProperties
numericFields = []
for k, v in dict(df.dtypes).items():
if v not in [np.float64, np.int64]:
df[k] = df[k].astype('category')
else:
numericFields.append(k)
assert BioMongoDB.GENE_FIELD_IMPORT in df.columns
for linenum, fields in tqdm(df.iterrows()):
gene = fields[BioMongoDB.GENE_FIELD_IMPORT]
if not gene:
text = str(linenum) + " gene field is empty"
errors.append(text)
continue
count = Protein.objects(organism=organism_name,
alias=gene).count()
if not count:
text = str(
linenum
) + " " + gene + " does not exists in " + organism_name
print(text)
errors.append(text)
continue
prots = Protein.objects(organism=organism_name, alias=gene)
for propertyName in headerProperties:
prop = {"_type": uploader, "value": fields[propertyName]}
prop["property"] = propertyName
prots.update(
__raw__={
"$push": {
"properties": prop
},
"$set": {
"search." + propertyName: fields[propertyName]
}
})
for p in headerProperties:
dpType = "number" if p in numericFields else "value"
options = [] if p in numericFields else list(set(df[p]))
currentDp = seqCollection.druggabilityParam(p, uploader)
if currentDp:
currentDp = currentDp[0]
currentDp.options = options
currentDp.type = dpType
else:
dp = SeqColDruggabilityParam(type=dpType,
name=p,
options=options,
uploader=uploader,
target="protein")
seqCollection.druggabilityParams.append(dp)
seqCollection.uploads.append(upload)
seqCollection.save()
def protein_fasta(outfile_path, organism):
with open(outfile_path, "w") as h:
for p in Protein.objects(organism=organism).no_cache():
r = SeqRecord(id=p.gene[0], description="", seq=Seq(p.seq))
bpio.write(r, h, "fasta")