예제 #1
0
def get_variants_from_variant_tuples(project, variant_tuples, user=None):
    datastore = get_datastore(project)
    population_slugs = project.get_reference_population_slugs()

    variant_tuples_by_family_id = {}
    for xpos, ref, alt, family_id in variant_tuples:
        if family_id not in variant_tuples_by_family_id:
            variant_tuples_by_family_id[family_id] = []
        variant_tuples_by_family_id[family_id].append((xpos, ref, alt))

    variants = []
    for family_id, variant_tuples in variant_tuples_by_family_id.items():
        variants_for_family = datastore.get_multiple_variants(
            project.project_id, family_id, variant_tuples, user=user)
        for (xpos, ref, alt), variant in zip(variant_tuples,
                                             variants_for_family):
            if not variant:
                variant = Variant(xpos, ref, alt)
                get_annotator().annotate_variant(variant, population_slugs)
                variant.set_extra('created_variant', True)

            variant.set_extra('family_id', family_id)
            variant.set_extra('project_id', project.project_id)
            variants.append(variant)

    return variants
예제 #2
0
파일: lookups.py 프로젝트: batsal/xbrowse
def get_variants_from_note_tuples(project, note_tuples):
    variants = []
    for note_t in note_tuples:
        variant = get_datastore(project.project_id).get_single_variant(
            project.project_id, note_t[3], note_t[0], note_t[1], note_t[2]
        )
        if not variant:
            variant = Variant(note_t[0], note_t[1], note_t[2])
            get_annotator().annotate_variant(variant, project.get_reference_population_slugs())
            # variant.annotation = get_annotator().get_variant(note_t[0], note_t[1], note_t[2])
        variant.set_extra("family_id", note_t[3])
        variant.set_extra("project_id", project.project_id)
        variants.append(variant)
    return variants
예제 #3
0
    def get_variants(self,
                     project_id,
                     family_id,
                     genotype_filter=None,
                     variant_filter=None,
                     quality_filter=None,
                     indivs_to_consider=None,
                     user=None):
        db_query = self._make_db_query(genotype_filter, variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            logger.error(
                "Error: mongodb collection not found for project %s family %s "
                % (project_id, family_id))
            return
        for i, variant_dict in enumerate(
                collection.find({
                    '$and': [{
                        k: v
                    } for k, v in db_query.items()]
                }).sort('xpos').limit(settings.VARIANT_QUERY_RESULTS_LIMIT +
                                      5)):
            if i >= settings.VARIANT_QUERY_RESULTS_LIMIT:
                raise Exception(
                    "ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again."
                    % settings.VARIANT_QUERY_RESULTS_LIMIT)

            variant = Variant.fromJSON(variant_dict)
            variant.set_extra('project_id', project_id)
            variant.set_extra('family_id', family_id)
            self.add_annotations_to_variants([variant], project_id)

            if passes_variant_filter(variant, variant_filter)[0]:
                yield variant
예제 #4
0
파일: views.py 프로젝트: rpete/seqr
def mendelian_variant_search_spec(request):

    project, family = get_project_and_family_for_user(request.user, request.GET)

    search_hash = request.GET.get('search_hash')
    search_spec_dict, variants = cache_utils.get_cached_results(project.project_id, search_hash)
    search_spec = MendelianVariantSearchSpec.fromJSON(search_spec_dict)
    if variants is None:
        variants = api_utils.calculate_mendelian_variant_search(search_spec, family.xfamily())
    else:
        variants = [Variant.fromJSON(v) for v in variants]
    add_extra_info_to_variants_family(get_reference(), family, variants)
    return_type = request.GET.get('return_type')
    if return_type == 'json' or not return_type:
        return JSONResponse({
            'is_error': False,
            'variants': [v.toJSON() for v in variants],
            'search_spec': search_spec_dict,
        })
    elif request.GET.get('return_type') == 'csv':
        response = HttpResponse(content_type='text/csv')
        response['Content-Disposition'] = 'attachment; filename="results_{}.csv"'.format(search_hash)
        writer = csv.writer(response)
        indiv_ids = family.indiv_ids_with_variant_data()
        headers = xbrowse_displays.get_variant_display_headers(get_mall(project.project_id), project, indiv_ids)
        writer.writerow(headers)
        for variant in variants:
            fields = xbrowse_displays.get_display_fields_for_variant(get_mall(project.project_id), project, variant, indiv_ids)
            writer.writerow(fields)
        return response
예제 #5
0
    def get_variants_in_gene(self,
                             project_id,
                             family_id,
                             gene_id,
                             genotype_filter=None,
                             variant_filter=None):

        if variant_filter is None:
            modified_variant_filter = VariantFilter()
        else:
            modified_variant_filter = copy.deepcopy(variant_filter)
        modified_variant_filter.add_gene(gene_id)

        db_query = self._make_db_query(genotype_filter,
                                       modified_variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            return

        # we have to collect list in memory here because mongo can't sort on xpos,
        # as result size can get too big.
        # need to find a better way to do this.
        variants = []
        for variant_dict in collection.find(db_query).hint([
            ('db_gene_ids', pymongo.ASCENDING), ('xpos', pymongo.ASCENDING)
        ]):
            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            if passes_variant_filter(variant, modified_variant_filter):
                variants.append(variant)
        variants = sorted(variants, key=lambda v: v.unique_tuple())
        for v in variants:
            yield v
예제 #6
0
    def get_variants_in_range(self, project_id, family_id, xpos_start,
                              xpos_end):
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            raise ValueError("Family not found: " + str(family_id))
        for i, variant_dict in enumerate(
                collection.find({
                    '$and': [{
                        'xpos': {
                            '$gte': xpos_start
                        }
                    }, {
                        'xpos': {
                            '$lte': xpos_end
                        }
                    }]
                }).limit(MONGO_QUERY_RESULTS_LIMIT + 5)):
            if i > MONGO_QUERY_RESULTS_LIMIT:
                raise Exception(
                    "ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again."
                    % MONGO_QUERY_RESULTS_LIMIT)

            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            yield variant
예제 #7
0
    def get_project_variants_in_gene(self,
                                     project_id,
                                     gene_id,
                                     variant_filter=None):

        if variant_filter is None:
            modified_variant_filter = VariantFilter()
        else:
            modified_variant_filter = copy.deepcopy(variant_filter)
        modified_variant_filter.add_gene(gene_id)

        db_query = self._make_db_query(None, modified_variant_filter)
        sys.stderr.write("Project Gene Search: " + str(project_id) +
                         " all variants query: " + str(db_query))
        collection = self._get_project_collection(project_id)
        # we have to collect list in memory here because mongo can't sort on xpos,
        # as result size can get too big.
        # need to find a better way to do this.
        variants = []
        for variant_dict in collection.find(db_query).hint([
            ('db_gene_ids', pymongo.ASCENDING), ('xpos', pymongo.ASCENDING)
        ]):
            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            if passes_variant_filter(variant, modified_variant_filter):
                variants.append(variant)
        variants = sorted(variants, key=lambda v: v.unique_tuple())
        return variants
예제 #8
0
파일: views.py 프로젝트: frichter/seqr
def mendelian_variant_search_spec(request):

    project, family = get_project_and_family_for_user(request.user, request.GET)

    # TODO: use form

    search_hash = request.GET.get('search_hash')
    search_spec_dict, variants = cache_utils.get_cached_results(project.project_id, search_hash)
    search_spec = MendelianVariantSearchSpec.fromJSON(search_spec_dict)
    if variants is None:
        variants = api_utils.calculate_mendelian_variant_search(search_spec, family.xfamily())
    else:
        variants = [Variant.fromJSON(v) for v in variants]
    add_extra_info_to_variants_family(get_reference(), family, variants)
    return_type = request.GET.get('return_type')
    if return_type == 'json' or not return_type:
        return JSONResponse({
            'is_error': False,
            'variants': [v.toJSON() for v in variants],
            'search_spec': search_spec_dict,
        })
    elif request.GET.get('return_type') == 'csv':
        response = HttpResponse(content_type='text/csv')
        response['Content-Disposition'] = 'attachment; filename="results_{}.csv"'.format(search_hash)
        writer = csv.writer(response)
        indiv_ids = family.indiv_ids_with_variant_data()
        headers = xbrowse_displays.get_variant_display_headers(get_mall(project.project_id), project, indiv_ids)
        writer.writerow(headers)
        for variant in variants:
            fields = xbrowse_displays.get_display_fields_for_variant(get_mall(project.project_id), project, variant, indiv_ids)
            writer.writerow(fields)
        return response
예제 #9
0
    def get_variants(self,
                     project_id,
                     family_id,
                     genotype_filter=None,
                     variant_filter=None):

        db_query = self._make_db_query(genotype_filter, variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            print(
                "Error: mongodb collection not found for project %s family %s "
                % (project_id, family_id))
            return
        for i, variant_dict in enumerate(
                collection.find(db_query).sort('xpos').limit(
                    MONGO_QUERY_RESULTS_LIMIT + 5)):
            if i >= MONGO_QUERY_RESULTS_LIMIT:
                raise Exception(
                    "ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again."
                    % MONGO_QUERY_RESULTS_LIMIT)

            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            if passes_variant_filter(variant, variant_filter)[0]:
                yield variant
예제 #10
0
 def get_variants_in_range(self, project_id, family_id, xpos_start, xpos_end):
     collection = self._get_family_collection(project_id, family_id)
     if not collection:
         raise ValueError("Family not found: " + str(family_id))
     for variant_dict in collection.find({'$and': [{'xpos': {'$gte': xpos_start}}, {'xpos': {'$lte': xpos_end}}]}):
         variant = Variant.fromJSON(variant_dict)
         self.add_annotations_to_variant(variant, project_id)
         yield variant
예제 #11
0
    def get_single_variant(self, project_id, family_id, xpos, ref, alt):

        collection = self._get_family_collection(project_id, family_id)
        variant = collection.find_one({'xpos': xpos, 'ref': ref, 'alt': alt})
        if variant:
            return Variant.fromJSON(variant)
        else:
            return None
예제 #12
0
    def get_variants(self, project_id, family_id, genotype_filter=None, variant_filter=None):

        db_query = _make_db_query(genotype_filter, variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        for variant_dict in collection.find(db_query).sort('xpos'):
            variant = Variant.fromJSON(variant_dict)
            if passes_variant_filter(variant, variant_filter)[0]:
                yield variant
예제 #13
0
def get_variants_from_variant_tuples(project, variant_tuples, user=None):
    datastore = get_datastore(project)
    population_slugs = project.get_reference_population_slugs()

    variant_tuples_by_family_id = {}
    for xpos, ref, alt, family_id in variant_tuples:
        if family_id not in variant_tuples_by_family_id:
            variant_tuples_by_family_id[family_id] = []
        variant_tuples_by_family_id[family_id].append((xpos, ref, alt))

    variants = []
    for family_id, variant_tuples in variant_tuples_by_family_id.items():
        variants_for_family = datastore.get_multiple_variants(
            project.project_id,
            family_id,
            variant_tuples,
            user=user
        )
        for (xpos, ref, alt), variant in zip(variant_tuples, variants_for_family):
            if not variant:
                variant = Variant(xpos, ref, alt)
                get_annotator().annotate_variant(variant, population_slugs)
                variant.set_extra('created_variant', True)

            variant.set_extra('family_id', family_id)
            variant.set_extra('project_id', project.project_id)
            variants.append(variant)

    return variants
예제 #14
0
def get_variants_from_variant_tuples(project, variant_tuples):
    variants = []
    for t in variant_tuples:
        variant = get_datastore(project.project_id).get_single_variant(
            project.project_id,
            t[3],
            t[0],
            t[1],
            t[2]
        )
        if not variant:
            variant = Variant(t[0], t[1], t[2])
            get_annotator().annotate_variant(variant, project.get_reference_population_slugs())
            
        variant.set_extra('family_id', t[3])
        variant.set_extra('project_id', project.project_id)
        variants.append(variant)
    return variants
예제 #15
0
    def get_variants_cohort(self, project_id, cohort_id, variant_filter=None):

        db_query = self._make_db_query(None, variant_filter)
        collection = self._get_family_collection(project_id, cohort_id)
        for i, variant in enumerate(collection.find(db_query).sort('xpos').limit(settings.VARIANT_QUERY_RESULTS_LIMIT+5)):
            if i > settings.VARIANT_QUERY_RESULTS_LIMIT:
                raise Exception("ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again." % settings.VARIANT_QUERY_RESULTS_LIMIT)

            yield Variant.fromJSON(variant)
예제 #16
0
    def get_variants_cohort(self, project_id, cohort_id, variant_filter=None):

        db_query = self._make_db_query(None, variant_filter)
        collection = self._get_family_collection(project_id, cohort_id)
        for i, variant in enumerate(collection.find(db_query).sort('xpos').limit(MONGO_QUERY_RESULTS_LIMIT+5)):
            if i > MONGO_QUERY_RESULTS_LIMIT:
                raise Exception("ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again." % MONGO_QUERY_RESULTS_LIMIT)

            yield Variant.fromJSON(variant)
예제 #17
0
    def get_variants_in_range(self, project_id, family_id, xpos_start, xpos_end):
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            raise ValueError("Family not found: " + str(family_id))
        for i, variant_dict in enumerate(collection.find({'$and': [{'xpos': {'$gte': xpos_start}}, {'xpos': {'$lte': xpos_end}}]}).limit(MONGO_QUERY_RESULTS_LIMIT+5)):
            if i > MONGO_QUERY_RESULTS_LIMIT:
                raise Exception("ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again." % MONGO_QUERY_RESULTS_LIMIT)

            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            yield variant
예제 #18
0
 def get_single_variant(self, project_id, family_id, xpos, ref, alt, user=None):
     collection = self._get_family_collection(project_id, family_id)
     if not collection:
         return None
     variant_dict = collection.find_one({'xpos': xpos, 'ref': ref, 'alt': alt})
     if variant_dict:
         variant = Variant.fromJSON(variant_dict)
         self.add_annotations_to_variants([variant], project_id, family_id=family_id)
         return variant
     else:
         return None
예제 #19
0
    def get_single_variant(self, project_id, family_id, xpos, ref, alt):

        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            return None
        variant_dict = collection.find_one({'xpos': xpos, 'ref': ref, 'alt': alt})
        if variant_dict:
            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            return variant
        else:
            return None
예제 #20
0
    def get_variants(self, project_id, family_id, genotype_filter=None, variant_filter=None):

        db_query = self._make_db_query(genotype_filter, variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            print("Error: mongodb collection not found for project %s family %s " % (project_id, family_id))
            return
        for variant_dict in collection.find(db_query).sort('xpos'):
            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            if passes_variant_filter(variant, variant_filter)[0]:
                yield variant
예제 #21
0
    def get_variants(self, project_id, variant_filter=None):

        variant_filter_t = VariantFilter(**(variant_filter if variant_filter else {}))

        db_query = self._make_db_query(None, variant_filter)
        collection = self._get_project_collection(project_id)
        for variant_dict in collection.find(db_query).sort('xpos'):
            variant = Variant.fromJSON(variant_dict)
            if variant_filter is None:
                yield variant
            if passes_variant_filter(variant, variant_filter_t)[0]:
                yield variant
예제 #22
0
    def get_variants(self, project_id, family_id, genotype_filter=None, variant_filter=None):

        db_query = self._make_db_query(genotype_filter, variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            print("Error: mongodb collection not found for project %s family %s " % (project_id, family_id))
            return
        for i, variant_dict in enumerate(collection.find(db_query).sort('xpos').limit(MONGO_QUERY_RESULTS_LIMIT+5)):
            if i >= MONGO_QUERY_RESULTS_LIMIT:
                raise Exception("ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again." % MONGO_QUERY_RESULTS_LIMIT)

            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            if passes_variant_filter(variant, variant_filter)[0]:
                yield variant
예제 #23
0
    def get_variants(self, project_id, family_id, genotype_filter=None, variant_filter=None, quality_filter=None, indivs_to_consider=None, user=None):
        db_query = self._make_db_query(genotype_filter, variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            logger.error("Error: mongodb collection not found for project %s family %s " % (project_id, family_id))
            return
        for i, variant_dict in enumerate(collection.find({'$and' : [{k: v} for k, v in db_query.items()]}).sort('xpos').limit(settings.VARIANT_QUERY_RESULTS_LIMIT+5)):
            if i >= settings.VARIANT_QUERY_RESULTS_LIMIT:
                raise Exception("ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again." % settings.VARIANT_QUERY_RESULTS_LIMIT)

            variant = Variant.fromJSON(variant_dict)
            variant.set_extra('project_id', project_id)
            variant.set_extra('family_id', family_id)
            self.add_annotations_to_variants([variant], project_id)

            if passes_variant_filter(variant, variant_filter)[0]:
                yield variant
예제 #24
0
    def get_variants_in_gene(self, project_id, gene_id, variant_filter=None):

        if variant_filter is None:
            modified_variant_filter = VariantFilter()
        else:
            modified_variant_filter = copy.deepcopy(variant_filter)
        modified_variant_filter.add_gene(gene_id)

        db_query = self._make_db_query(None, modified_variant_filter)
        collection = self._get_project_collection(project_id)

        variants = []
        for variant_dict in collection.find(db_query).hint([('gene_ids', pymongo.ASCENDING), ('xpos', pymongo.ASCENDING)]):
            variant = Variant.fromJSON(variant_dict)
            if passes_variant_filter(variant, modified_variant_filter):
                variants.append(variant)
        variants = sorted(variants, key=lambda v: v.unique_tuple())
        return variants
예제 #25
0
    def get_project_variants_in_gene(self, project_id, gene_id, variant_filter=None):

        if variant_filter is None:
            modified_variant_filter = VariantFilter()
        else:
            modified_variant_filter = copy.deepcopy(variant_filter)
        modified_variant_filter.add_gene(gene_id)

        db_query = self._make_db_query(None, modified_variant_filter)
        logger.info("Project Gene Search: " + str(project_id) + " all variants query: " + str(db_query))
        collection = self._get_project_collection(project_id)
        # we have to collect list in memory here because mongo can't sort on xpos,
        # as result size can get too big.
        # need to find a better way to do this.
        variants = [Variant.fromJSON(variant_dict) for variant_dict in collection.find(db_query).hint([('db_gene_ids', pymongo.ASCENDING), ('xpos', pymongo.ASCENDING)])]
        self.add_annotations_to_variants(variants, project_id)
        variants = filter(lambda variant: passes_variant_filter(variant, modified_variant_filter), variants)
        variants = sorted(variants, key=lambda v: v.unique_tuple())
        return variants
예제 #26
0
파일: views.py 프로젝트: frichter/seqr
def cohort_variant_search_spec(request):

    project, cohort = get_project_and_cohort_for_user(request.user, request.GET)

    # TODO: use form

    search_spec_dict, variants = cache_utils.get_cached_results(project.project_id, request.GET.get('search_hash'))
    search_spec = MendelianVariantSearchSpec.fromJSON(search_spec_dict)
    if variants is None:
        variants = api_utils.calculate_mendelian_variant_search(search_spec, cohort.xfamily())
    else:
        variants = [Variant.fromJSON(v) for v in variants]
    api_utils.add_extra_info_to_variants_cohort(get_reference(), cohort, variants)

    return JSONResponse({
        'is_error': False,
        'variants': [v.toJSON() for v in variants],
        'search_spec': search_spec.toJSON(),
    })
예제 #27
0
파일: views.py 프로젝트: rpete/seqr
def cohort_variant_search_spec(request):

    project, cohort = get_project_and_cohort_for_user(request.user, request.GET)

    # TODO: use form

    search_spec_dict, variants = cache_utils.get_cached_results(project.project_id, request.GET.get('search_hash'))
    search_spec = MendelianVariantSearchSpec.fromJSON(search_spec_dict)
    if variants is None:
        variants = api_utils.calculate_mendelian_variant_search(search_spec, cohort.xfamily())
    else:
        variants = [Variant.fromJSON(v) for v in variants]
    api_utils.add_extra_info_to_variants_cohort(get_reference(), cohort, variants)

    return JSONResponse({
        'is_error': False,
        'variants': [v.toJSON() for v in variants],
        'search_spec': search_spec.toJSON(),
    })
예제 #28
0
파일: lookups.py 프로젝트: obonyojimmy/seqr
def get_variants_from_variant_tuples(project, variant_tuples):
    variants = []
    for t in variant_tuples:
        variant = get_datastore(project.project_id).get_single_variant(
            project.project_id, t[3], t[0], t[1], t[2])
        if not variant:
            variant = Variant(t[0], t[1], t[2])
            get_annotator().annotate_variant(
                variant, project.get_reference_population_slugs())

        variant.set_extra('family_id', t[3])
        variant.set_extra('project_id', project.project_id)
        variants.append(variant)
    return variants
예제 #29
0
    def get_variants_in_gene(self, project_id, family_id, gene_id, genotype_filter=None, variant_filter=None):

        if variant_filter is None:
            modified_variant_filter = VariantFilter()
        else:
            modified_variant_filter = copy.deepcopy(variant_filter)
        modified_variant_filter.add_gene(gene_id)

        db_query = _make_db_query(genotype_filter, modified_variant_filter)
        collection = self._get_family_collection(project_id, family_id)

        # we have to collect list in memory here because mongo can't sort on xpos,
        # as result size can get too big.
        # need to find a better way to do this.
        variants = []
        for variant_dict in collection.find(db_query).hint([('gene_ids', pymongo.ASCENDING), ('xpos', pymongo.ASCENDING)]):
            variant = Variant.fromJSON(variant_dict)
            if passes_variant_filter(variant, modified_variant_filter):
                variants.append(variant)
        variants = sorted(variants, key=lambda v: v.unique_tuple())
        for v in variants:
            yield v
예제 #30
0
    def get_variants(self, project_id, family_id, genotype_filter=None, variant_filter=None):

        db_query = self._make_db_query(genotype_filter, variant_filter)
        collection = self._get_family_collection(project_id, family_id)
        if not collection:
            print("Error: mongodb collection not found for project %s family %s " % (project_id, family_id))
            return
        
        counters = OrderedDict([('returned_by_query', 0), ('passes_variant_filter', 0)])
        for i, variant_dict in enumerate(collection.find({'$and' : [{k: v} for k, v in db_query.items()]}).sort('xpos').limit(MONGO_QUERY_RESULTS_LIMIT+5)):
            if i >= MONGO_QUERY_RESULTS_LIMIT:
                raise Exception("ERROR: this search exceeded the %s variant result size limit. Please set additional filters and try again." % MONGO_QUERY_RESULTS_LIMIT)

            variant = Variant.fromJSON(variant_dict)
            self.add_annotations_to_variant(variant, project_id)
            counters["returned_by_query"] += 1
            if passes_variant_filter(variant, variant_filter)[0]:
                counters["passes_variant_filter"] += 1
                yield variant

        for k, v in counters.items():
            sys.stderr.write("    %s: %s\n" % (k,v))
예제 #31
0
    def get_elasticsearch_variants(
            self,
            project_id,
            family_id=None,
            variant_filter=None,
            genotype_filter=None,
            variant_id_filter=None,
            quality_filter=None,
            indivs_to_consider=None,
            include_all_consequences=False,
            user=None,
            max_results_limit=settings.VARIANT_QUERY_RESULTS_LIMIT,
        ):
        from xbrowse_server.base.models import Project, Family, Individual
        from seqr.models import Sample
        from seqr.utils.es_utils import _liftover_grch38_to_grch37
        from xbrowse_server.mall import get_reference

        redis_client = None
        if settings.REDIS_SERVICE_HOSTNAME:
            try:
                redis_client = redis.StrictRedis(host=settings.REDIS_SERVICE_HOSTNAME, socket_connect_timeout=3)
                redis_client.ping()
            except redis.exceptions.TimeoutError as e:
                logger.warn("Unable to connect to redis host: {}".format(settings.REDIS_SERVICE_HOSTNAME) + str(e))
                redis_client = None

        cache_key = "Variants___%s___%s___%s" % (
            project_id,
            family_id,
            json.dumps([
                variant_filter.toJSON() if variant_filter else None,
                genotype_filter,
                quality_filter,
                variant_id_filter,
                indivs_to_consider,
                include_all_consequences,
            ])
        )

        cached_results = redis_client and redis_client.get(cache_key)
        if cached_results is not None:
            variant_results = json.loads(cached_results)
            return [Variant.fromJSON(variant_json) for variant_json in variant_results]

        if family_id is None:
            project = Project.objects.get(project_id=project_id)
            elasticsearch_index = project.get_elasticsearch_index()
            logger.info("Searching in project elasticsearch index: " + str(elasticsearch_index))
        else:
            family = Family.objects.get(project__project_id=project_id, family_id=family_id)
            elasticsearch_index = family.get_elasticsearch_index()
            project = family.project
            logger.info("Searching in family elasticsearch index: " + str(elasticsearch_index))

        if indivs_to_consider is None and genotype_filter and not family_id:
            indivs_to_consider = genotype_filter.keys()

        individuals = Individual.objects.filter(family__project__project_id=project_id).only("indiv_id", "seqr_individual")
        if indivs_to_consider:
            individuals = individuals.filter(indiv_id__in=indivs_to_consider)
        if family_id is not None:
            individuals = individuals.filter(family__family_id=family_id)
            if not indivs_to_consider:
                indivs_to_consider = [i.indiv_id for i in individuals]
        prefetch_related_objects(individuals, "seqr_individual")

        es_indices = [index.rstrip('*') for index in elasticsearch_index.split(',')]

        samples = Sample.objects.filter(
            individual__in=[i.seqr_individual for i in individuals if i.seqr_individual],
            dataset_type=Sample.DATASET_TYPE_VARIANT_CALLS,
            sample_status=Sample.SAMPLE_STATUS_LOADED,
            elasticsearch_index__startswith=es_indices[0],
            loaded_date__isnull=False,
        ).order_by('-loaded_date')
        prefetch_related_objects(samples, "individual")

        family_individual_ids_to_sample_ids = {}
        for i in individuals:
            indiv_id = i.indiv_id
            sample_id = None
            if i.seqr_individual:
                sample_id = next((
                    sample.sample_id for sample in samples
                    if sample.individual == i.seqr_individual and sample.elasticsearch_index.startswith(tuple(es_indices))
                ), None)
            family_individual_ids_to_sample_ids[indiv_id] = sample_id or indiv_id

        query_json = self._make_db_query(genotype_filter, variant_filter)

        es_client = elasticsearch.Elasticsearch(host=settings.ELASTICSEARCH_SERVICE_HOSTNAME, timeout=30)
        mapping = es_client.indices.get_mapping(str(elasticsearch_index) + "*")
        index_fields = {}
        is_parent_child = False
        is_nested = False
        if elasticsearch_index in mapping and 'join_field' in mapping[elasticsearch_index]["mappings"]["variant"]["properties"]:
            # Nested indices are not sharded so all samples are in the single index
            logger.info("matching indices: " + str(elasticsearch_index))
            is_parent_child = True
        elif elasticsearch_index in mapping and 'genotypes' in mapping[elasticsearch_index]["mappings"]["variant"]["properties"]:
            # Nested indices are not sharded so all samples are in the single index
            logger.info("matching indices: " + str(elasticsearch_index))
            is_nested = True
        elif family_id is not None and len(family_individual_ids_to_sample_ids) > 0:
            # figure out which index to use
            # TODO add caching

            matching_indices = []

            for raw_sample_id in family_individual_ids_to_sample_ids.values():
                sample_id = _encode_name(raw_sample_id)
                for index_name, index_mapping in mapping.items():
                    if sample_id+"_num_alt" in index_mapping["mappings"]["variant"]["properties"]:
                        matching_indices.append(index_name)
                        index_fields.update(index_mapping["mappings"]["variant"]["properties"])
                if len(matching_indices) > 0:
                    break

            if not matching_indices:
                if family_id is not None and not family_individual_ids_to_sample_ids:
                    logger.error("no individuals found for family %s" % (family_id))
                elif not mapping:
                    logger.error("no es mapping found for found with prefix %s" % (elasticsearch_index))
                else:
                    logger.error("%s not found in %s:\n%s" % (indiv_id, elasticsearch_index, pformat(index_mapping["mappings"]["variant"]["properties"])))
            else:
                elasticsearch_index = ",".join(matching_indices)
                logger.info("matching indices: " + str(elasticsearch_index))
        else:
            elasticsearch_index = str(elasticsearch_index)+"*"
                
        if not index_fields:
            for index_mapping in mapping.values():
                index_fields.update(index_mapping["mappings"]["variant"]["properties"])

        s = elasticsearch_dsl.Search(using=es_client, index=elasticsearch_index) #",".join(indices))

        if variant_id_filter is not None:
            variant_id_filter_term = None
            for variant_id in variant_id_filter:
                q_obj = Q('term', **{"variantId": variant_id})
                if variant_id_filter_term is None:
                    variant_id_filter_term = q_obj
                else:
                    variant_id_filter_term |= q_obj
            s = s.filter(variant_id_filter_term)

        genotype_filters = {}
        for key, value in query_json.items():
            if key.startswith("genotypes"):
                indiv_id = ".".join(key.split(".")[1:-1])
                sample_id = family_individual_ids_to_sample_ids.get(indiv_id) or indiv_id
                genotype_filter = value
                if type(genotype_filter) == int or type(genotype_filter) == basestring:
                    genotype_filters[sample_id] = [('term', genotype_filter)]
                elif '$gte' in genotype_filter:
                    genotype_filter = {k.replace("$", ""): v for k, v in genotype_filter.items()}
                    genotype_filters[sample_id] = [('range', genotype_filter)]
                elif "$in" in genotype_filter:
                    num_alt_values = genotype_filter['$in']
                    genotype_filters[sample_id] = [('term', num_alt_value) for num_alt_value in num_alt_values]

        sample_ids = [family_individual_ids_to_sample_ids.get(indiv_id) or indiv_id for indiv_id in (indivs_to_consider or [])]

        min_ab = None
        min_gq = None
        if quality_filter is not None and indivs_to_consider:
            min_ab = quality_filter.get('min_ab')
            if min_ab is not None and not is_nested:
                min_ab /= 100.0  # convert to fraction
            min_gq = quality_filter.get('min_gq')
            vcf_filter = quality_filter.get('vcf_filter')
            if vcf_filter is not None:
                s = s.filter(~Q('exists', field='filters'))

        if is_parent_child:
            quality_q = Q()
            if min_ab or min_gq:
                if min_ab is not None:
                    #  AB only relevant for hets
                    quality_q &= Q(~Q('term', num_alt=1) | Q('range', ab={'gte': min_ab}))
                if min_gq is not None:
                    quality_q &= Q('range', gq={'gte': min_gq})

            if genotype_filters:
                # Return inner hits for all requested samples, even those without a specified genotype
                genotype_sample_ids = sample_ids or genotype_filters.keys()
                genotype_q = None
                for sample_id in genotype_sample_ids:
                    sample_q = Q(Q('term', sample_id=sample_id) & quality_q)
                    if genotype_filters.get(sample_id):
                        q = None
                        for (op, val) in genotype_filters[sample_id]:
                            if q:
                                q |= Q(op, num_alt=val)
                            else:
                                q = Q(op, num_alt=val)
                        sample_q &= q
                    if not genotype_q:
                        genotype_q = sample_q
                    else:
                        genotype_q |= sample_q
                genotype_kwargs = {'query': genotype_q, 'min_children': len(genotype_sample_ids)}
            elif sample_ids:
                # Subquery for child docs with the requested sample IDs and quality metrics
                sample_id_q = Q('terms', sample_id=sample_ids) & quality_q
                # Only return variants where at least one of the requested samples has an alt allele
                s = s.filter(Q('has_child', type='genotype', query=(Q(Q('range', num_alt={'gte': 1}) & sample_id_q))))
                # Return inner hits for all the requested samples regardless of genotype
                genotype_kwargs = {'query': sample_id_q, 'min_children': len(sample_ids)}
            else:
                # Return all inner hits for the variant
                # This case is only used by gene search, which also does not use quality filters
                genotype_kwargs = {'query': Q()}

            s = s.filter(Q('has_child', type='genotype',
                           inner_hits={'size': genotype_kwargs.get('min_children', MAX_INNER_HITS)}, **genotype_kwargs))

        if is_nested:
            if sample_ids and min_ab is not None:
                min_ab_filter_val = int(min_ab) - int(min_ab % 5)
                for sample_id in sample_ids:
                    q = Q('term', samples_ab_0_to_5=sample_id)
                    for i in range(5, min_ab_filter_val, 5):
                        q = q | Q('term', **{'samples_ab_{}_to_{}'.format(i, i+5): sample_id})
                    #  AB only relevant for hets
                    s = s.filter(~Q(q) | ~Q('term', samples_num_alt_1=sample_id))
            if sample_ids and min_gq is not None:
                min_gq_filter_val = int(min_gq) - int(min_gq % 5)
                for sample_id in sample_ids:
                    q = Q('term', samples_gq_0_to_5=sample_id)
                    for i in range(5, min_gq_filter_val, 5):
                        q = q | Q('term', **{'samples_gq_{}_to_{}'.format(i, i+5): sample_id})
                    s = s.filter(~Q(q))

            if genotype_filters:
                for sample_id, queries in genotype_filters.items():
                    if queries[0][0] == 'range':
                        allowed_num_alt = range(queries[0][1]['gte'], 3)
                    else:
                        allowed_num_alt = [query[1] for query in queries]

                    if 0 in allowed_num_alt:
                        q = Q('term', samples_no_call=sample_id)
                        if 1 not in allowed_num_alt:
                            q = q | Q('term', samples_num_alt_1=sample_id)
                        if 2 not in allowed_num_alt:
                            q = q | Q('term', samples_num_alt_2=sample_id)
                        s = s.filter(~q)
                    else:
                        q = Q('term', **{'samples_num_alt_{}'.format(allowed_num_alt[0]): sample_id})
                        for num_alt in allowed_num_alt[1:]:
                            q = q | Q('term', **{'samples_num_alt_{}'.format(num_alt): sample_id})
                        s = s.filter(q)

            elif sample_ids:
                s = s.filter(Q('terms', samples_num_alt_1=sample_ids) | Q('terms', samples_num_alt_2=sample_ids))

        else:
            for sample_id, queries in genotype_filters.items():
                encoded_sample_id = _encode_name(sample_id)
                q = Q(queries[0][0], **{encoded_sample_id + "_num_alt": queries[0][1]})
                for (op, val) in queries[1:]:
                    q = q | Q(op, **{encoded_sample_id + "_num_alt": val})
                s = s.filter(q)

            if sample_ids:
                atleast_one_nonref_genotype_filter = None
                for sample_id in sample_ids:
                    encoded_sample_id = _encode_name(sample_id)
                    q = Q('range', **{encoded_sample_id+"_num_alt": {'gte': 1}})
                    if atleast_one_nonref_genotype_filter is None:
                        atleast_one_nonref_genotype_filter = q
                    else:
                        atleast_one_nonref_genotype_filter |= q

                s = s.filter(atleast_one_nonref_genotype_filter)

            if min_ab or min_gq:
                for sample_id in sample_ids:
                    encoded_sample_id = _encode_name(sample_id)

                    if min_ab:
                        s = s.filter(
                            ~Q('term', **{encoded_sample_id+"_num_alt": 1}) |
                            Q('range', **{encoded_sample_id+"_ab": {'gte': min_ab}}))
                        #logger.info("### ADDED FILTER: " + str({encoded_sample_id+"_ab": {'gte': min_ab}}))
                    if min_gq:
                        s = s.filter('range', **{encoded_sample_id+"_gq": {'gte': min_gq}})
                        #logger.info("### ADDED FILTER: " + str({encoded_sample_id+"_gq": {'gte': min_gq}}))

        # parse variant query
        annotation_groups_map = ANNOTATION_GROUPS_MAP_INTERNAL if user and user.is_staff else ANNOTATION_GROUPS_MAP

        for key, value in query_json.items():
            if key == 'db_tags':
                so_annotations = query_json.get('db_tags', {}).get('$in', [])

                # handle clinvar filters
                selected_so_annotations_set = set(so_annotations)

                all_clinvar_filters_set = set(annotation_groups_map.get("clinvar", {}).get("children", []))
                selected_clinvar_filters_set = all_clinvar_filters_set & selected_so_annotations_set

                all_hgmd_filters_set = set(annotation_groups_map.get("hgmd", {}).get("children", []))
                selected_hgmd_filters_set = all_hgmd_filters_set & selected_so_annotations_set

                vep_consequences = list(selected_so_annotations_set - selected_clinvar_filters_set - selected_hgmd_filters_set)
                consequences_filter = Q("terms", transcriptConsequenceTerms=vep_consequences)

                if selected_clinvar_filters_set:
                    clinvar_clinical_significance_terms = set()
                    for clinvar_filter in selected_clinvar_filters_set:
                        # translate selected filters to the corresponding clinvar clinical consequence terms
                        if clinvar_filter == "pathogenic":
                            clinvar_clinical_significance_terms.update(["Pathogenic", "Pathogenic/Likely_pathogenic"])
                        elif clinvar_filter == "likely_pathogenic":
                            clinvar_clinical_significance_terms.update(["Likely_pathogenic", "Pathogenic/Likely_pathogenic"])
                        elif clinvar_filter == "benign":
                            clinvar_clinical_significance_terms.update(["Benign", "Benign/Likely_benign"])
                        elif clinvar_filter == "likely_benign":
                            clinvar_clinical_significance_terms.update(["Likely_benign", "Benign/Likely_benign"])
                        elif clinvar_filter == "vus_or_conflicting":
                            clinvar_clinical_significance_terms.update([
                                "Conflicting_interpretations_of_pathogenicity",
                                "Uncertain_significance",
                                "not_provided",
                                "other"])
                        else:
                            raise ValueError("Unexpected clinvar filter: " + str(clinvar_filter))

                    consequences_filter = consequences_filter | Q("terms", clinvar_clinical_significance=list(clinvar_clinical_significance_terms))

                if selected_hgmd_filters_set:
                    hgmd_class = set()
                    for hgmd_filter in selected_hgmd_filters_set:
                        # translate selected filters to the corresponding hgmd clinical consequence terms
                        if hgmd_filter == "disease_causing":
                            hgmd_class.update(["DM"])
                        elif hgmd_filter == "likely_disease_causing":
                            hgmd_class.update(["DM?"])
                        elif hgmd_filter == "hgmd_other":
                            hgmd_class.update(["DP", "DFP", "FP", "FTV"])
                        else:
                            raise ValueError("Unexpected hgmd filter: " + str(hgmd_filter))

                    consequences_filter = consequences_filter | Q("terms", hgmd_class=list(hgmd_class))

                if 'intergenic_variant' in vep_consequences:
                    # for many intergenic variants VEP doesn't add any annotations, so if user selected 'intergenic_variant', also match variants where transcriptConsequenceTerms is emtpy
                    consequences_filter = consequences_filter | ~Q('exists', field='transcriptConsequenceTerms')

                s = s.filter(consequences_filter)
                #logger.info("==> transcriptConsequenceTerms: %s" % str(vep_consequences))

            if key.startswith("genotypes"):
                continue

            if key == "db_gene_ids":
                db_gene_ids = query_json.get('db_gene_ids', {})

                exclude_genes = db_gene_ids.get('$nin', [])
                gene_ids = exclude_genes or db_gene_ids.get('$in', [])

                if exclude_genes:
                    s = s.exclude("terms", geneIds=gene_ids)
                else:
                    s = s.filter("terms",  geneIds=gene_ids)
                #logger.info("==> %s %s" % ("exclude" if exclude_genes else "include", "geneIds: " + str(gene_ids)))

            if key == "$or" and type(value) == list:
                q_terms = None
                for region_filter in value:
                    xpos_filters = region_filter.get("$and", {})

                    # for example: $or : [{'$and': [{'xpos': {'$gte': 12345}}, {'xpos': {'$lte': 54321}}]}]
                    xpos_filters_dict = {}
                    for xpos_filter in xpos_filters:
                        xpos_filter_setting = xpos_filter["xpos"]  # for example {'$gte': 12345} or {'$lte': 54321}
                        xpos_filters_dict.update(xpos_filter_setting)

                    xpos_filter_setting = {k.replace("$", ""): v for k, v in xpos_filters_dict.items()}
                    q = Q('range', **{"xpos": xpos_filter_setting})
                    if q_terms is None:
                        q_terms = q
                    else:
                        q_terms |= q
                if q_terms is not None:
                    s = s.filter(q_terms)

                #logger.info("==> xpos range: " + str({"xpos": xpos_filter_setting}))

            af_key_map = {
                "db_freqs.AF": ["AF"],
                "db_freqs.1kg_wgs_phase3": ["g1k_POPMAX_AF"],
                "db_freqs.exac_v3": ["exac_AF_POPMAX"],
                "db_freqs.topmed": ["topmed_AF"],
                "db_freqs.gnomad_exomes": ["gnomad_exomes_AF_POPMAX", "gnomad_exomes_AF_POPMAX_OR_GLOBAL"],
                "db_freqs.gnomad_genomes": ["gnomad_genomes_AF_POPMAX", "gnomad_genomes_AF_POPMAX_OR_GLOBAL"],
                "db_freqs.gnomad-exomes2": ["gnomad_exomes_AF_POPMAX", "gnomad_exomes_AF_POPMAX_OR_GLOBAL"],
                "db_freqs.gnomad-genomes2": ["gnomad_genomes_AF_POPMAX", "gnomad_genomes_AF_POPMAX_OR_GLOBAL"],
            }

            if key in af_key_map:
                for filter_key in af_key_map[key]:
                    af_filter_setting = {k.replace("$", ""): v for k, v in value.items()}
                    s = s.filter(Q('range', **{filter_key: af_filter_setting}) | ~Q('exists', field=filter_key))
                #logger.info("==> %s: %s" % (filter_key, af_filter_setting))

            ac_key_map = {
                "db_acs.AF": "AC",
                "db_acs.1kg_wgs_phase3": "g1k_AC",
                "db_acs.exac_v3": "exac_AC",
                "db_acs.topmed": "topmed_AC",
                "db_acs.gnomad_exomes": "gnomad_exomes_AC",
                "db_acs.gnomad_genomes": "gnomad_genomes_AC",
                "db_acs.gnomad-exomes2": "gnomad_exomes_AC",
                "db_acs.gnomad-genomes2": "gnomad_genomes_AC",
            }

            if key in ac_key_map:
                filter_key = ac_key_map[key]
                ac_filter_setting = {k.replace("$", ""): v for k, v in value.items()}
                s = s.filter(Q('range', **{filter_key: ac_filter_setting}) | ~Q('exists', field=filter_key))

            hemi_key_map = {
                "db_hemi.exac_v3": "exac_AC_Hemi",
                "db_hemi.gnomad_exomes": "gnomad_exomes_Hemi",
                "db_hemi.gnomad_genomes": "gnomad_genomes_Hemi",
                "db_hemi.gnomad-exomes2": "gnomad_exomes_Hemi",
                "db_hemi.gnomad-genomes2": "gnomad_genomes_Hemi",
            }

            if key in hemi_key_map:
                filter_key = hemi_key_map[key]
                hemi_filter_setting = {k.replace("$", ""): v for k, v in value.items()}
                s = s.filter(Q('range', **{filter_key: hemi_filter_setting}) | ~Q('exists', field=filter_key))

            hom_key_map = {
                "db_hom.exac_v3": "exac_AC_Hom",
                "db_hom.gnomad_exomes": "gnomad_exomes_Hom",
                "db_hom.gnomad_genomes": "gnomad_genomes_Hom",
                "db_hom.gnomad-exomes2": "gnomad_exomes_Hom",
                "db_hom.gnomad-genomes2": "gnomad_genomes_Hom",
            }

            if key in hom_key_map:
                filter_key = hom_key_map[key]
                hom_filter_setting = {k.replace("$", ""): v for k, v in value.items()}
                s = s.filter(Q('range', **{filter_key: hom_filter_setting}) | ~Q('exists', field=filter_key))

            #s = s.sort("xpos")

        #logger.info("=====")
        #logger.info("FULL QUERY OBJ: " + pformat(s.__dict__))
        #logger.info("FILTERS: " + pformat(s.to_dict()))

        # https://elasticsearch-py.readthedocs.io/en/master/helpers.html#elasticsearch.helpers.scan
        start = time.time()

        s = s.params(size=max_results_limit + 1)
        #if not include_all_consequences:
        #    s = s.source(exclude=["sortedTranscriptConsequences"])
        response = s.execute()
        logger.info("=====")

        logger.info("TOTAL: %s. Query took %s seconds" % (response.hits.total, time.time() - start))

        if response.hits.total > max_results_limit + 1:
            raise Exception("This search matched too many variants. Please set additional filters and try again.")

        #print(pformat(response.to_dict()))

        project = Project.objects.get(project_id=project_id)

        #gene_list_map = project.get_gene_list_map()

        reference = get_reference()

        #for i, hit in enumerate(response.hits):
        variant_results = []
        for i, hit in enumerate(response):  # preserve_order=True
            #logger.info("HIT %s: %s %s %s" % (i, hit["variantId"], hit["geneIds"], pformat(hit.__dict__)))
            #print("HIT %s: %s" % (i, pformat(hit.to_dict())))
            filters = ",".join(hit["filters"] or []) if "filters" in hit else ""
            genotypes = {}
            all_num_alt = []

            if is_parent_child:
                genotypes_by_sample_id = {gen_hit['sample_id']: gen_hit for gen_hit in hit.meta.inner_hits.genotype}
            elif is_nested:
                genotypes_by_sample_id = {gen_hit['sample_id']: gen_hit for gen_hit in hit['genotypes']}

            for individual_id, sample_id in family_individual_ids_to_sample_ids.items():
                def _get_hit_field(field):
                    if is_parent_child or is_nested:
                        gen_hit = genotypes_by_sample_id.get(sample_id, {})
                        key = field
                    else:
                        gen_hit = hit
                        key = '{}_{}'.format(_encode_name(sample_id), field)
                    return gen_hit[key] if key in gen_hit else None

                num_alt = _get_hit_field('num_alt')
                if num_alt is None:
                    num_alt = -1
                all_num_alt.append(num_alt)

                alleles = []
                if num_alt == 0:
                    alleles = [hit["ref"], hit["ref"]]
                elif num_alt == 1:
                    alleles = [hit["ref"], hit["alt"]]
                elif num_alt == 2:
                    alleles = [hit["alt"], hit["alt"]]
                elif num_alt == -1 or num_alt == None:
                    alleles = []
                else:
                    raise ValueError("Invalid num_alt: " + str(num_alt))

                genotypes[individual_id] = {
                    'ab': _get_hit_field('ab'),
                    'alleles': map(str, alleles),
                    'extras': {
                        'ad': _get_hit_field('ad'),
                        'dp': _get_hit_field('dp'),
                        #'pl': '',
                    },
                    'filter': filters or "pass",
                    'gq': _get_hit_field('gq') or '',
                    'num_alt': num_alt,
                }

            vep_annotation = hit['sortedTranscriptConsequences'] if 'sortedTranscriptConsequences' in hit else None
            if vep_annotation is not None:
                if is_parent_child or is_nested:
                    vep_annotation = [annot.to_dict() for annot in vep_annotation]
                else:
                    vep_annotation = json.loads(str(vep_annotation))

            gene_ids = list(hit['geneIds'] or [])
            worst_vep_index_per_gene = {
                gene_id: next((i for i, annot in enumerate(vep_annotation) if annot['gene_id'] == gene_id), None)
                for gene_id in gene_ids
            }

            if project.genome_version == GENOME_VERSION_GRCh37:
                grch38_coord = None
                if self.liftover_grch37_to_grch38:
                    grch38_coord = self.liftover_grch37_to_grch38.convert_coordinate("chr%s" % hit["contig"].replace("chr", ""), int(hit["start"]))
                    if grch38_coord and grch38_coord[0]:
                        grch38_coord = "%s-%s-%s-%s "% (grch38_coord[0][0], grch38_coord[0][1], hit["ref"], hit["alt"])
                    else:
                        grch38_coord = None
            else:
                grch38_coord = hit["variantId"]

            if project.genome_version == GENOME_VERSION_GRCh38:
                grch37_coord = None
                liftover_grch38_to_grch37 = _liftover_grch38_to_grch37()
                if liftover_grch38_to_grch37:
                    grch37_coord = liftover_grch38_to_grch37.convert_coordinate("chr%s" % hit["contig"].replace("chr", ""), int(hit["start"]))
                    if grch37_coord and grch37_coord[0]:
                        grch37_coord = "%s-%s-%s-%s "% (grch37_coord[0][0], grch37_coord[0][1], hit["ref"], hit["alt"])
                    else:
                        grch37_coord = None
            else:
                grch37_coord = hit["variantId"]

            freq_fields = {
                'AF': "AF" if "AF" in index_fields else None,
                '1kg_wgs_AF': "g1k_AF" if "g1k_AF" in index_fields else None,
                '1kg_wgs_popmax_AF': "g1k_POPMAX_AF" if "g1k_POPMAX_AF" in index_fields else None,
                'exac_v3_AF': "exac_AF" if "exac_AF" in index_fields else None,
                'exac_v3_popmax_AF': "exac_AF_POPMAX" if "exac_AF_POPMAX" in index_fields else None,
                'gnomad_exomes_AF': "gnomad_exomes_AF" if "gnomad_exomes_AF" in index_fields else None,
                'gnomad_exomes_popmax_AF': "gnomad_exomes_AF_POPMAX_OR_GLOBAL" if "gnomad_exomes_AF_POPMAX_OR_GLOBAL" in index_fields else (
                     "gnomad_exomes_AF_POPMAX" if "gnomad_exomes_AF_POPMAX" in index_fields else None),
                'gnomad_genomes_AF': "gnomad_genomes_AF" if "gnomad_genomes_AF" in index_fields else None,
                'gnomad_genomes_popmax_AF': "gnomad_genomes_AF_POPMAX_OR_GLOBAL" if "gnomad_genomes_AF_POPMAX_OR_GLOBAL" in index_fields else (
                    "gnomad_genomes_AF_POPMAX" if "gnomad_genomes_AF_POPMAX" in index_fields else None),
                'topmed_AF': "topmed_AF" if "topmed_AF" in index_fields else None,
            }

            result = {
                #u'_id': ObjectId('596d2207ff66f729285ca588'),
                'alt': str(hit["alt"]) if "alt" in hit else None,
                'annotation': {
                    'fathmm': fathmm_map.get(hit["dbnsfp_FATHMM_pred"].split(';')[0]) if "dbnsfp_FATHMM_pred" in hit and hit["dbnsfp_FATHMM_pred"] else None,
                    'muttaster': muttaster_map.get(hit["dbnsfp_MutationTaster_pred"].split(';')[0]) if "dbnsfp_MutationTaster_pred" in hit and hit["dbnsfp_MutationTaster_pred"] else None,
                    'polyphen': polyphen_map.get(hit["dbnsfp_Polyphen2_HVAR_pred"].split(';')[0]) if "dbnsfp_Polyphen2_HVAR_pred" in hit and hit["dbnsfp_Polyphen2_HVAR_pred"] else None,
                    'sift': sift_map.get(hit["dbnsfp_SIFT_pred"].split(';')[0]) if "dbnsfp_SIFT_pred" in hit and hit["dbnsfp_SIFT_pred"] else None,
                    'metasvm': metasvm_map.get(hit["dbnsfp_MetaSVM_pred"].split(';')[0]) if "dbnsfp_MetaSVM_pred" in hit and hit["dbnsfp_MetaSVM_pred"] else None,

                    'GERP_RS': float(hit["dbnsfp_GERP_RS"]) if "dbnsfp_GERP_RS" in hit and hit["dbnsfp_GERP_RS"] else None,
                    'phastCons100way_vertebrate': float(hit["dbnsfp_phastCons100way_vertebrate"]) if "dbnsfp_phastCons100way_vertebrate" in hit and hit["dbnsfp_phastCons100way_vertebrate"] else None,

                    'cadd_phred': hit["cadd_PHRED"] if "cadd_PHRED" in hit else None,
                    'dann_score': hit["dbnsfp_DANN_score"] if "dbnsfp_DANN_score" in hit else None,
                    'revel_score': hit["dbnsfp_REVEL_score"] if "dbnsfp_REVEL_score" in hit else None,
                    'eigen_phred': hit["eigen_Eigen_phred"] if "eigen_Eigen_phred" in hit else (hit["dbnsfp_Eigen_phred"] if "dbnsfp_Eigen_phred" in hit else None),
                    'mpc_score': hit["mpc_MPC"] if "mpc_MPC" in hit else None,
                    'primate_ai_score': hit["primate_ai_score"] if "primate_ai_score" in hit else None,
                    'splice_ai_delta_score': hit["splice_ai_delta_score"] if "splice_ai_delta_score" in hit else None,
                    'rsid': hit["rsid"] if "rsid" in hit else None,
                    'annotation_tags': list(hit["transcriptConsequenceTerms"] or []) if "transcriptConsequenceTerms" in hit else None,
                    'coding_gene_ids': list(hit['codingGeneIds'] or []),
                    'gene_ids': list(hit['geneIds'] or []),
                    'vep_annotation': vep_annotation,
                    'vep_group': str(hit['mainTranscript_major_consequence'] or "") if "mainTranscript_major_consequence" in hit else "",
                    'vep_consequence': str(hit['mainTranscript_major_consequence'] or "") if "mainTranscript_major_consequence" in hit else "",
                    'main_transcript': {k.replace('mainTranscript_', ''): hit[k] for k in dir(hit) if k.startswith('mainTranscript_')},
                    'worst_vep_annotation_index': 0,
                    'worst_vep_index_per_gene': worst_vep_index_per_gene,
                },
                'chr': hit["contig"],
                'coding_gene_ids': list(hit['codingGeneIds'] or []),
                'gene_ids': gene_ids,
                'coverage': {
                    'gnomad_exome_coverage': float(hit["gnomad_exome_coverage"] or -1) if "gnomad_exome_coverage" in hit else -1,
                    'gnomad_genome_coverage': float(hit["gnomad_genome_coverage"] or -1) if "gnomad_genome_coverage" in hit else -1,
                },
                'pop_counts': {
                    'AC': int(hit['AC'] or 0) if 'AC' in hit else None,
                    'AN': int(hit['AN'] or 0) if 'AN' in hit else None,

                    'g1kAC': int(hit['g1k_AC'] or 0) if 'g1k_AC' in hit else None,
                    'g1kAN': int(hit['g1k_AN'] or 0) if 'g1k_AN' in hit else None,

                    'exac_v3_AC': int(hit["exac_AC_Adj"] or 0) if "exac_AC_Adj" in hit else None,
                    'exac_v3_Het': int(hit["exac_AC_Het"] or 0) if "exac_AC_Het" in hit else None,
                    'exac_v3_Hom': int(hit["exac_AC_Hom"] or 0) if "exac_AC_Hom" in hit else None,
                    'exac_v3_Hemi': int(hit["exac_AC_Hemi"] or 0) if "exac_AC_Hemi" in hit else None,
                    'exac_v3_AN': int(hit["exac_AN_Adj"] or 0) if "exac_AN_Adj" in hit else None,

                    'gnomad_exomes_AC': int(hit["gnomad_exomes_AC"] or 0) if "gnomad_exomes_AC" in hit else None,
                    'gnomad_exomes_Hom': int(hit["gnomad_exomes_Hom"] or 0) if "gnomad_exomes_Hom" in hit else None,
                    'gnomad_exomes_Hemi': int(hit["gnomad_exomes_Hemi"] or 0) if "gnomad_exomes_Hemi" in hit else None,
                    'gnomad_exomes_AN': int(hit["gnomad_exomes_AN"] or 0) if "gnomad_exomes_AN" in hit else None,

                    'gnomad_genomes_AC': int(hit["gnomad_genomes_AC"] or 0) if "gnomad_genomes_AC" in hit else None,
                    'gnomad_genomes_Hom': int(hit["gnomad_genomes_Hom"] or 0) if "gnomad_genomes_Hom" in hit else None,
                    'gnomad_genomes_Hemi': int(hit["gnomad_genomes_Hemi"] or 0) if "gnomad_genomes_Hemi" in hit else None,
                    'gnomad_genomes_AN': int(hit["gnomad_genomes_AN"] or 0) if "gnomad_genomes_AN" in hit else None,

                    'topmed_AC': float(hit["topmed_AC"] or 0) if "topmed_AC" in hit else None,
                    'topmed_Het': float(hit["topmed_Het"] or 0) if "topmed_Het" in hit else None,
                    'topmed_Hom': float(hit["topmed_Hom"] or 0) if "topmed_Hom" in hit else None,
                    'topmed_AN': float(hit["topmed_AN"] or 0) if "topmed_AN" in hit else None,
                },
                'db_freqs': {k: float(hit[v] or 0.0) if v in hit else (0.0 if v else None) for k, v in freq_fields.items()},
                #'popmax_populations': {
                #    'exac_popmax': hit["exac_POPMAX"] or None,
                #    'gnomad_exomes_popmax': hit["gnomad_exomes_POPMAX"] or None,
                #    'gnomad_genomes_popmax': hit["gnomad_genomes_POPMAX"] or None,
                #},
                'db_gene_ids': list((hit["geneIds"] or []) if "geneIds" in hit else []),
                'db_tags': str(hit["transcriptConsequenceTerms"] or "") if "transcriptConsequenceTerms" in hit else None,
                'extras': {
                    'clinvar_variant_id': hit['clinvar_variation_id'] if 'clinvar_variation_id' in hit and hit['clinvar_variation_id'] else None,
                    'clinvar_allele_id': hit['clinvar_allele_id'] if 'clinvar_allele_id' in hit and hit['clinvar_allele_id'] else None,
                    'clinvar_clinsig': hit['clinvar_clinical_significance'].lower() if ('clinvar_clinical_significance' in hit) and hit['clinvar_clinical_significance'] else None,
                    'clinvar_gold_stars': hit['clinvar_gold_stars'] if 'clinvar_gold_stars' in hit and hit['clinvar_gold_stars'] else None,
                    'hgmd_class': hit['hgmd_class'] if 'hgmd_class' in hit and user and user.is_staff else None,
                    'hgmd_accession': hit['hgmd_accession'] if 'hgmd_accession' in hit else None,
                    'genome_version': project.genome_version,
                    'grch37_coords': grch37_coord,
                    'grch38_coords': grch38_coord,
                    'alt_allele_pos': 0,
                    'orig_alt_alleles': map(str, [a.split("-")[-1] for a in hit["originalAltAlleles"]]) if "originalAltAlleles" in hit else None
                },
                'genotypes': genotypes,
                'pos': long(hit['start']),
                'pos_end': str(hit['end']),
                'ref': str(hit['ref']),
                'vartype': 'snp' if len(hit['ref']) == len(hit['alt']) else "indel",
                'vcf_id': None,
                'xpos': long(hit["xpos"]),
                'xposx': long(hit["xpos"]),
            }

            result["annotation"]["freqs"] = result["db_freqs"]
            result["annotation"]["pop_counts"] = result["pop_counts"]
            result["annotation"]["db"] = "elasticsearch"

            result["extras"]["svlen"] = hit["SVLEN"] if "SVLEN" in hit else None
            result["extras"]["svtype"] = hit["SVTYPE"] if "SVTYPE" in hit else None


            logger.info("Result %s: GRCh37: %s GRCh38: %s - gene ids: %s, coding gene_ids: %s" % (
                i, grch37_coord, grch38_coord,
                result["gene_ids"],
                result["coding_gene_ids"]))

            result["extras"]["project_id"] = project_id
            result["extras"]["family_id"] = family_id

            # add gene info
            gene_names = {}
            if vep_annotation is not None:
                gene_names = {vep_anno["gene_id"]: vep_anno.get("gene_symbol") for vep_anno in vep_annotation if vep_anno.get("gene_symbol")}
            result["extras"]["gene_names"] = gene_names

            try:
                genes = {}
                for gene_id in result["gene_ids"]:
                    if gene_id:
                        genes[gene_id] = reference.get_gene_summary(gene_id) or {}

                #if not genes:
                #    genes =  {vep_anno["gene_id"]: {"symbol": vep_anno["gene_symbol"]} for vep_anno in vep_annotation}

                result["extras"]["genes"] = genes
            except Exception as e:
                exc_type, exc_obj, exc_tb = sys.exc_info()
                logger.warn("WARNING: got unexpected error in add_gene_names_to_variants: %s : line %s" % (e, exc_tb.tb_lineno))

            variant_results.append(result)

        logger.info("Finished returning the %s variants: %s seconds" % (response.hits.total, time.time() - start))

        if redis_client:
            redis_client.set(cache_key, json.dumps(variant_results))

        return [Variant.fromJSON(variant_json) for variant_json in variant_results]
예제 #32
0
    def get_variants_cohort(self, project_id, cohort_id, variant_filter=None):

        db_query = _make_db_query(None, variant_filter)
        collection = self._get_family_collection(project_id, cohort_id)
        for variant in collection.find(db_query).sort('xpos'):
            yield Variant.fromJSON(variant)
예제 #33
0
파일: annotator.py 프로젝트: zmcv/seqr
 def get_variant(self, xpos, ref, alt):
     variant = Variant(xpos, ref, alt)
     self.annotate_variant(variant)
     return variant
예제 #34
0
파일: views.py 프로젝트: rpete/seqr
def add_or_edit_variant_note(request):
    """Add a variant note"""
    family = None
    if 'family_id' in request.GET:
        project, family = get_project_and_family_for_user(request.user, request.GET)
    else:
        project = utils.get_project_for_user(request.user, request.GET)

    form = api_forms.VariantNoteForm(project, request.GET)
    if not form.is_valid():
        return JSONResponse({
            'is_error': True,
            'error': server_utils.form_error_string(form)
        })

    variant = get_datastore(project.project_id).get_single_variant(
        project.project_id,
        family.family_id,
        form.cleaned_data['xpos'],
        form.cleaned_data['ref'],
        form.cleaned_data['alt'],
    )

    if not variant:
        variant = Variant.fromJSON({
            'xpos' : form.cleaned_data['xpos'], 'ref': form.cleaned_data['ref'], 'alt': form.cleaned_data['alt'],
            'genotypes': {}, 'extras': {},
        })

    if 'note_id' in form.cleaned_data and form.cleaned_data['note_id']:
        event_type = "edit_variant_note"

        notes = VariantNote.objects.filter(
            id=form.cleaned_data['note_id'],
            project=project,
            xpos=form.cleaned_data['xpos'],
            ref=form.cleaned_data['ref'],
            alt=form.cleaned_data['alt'],
        )
        if not notes:
            return JSONResponse({
                'is_error': True,
                'error': 'note id %s not found' % form.cleaned_data['note_id']
            })

        note = notes[0]
        note.user = request.user
        note.note = form.cleaned_data['note_text']
        note.date_saved = timezone.now()
        if family:
            note.family = family
        note.save()
    else:
        event_type = "add_variant_note"

        VariantNote.objects.create(
            user=request.user,
            project=project,
            xpos=form.cleaned_data['xpos'],
            ref=form.cleaned_data['ref'],
            alt=form.cleaned_data['alt'],
            note=form.cleaned_data['note_text'],
            date_saved=timezone.now(),
            family=family,
        )

    add_extra_info_to_variants_family(get_reference(), family, [variant,])

    try:
        settings.EVENTS_COLLECTION.insert({
            'event_type': event_type,
            'date': timezone.now(),
            'project_id': ''.join(project.project_id),
            'family_id': family.family_id,
            'note': form.cleaned_data['note_text'],

            'xpos':form.cleaned_data['xpos'],
            'pos':variant.pos,
            'chrom': variant.chr,
            'ref':form.cleaned_data['ref'],
            'alt':form.cleaned_data['alt'],
            'gene_names': ", ".join(variant.extras['gene_names'].values()),
            'username': request.user.username,
            'email': request.user.email,
        })
    except Exception as e:
        logging.error("Error while logging %s event: %s" % (event_type, e))


    return JSONResponse({
        'is_error': False,
        'variant': variant.toJSON(),
    })
예제 #35
0
    def get_single_variant_cohort(self, project_id, cohort_id, xpos, ref, alt):

        collection = self._get_family_collection(project_id, cohort_id)
        variant = collection.find_one({'xpos': xpos, 'ref': ref, 'alt': alt})
        return Variant.fromJSON(variant)
예제 #36
0
파일: family.py 프로젝트: batsal/xbrowse
def get_de_novo_variants(datastore, reference, family, variant_filter=None, quality_filter=None):
    """
    Returns variants that follow homozygous recessive inheritance in family
    """
    de_novo_filter = inheritance.get_de_novo_filter(family)
    db_query = datastore._make_db_query(de_novo_filter, variant_filter)

    collection = datastore._get_family_collection(family.project_id, family.family_id)
    if not collection:
        raise ValueError("Error: mongodb collection not found for project %s family %s " % (family.project_id, family.family_id))

    variant_iter = collection.find(db_query).sort('xpos')

    # get ids of parents in this family
    valid_ids = set(indiv_id for indiv_id in family.individuals)
    paternal_ids = set(i.paternal_id for i in family.get_individuals() if i.paternal_id in valid_ids)
    maternal_ids = set(i.maternal_id for i in family.get_individuals() if i.maternal_id in valid_ids)
    parental_ids = paternal_ids | maternal_ids

    # loop over all variants returned
    for variant_dict in variant_iter:
        variant = Variant.fromJSON(variant_dict)
        datastore.add_annotations_to_variant(variant, family.project_id)
        if not passes_variant_filter(variant, variant_filter)[0]:
            continue

        # handle genotype filters
        if len(parental_ids) != 2:
            # ordinary filters for non-trios
            for indiv_id in de_novo_filter.keys():
                genotype = variant.get_genotype(indiv_id)
                if not passes_genotype_filter(genotype, quality_filter):
                    break
            else:
                yield variant
        else:
            # for trios use Mark's recommended filters for de-novo variants:
            # Hard-coded thresholds:
            #   1) Child must have > 10% of combined Parental Read Depth
            #   2) MinimumChildGQscore >= 20
            #   3) MaximumParentAlleleBalance <= 5%
            # Adjustable filters:
            #   Variants should PASS
            #   Child AB should be >= 20

            # compute parental read depth for filter 1
            total_parental_read_depth = 0
            for indiv_id in parental_ids:
                genotype = variant.get_genotype(indiv_id)
                if genotype.extras and 'dp' in genotype.extras:
                    total_parental_read_depth += int(genotype.extras['dp'])
                else:
                    total_parental_read_depth = None  # both parents must have DP to use the parental_read_depth filters 
                    break
                
            for indiv_id in de_novo_filter.keys():            
                quality_filter_temp = quality_filter.copy()  # copy before modifying
                if indiv_id in parental_ids:
                    # handle one of the parents
                    quality_filter_temp['max_ab'] = 5
                else: 
                    # handle child
                    quality_filter_temp['min_gq'] = 20
                    if total_parental_read_depth is not None:
                        quality_filter_temp['min_dp'] = total_parental_read_depth * 0.1

                genotype = variant.get_genotype(indiv_id)
                if not passes_genotype_filter(genotype, quality_filter_temp):
                    #print("%s: %s " % (variant.chr, variant.pos))
                    break
            else:
                yield variant
예제 #37
0
def get_de_novo_variants(datastore,
                         reference,
                         family,
                         variant_filter=None,
                         quality_filter=None):
    """
    Returns variants that follow homozygous recessive inheritance in family
    """
    de_novo_filter = inheritance.get_de_novo_filter(family)
    db_query = datastore._make_db_query(de_novo_filter, variant_filter)

    collection = datastore._get_family_collection(family.project_id,
                                                  family.family_id)
    if not collection:
        raise ValueError(
            "Error: mongodb collection not found for project %s family %s " %
            (family.project_id, family.family_id))

    MONGO_QUERY_RESULTS_LIMIT = 5000
    variant_iter = collection.find(db_query).sort('xpos').limit(
        MONGO_QUERY_RESULTS_LIMIT + 5)

    # get ids of parents in this family
    valid_ids = set(indiv_id for indiv_id in family.individuals)
    paternal_ids = set(i.paternal_id for i in family.get_individuals()
                       if i.paternal_id in valid_ids)
    maternal_ids = set(i.maternal_id for i in family.get_individuals()
                       if i.maternal_id in valid_ids)
    parental_ids = paternal_ids | maternal_ids

    # loop over all variants returned
    for i, variant_dict in enumerate(variant_iter):
        if i > MONGO_QUERY_RESULTS_LIMIT:
            raise Exception(
                "MONGO_QUERY_RESULTS_LIMIT of %s exceeded for query: %s" %
                (MONGO_QUERY_RESULTS_LIMIT, db_query))

        variant = Variant.fromJSON(variant_dict)
        datastore.add_annotations_to_variant(variant, family.project_id)
        if not passes_variant_filter(variant, variant_filter)[0]:
            continue

        # handle genotype filters
        if len(parental_ids) != 2:
            # ordinary filters for non-trios
            for indiv_id in de_novo_filter.keys():
                genotype = variant.get_genotype(indiv_id)
                if not passes_genotype_filter(genotype, quality_filter):
                    break
            else:
                yield variant
        else:
            # for trios use Mark's recommended filters for de-novo variants:
            # Hard-coded thresholds:
            #   1) Child must have > 10% of combined Parental Read Depth
            #   2) MinimumChildGQscore >= 20
            #   3) MaximumParentAlleleBalance <= 5%
            # Adjustable filters:
            #   Variants should PASS
            #   Child AB should be >= 20

            # compute parental read depth for filter 1
            total_parental_read_depth = 0
            for indiv_id in parental_ids:
                genotype = variant.get_genotype(indiv_id)
                if genotype.extras and 'dp' in genotype.extras and genotype.extras[
                        'dp'] != '.':
                    total_parental_read_depth += int(genotype.extras['dp'])
                else:
                    total_parental_read_depth = None  # both parents must have DP to use the parental_read_depth filters
                    break

            for indiv_id in de_novo_filter.keys():
                quality_filter_temp = quality_filter.copy(
                )  # copy before modifying
                if indiv_id in parental_ids:
                    # handle one of the parents
                    quality_filter_temp['max_ab'] = 5
                else:
                    # handle child
                    quality_filter_temp['min_gq'] = 20
                    if total_parental_read_depth is not None:
                        quality_filter_temp[
                            'min_dp'] = total_parental_read_depth * 0.1

                genotype = variant.get_genotype(indiv_id)
                if not passes_genotype_filter(genotype, quality_filter_temp):
                    #print("%s: %s " % (variant.chr, variant.pos))
                    break
            else:
                yield variant
예제 #38
0
    def get_elasticsearch_variants(
        self,
        project_id,
        family_id=None,
        variant_filter=None,
        genotype_filter=None,
        variant_id_filter=None,
        quality_filter=None,
        indivs_to_consider=None,
        include_all_consequences=False,
        user=None,
        max_results_limit=settings.VARIANT_QUERY_RESULTS_LIMIT,
    ):
        from xbrowse_server.base.models import Individual
        from xbrowse_server.mall import get_reference

        cache_key = "Variants___%s___%s___%s" % (project_id, family_id,
                                                 json.dumps([
                                                     variant_filter.toJSON() if
                                                     variant_filter else None,
                                                     genotype_filter,
                                                     quality_filter,
                                                     variant_id_filter,
                                                     indivs_to_consider,
                                                     include_all_consequences,
                                                 ]))

        cached_results = self._redis_client and self._redis_client.get(
            cache_key)
        if cached_results is not None:
            variant_results = json.loads(cached_results)
            return [
                Variant.fromJSON(variant_json)
                for variant_json in variant_results
            ]

        if indivs_to_consider is None:
            if genotype_filter:
                indivs_to_consider = genotype_filter.keys()
            else:
                indivs_to_consider = []

        if family_id is not None:
            family_individual_ids = [
                i.indiv_id for i in Individual.objects.filter(
                    family__family_id=family_id).only("indiv_id")
            ]
        else:
            family_individual_ids = [
                i.indiv_id for i in Individual.objects.filter(
                    family__project__project_id=project_id).only("indiv_id")
            ]

        from xbrowse_server.base.models import Project, Family
        from pyliftover.liftover import LiftOver

        query_json = self._make_db_query(genotype_filter, variant_filter)

        try:
            if self.liftover_grch38_to_grch37 is None:
                self.liftover_grch38_to_grch37 = LiftOver('hg38', 'hg19')

            if self.liftover_grch37_to_grch38 is None:
                self.liftover_grch37_to_grch38 = None  # LiftOver('hg19', 'hg38')
        except Exception as e:
            logger.info(
                "WARNING: Unable to set up liftover. Is there a working internet connection? "
                + str(e))

        if family_id is None:
            project = Project.objects.get(project_id=project_id)
            elasticsearch_index = project.get_elasticsearch_index()
            logger.info("Searching in project elasticsearch index: " +
                        str(elasticsearch_index))
        else:
            family = Family.objects.get(project__project_id=project_id,
                                        family_id=family_id)
            elasticsearch_index = family.get_elasticsearch_index()
            project = family.project
            logger.info("Searching in family elasticsearch index: " +
                        str(elasticsearch_index))

        if family_id is not None and len(family_individual_ids) > 0:
            # figure out which index to use
            # TODO add caching
            matching_indices = []
            mapping = self._es_client.indices.get_mapping(
                str(elasticsearch_index) + "*")

            if family_individual_ids:
                indiv_id = _encode_name(family_individual_ids[0])
                for index_name, index_mapping in mapping.items():
                    if indiv_id + "_num_alt" in index_mapping["mappings"][
                            "variant"]["properties"]:
                        matching_indices.append(index_name)

            if not matching_indices:
                if not family_individual_ids:
                    logger.error("no individuals found for family %s" %
                                 (family_id))
                elif not mapping:
                    logger.error(
                        "no es mapping found for found with prefix %s" %
                        (elasticsearch_index))
                else:
                    logger.error("%s not found in %s:\n%s" %
                                 (indiv_id, elasticsearch_index,
                                  pformat(index_mapping["mappings"]["variant"]
                                          ["properties"])))
            else:
                logger.info("matching indices: " + str(elasticsearch_index))
                elasticsearch_index = ",".join(matching_indices)

        s = elasticsearch_dsl.Search(using=self._es_client,
                                     index=str(elasticsearch_index) +
                                     "*")  #",".join(indices))

        if variant_id_filter is not None:
            variant_id_filter_term = None
            for variant_id in variant_id_filter:
                q_obj = Q('term', **{"variantId": variant_id})
                if variant_id_filter_term is None:
                    variant_id_filter_term = q_obj
                else:
                    variant_id_filter_term |= q_obj
            s = s.filter(variant_id_filter_term)

        if indivs_to_consider:
            atleast_one_nonref_genotype_filter = None
            for sample_id in indivs_to_consider:
                encoded_sample_id = _encode_name(sample_id)
                q = Q('range', **{encoded_sample_id + "_num_alt": {'gte': 1}})
                if atleast_one_nonref_genotype_filter is None:
                    atleast_one_nonref_genotype_filter = q
                else:
                    atleast_one_nonref_genotype_filter |= q

            s = s.filter(atleast_one_nonref_genotype_filter)

        if quality_filter is not None and indivs_to_consider:
            #'vcf_filter': u'pass', u'min_ab': 17, u'min_gq': 46
            min_ab = quality_filter.get('min_ab')
            if min_ab is not None:
                min_ab /= 100.0  # convert to fraction
            min_gq = quality_filter.get('min_gq')
            vcf_filter = quality_filter.get('vcf_filter')
            for sample_id in indivs_to_consider:
                encoded_sample_id = _encode_name(sample_id)

                #'vcf_filter': u'pass', u'min_ab': 17, u'min_gq': 46
                if min_ab:
                    s = s.filter(
                        ~Q('term', **{encoded_sample_id + "_num_alt": 1})
                        | Q('range', **
                            {encoded_sample_id + "_ab": {
                                'gte': min_ab
                            }}))
                    #logger.info("### ADDED FILTER: " + str({encoded_sample_id+"_ab": {'gte': min_ab}}))
                if min_gq:
                    s = s.filter(
                        'range',
                        **{encoded_sample_id + "_gq": {
                            'gte': min_gq
                        }})
                    #logger.info("### ADDED FILTER: " + str({encoded_sample_id+"_gq": {'gte': min_gq}}))
                if vcf_filter is not None:
                    s = s.filter(~Q('exists', field='filters'))
                    #logger.info("### ADDED FILTER: " + str(~Q('exists', field='filters')))

        # parse variant query
        annotation_groups_map = ANNOTATION_GROUPS_MAP_INTERNAL if user and user.is_staff else ANNOTATION_GROUPS_MAP

        for key, value in query_json.items():
            if key == 'db_tags':
                so_annotations = query_json.get('db_tags', {}).get('$in', [])

                # handle clinvar filters
                selected_so_annotations_set = set(so_annotations)

                all_clinvar_filters_set = set(
                    annotation_groups_map.get("clinvar",
                                              {}).get("children", []))
                selected_clinvar_filters_set = all_clinvar_filters_set & selected_so_annotations_set

                all_hgmd_filters_set = set(
                    annotation_groups_map.get("hgmd", {}).get("children", []))
                selected_hgmd_filters_set = all_hgmd_filters_set & selected_so_annotations_set

                vep_consequences = list(selected_so_annotations_set -
                                        selected_clinvar_filters_set -
                                        selected_hgmd_filters_set)
                consequences_filter = Q(
                    "terms", transcriptConsequenceTerms=vep_consequences)

                if selected_clinvar_filters_set:
                    clinvar_clinical_significance_terms = set()
                    for clinvar_filter in selected_clinvar_filters_set:
                        # translate selected filters to the corresponding clinvar clinical consequence terms
                        if clinvar_filter == "pathogenic":
                            clinvar_clinical_significance_terms.update(
                                ["Pathogenic", "Pathogenic/Likely_pathogenic"])
                        elif clinvar_filter == "likely_pathogenic":
                            clinvar_clinical_significance_terms.update([
                                "Likely_pathogenic",
                                "Pathogenic/Likely_pathogenic"
                            ])
                        elif clinvar_filter == "benign":
                            clinvar_clinical_significance_terms.update(
                                ["Benign", "Benign/Likely_benign"])
                        elif clinvar_filter == "likely_benign":
                            clinvar_clinical_significance_terms.update(
                                ["Likely_benign", "Benign/Likely_benign"])
                        elif clinvar_filter == "vus_or_conflicting":
                            clinvar_clinical_significance_terms.update([
                                "Conflicting_interpretations_of_pathogenicity",
                                "Uncertain_significance", "not_provided",
                                "other"
                            ])
                        else:
                            raise ValueError("Unexpected clinvar filter: " +
                                             str(clinvar_filter))

                    consequences_filter = consequences_filter | Q(
                        "terms",
                        clinvar_clinical_significance=list(
                            clinvar_clinical_significance_terms))

                if selected_hgmd_filters_set:
                    hgmd_class = set()
                    for hgmd_filter in selected_hgmd_filters_set:
                        # translate selected filters to the corresponding hgmd clinical consequence terms
                        if hgmd_filter == "disease_causing":
                            hgmd_class.update(["DM"])
                        elif hgmd_filter == "likely_disease_causing":
                            hgmd_class.update(["DM?"])
                        elif hgmd_filter == "hgmd_other":
                            hgmd_class.update(["DP", "DFP", "FP", "FTV"])
                        else:
                            raise ValueError("Unexpected hgmd filter: " +
                                             str(hgmd_filter))

                    consequences_filter = consequences_filter | Q(
                        "terms", hgmd_class=list(hgmd_class))

                if 'intergenic_variant' in vep_consequences:
                    # for many intergenic variants VEP doesn't add any annotations, so if user selected 'intergenic_variant', also match variants where transcriptConsequenceTerms is emtpy
                    consequences_filter = consequences_filter | ~Q(
                        'exists', field='transcriptConsequenceTerms')

                s = s.filter(consequences_filter)
                #logger.info("==> transcriptConsequenceTerms: %s" % str(vep_consequences))

            if key.startswith("genotypes"):
                sample_id = ".".join(key.split(".")[1:-1])
                encoded_sample_id = _encode_name(sample_id)
                genotype_filter = value
                #logger.info("==> genotype filter: " + str(genotype_filter))
                if type(genotype_filter) == int or type(
                        genotype_filter) == basestring:
                    #logger.info("==> genotypes: %s" % str({encoded_sample_id+"_num_alt": genotype_filter}))
                    s = s.filter(
                        'term',
                        **{encoded_sample_id + "_num_alt": genotype_filter})

                elif '$gte' in genotype_filter:
                    genotype_filter = {
                        k.replace("$", ""): v
                        for k, v in genotype_filter.items()
                    }
                    s = s.filter(
                        'range',
                        **{encoded_sample_id + "_num_alt": genotype_filter})
                    #logger.info("==> genotypes: %s" % str({encoded_sample_id+"_num_alt": genotype_filter}))
                elif "$in" in genotype_filter:
                    num_alt_values = genotype_filter['$in']
                    q = Q(
                        'term',
                        **{encoded_sample_id + "_num_alt": num_alt_values[0]})
                    #logger.info("==> genotypes: %s" % str({encoded_sample_id+"_num_alt": num_alt_values[0]}))
                    for num_alt_value in num_alt_values[1:]:
                        q = q | Q(
                            'term', **
                            {encoded_sample_id + "_num_alt": num_alt_value})
                        #logger.info("==> genotypes: %s" % str({encoded_sample_id+"_num_alt": num_alt_value}))
                    s = s.filter(q)

            if key == "db_gene_ids":
                db_gene_ids = query_json.get('db_gene_ids', {})

                exclude_genes = db_gene_ids.get('$nin', [])
                gene_ids = exclude_genes or db_gene_ids.get('$in', [])

                if exclude_genes:
                    s = s.exclude("terms", geneIds=gene_ids)
                else:
                    s = s.filter("terms", geneIds=gene_ids)
                #logger.info("==> %s %s" % ("exclude" if exclude_genes else "include", "geneIds: " + str(gene_ids)))

            if key == "$or" and type(value) == list:
                q_terms = None
                for region_filter in value:
                    xpos_filters = region_filter.get("$and", {})

                    # for example: $or : [{'$and': [{'xpos': {'$gte': 12345}}, {'xpos': {'$lte': 54321}}]}]
                    xpos_filters_dict = {}
                    for xpos_filter in xpos_filters:
                        xpos_filter_setting = xpos_filter[
                            "xpos"]  # for example {'$gte': 12345} or {'$lte': 54321}
                        xpos_filters_dict.update(xpos_filter_setting)

                    xpos_filter_setting = {
                        k.replace("$", ""): v
                        for k, v in xpos_filters_dict.items()
                    }
                    q = Q('range', **{"xpos": xpos_filter_setting})
                    if q_terms is None:
                        q_terms = q
                    else:
                        q_terms |= q
                if q_terms is not None:
                    s = s.filter(q_terms)

                #logger.info("==> xpos range: " + str({"xpos": xpos_filter_setting}))

            af_key_map = {
                "db_freqs.AF": "AF",
                "db_freqs.1kg_wgs_phase3": "g1k_POPMAX_AF",
                "db_freqs.exac_v3": "exac_AF_POPMAX",
                "db_freqs.topmed": "topmed_AF",
                "db_freqs.gnomad_exomes": "gnomad_exomes_AF_POPMAX",
                "db_freqs.gnomad_genomes": "gnomad_genomes_AF_POPMAX",
                "db_freqs.gnomad-exomes2": "gnomad_exomes_AF_POPMAX",
                "db_freqs.gnomad-genomes2": "gnomad_genomes_AF_POPMAX",
            }

            if key in af_key_map:
                filter_key = af_key_map[key]
                af_filter_setting = {
                    k.replace("$", ""): v
                    for k, v in value.items()
                }
                s = s.filter(
                    Q('range', **{filter_key: af_filter_setting})
                    | ~Q('exists', field=filter_key))
                #logger.info("==> %s: %s" % (filter_key, af_filter_setting))

            ac_key_map = {
                "db_acs.AF": "AC",
                "db_acs.1kg_wgs_phase3": "g1k_AC",
                "db_acs.exac_v3": "exac_AC",
                "db_acs.topmed": "topmed_AC",
                "db_acs.gnomad_exomes": "gnomad_exomes_AC",
                "db_acs.gnomad_genomes": "gnomad_genomes_AC",
                "db_acs.gnomad-exomes2": "gnomad_exomes_AC",
                "db_acs.gnomad-genomes2": "gnomad_genomes_AC",
            }

            if key in ac_key_map:
                filter_key = ac_key_map[key]
                ac_filter_setting = {
                    k.replace("$", ""): v
                    for k, v in value.items()
                }
                s = s.filter(
                    Q('range', **{filter_key: ac_filter_setting})
                    | ~Q('exists', field=filter_key))

            hemi_key_map = {
                "db_hemi.exac_v3": "exac_AC_Hemi",
                "db_hemi.gnomad_exomes": "gnomad_exomes_Hemi",
                "db_hemi.gnomad_genomes": "gnomad_genomes_Hemi",
                "db_hemi.gnomad-exomes2": "gnomad_exomes_Hemi",
                "db_hemi.gnomad-genomes2": "gnomad_genomes_Hemi",
            }

            if key in hemi_key_map:
                filter_key = hemi_key_map[key]
                hemi_filter_setting = {
                    k.replace("$", ""): v
                    for k, v in value.items()
                }
                s = s.filter(
                    Q('range', **{filter_key: hemi_filter_setting})
                    | ~Q('exists', field=filter_key))

            hom_key_map = {
                "db_hom.exac_v3": "exac_AC_Hom",
                "db_hom.gnomad_exomes": "gnomad_exomes_Hom",
                "db_hom.gnomad_genomes": "gnomad_genomes_Hom",
                "db_hom.gnomad-exomes2": "gnomad_exomes_Hom",
                "db_hom.gnomad-genomes2": "gnomad_genomes_Hom",
            }

            if key in hom_key_map:
                filter_key = hom_key_map[key]
                hom_filter_setting = {
                    k.replace("$", ""): v
                    for k, v in value.items()
                }
                s = s.filter(
                    Q('range', **{filter_key: hom_filter_setting})
                    | ~Q('exists', field=filter_key))

            #s = s.sort("xpos")

        #logger.info("=====")
        #logger.info("FULL QUERY OBJ: " + pformat(s.__dict__))
        #logger.info("FILTERS: " + pformat(s.to_dict()))

        # https://elasticsearch-py.readthedocs.io/en/master/helpers.html#elasticsearch.helpers.scan
        start = time.time()

        s = s.params(size=max_results_limit + 1)
        #if not include_all_consequences:
        #    s = s.source(exclude=["sortedTranscriptConsequences"])
        response = s.execute()
        logger.info("=====")

        logger.info("TOTAL: %s. Query took %s seconds" %
                    (response.hits.total, time.time() - start))

        if response.hits.total > max_results_limit + 1:
            raise Exception(
                "This search matched too many variants. Please set additional filters and try again."
            )

        #print(pformat(response.to_dict()))

        project = Project.objects.get(project_id=project_id)

        #gene_list_map = project.get_gene_list_map()

        reference = get_reference()

        #for i, hit in enumerate(response.hits):
        variant_results = []
        for i, hit in enumerate(s.scan()):  # preserve_order=True
            #logger.info("HIT %s: %s %s %s" % (i, hit["variantId"], hit["geneIds"], pformat(hit.__dict__)))
            #print("HIT %s: %s" % (i, pformat(hit.to_dict())))
            filters = ",".join(hit["filters"]
                               or []) if "filters" in hit else ""
            genotypes = {}
            all_num_alt = []
            for individual_id in family_individual_ids:
                encoded_individual_id = _encode_name(individual_id)
                num_alt = int(hit["%s_num_alt" % encoded_individual_id]) if (
                    "%s_num_alt" % encoded_individual_id) in hit else -1
                if num_alt is not None:
                    all_num_alt.append(num_alt)

                alleles = []
                if num_alt == 0:
                    alleles = [hit["ref"], hit["ref"]]
                elif num_alt == 1:
                    alleles = [hit["ref"], hit["alt"]]
                elif num_alt == 2:
                    alleles = [hit["alt"], hit["alt"]]
                elif num_alt == -1 or num_alt == None:
                    alleles = []
                else:
                    raise ValueError("Invalid num_alt: " + str(num_alt))

                genotypes[individual_id] = {
                    'ab':
                    hit["%s_ab" % encoded_individual_id] if
                    ("%s_ab" % encoded_individual_id) in hit else None,
                    'alleles':
                    map(str, alleles),
                    'extras': {
                        'ad':
                        hit["%s_ab" % encoded_individual_id] if
                        ("%s_ad" % encoded_individual_id) in hit else None,
                        'dp':
                        hit["%s_dp" % encoded_individual_id] if
                        ("%s_dp" % encoded_individual_id) in hit else None,
                        #'pl': '',
                    },
                    'filter':
                    filters or "pass",
                    'gq':
                    hit["%s_gq" % encoded_individual_id] if
                    ("%s_gq" % encoded_individual_id in hit
                     and hit["%s_gq" % encoded_individual_id] is not None) else
                    '',
                    'num_alt':
                    num_alt,
                }

            if all([num_alt <= 0 for num_alt in all_num_alt]):
                #logger.info("Filtered out due to genotype: " + str(genotypes))
                #print("Filtered all_num_alt <= 0 - Result %s: GRCh38: %s:%s,  cadd: %s  %s - %s" % (i, hit["contig"], hit["start"], hit["cadd_PHRED"] if "cadd_PHRED" in hit else "", hit["transcriptConsequenceTerms"], all_num_alt))
                continue

            vep_annotation = json.loads(
                str(hit['sortedTranscriptConsequences'])
            ) if 'sortedTranscriptConsequences' in hit else None

            if project.genome_version == GENOME_VERSION_GRCh37:
                grch38_coord = None
                if self.liftover_grch37_to_grch38:
                    grch38_coord = self.liftover_grch37_to_grch38.convert_coordinate(
                        "chr%s" % hit["contig"].replace("chr", ""),
                        int(hit["start"]))
                    if grch38_coord and grch38_coord[0]:
                        grch38_coord = "%s-%s-%s-%s " % (
                            grch38_coord[0][0], grch38_coord[0][1], hit["ref"],
                            hit["alt"])
                    else:
                        grch38_coord = None
            else:
                grch38_coord = hit["variantId"]

            if project.genome_version == GENOME_VERSION_GRCh38:
                grch37_coord = None
                if self.liftover_grch38_to_grch37:
                    grch37_coord = self.liftover_grch38_to_grch37.convert_coordinate(
                        "chr%s" % hit["contig"].replace("chr", ""),
                        int(hit["start"]))
                    if grch37_coord and grch37_coord[0]:
                        grch37_coord = "%s-%s-%s-%s " % (
                            grch37_coord[0][0], grch37_coord[0][1], hit["ref"],
                            hit["alt"])
                    else:
                        grch37_coord = None
            else:
                grch37_coord = hit["variantId"]

            result = {
                #u'_id': ObjectId('596d2207ff66f729285ca588'),
                'alt':
                str(hit["alt"]) if "alt" in hit else None,
                'annotation': {
                    'fathmm':
                    fathmm_map.get(hit["dbnsfp_FATHMM_pred"].split(';')[0])
                    if "dbnsfp_FATHMM_pred" in hit
                    and hit["dbnsfp_FATHMM_pred"] else None,
                    'muttaster':
                    muttaster_map.get(
                        hit["dbnsfp_MutationTaster_pred"].split(';')[0])
                    if "dbnsfp_MutationTaster_pred" in hit
                    and hit["dbnsfp_MutationTaster_pred"] else None,
                    'polyphen':
                    polyphen_map.get(
                        hit["dbnsfp_Polyphen2_HVAR_pred"].split(';')[0])
                    if "dbnsfp_Polyphen2_HVAR_pred" in hit
                    and hit["dbnsfp_Polyphen2_HVAR_pred"] else None,
                    'sift':
                    sift_map.get(hit["dbnsfp_SIFT_pred"].split(';')[0])
                    if "dbnsfp_SIFT_pred" in hit and hit["dbnsfp_SIFT_pred"]
                    else None,
                    'GERP_RS':
                    hit["dbnsfp_GERP_RS"] if "dbnsfp_GERP_RS" in hit else None,
                    'phastCons100way_vertebrate':
                    hit["dbnsfp_phastCons100way_vertebrate"]
                    if "dbnsfp_phastCons100way_vertebrate" in hit else None,
                    'cadd_phred':
                    hit["cadd_PHRED"] if "cadd_PHRED" in hit else None,
                    'dann_score':
                    hit["dbnsfp_DANN_score"]
                    if "dbnsfp_DANN_score" in hit else None,
                    'revel_score':
                    hit["dbnsfp_REVEL_score"]
                    if "dbnsfp_REVEL_score" in hit else None,
                    'eigen_phred':
                    hit["eigen_Eigen_phred"] if "eigen_Eigen_phred" in hit else
                    (hit["dbnsfp_Eigen_phred"]
                     if "dbnsfp_Eigen_phred" in hit else None),
                    'mpc_score':
                    hit["mpc_MPC"] if "mpc_MPC" in hit else None,
                    'annotation_tags':
                    list(hit["transcriptConsequenceTerms"] or [])
                    if "transcriptConsequenceTerms" in hit else None,
                    'coding_gene_ids':
                    list(hit['codingGeneIds'] or []),
                    'gene_ids':
                    list(hit['geneIds'] or []),
                    'vep_annotation':
                    vep_annotation,
                    'vep_group':
                    str(hit['mainTranscript_major_consequence'] or ""),
                    'vep_consequence':
                    str(hit['mainTranscript_major_consequence'] or ""),
                    'main_transcript': {
                        k.replace('mainTranscript_', ''): hit[k]
                        for k in dir(hit) if k.startswith('mainTranscript_')
                    },
                    'worst_vep_annotation_index':
                    0,
                    'worst_vep_index_per_gene': {
                        str(hit['mainTranscript_gene_id']): 0
                    },
                },
                'chr':
                hit["contig"],
                'coding_gene_ids':
                list(hit['codingGeneIds'] or []),
                'gene_ids':
                list(hit['geneIds'] or []),
                'coverage': {
                    'gnomad_exome_coverage':
                    float(hit["gnomad_exome_coverage"] or -1)
                    if "gnomad_exome_coverage" in hit else -1,
                    'gnomad_genome_coverage':
                    float(hit["gnomad_genome_coverage"] or -1)
                    if "gnomad_genome_coverage" in hit else -1,
                },
                'pop_counts': {
                    'AC':
                    int(hit['AC'] or 0) if 'AC' in hit else None,
                    'AN':
                    int(hit['AN'] or 0) if 'AN' in hit else None,
                    '1kg_AC':
                    int(hit['g1k_AC'] or 0) if 'g1k_AC' in hit else None,
                    '1kg_AN':
                    int(hit['g1k_AN'] or 0) if 'g1k_AN' in hit else None,
                    'exac_v3_AC':
                    int(hit["exac_AC_Adj"] or 0)
                    if "exac_Adj_AC" in hit else None,
                    'exac_v3_Het':
                    int(hit["exac_AC_Het"] or 0)
                    if "exac_AC_Het" in hit else None,
                    'exac_v3_Hom':
                    int(hit["exac_AC_Hom"] or 0)
                    if "exac_AC_Hom" in hit else None,
                    'exac_v3_Hemi':
                    int(hit["exac_AC_Hemi"] or 0)
                    if "exac_AC_Hemi" in hit else None,
                    'gnomad_exomes_AC':
                    int(hit["gnomad_exomes_AC"] or 0)
                    if "gnomad_exomes_AC" in hit else None,
                    'gnomad_exomes_Hom':
                    int(hit["gnomad_exomes_Hom"] or 0)
                    if "gnomad_exomes_Hom" in hit else None,
                    'gnomad_exomes_Hemi':
                    int(hit["gnomad_exomes_Hemi"] or 0)
                    if "gnomad_exomes_Hemi" in hit else None,
                    'gnomad_exomes_AN':
                    int(hit["gnomad_exomes_AN"] or 0)
                    if "gnomad_exomes_AN" in hit else None,
                    'gnomad_genomes_AC':
                    int(hit["gnomad_genomes_AC"] or 0)
                    if "gnomad_genomes_AC" in hit else None,
                    'gnomad_genomes_Hom':
                    int(hit["gnomad_genomes_Hom"] or 0)
                    if "gnomad_genomes_Hom" in hit else None,
                    'gnomad_genomes_Hemi':
                    int(hit["gnomad_genomes_Hemi"] or 0)
                    if "gnomad_genomes_Hemi" in hit else None,
                    'gnomad_genomes_AN':
                    int(hit["gnomad_genomes_AN"] or 0)
                    if "gnomad_genomes_AN" in hit else None,
                    'topmed_AC':
                    float(hit["topmed_AC"] or 0)
                    if "topmed_AC" in hit else None,
                    'topmed_Het':
                    float(hit["topmed_Het"] or 0)
                    if "topmed_Het" in hit else None,
                    'topmed_Hom':
                    float(hit["topmed_Hom"] or 0)
                    if "topmed_Hom" in hit else None,
                    'topmed_AN':
                    float(hit["topmed_AN"] or 0)
                    if "topmed_AN" in hit else None,
                },
                'db_freqs': {
                    'AF':
                    float(hit["AF"] or 0.0) if "AF" in hit else None,
                    '1kg_wgs_AF':
                    float(hit["g1k_AF"] or 0.0) if "g1k_AF" in hit else None,
                    '1kg_wgs_popmax_AF':
                    float(hit["g1k_POPMAX_AF"] or 0.0)
                    if "g1k_POPMAX_AF" in hit else None,
                    'exac_v3_AF':
                    float(hit["exac_AF"] or 0.0) if "exac_AF" in hit else
                    (hit["exac_AC_Adj"] / float(hit["exac_AN_Adj"])
                     if "exac_AC_Adj" in hit and "exac_AN_Adj" in hit
                     and int(hit["exac_AN_Adj"] or 0) > 0 else None),
                    'exac_v3_popmax_AF':
                    float(hit["exac_AF_POPMAX"] or 0.0)
                    if "exac_AF_POPMAX" in hit else None,
                    'gnomad_exomes_AF':
                    float(hit["gnomad_exomes_AF"] or 0.0)
                    if "gnomad_exomes_AF" in hit else None,
                    'gnomad_exomes_popmax_AF':
                    float(hit["gnomad_exomes_AF_POPMAX"] or 0.0)
                    if "gnomad_exomes_AF_POPMAX" in hit else None,
                    'gnomad_genomes_AF':
                    float(hit["gnomad_genomes_AF"] or 0.0)
                    if "gnomad_genomes_AF" in hit else None,
                    'gnomad_genomes_popmax_AF':
                    float(hit["gnomad_genomes_AF_POPMAX"] or 0.0)
                    if "gnomad_genomes_AF_POPMAX" in hit else None,
                    'topmed_AF':
                    float(hit["topmed_AF"] or 0.0)
                    if "topmed_AF" in hit else None,
                },
                #'popmax_populations': {
                #    'exac_popmax': hit["exac_POPMAX"] or None,
                #    'gnomad_exomes_popmax': hit["gnomad_exomes_POPMAX"] or None,
                #    'gnomad_genomes_popmax': hit["gnomad_genomes_POPMAX"] or None,
                #},
                'db_gene_ids':
                list((hit["geneIds"] or []) if "geneIds" in hit else []),
                'db_tags':
                str(hit["transcriptConsequenceTerms"] or "")
                if "transcriptConsequenceTerms" in hit else None,
                'extras': {
                    'clinvar_variant_id':
                    hit['clinvar_variation_id']
                    if 'clinvar_variation_id' in hit
                    and hit['clinvar_variation_id'] else None,
                    'clinvar_allele_id':
                    hit['clinvar_allele_id'] if 'clinvar_allele_id' in hit
                    and hit['clinvar_allele_id'] else None,
                    'clinvar_clinsig':
                    hit['clinvar_clinical_significance'].lower() if
                    ('clinvar_clinical_significance' in hit)
                    and hit['clinvar_clinical_significance'] else None,
                    'hgmd_class':
                    hit['hgmd_class'] if 'hgmd_class' in hit and user
                    and user.is_staff else None,
                    'hgmd_accession':
                    hit['hgmd_accession'] if 'hgmd_accession' in hit else None,
                    'genome_version':
                    project.genome_version,
                    'grch37_coords':
                    grch37_coord,
                    'grch38_coords':
                    grch38_coord,
                    'alt_allele_pos':
                    0,
                    'orig_alt_alleles':
                    map(str,
                        [a.split("-")[-1] for a in hit["originalAltAlleles"]])
                    if "originalAltAlleles" in hit else None
                },
                'genotypes':
                genotypes,
                'pos':
                long(hit['start']),
                'pos_end':
                str(hit['end']),
                'ref':
                str(hit['ref']),
                'vartype':
                'snp' if len(hit['ref']) == len(hit['alt']) else "indel",
                'vcf_id':
                None,
                'xpos':
                long(hit["xpos"]),
                'xposx':
                long(hit["xpos"]),
            }

            result["annotation"]["freqs"] = result["db_freqs"]
            result["annotation"]["pop_counts"] = result["pop_counts"]
            result["annotation"]["db"] = "elasticsearch"

            result["extras"][
                "svlen"] = hit["SVLEN"] if "SVLEN" in hit else None
            result["extras"][
                "svtype"] = hit["SVTYPE"] if "SVTYPE" in hit else None

            logger.info(
                "Result %s: GRCh37: %s GRCh38: %s:,  cadd: %s  %s - gene ids: %s, coding gene_ids: %s"
                % (i, grch37_coord, grch38_coord,
                   hit["cadd_PHRED"] if "cadd_PHRED" in hit else "",
                   hit["transcriptConsequenceTerms"], result["gene_ids"],
                   result["coding_gene_ids"]))

            result["extras"]["project_id"] = project_id
            result["extras"]["family_id"] = family_id

            # add gene info
            gene_names = {}
            if vep_annotation is not None:
                gene_names = {
                    vep_anno["gene_id"]: vep_anno.get("gene_symbol")
                    for vep_anno in vep_annotation
                    if vep_anno.get("gene_symbol")
                }
            result["extras"]["gene_names"] = gene_names

            try:
                genes = {}
                for gene_id in result["coding_gene_ids"]:
                    if gene_id:
                        genes[gene_id] = reference.get_gene_summary(
                            gene_id) or {}

                if not genes:
                    for gene_id in result["gene_ids"]:
                        if gene_id:
                            genes[gene_id] = reference.get_gene_summary(
                                gene_id) or {}

                #if not genes:
                #    genes =  {vep_anno["gene_id"]: {"symbol": vep_anno["gene_symbol"]} for vep_anno in vep_annotation}

                result["extras"]["genes"] = genes
            except Exception as e:
                exc_type, exc_obj, exc_tb = sys.exc_info()
                logger.warn(
                    "WARNING: got unexpected error in add_gene_names_to_variants: %s : line %s"
                    % (e, exc_tb.tb_lineno))

            variant_results.append(result)

        logger.info("Finished returning the %s variants: %s seconds" %
                    (response.hits.total, time.time() - start))

        if self._redis_client:
            self._redis_client.set(cache_key, json.dumps(variant_results))

        return [
            Variant.fromJSON(variant_json) for variant_json in variant_results
        ]
예제 #39
0
파일: views.py 프로젝트: rpete/seqr
def add_or_edit_variant_tags(request):

    family = None
    if 'family_id' in request.GET:
        project, family = get_project_and_family_for_user(request.user, request.GET)
    else:
        project = utils.get_project_for_user(request.user, request.GET)

    form = api_forms.VariantTagsForm(project, request.GET)
    if not form.is_valid():
        ret = {
            'is_error': True,
            'error': server_utils.form_error_string(form)
        }
        return JSONResponse(ret)

    variant = get_datastore(project.project_id).get_single_variant(
            project.project_id,
            family.family_id,
            form.cleaned_data['xpos'],
            form.cleaned_data['ref'],
            form.cleaned_data['alt'],
    )

    if not variant:
        variant = Variant(form.cleaned_data['xpos'], form.cleaned_data['ref'], form.cleaned_data['alt'])

    variant_tags_to_delete = {
        variant_tag.id: variant_tag for variant_tag in VariantTag.objects.filter(
            family=family,
            xpos=form.cleaned_data['xpos'],
            ref=form.cleaned_data['ref'],
            alt=form.cleaned_data['alt'])
    }

    project_tag_events = {}
    for project_tag in form.cleaned_data['project_tags']:
        # retrieve tags
        tag, created = VariantTag.objects.get_or_create(
            project_tag=project_tag,
            family=family,
            xpos=form.cleaned_data['xpos'],
            ref=form.cleaned_data['ref'],
            alt=form.cleaned_data['alt'],
        )

        if not created:
            # this tag already exists so just keep it (eg. remove it from the set of tags that will be deleted)
            del variant_tags_to_delete[tag.id]
            continue

        # this a new tag, so update who saved it and when
        project_tag_events[project_tag] = "add_variant_tag"

        tag.user = request.user
        tag.date_saved = timezone.now()
        tag.search_url = form.cleaned_data['search_url']
        tag.save()

    # delete the tags that are no longer checked.
    for variant_tag in variant_tags_to_delete.values():
        project_tag_events[variant_tag.project_tag] = "delete_variant_tag"
        variant_tag.delete()


    # add the extra info after updating the tag info in the database, so that the new tag info is added to the variant JSON
    add_extra_info_to_variants_family(get_reference(), family, [variant,])

    # log tag creation
    for project_tag, event_type in project_tag_events.items():
        try:
            settings.EVENTS_COLLECTION.insert({
                'event_type': event_type,
                'date': timezone.now(),
                'project_id': ''.join(project.project_id),
                'family_id': family.family_id,
                'tag': project_tag.tag,
                'title': project_tag.title,

                'xpos':form.cleaned_data['xpos'],
                'pos':variant.pos,
                'chrom': variant.chr,
                'ref':form.cleaned_data['ref'],
                'alt':form.cleaned_data['alt'],
                'gene_names': ", ".join(variant.extras['gene_names'].values()),
                'username': request.user.username,
                'email': request.user.email,
                'search_url': form.cleaned_data.get('search_url'),
            })
        except Exception as e:
            logging.error("Error while logging add_variant_tag event: %s" % e)

    return JSONResponse({
        'is_error': False,
        'variant': variant.toJSON(),
    })