def generate_cross_species_template(taxonomy_file_path, output_filepath):
taxon = extract_taxonomy_name_from_path(taxonomy_file_path)
taxonomy_config = read_taxonomy_config(taxon)
if taxonomy_config:
if str(taxonomy_file_path).endswith(".json"):
dend = dend_json_2_nodes_n_edges(taxonomy_file_path)
else:
dend = nomenclature_2_nodes_n_edges(taxonomy_file_path)
dend_tree = generate_dendrogram_tree(dend)
subtrees = get_subtrees(dend_tree, taxonomy_config)
cross_species_template = []
headers, cs_by_preferred_alias = read_csv_to_dict(
CROSS_SPECIES_PATH,
id_column_name="cell_set_preferred_alias",
id_to_lower=True)
headers, cs_by_aligned_alias = read_csv_to_dict(
CROSS_SPECIES_PATH,
id_column_name="cell_set_aligned_alias",
id_to_lower=True)
for o in dend['nodes']:
if o['cell_set_accession'] in set.union(
*subtrees) and (o['cell_set_preferred_alias']
or o['cell_set_additional_aliases']):
cross_species_classes = set()
if o["cell_set_aligned_alias"] and str(
o["cell_set_aligned_alias"]).lower(
) in cs_by_aligned_alias:
cross_species_classes.add(PCL_BASE + get_class_id(
cs_by_aligned_alias[str(o["cell_set_aligned_alias"]).
lower()]["cell_set_accession"]))
if "cell_set_additional_aliases" in o and o[
"cell_set_additional_aliases"]:
additional_aliases = str(
o["cell_set_additional_aliases"]).lower().split(
EXPRESSION_SEPARATOR)
for additional_alias in additional_aliases:
if additional_alias in cs_by_preferred_alias:
cross_species_classes.add(PCL_BASE + get_class_id(
cs_by_preferred_alias[additional_alias]
["cell_set_accession"]))
if len(cross_species_classes):
d = dict()
d['defined_class'] = PCL_BASE + get_class_id(
o['cell_set_accession'])
d['cross_species_classes'] = EXPRESSION_SEPARATOR.join(
cross_species_classes)
cross_species_template.append(d)
class_robot_template = pd.DataFrame.from_records(
cross_species_template)
class_robot_template.to_csv(output_filepath, sep="\t", index=False)
def test_base_class_template_generation_with_nomenclature(self):
generate_base_class_template(PATH_NOMENCLATURE_TABLE,
PATH_OUTPUT_CLASS_TSV)
output = read_tsv(PATH_OUTPUT_CLASS_TSV)
# assert only descendants of the root nodes (except root nodes itself) exist
self.assertFalse(PCL_BASE + get_class_id("CS201912131_149")
in output) # root
self.assertTrue(PCL_BASE + get_class_id("CS201912131_22")
in output) # child
self.assertTrue(PCL_BASE + get_class_id("CS201912131_70")
in output) # child
self.assertTrue(PCL_BASE + get_class_id("CS201912131_125")
in output) # root & leaf
self.assertFalse(PCL_BASE + get_class_id("CS201912131_148")
in output) # parent
def generate_homologous_to_template(taxonomy_file_path, all_base_files,
output_filepath):
"""
Homologous_to relations require a separate template. If this operation is driven by the nomenclature tables,
some dangling classes may be generated due to root classes that don't have a class and should not be aligned.
So, instead of nomenclature tables, base files are used for populating homologous to relations. This ensures all
alignments has a corresponding class.
Args:
taxonomy_file_path: path of the taxonomy file
all_base_files: paths of the all class template base files
output_filepath: template output file path
"""
taxon = extract_taxonomy_name_from_path(taxonomy_file_path)
taxonomy_config = read_taxonomy_config(taxon)
other_taxonomy_aliases = index_base_files(
[t for t in all_base_files if taxon not in t])
if taxonomy_config:
if str(taxonomy_file_path).endswith(".json"):
dend = dend_json_2_nodes_n_edges(taxonomy_file_path)
else:
dend = nomenclature_2_nodes_n_edges(taxonomy_file_path)
dend_tree = generate_dendrogram_tree(dend)
subtrees = get_subtrees(dend_tree, taxonomy_config)
data_template = []
for o in dend['nodes']:
if o['cell_set_accession'] in set.union(
*subtrees) and (o['cell_set_preferred_alias']
or o['cell_set_additional_aliases']):
d = dict()
d['defined_class'] = PCL_BASE + get_class_id(
o['cell_set_accession'])
homologous_to = list()
for other_aliases in other_taxonomy_aliases:
if "cell_set_aligned_alias" in o and o["cell_set_aligned_alias"] \
and str(o["cell_set_aligned_alias"]).lower() in other_aliases:
homologous_to.append(other_aliases[str(
o["cell_set_aligned_alias"]).lower()]
["defined_class"])
d['homologous_to'] = "|".join(homologous_to)
data_template.append(d)
robot_template = pd.DataFrame.from_records(data_template)
robot_template.to_csv(output_filepath, sep="\t", index=False)
def generate_non_taxonomy_classification_template(taxonomy_file_path,
output_filepath):
taxon = extract_taxonomy_name_from_path(taxonomy_file_path)
cell_set_accession = 3
child_cell_set_accessions = 14
nomenclature_path = os.path.join(
os.path.dirname(os.path.realpath(__file__)),
NOMENCLATURE_TABLE_PATH.format(taxon))
taxonomy_config = read_taxonomy_config(taxon)
if taxonomy_config and os.path.exists(nomenclature_path):
nomenclature_records = read_csv(nomenclature_path,
id_column=cell_set_accession)
nomenclature_template = []
non_taxo_roots = {}
for root in taxonomy_config['non_taxonomy_roots']:
non_taxo_roots[root["Node"]] = root["Cell_type"]
for record in nomenclature_records:
columns = nomenclature_records[record]
if columns[cell_set_accession] in non_taxo_roots:
# dendrogram is not mandatory for human & marmoset
# if columns[cell_set_accession] in dend_nodes:
# raise Exception("Node {} exists both in dendrogram and nomenclature of the taxonomy: {}."
# .format(columns[cell_set_accession], taxon))
children = columns[child_cell_set_accessions].split("|")
for child in children:
# child of root with cell_set_preferred_alias
if child not in non_taxo_roots and nomenclature_records[
child][0]:
d = dict()
d['defined_class'] = PCL_BASE + get_class_id(child)
d['Classification'] = non_taxo_roots[
columns[cell_set_accession]]
nomenclature_template.append(d)
class_robot_template = pd.DataFrame.from_records(nomenclature_template)
class_robot_template.to_csv(output_filepath, sep="\t", index=False)
def generate_curated_class_template(taxonomy_file_path, output_filepath):
taxon = extract_taxonomy_name_from_path(taxonomy_file_path)
taxonomy_config = read_taxonomy_config(taxon)
if taxonomy_config:
if str(taxonomy_file_path).endswith(".json"):
dend = dend_json_2_nodes_n_edges(taxonomy_file_path)
else:
dend = nomenclature_2_nodes_n_edges(taxonomy_file_path)
dend_tree = generate_dendrogram_tree(dend)
subtrees = get_subtrees(dend_tree, taxonomy_config)
class_curation_seed = [
'defined_class', 'Curated_synonyms', 'Classification',
'Classification_comment', 'Classification_pub', 'Expresses',
'Expresses_comment', 'Expresses_pub', 'Projection_type', 'Layers',
'Cross_species_text', 'Comment'
]
class_template = []
for o in dend['nodes']:
if o['cell_set_accession'] in set.union(
*subtrees) and (o['cell_set_preferred_alias']
or o['cell_set_additional_aliases']):
d = dict()
d['defined_class'] = PCL_BASE + get_class_id(
o['cell_set_accession'])
if o['cell_set_preferred_alias']:
d['prefLabel'] = o['cell_set_preferred_alias']
elif o['cell_set_additional_aliases']:
d['prefLabel'] = str(
o['cell_set_additional_aliases']).split(
EXPRESSION_SEPARATOR)[0]
for k in class_curation_seed:
if not (k in d.keys()):
d[k] = ''
class_template.append(d)
class_robot_template = pd.DataFrame.from_records(class_template)
class_robot_template.to_csv(output_filepath, sep="\t", index=False)
def test_human_mtg_ids(self):
self.assertEqual(get_class_id("CS1908210001"), "0023001")
self.assertEqual(get_class_id("CS1908210148"), "0023148")
self.assertEqual(get_individual_id("CS1908210001"), "0023501")
self.assertEqual(get_individual_id("CS1908210148"), "0023648")
def test_marmoset_ids(self):
self.assertEqual(get_class_id("CS201912132_1"), "0019001")
self.assertEqual(get_class_id("CS201912132_121"), "0019121")
self.assertEqual(get_individual_id("CS201912132_1"), "0019501")
self.assertEqual(get_individual_id("CS201912132_121"), "0019621")
def test_human_ids(self):
self.assertEqual(get_class_id("CS201912131_1"), "0015001")
self.assertEqual(get_class_id("CS201912131_121"), "0015121")
self.assertEqual(get_individual_id("CS201912131_1"), "0015501")
self.assertEqual(get_individual_id("CS201912131_121"), "0015621")
def test_mouse_ids(self):
self.assertEqual(get_class_id("CS202002013_1"), "0011001")
self.assertEqual(get_class_id("CS202002013_121"), "0011121")
self.assertEqual(get_individual_id("CS202002013_1"), "0011501")
self.assertEqual(get_individual_id("CS202002013_121"), "0011621")
def test_base_class_template_generation(self):
generate_base_class_template(PATH_MOUSE_NOMENCLATURE,
PATH_OUTPUT_CLASS_TSV)
output = read_tsv(PATH_OUTPUT_CLASS_TSV)
self.assertTrue(PCL_BASE + get_class_id("CS202002013_150")
in output) # child
# assert only descendants of the root nodes (except root nodes itself) exist
self.assertFalse(PCL_BASE + get_class_id("CS202002013_117")
in output) # root
self.assertFalse(PCL_BASE + get_class_id("CS202002013_123")
in output) # root
self.assertTrue(PCL_BASE + get_class_id("CS202002013_150")
in output) # child
self.assertTrue(PCL_BASE + get_class_id("CS202002013_124")
in output) # child
self.assertFalse(
PCL_BASE + get_class_id("CS202002013_158")
in output) # grand child, empty cell_set_preferred_alias
self.assertTrue(PCL_BASE + get_class_id("CS202002013_3")
in output) # grand child
self.assertFalse(PCL_BASE + get_class_id("CS202002013_122")
in output) # parent
self.assertFalse(PCL_BASE + get_class_id("CS202002013_120")
in output) # grand parent
self.assertTrue(PCL_BASE + get_class_id("CS202002013_103")
in output) # root & leaf
self.assertFalse(PCL_BASE + get_class_id("CS202002013_220")
in output) # parent
self.assertFalse(PCL_BASE + get_class_id("CS202002013_179")
in output) # root
self.assertFalse(PCL_BASE + get_class_id("CS202002013_180")
in output) # child, empty cell_set_preferred_alias
self.assertTrue(PCL_BASE + get_class_id("CS202002013_207")
in output) # child
self.assertFalse(
PCL_BASE + get_class_id("CS202002013_208")
in output) # grand child, empty cell_set_preferred_alias
self.assertTrue(PCL_BASE + get_class_id("CS202002013_83")
in output) # grand child
self.assertFalse(PCL_BASE + get_class_id("CS202002013_219")
in output) # parent
self.assertFalse(PCL_BASE + get_class_id("CS202002013_220")
in output) # grand parent
def test_homologous_to_template_generation(self):
generate_homologous_to_template(PATH_NOMENCLATURE_TABLE,
ALL_BASE_FILES, PATH_OUTPUT_CLASS_TSV)
output = read_tsv(PATH_OUTPUT_CLASS_TSV)
self.assertTrue(PCL_BASE + get_class_id("CS201912131_164") in output)
test_node = output[PCL_BASE + get_class_id("CS201912131_164")]
homologous_to = test_node[1].split("|")
self.assertEqual(2, len(homologous_to))
self.assertTrue(PCL_BASE +
get_class_id("CS201912132_039") in homologous_to)
self.assertTrue(PCL_BASE +
get_class_id("CS202002013_244") in homologous_to)
self.assertTrue(PCL_BASE + get_class_id("CS201912131_176") in output)
test_node = output[PCL_BASE + get_class_id("CS201912131_176")]
homologous_to = test_node[1].split("|")
self.assertEqual(2, len(homologous_to))
self.assertTrue(PCL_BASE +
get_class_id("CS201912132_060") in homologous_to)
self.assertTrue(PCL_BASE +
get_class_id("CS202002013_067") in homologous_to)
self.assertTrue(PCL_BASE + get_class_id("CS201912131_157") in output)
test_node = output[PCL_BASE + get_class_id("CS201912131_157")]
homologous_to = test_node[1].split("|")
self.assertEqual(1, len(homologous_to))
self.assertTrue(PCL_BASE +
get_class_id("CS201912132_002") in homologous_to)
self.assertTrue(PCL_BASE + get_class_id("CS201912131_142")
in output) # human Astrocyte
test_node = output[PCL_BASE + get_class_id("CS201912131_142")]
homologous_to = test_node[1]
self.assertFalse(homologous_to) # mouse and marmoset astro not exists
def generate_ind_template(taxonomy_file_path, output_filepath):
path_parts = taxonomy_file_path.split(os.path.sep)
taxon = path_parts[len(path_parts) - 1].split(".")[0]
if str(taxonomy_file_path).endswith(".json"):
dend = dend_json_2_nodes_n_edges(taxonomy_file_path)
else:
dend = nomenclature_2_nodes_n_edges(taxonomy_file_path)
taxon = path_parts[len(path_parts) - 1].split(".")[0].replace(
"nomenclature_table_", "")
dend_tree = generate_dendrogram_tree(dend)
taxonomy_config = read_taxonomy_config(taxon)
allen_descriptions = read_allen_descriptions(
ALLEN_DESCRIPTIONS_PATH, taxonomy_config['Species_abbv'][0])
subtrees = get_subtrees(dend_tree, taxonomy_config)
robot_template_seed = {
'ID': 'ID',
'Label': 'LABEL',
'PrefLabel': 'A skos:prefLabel',
'Entity Type': 'TI %',
'TYPE': 'TYPE',
'Property Assertions': "I 'subcluster of' SPLIT=|",
'Synonyms': 'A oboInOwl:hasExactSynonym SPLIT=|',
'Cluster_ID': "A 'cluster id'",
'Function': 'TI capable_of some %',
'cell_set_preferred_alias': "A n2o:cell_set_preferred_alias",
'original_label': "A n2o:original_label",
'cell_set_label': "A n2o:cell_set_label",
'cell_set_aligned_alias': "A n2o:cell_set_aligned_alias",
'cell_set_additional_aliases':
"A n2o:cell_set_additional_aliases SPLIT=|",
'cell_set_alias_assignee': "A n2o:cell_set_alias_assignee SPLIT=|",
'cell_set_alias_citation': "A n2o:cell_set_alias_citation SPLIT=|",
'Metadata': "A n2o:node_metadata",
'Exemplar_of': "TI 'exemplar data of' some %",
'Comment': "A rdfs:comment",
'Aliases': "A oboInOwl:hasRelatedSynonym SPLIT=|",
'Rank': "A 'cell_type_rank' SPLIT=|"
}
dl = [robot_template_seed]
synonym_properties = [
'cell_set_aligned_alias', 'cell_set_additional_aliases'
]
for o in dend['nodes']:
d = dict()
d['ID'] = 'PCL:' + get_individual_id(o['cell_set_accession'])
d['TYPE'] = 'owl:NamedIndividual'
d['Label'] = o['cell_set_label'] + ' - ' + o['cell_set_accession']
if 'cell_set_preferred_alias' in o and o['cell_set_preferred_alias']:
d['PrefLabel'] = o['cell_set_preferred_alias']
else:
d['PrefLabel'] = o['cell_set_accession']
d['Entity Type'] = 'PCL:0010001' # Cluster
d['Metadata'] = json.dumps(o)
d['Synonyms'] = '|'.join([
o[prop] for prop in synonym_properties
if prop in o.keys() and o[prop]
])
d['Property Assertions'] = '|'.join(
sorted([
'PCL:' + get_individual_id(e[1]) for e in dend['edges']
if e[0] == o['cell_set_accession']
]))
meta_properties = [
'cell_set_preferred_alias', 'original_label', 'cell_set_label',
'cell_set_aligned_alias', 'cell_set_additional_aliases',
'cell_set_alias_assignee', 'cell_set_alias_citation'
]
for prop in meta_properties:
if prop in o.keys():
d[prop] = '|'.join([
prop_val.strip() for prop_val in str(o[prop]).split("|")
if prop_val
])
else:
d[prop] = ''
d['Cluster_ID'] = o['cell_set_accession']
if o['cell_set_accession'] in set().union(
*subtrees) and o['cell_set_preferred_alias']:
d['Exemplar_of'] = PCL_BASE + get_class_id(o['cell_set_accession'])
if "cell_type_card" in o:
d['Rank'] = '|'.join([
cell_type.strip().replace("No", "None")
for cell_type in str(o["cell_type_card"]).split(",")
])
if o['cell_set_accession'] in allen_descriptions:
allen_data = allen_descriptions[o['cell_set_accession']]
d['Comment'] = allen_data["summary"][0]
if allen_data["aliases"][0]:
d['Aliases'] = '|'.join([
alias.strip()
for alias in str(allen_data["aliases"][0]).split("|")
])
# There should only be one!
dl.append(d)
robot_template = pd.DataFrame.from_records(dl)
robot_template.to_csv(output_filepath, sep="\t", index=False)
def generate_base_class_template(taxonomy_file_path, output_filepath):
taxon = extract_taxonomy_name_from_path(taxonomy_file_path)
taxonomy_config = read_taxonomy_config(taxon)
if taxonomy_config:
if str(taxonomy_file_path).endswith(".json"):
dend = dend_json_2_nodes_n_edges(taxonomy_file_path)
else:
dend = nomenclature_2_nodes_n_edges(taxonomy_file_path)
dend_tree = generate_dendrogram_tree(dend)
subtrees = get_subtrees(dend_tree, taxonomy_config)
if "Reference_gene_list" in taxonomy_config:
gene_db_path = ENSEMBLE_PATH.format(
str(taxonomy_config["Reference_gene_list"][0]).strip().lower())
gene_names = read_gene_data(gene_db_path)
minimal_markers = read_markers(
MARKER_PATH.format(taxon.replace("CCN", "").replace("CS", "")),
gene_names)
allen_markers = read_markers(
ALLEN_MARKER_PATH.format(
taxon.replace("CCN", "").replace("CS", "")), gene_names)
else:
minimal_markers = {}
allen_markers = {}
class_seed = [
'defined_class', 'prefLabel', 'Alias_citations',
'Synonyms_from_taxonomy', 'Gross_cell_type', 'Taxon',
'Brain_region', 'Minimal_markers', 'Allen_markers', 'Individual',
'Brain_region_abbv', 'Species_abbv', 'Cluster_ID', 'part_of',
'has_soma_location', 'aligned_alias', 'marker_gene_set'
]
class_template = []
for o in dend['nodes']:
if o['cell_set_accession'] in set.union(
*subtrees) and (o['cell_set_preferred_alias']
or o['cell_set_additional_aliases']):
d = dict()
d['defined_class'] = PCL_BASE + get_class_id(
o['cell_set_accession'])
if o['cell_set_preferred_alias']:
d['prefLabel'] = o['cell_set_preferred_alias']
elif o['cell_set_additional_aliases']:
d['prefLabel'] = str(
o['cell_set_additional_aliases']).split(
EXPRESSION_SEPARATOR)[0]
d['Synonyms_from_taxonomy'] = get_synonyms_from_taxonomy(o)
d['Gross_cell_type'] = get_gross_cell_type(
o['cell_set_accession'], subtrees, taxonomy_config)
d['Taxon'] = taxonomy_config['Species'][0]
d['Brain_region'] = taxonomy_config['Brain_region'][0]
d['Cluster_ID'] = o['cell_set_accession']
if 'cell_set_alias_citation' in o and o[
'cell_set_alias_citation']:
alias_citations = [
citation.strip() for citation in str(
o["cell_set_alias_citation"]).split("|")
if citation and citation.strip()
]
d["Alias_citations"] = "|".join(alias_citations)
if o['cell_set_accession'] in minimal_markers:
d['Minimal_markers'] = minimal_markers[
o['cell_set_accession']]
if o['cell_set_accession'] in allen_markers:
d['Allen_markers'] = allen_markers[o['cell_set_accession']]
else:
d['Allen_markers'] = ''
if 'Brain_region_abbv' in taxonomy_config:
d['Brain_region_abbv'] = taxonomy_config[
'Brain_region_abbv'][0]
if 'Species_abbv' in taxonomy_config:
d['Species_abbv'] = taxonomy_config['Species_abbv'][0]
d['Individual'] = PCL_BASE + get_individual_id(
o['cell_set_accession'])
for index, subtree in enumerate(subtrees):
if o['cell_set_accession'] in subtree:
location_rel = taxonomy_config['Root_nodes'][index][
'Location_relation']
if location_rel == "part_of":
d['part_of'] = taxonomy_config['Brain_region'][0]
d['has_soma_location'] = ''
elif location_rel == "has_soma_location":
d['part_of'] = ''
d['has_soma_location'] = taxonomy_config[
'Brain_region'][0]
if "cell_set_aligned_alias" in o and o[
"cell_set_aligned_alias"]:
d['aligned_alias'] = o["cell_set_aligned_alias"]
if o['cell_set_accession'] in minimal_markers:
d['marker_gene_set'] = PCL_PREFIX + get_marker_gene_set_id(
o['cell_set_accession'])
for k in class_seed:
if not (k in d.keys()):
d[k] = ''
class_template.append(d)
class_robot_template = pd.DataFrame.from_records(class_template)
class_robot_template.to_csv(output_filepath, sep="\t", index=False)