aboutsummaryrefslogtreecommitdiff
path: root/workflows
diff options
context:
space:
mode:
Diffstat (limited to 'workflows')
-rw-r--r--workflows/pull-data/genbank/genbank.py36
-rw-r--r--workflows/pull-data/genbank/ref.py237
2 files changed, 66 insertions, 207 deletions
diff --git a/workflows/pull-data/genbank/genbank.py b/workflows/pull-data/genbank/genbank.py
index 90f5a14..80432de 100644
--- a/workflows/pull-data/genbank/genbank.py
+++ b/workflows/pull-data/genbank/genbank.py
@@ -52,18 +52,28 @@ Example of an output JSON:
}
}
-Note: missing data should be None! Do not fill in other data by
-'guessing'.
-
-When data is malformed an warning should be issued.
"""
def get_metadata(id, gbseq):
+ """This is a minimal data parser from genbank XML records. Inference
+ on, for example geo location, is not allowed in this function and
+ happens downstream.
+
+ That is to keep the parsing simple.
+
+ Important: missing data should be missing or None! Do not fill in
+ data by 'guessing'.
+
+ When data is malformed a warning should be logged and added to the
+ warning list.
+
+ """
host = types.SimpleNamespace()
sample = types.SimpleNamespace()
submitter = types.SimpleNamespace()
technology = types.SimpleNamespace()
+ virus = types.SimpleNamespace()
warnings = []
def warn(msg):
@@ -81,13 +91,14 @@ def get_metadata(id, gbseq):
sample.sample_id = id
sample.database = "https://www.ncbi.nlm.nih.gov/genbank/"
sample.source_database_accession = f"http://identifiers.org/insdc/{id}#sequence"
- # <GBQualifier>
- # <GBQualifier_name>country</GBQualifier_name>
# <GBQualifier_value>USA: Cruise_Ship_1, California</GBQualifier_value>
- # </GBQualifier>
- sample.collection_location = "FIXME"
+ n = fetch("host_species", ".//GBQualifier/GBQualifier_name/[.='country']/../GBQualifier_value")
+ if n: sample.collection_location = n
+ else: warn("Missing collection_location")
submitter.authors = [n.text for n in gbseq.findall(".//GBAuthor")]
+ if not len(submitter.authors): warn("Missing authors")
+
# <GBReference_journal>Submitted (28-OCT-2020) MDU-PHL, The Peter
# Doherty Institute for Infection and Immunity, 792 Elizabeth
# Street, Melbourne, Vic 3000, Australia
@@ -103,7 +114,6 @@ def get_metadata(id, gbseq):
pass
except ValueError:
submitter.additional_submitter_information = n
- pass
try:
n = gbseq.find("./GBSeq_comment").text
@@ -122,6 +132,7 @@ def get_metadata(id, gbseq):
p = re.compile(r'.*Sequencing Technology :: ([^;]+).*')
m = p.match(n)
if m: technology.sample_sequencing_technology = m.group(1).strip()
+ else: warn("Missing sample_sequencing_technology")
# --- Dates
n = gbseq.find("./GBSeq_create-date")
@@ -171,13 +182,18 @@ def get_metadata(id, gbseq):
host.host_age = int(m.group(1))
host.host_age_unit = 'http://purl.obolibrary.org/obo/UO_0000036'
# sys.exit(1)
+ n = fetch("virus_strain", ".//GBQualifier/GBQualifier_name/[.='isolate']/../GBQualifier_value")
+ if n: virus.virus_strain = n
+ n = fetch("virus_species", ".//GBQualifier/GBQualifier_name/[.='db_xref']/../GBQualifier_value")
+ if n: virus.virus_species = "http://purl.obolibrary.org/obo/NCBITaxon_"+n.split('taxon:')[1]
+
info = {
'id': 'placeholder',
'update_date': str(update_date),
'host': host.__dict__,
'sample': sample.__dict__,
- #'virus': virus,
+ 'virus': virus.__dict__,
'technology': technology.__dict__,
'submitter': submitter.__dict__,
'warnings': warnings,
diff --git a/workflows/pull-data/genbank/ref.py b/workflows/pull-data/genbank/ref.py
index a9b4f06..4d4df48 100644
--- a/workflows/pull-data/genbank/ref.py
+++ b/workflows/pull-data/genbank/ref.py
@@ -1,199 +1,42 @@
# ---- BELOW IS JUST FOR REFERENCE ----
- if field_in_yaml == 'sequencing_coverage':
- # A regular expression would be better!
- try:
- technology[field_in_yaml] = [
- float(tech_info_to_parse.replace('(average)', '').replace("reads/nt", '').
- replace('(average for 6 sequences)', '').replace(',', '.').strip(' xX>'))
- ]
- except ValueError:
- print(accession_version, "Couldn't make sense of Coverage '%s'" % tech_info_to_parse)
- pass
- elif field_in_yaml == 'sample_sequencing_technology':
- new_seq_tec_list = []
- for seq_tec in tech_info_to_parse.split(';'):
- seq_tec = seq_tec.strip()
- if seq_tec in field_to_term_to_uri_dict['ncbi_sequencing_technology']:
- seq_tec = field_to_term_to_uri_dict['ncbi_sequencing_technology'][seq_tec]
- new_seq_tec_list.append(seq_tec)
- else:
- missing_value_list.append('\t'.join([accession_version, 'sample_sequencing_technology', seq_tec]))
-
- if len(new_seq_tec_list) > 0:
- technology['sample_sequencing_technology'] = [x for x in new_seq_tec_list]
- else:
- technology[field_in_yaml] = tech_info_to_parse
-
-
-
-
- elif 'gender' in GBQualifier_value_text_list[1]:
- host_sex_one_lecter = GBQualifier_value_text_list[1].split(':')[-1].strip()
- if host_sex_one_lecter in ['F', 'M']:
- host_sex = 'female' if host_sex_one_lecter == 'F' else 'male'
-
- if host_sex in ['male', 'female']:
- host['host_sex'] = "http://purl.obolibrary.org/obo/PATO_0000384" if host_sex == 'male' else "http://purl.obolibrary.org/obo/PATO_0000383"
- elif GBQualifier_value_text_list[1] in field_to_term_to_uri_dict['ncbi_host_health_status']:
- host['host_health_status'] = field_to_term_to_uri_dict['ncbi_host_health_status'][GBQualifier_value_text_list[1]]
- else:
- missing_value_list.append('\t'.join([accession_version, 'host_sex or host_health_status', GBQualifier_value_text_list[1]]))
-
- # Host age
- host_age = -1
- if len(GBQualifier_value_text_list[1].split(' ')) > 1 and is_integer(GBQualifier_value_text_list[1].split(' ')[-1]):
- host_age = int(GBQualifier_value_text_list[1].split(' ')[-1])
- elif len(GBQualifier_value_text_list) > 2 and is_integer(GBQualifier_value_text_list[2].split(' ')[-1]):
- host_age = int(GBQualifier_value_text_list[2].split(' ')[-1])
-
- if host_age >= 0 and host_age < 110:
- host['host_age'] = host_age
- host['host_age_unit'] = 'http://purl.obolibrary.org/obo/UO_0000036'
- elif len(GBQualifier_value_text_list) > 2:
- missing_value_list.append('\t'.join([accession_version, 'host_age', GBQualifier_value_text_list[2]]))
- elif GBQualifier_name_text == 'collected_by':
- if any([x in GBQualifier_value_text.lower() for x in ['institute', 'hospital', 'city', 'center']]):
- sample['collecting_institution'] = GBQualifier_value_text
- else:
- sample['collector_name'] = GBQualifier_value_text
- elif GBQualifier_name_text == 'isolation_source':
- if GBQualifier_value_text.upper() in field_to_term_to_uri_dict['ncbi_speciesman_source']:
- GBQualifier_value_text = GBQualifier_value_text.upper() # For example, in case of 'usa: wa'
-
- # Little cleaning
- GBQualifier_value_text = GBQualifier_value_text.strip("/'")
-
- if GBQualifier_value_text in field_to_term_to_uri_dict['ncbi_speciesman_source']:
- sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source'][GBQualifier_value_text]]
- else:
- if GBQualifier_value_text.lower() in ['np/op', 'np-op', 'np/op swab', 'np/np swab', 'nasopharyngeal and oropharyngeal swab', 'nasopharyngeal/oropharyngeal swab', 'combined nasopharyngeal and oropharyngeal swab', 'naso and/or oropharyngeal swab']:
- sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasopharyngeal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['oropharyngeal swab']]
- elif GBQualifier_value_text.lower() in ['nasopharyngeal swab/throat swab', 'nasopharyngeal/throat swab', 'nasopharyngeal swab and throat swab', 'nasal swab and throat swab', 'nasopharyngeal aspirate/throat swab', 'Nasopharyngeal/Throat']:
- sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasopharyngeal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['throat swab']]
- elif GBQualifier_value_text.lower() in ['nasopharyngeal aspirate & throat swab', 'nasopharyngeal aspirate and throat swab']:
- sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasopharyngeal aspirate'], field_to_term_to_uri_dict['ncbi_speciesman_source']['throat swab']]
- elif GBQualifier_value_text.lower() in ['nasal swab and throat swab']:
- sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['throat swab']]
- elif GBQualifier_value_text.lower() in ['nasal-swab and oro-pharyngeal swab']:
- sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['oropharyngeal swab']]
- else:
- missing_value_list.append('\t'.join([accession_version, 'specimen_source', GBQualifier_value_text]))
- elif GBQualifier_name_text == 'collection_date':
- # TO_DO: which format we will use?
- date_to_write = GBQualifier_value_text
-
- if len(GBQualifier_value_text.split('-')) == 1:
- if int(GBQualifier_value_text) < 2020:
- date_to_write = "{}-12-15".format(GBQualifier_value_text)
- else:
- date_to_write = "{}-01-15".format(GBQualifier_value_text)
-
- if 'additional_collection_information' in sample:
- sample['additional_collection_information'] += "; The 'collection_date' is estimated (the original date was: {})".format(GBQualifier_value_text)
- else:
- sample['additional_collection_information'] = "The 'collection_date' is estimated (the original date was: {})".format(GBQualifier_value_text)
- elif len(GBQualifier_value_text.split('-')) == 2:
- date_to_write = parse(GBQualifier_value_text).strftime('%Y-%m') + '-15'
-
- if 'additional_collection_information' in sample:
- sample['additional_collection_information'] += "; The 'collection_date' is estimated (the original date was: {})".format(GBQualifier_value_text)
- else:
- sample['additional_collection_information'] = "The 'collection_date' is estimated (the original date was: {})".format(GBQualifier_value_text)
- elif len(GBQualifier_value_text.split('-')) == 3:
- GBQualifier_value_text_list = GBQualifier_value_text.split('-')
-
- if GBQualifier_value_text_list[1].isalpha():
- date_to_write = parse(GBQualifier_value_text).strftime('%Y-%m-%d')
-
- sample['collection_date'] = date_to_write
- elif GBQualifier_name_text in ['lat_lon', 'country']:
- if GBQualifier_name_text == 'country' and ': ' in GBQualifier_value_text:
- GBQualifier_value_text = GBQualifier_value_text.replace(': ', ':')
-
- if GBQualifier_value_text in field_to_term_to_uri_dict['ncbi_countries']:
- sample['collection_location'] = field_to_term_to_uri_dict['ncbi_countries'][GBQualifier_value_text]
- else:
- missing_value_list.append('\t'.join([accession_version, GBQualifier_name_text, GBQualifier_value_text]))
- elif GBQualifier_name_text == 'note':
- if 'additional_collection_information' in sample:
- sample['additional_collection_information'] += '; ' + GBQualifier_value_text
- else:
- sample['additional_collection_information'] = GBQualifier_value_text
- elif GBQualifier_name_text == 'isolate':
- virus['virus_strain'] = GBQualifier_value_text
- elif GBQualifier_name_text == 'db_xref':
- virus['virus_species'] = "http://purl.obolibrary.org/obo/NCBITaxon_"+GBQualifier_value_text.split('taxon:')[1]
-
- # Check if mandatory fields are missing
- if 'sample_sequencing_technology' not in technology:
- # print(accession_version, ' - technology not found')
- if accession_version not in not_created_accession_dict:
- not_created_accession_dict[accession_version] = []
- not_created_accession_dict[accession_version].append('sample_sequencing_technology not found')
-
- if 'collection_location' not in sample:
- if accession_version not in not_created_accession_dict:
- not_created_accession_dict[accession_version] = []
- not_created_accession_dict[accession_version].append('collection_location not found')
-
- if 'collection_date' not in sample:
- if accession_version not in not_created_accession_dict:
- not_created_accession_dict[accession_version] = []
- not_created_accession_dict[accession_version].append('collection_date not found')
- else:
- year, month, day = [int(x) for x in sample['collection_date'].split('-')]
-
- collection_date_in_yaml = datetime(year, month, day)
- if collection_date_in_yaml < min_acceptable_collection_date:
- if accession_version not in not_created_accession_dict:
- not_created_accession_dict[accession_version] = []
- not_created_accession_dict[accession_version].append('collection_date too early')
-
- if 'authors' not in submitter:
- if accession_version not in not_created_accession_dict:
- not_created_accession_dict[accession_version] = []
- not_created_accession_dict[accession_version].append('authors not found')
-
- if 'host_species' not in host:
- if accession_version not in not_created_accession_dict:
- not_created_accession_dict[accession_version] = []
- not_created_accession_dict[accession_version].append('host_species not found')
-
- if len(GBSeq_sequence.text) < min_len_to_count:
- if accession_version not in not_created_accession_dict:
- not_created_accession_dict[accession_version] = []
- not_created_accession_dict[accession_version].append('sequence shorter than {} bp'.format(min_len_to_count))
-
- if accession_version not in not_created_accession_dict:
- num_seq_with_len_ge_X_bp += 1
-
- # with open(os.path.join(dir_fasta_and_yaml, '{}.fasta'.format(accession_version)), 'w') as fw:
- # fw.write('>{}\n{}'.format(accession_version, GBSeq_sequence.text.upper()))
-
- with open(os.path.join(dir_fasta_and_yaml, '{}.yaml'.format(accession_version)), 'w') as fw:
- json.dump(info, fw, indent=2)
- except:
- print("Unexpected error for the ID {}: {}".format(accession_version, sys.exc_info()[0]))
- accession_with_errors_list.append(accession_version)
- continue
-
-if len(missing_value_list) > 0:
- path_missing_terms_tsv = 'missing_terms.genbank.tsv'
- print('Written missing terms in {}'.format(path_missing_terms_tsv))
- with open(path_missing_terms_tsv, 'w') as fw:
- fw.write('\n'.join(missing_value_list))
-
-if len(accession_with_errors_list) > 0:
- path_accession_with_errors_tsv = 'accession_with_errors.genbank.tsv'
- print('Written the accession with errors in {}'.format(path_accession_with_errors_tsv))
- with open(path_accession_with_errors_tsv, 'w') as fw:
- fw.write('\n'.join(accession_with_errors_list))
-
-if len(not_created_accession_dict) > 0:
- path_not_created_accession_tsv = 'not_created_accession.genbank.tsv'
- print('Written not created accession in {}'.format(path_not_created_accession_tsv))
- with open(path_not_created_accession_tsv, 'w') as fw:
- fw.write('\n'.join(['\t'.join([accession_version, ','.join(missing_info_list)]) for accession_version, missing_info_list in not_created_accession_dict.items()]))
-
-print('Num. new sequences with length >= {} bp: {}'.format(min_len_to_count, num_seq_with_len_ge_X_bp))
+if field_in_yaml == 'sequencing_coverage':
+ # A regular expression would be better!
+ try:
+ technology[field_in_yaml] = [
+ float(tech_info_to_parse.replace('(average)', '').replace("reads/nt", '').
+ replace('(average for 6 sequences)', '').replace(',', '.').strip(' xX>'))
+ ]
+ except ValueError:
+ print(accession_version, "Couldn't make sense of Coverage '%s'" % tech_info_to_parse)
+ pass
+
+
+elif GBQualifier_name_text == 'collected_by':
+ if any([x in GBQualifier_value_text.lower() for x in ['institute', 'hospital', 'city', 'center']]):
+ sample['collecting_institution'] = GBQualifier_value_text
+ else:
+ sample['collector_name'] = GBQualifier_value_text
+
+elif GBQualifier_name_text == 'isolation_source':
+if GBQualifier_value_text.upper() in field_to_term_to_uri_dict['ncbi_speciesman_source']:
+ GBQualifier_value_text = GBQualifier_value_text.upper() # For example, in case of 'usa: wa'
+
+# Little cleaning
+GBQualifier_value_text = GBQualifier_value_text.strip("/'")
+
+if GBQualifier_value_text in field_to_term_to_uri_dict['ncbi_speciesman_source']:
+ sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source'][GBQualifier_value_text]]
+else:
+ if GBQualifier_value_text.lower() in ['np/op', 'np-op', 'np/op swab', 'np/np swab', 'nasopharyngeal and oropharyngeal swab', 'nasopharyngeal/oropharyngeal swab', 'combined nasopharyngeal and oropharyngeal swab', 'naso and/or oropharyngeal swab']:
+ sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasopharyngeal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['oropharyngeal swab']]
+ elif GBQualifier_value_text.lower() in ['nasopharyngeal swab/throat swab', 'nasopharyngeal/throat swab', 'nasopharyngeal swab and throat swab', 'nasal swab and throat swab', 'nasopharyngeal aspirate/throat swab', 'Nasopharyngeal/Throat']:
+ sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasopharyngeal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['throat swab']]
+ elif GBQualifier_value_text.lower() in ['nasopharyngeal aspirate & throat swab', 'nasopharyngeal aspirate and throat swab']:
+ sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasopharyngeal aspirate'], field_to_term_to_uri_dict['ncbi_speciesman_source']['throat swab']]
+ elif GBQualifier_value_text.lower() in ['nasal swab and throat swab']:
+ sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['throat swab']]
+ elif GBQualifier_value_text.lower() in ['nasal-swab and oro-pharyngeal swab']:
+ sample['specimen_source'] = [field_to_term_to_uri_dict['ncbi_speciesman_source']['nasal swab'], field_to_term_to_uri_dict['ncbi_speciesman_source']['oropharyngeal swab']]
+ else:
+ missing_value_list.append('\t'.join([accession_version, 'specimen_source', GBQualifier_value_text]))