# Genbank XML parser
#
# Pjotr Prins (c) 2021
from collections import namedtuple
import dateutil
from dateutil.parser import parse as dateparse
from dateutil.tz import gettz
import os
import re
import sys
import types
import xml.etree.ElementTree as ET
class GBError(Exception):
pass
"""
Example of an output JSON:
{
"id": "placeholder",
"host": {
"host_species": "http://purl.obolibrary.org/obo/NCBITaxon_9606"
},
"sample": {
"sample_id": "MT890462.1",
"source_database_accession": [
"http://identifiers.org/insdc/MT890462.1#sequence"
],
"collection_location": "http://www.wikidata.org/entity/Q649",
"collection_date": "2020-04-17",
"collecting_institution": "N.A.Kovtun Clinical Hospital 1 of Departament of President Affairs",
"specimen_source": ["http://purl.obolibrary.org/obo/NCIT_C155831"]
},
"virus": {
"virus_strain": "SARS-CoV-2/human/RUS/20200417_10/2020",
"virus_species": "http://purl.obolibrary.org/obo/NCBITaxon_2697049"
},
"technology": {
"assembly_method": "http://purl.obolibrary.org/obo/GENEPIO_0001628",
"alignment_protocol": "bowtie2 v. 2.3.4",
"sample_sequencing_technology": [
"http://purl.obolibrary.org/obo/OBI_0000759"
]
},
"submitter": {
"authors": [
"Blagodatskikh,K.A."
],
"submitter_name": [
"R&D"
],
"submitter_address": "Pirogov Russian National Research Medical University, Ostrovityanova 1, Moscow 117997, Russia"
}
}
"""
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 (in normalize).
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):
print(f"WARNING: {msg}",file=sys.stderr)
warnings.append(msg)
def fetch(msg, xpath):
try:
n = gbseq.find(xpath).text
return n
except AttributeError:
warn("Missing "+msg)
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_value>USA: Cruise_Ship_1, California</GBQualifier_value>
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
# </GBReference_journal>
try:
n = gbseq.find(".//GBReference_journal").text
# print(n,file=sys.stderr)
if n != 'Unpublished':
institute,address = n.split(',',1)
if ")" in institute:
submitter.submitter_name = [institute.split(')')[1].strip()]
submitter.submitter_address = address.strip()
except AttributeError:
pass
except ValueError:
submitter.additional_submitter_information = n
try:
n = gbseq.find("./GBSeq_comment").text
except AttributeError:
pass
if 'Assembly-Data' in n:
# print(n,file=sys.stderr)
# the following is wrong (de novo by default)
# technology.assembly_method = 'http://purl.obolibrary.org/obo/GENEPIO_0001628'
p = re.compile(r'.*Assembly Method :: ([^;]+).*')
m = p.match(n)
if m: technology.alignment_protocol = m.group(1).strip()
p = re.compile(r'.*Coverage :: ([^;]+).*')
m = p.match(n)
if m: technology.sequencing_coverage = m.group(1)
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")
creation_date = dateparse(n.text).date()
n = gbseq.find("./GBSeq_update-date")
update_date = dateparse(n.text).date()
n = gbseq.find(".//GBQualifier/GBQualifier_name/[.='collection_date']/../GBQualifier_value")
try:
date = dateparse(n.text).date()
sample.collection_date = str(date)
except dateutil.parser._parser.ParserError as e:
warn("No collection_date: ",str(e))
except AttributeError:
warn("Missing collection_date")
# --- Host info
# - Homo sapiens
# - Homo sapiens; female
# - Homo sapiens; female 63
# - Homo sapiens; female; age 40
# - Homo sapiens; gender: F; age: 61
# - Homo sapiens; gender: M; age: 68
# - Homo sapiens; hospitalized patient
# - Homo sapiens; male
# - Homo sapiens; male; 63
# - Homo sapiens; male; age 29
# - Homo sapiens; symptomatic
n = fetch("host_species", ".//GBQualifier/GBQualifier_name/[.='host']/../GBQualifier_value")
if n:
list = n.split('; ')
host.host_species = list[0]
if len(list)>1:
sex = list[1]
if 'male' in sex or 'gender: M' in sex: host.host_sex = 'male'
if 'female' in sex or 'gender: F' in sex: host.host_sex = 'female'
if len(list)>2:
age = list[2]
p = re.compile(r'[^\d]+(\d+)')
m = p.match(n)
print(m.group(1))
if m:
host.host_age = int(m.group(1))
host.host_age_unit = 'http://purl.obolibrary.org/obo/UO_0000036'
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]
n = fetch("specimen_source", ".//GBQualifier/GBQualifier_name/[.='isolation_source']/../GBQualifier_value")
if n: sample.specimen_source = [n]
info = {
'id': 'placeholder',
'update_date': str(update_date),
'host': host.__dict__,
'sample': sample.__dict__,
'virus': virus.__dict__,
'technology': technology.__dict__,
'submitter': submitter.__dict__,
'warnings': warnings,
}
print(info)
return True,info
def get_sequence(id, gbseq):
seq = None
count = 0
for gbseq_sequence in gbseq.findall('./GBSeq_sequence'):
count += 1
if count > 1:
raise GBError(f"Expected one sequence for {id}")
seq = gbseq_sequence.text.upper()
print(f"SEQ: size={len(seq)}",seq[0:30])
if len(seq) < 20_000:
raise GBError(f"Sequence too short")
return seq
