def build_filesystem():
# DIRECTORY STRUCTURE
global SCRATCH, BASEDIR, REFERENCES, DATA, TRIM_DIR, ASSEMBLY_DIR, ANNOTATION_DIR, MAP_DIR
SCRATCH = Dir.make('/scratch/' + os.environ['USER'])
BASEDIR = Dir('/Strong/proj/.data/Project_NTM')
REFERENCES = Dir(BASEDIR.join("lib", "reference_genomes"))
DATA = BASEDIR.make_subdir("data")
# RAW_DIR = DATA.make_subdir("00_raw")
TRIM_DIR = DATA.make_subdir('trimmed_reads')
ASSEMBLY_DIR = DATA.make_subdir('assemblies')
ANNOTATION_DIR = DATA.make_subdir('annotations')
MAP_DIR = DATA.make_subdir('mapped_reads')
def identify(isolate,
delimiter="_",
species_threshold=0.97,
genus_threshold=0.80):
isolate.log('Identifying Isolate')
assembly = isolate.files.assembly
"""Blast fasta to reference genomes and store values in db"""
ani_script = "/Strong/proj/.data/Morty/.config/software/ani-script/ANI.pl"
blastall = "/software/cgeh/blast/2.2.22/bin/blastall"
formatdb = "/software/cgeh/blast/2.2.22/bin/formatdb"
# COMPARE TO REFERENCES
try:
matches = []
references = Fasta.get_all(REFERENCES)
isolate.log(
f"{isolate}: IDENTIFYING TAXON USING {len(references)} REFERENCES",
lvl='INFO')
for reference in references:
ref_id = reference.filename.split('.')[0]
scratch = SCRATCH.make_subdir('ani', isolate.name,
f"{isolate}_vs_{ref_id}")
command = f"perl {ani_script} -bl {blastall} -fd {formatdb} -qr {assembly} -sb {reference} -od {scratch}"
output, error = subprocess.Popen(
command.split(), stdout=subprocess.PIPE).communicate()
try:
ani = float(output) / 100
except (ValueError, TypeError):
ani = 0
finally:
record = {
'ani': ani,
'sample': isolate.name,
'reference': ref_id,
'taxon': ref_id.split('_')[0].split('-')[0]
}
matches.append(record)
# WRITE CSV
ani_csv = Dir.make(assembly.dir.join("ANI")).join(f"{isolate}_ANI.csv")
df = pd.DataFrame.from_records(matches)
df = df.sort_values('ani', ascending=False)
df.to_csv(ani_csv, index=False)
isolate.files.ani = File(ani_csv)
# ASSIGN TAXON
taxon = 'UNKNOWN'
possible_species = df[(df.ani >= species_threshold)]
if len(possible_species) > 0:
taxon = possible_species.iloc[0].taxon
elif len(df[(df.ani >= genus_threshold)]) > 0:
taxon = 'NTM'
isolate.taxon = taxon
isolate.log(f"taxon={isolate.taxon}", lvl='INFO')
return taxon
except Exception as e:
isolate.log(f"Identification failed: {e}", lvl='WARNING')
subprocess.call(f"rm error.log formatdb.log".split())