final_gbk = os.path.join(basedir, organism_name + '.gbk')
final_tbl = os.path.join(basedir, organism_name + '.tbl')
final_proteins = os.path.join(basedir, organism_name + '.proteins.fa')
final_transcripts = os.path.join(basedir, organism_name + '.transcripts.fa')
final_validation = os.path.join(basedir, organism_name + '.validation.txt')
final_error = os.path.join(basedir, organism_name + '.error.summary.txt')
final_fixes = os.path.join(basedir, organism_name + '.models-need-fixing.txt')
#retrieve files/reorganize
shutil.copyfile(os.path.join(basedir, 'tbl2asn', 'genome.gbf'), final_gbk)
shutil.copyfile(os.path.join(basedir, 'tbl2asn', 'genome.tbl'), final_tbl)
shutil.copyfile(os.path.join(basedir, 'tbl2asn', 'genome.val'),
final_validation)
shutil.copyfile(os.path.join(basedir, 'tbl2asn', 'errorsummary.val'),
final_error)
lib.gb2allout(final_gbk, final_gff, final_proteins, final_transcripts,
final_fasta)
errors = lib.ncbiCheckErrors(final_error, final_validation, locustag,
final_fixes)
if errors > 0:
lib.log.info(
"Manually edit the tbl file %s, then run:\n\nfunannotate fix -i %s -t %s\n"
% (final_tbl, final_gbk, final_tbl))
else:
contigs, genes, trnas = lib.countGenBank(final_gbk)
lib.log.info(
'Output genome consists of: {:,} contigs containing {:,} protein coding genes and {:,} tRNA genes'
.format(contigs, genes, trnas))
#clean up
shutil.rmtree(os.path.join(basedir, 'tbl2asn'))
os.makedirs(os.path.join(outputdir, 'annotate_results'))
if not os.path.isdir(os.path.join(outputdir, 'logfiles')):
os.makedirs(os.path.join(outputdir, 'logfiles'))
genbank = args.genbank
Scaffolds = os.path.join(outputdir, 'annotate_misc',
'genome.scaffolds.fasta')
Proteins = os.path.join(outputdir, 'annotate_misc',
'genome.proteins.fasta')
Transcripts = os.path.join(outputdir, 'annotate_misc',
'genome.transcripts.fasta')
GFF = os.path.join(outputdir, 'annotate_misc', 'genome.gff3')
lib.log.info("Checking GenBank file for annotation")
if not lib.checkGenBank(genbank):
lib.log.error("Found no annotation in GenBank file, exiting")
sys.exit(1)
lib.gb2allout(genbank, GFF, Proteins, Transcripts, Scaffolds)
else:
#should be a folder, with funannotate files, thus store results there, no need to create output folder
if not os.path.isdir(args.input):
lib.log.error("%s directory does not exist" % args.input)
sys.exit(1)
if os.path.isdir(os.path.join(
args.input,
'update_results')): #funannotate results should be here
inputdir = os.path.join(args.input, 'update_results')
outputdir = args.input
elif os.path.isdir(os.path.join(args.input, 'predict_results')):
inputdir = os.path.join(args.input, 'predict_results')
outputdir = args.input
else:
else:
lib.log.error("Output directory %s already exists, will use any existing data. If this is not what you want, exit, and provide a unique name for output folder" % (outputdir))
if not os.path.isdir(os.path.join(outputdir, 'annotate_misc')):
os.makedirs(os.path.join(outputdir, 'annotate_misc'))
if not os.path.isdir(os.path.join(outputdir, 'annotate_results')):
os.makedirs(os.path.join(outputdir, 'annotate_results'))
genbank = args.genbank
Scaffolds = os.path.join(outputdir, 'annotate_misc', 'genome.scaffolds.fasta')
Proteins = os.path.join(outputdir, 'annotate_misc', 'genome.proteins.fasta')
Transcripts = os.path.join(outputdir, 'annotate_misc', 'genome.transcripts.fasta')
GFF = os.path.join(outputdir, 'annotate_misc', 'genome.gff3')
lib.log.info("Checking GenBank file for annotation")
if not lib.checkGenBank(genbank):
lib.log.error("Found no annotation in GenBank file, exiting")
os._exit(1)
lib.gb2allout(genbank, GFF, Proteins, Transcripts, Scaffolds)
else:
#should be a folder, with funannotate files, thus store results there, no need to create output folder
if not os.path.isdir(args.input):
lib.log.error("%i directory does not exist" % args.input)
os._exit(1)
if os.path.isdir(os.path.join(args.input, 'predict_results')): #funannotate results should be here
inputdir = os.path.join(args.input, 'predict_results')
outputdir = args.input
else:
inputdir = os.path.join(args.input) #here user specified the predict_results folder, or it is all together wrong, find out in next few lines
#get files that you need
for file in os.listdir(inputdir):
if file.endswith('.gbk'):
genbank = os.path.join(inputdir, file)
