YuuMJ1997/EukPhylo
1
0
Fork 0
mirror of http://43.156.76.180:8026/YuuMJ/EukPhylo.git synced 2025-12-27 07:50:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update Cluster_v2.0.py

Browse Source
This commit is contained in:
Godwin Ani 2024-02-05 13:25:11 -05:00 committed by GitHub
parent f6c34517ae
commit 4ae1382815
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
Utilities/for_fastas/Cluster_v2.0.py
View File

@ -25,10 +25,10 @@ def input_validation(value, error_message):
print(error_message) print(error_message)
exit(1) exit(1)
def cluster_sequences(program, threshold, overlap, input_folder, output_folder): def cluster_sequences(program, identity, overlap, input_folder, output_folder):
for file in tqdm(os.listdir(input_folder)): for file in tqdm(os.listdir(input_folder)):
if file.endswith('.fasta'): if file.endswith('.fasta'):
subprocess.run([f'{program}', '-i', f'{input_folder}/{file}', '-o', f'{output_folder}/{file}', '-c', f'{threshold}', '-d', '0', '-aS', f'{overlap}']) subprocess.run([f'{program}', '-i', f'{input_folder}/{file}', '-o', f'{output_folder}/{file}', '-c', f'{identity}', '-d', '0', '-aS', f'{overlap}'])
for file in os.listdir(output_folder): for file in os.listdir(output_folder):
if file.endswith('.clstr'): if file.endswith('.clstr'):
@ -52,13 +52,13 @@ def main():
os.mkdir(args.output) os.mkdir(args.output)
if args.type == 'aa': if args.type == 'aa':
threshold = input_validation(args.identity, 'ERROR! Use format 0.## for Amino acids sequence identity threshold.') identity = input_validation(args.identity, 'ERROR! Use format 0.## for Amino acids sequence identity threshold.')
overlap = input_validation(args.overlap, 'ERROR! Use format 0.## for Amino acids sequence alignment overlap value.') overlap = input_validation(args.overlap, 'ERROR! Use format 0.## for Amino acids sequence alignment overlap value.')
cluster_sequences('cd-hit', threshold, overlap, args.input, args.output) cluster_sequences('cd-hit', identity, overlap, args.input, args.output)
elif args.type == 'dna': elif args.type == 'dna':
threshold = input_validation(args.identity, 'ERROR! Use format 0.## for DNA sequence identity threshold.') identity = input_validation(args.identity, 'ERROR! Use format 0.## for DNA sequence identity threshold.')
overlap = input_validation(args.overlap, 'ERROR! Use format 0.## for DNA sequence alignment overlap value.') overlap = input_validation(args.overlap, 'ERROR! Use format 0.## for DNA sequence alignment overlap value.')
cluster_sequences('cd-hit-est', threshold, overlap, args.input, args.output) cluster_sequences('cd-hit-est', identity, overlap, args.input, args.output)
else: else:
print('Invalid sequence type. Choose "aa" for Amino Acids or "dna" for DNA.') print('Invalid sequence type. Choose "aa" for Amino Acids or "dna" for DNA.')
exit(1) exit(1)