YuuMJ1997/EukPhylo
1
0
Fork 0
mirror of http://43.156.76.180:8026/YuuMJ/EukPhylo.git synced 2025-12-28 05:30:30 +08:00
Code Issues Packages Projects Releases Wiki Activity

OG6/G/A, Sim_filter by taxa, and Sim_removed_seqs

Browse Source
This commit is contained in:
Godwin Ani 2023-11-27 15:15:50 -05:00 committed by GitHub
parent bacb094028
commit 22bd67d2fa
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 14 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
PTL2/Scripts/preguidance.py
View File

@ -15,6 +15,11 @@ def run(params):
except (FileNotFoundError, TypeError) as e: except (FileNotFoundError, TypeError) as e:
Logger.Error('Unable to read taxon list file. Please make sure that the path is correct and that the file is formatted correctly.\n\n' + str(e)) Logger.Error('Unable to read taxon list file. Please make sure that the path is correct and that the file is formatted correctly.\n\n' + str(e))
try:
sim_taxa = list(dict.fromkeys([line.strip() for line in open(params.sim_taxa)]))
except (FileNotFoundError, TypeError) as e:
Logger.Error('Unable to read similarity taxa list file. Please make sure that the path is correct and that the file is formatted correctly.\n\n' + str(e))
try: try:
blacklist_seqs = list(dict.fromkeys([line.strip() for line in open(params.blacklist)])) blacklist_seqs = list(dict.fromkeys([line.strip() for line in open(params.blacklist)]))
except (FileNotFoundError, TypeError) as e: except (FileNotFoundError, TypeError) as e:
@ -31,19 +36,21 @@ def run(params):
os.mkdir(params.output + '/Output/Intermediate/SF_Diamond') os.mkdir(params.output + '/Output/Intermediate/SF_Diamond')
removed_file = open(params.output + '/Output/Pre-Guidance/SimFilter_removed.txt', 'w')
for og in ogs: for og in ogs:
Logger.Message('Processing ' + og) Logger.Message('Processing ' + og)
with open(params.output + '/Output/Pre-Guidance/' + og + '_preguidance.faa', 'w') as preguidance_file: with open(params.output + '/Output/Pre-Guidance/' + og + '_preguidance.faa', 'w') as preguidance_file:
for taxon_file in aa_files: for taxon_file in aa_files:
recs = [] recs = []
for rec in sorted([rec for rec in SeqIO.parse(params.data + '/' + taxon_file, 'fasta') if rec.id[-10:] == og and rec.id not in blacklist_seqs], key=lambda x: -len(x.seq)): for rec in sorted([rec for rec in SeqIO.parse(params.data + '/' + taxon_file, 'fasta') if rec.id[-10:] == og and rec.id not in blacklist_seqs and params.og_identifier in rec.id], key=lambda x: -len(x.seq)):
if(rec.id == rec.description): if(rec.id == rec.description):
recs.append(rec) recs.append(rec)
else: else:
Logger.Warning('\tThe sequence ID ' + rec.description + ' is invalid. Please make sure that sequence IDs contain no spaces, tabs, etc. This sequence is being excluded.') Logger.Warning('\tThe sequence ID ' + rec.description + ' is invalid. Please make sure that sequence IDs contain no spaces, tabs, etc. This sequence is being excluded.')
masters = []; removed = 0; flag = 0; cycle = 0 masters = []; removed = 0; flag = 0; cycle = 0
if params.similarity_filter: if params.similarity_filter and taxon_file[:10] in sim_taxa:
if len(recs) > 1: if len(recs) > 1:
while flag == 0: while flag == 0:
master_file_name = params.output + '/Output/Intermediate/SF_Diamond/' + og + '_' + taxon_file[:10] + '_master_' + str(cycle) master_file_name = params.output + '/Output/Intermediate/SF_Diamond/' + og + '_' + taxon_file[:10] + '_master_' + str(cycle)
@ -75,11 +82,16 @@ def run(params):
recs = [rec for rec in recs[1:] if rec.id not in recs_to_remove] recs = [rec for rec in recs[1:] if rec.id not in recs_to_remove]
cycle += 1 cycle += 1
for item in recs_to_remove:
removed_file.write(f"{item}\n")
Logger.Message('\t' + str(removed) + ' sequence(s) removed by the similarity filter (' + str(cycle + 1) + ' iterations) from ' + taxon_file[:10]) Logger.Message('\t' + str(removed) + ' sequence(s) removed by the similarity filter (' + str(cycle + 1) + ' iterations) from ' + taxon_file[:10])
for rec in recs + masters: for rec in recs + masters:
preguidance_file.write('>' + rec.id + '\n' + str(rec.seq) + '\n\n') preguidance_file.write('>' + rec.id + '\n' + str(rec.seq) + '\n\n')
removed_file.close()
if(not params.keep_temp): if(not params.keep_temp):
os.system('rm -r ' + params.output + '/Output/Intermediate/SF_Diamond') os.system('rm -r ' + params.output + '/Output/Intermediate/SF_Diamond')