#!python
import sys
import os
import os.path
from Hotpep.hotpep_data import hotpep_data_path
protein_collection = hotpep_data_path("fungus_fungus/Chaetomium_thermophilum")
if len(sys.argv) > 1:
protein_collection = sys.argv[1].replace("?", " ")
cazyme_array = ["AAexp"]
if len(sys.argv) > 2:
cazyme_array = sys.argv[2].split("_")
assension_family_hash = {}
for cazy_class in cazyme_array:
if os.path.exists(protein_collection + "/" + cazy_class):
for filename in os.listdir(protein_collection + "/" + cazy_class):
if "summary" not in filename and filename.endswith(".txt"):
fam = filename[:-4]
array = open(
protein_collection + "/" + cazy_class + "/" + filename,
'r').readlines()
array = array[1:]
for hit in array:
arr = hit.split('/t')
if len(arr) > 1:
if arr[1] not in assension_family_hash:
assension_family_hash[arr[1]] = []
assension_family_hash[arr[1]].append(fam + "_" +
#Start Add by Le Huang 12/24/2018
if os.path.exists(organism_array[0] + '/Results/output.txt'):
call(['rm', organism_array[0] + '/Results/output.txt'])
#End Add by Le Huang 12/24/2018
##Start Delete by Le Huang
#try:
# call(['rm', organism_array[0]+'/Results/output.txt'])
# except:
# pass
## End Delete by Le Huang
for protein_dir_name in organism_array:
print("Screening " + protein_dir_name + " for")
for cazy_class in cazyme_array:
print(cazy_class)
peptide_dir_name = hotpep_data_path("CAZY_PPR_patterns", cazy_class)
variables = [
threads, protein_dir_name, peptide_dir_name, peptide_length,
hit_cut_off, freq_cut_off
]
call("parallel_group_many_proteins_many_patterns_noDNA.py " +
" ".join(str(x) for x in variables),
shell=True)
#call(["add_functions_orf.py", protein_dir_name, peptide_dir_name])
var1 = 1
while var1 <= threads:
try:
os.remove(protein_dir_name + "/thread" + str(var1) + ".txt")
except:
pass
var1 += 1
#!/usr/bin/env python3
import sys
import os
import os.path
from subprocess import call
import natsort
from Hotpep.hotpep_data import hotpep_data_path
protein_dir_name = hotpep_data_path("fungus_fungus")
if len(sys.argv) > 1:
protein_dir_name = sys.argv[1].replace("?", " ")
function_significance_limit = 2 #Normally 2 for new patterns (no proteins with this function in group) and 10 for old patterns (sum freq conserved peptides for function in group )
score_file = "_group_ec"
peptide_dir_name = hotpep_data_path("CAZY_PPR_patterns/GH")
if len(sys.argv) > 2:
peptide_dir_name = sys.argv[2].replace("?", " ")
fam_group_score_hash = {}
with open(peptide_dir_name + "/fam_list.txt", 'r') as f: ####
for line in f:
fam = line.rstrip()
try:
with open(
peptide_dir_name + "/" + fam + "/" + fam + score_file +
".txt", 'r') as f2:
group_func_hash = {}
for row in f2:
row = row.rstrip()
arr = row.split('\t')