def test_count_to_transfac(self):
"""
Test if the count_to_transfac function works.
"""
ref_input = [
" a b c d\n", # header
"1 0 0 0 0\n",
"2 2 789 99999 89\n", # one value is written
# on 5 characters
"3 99999 8888 2 2\n", # the first value is
# written on 5 characters
"4 99 0 999999 0\n", # one value is written
# on more than 5
# characters
"5 0 0 0 0\n",
]
identifier = "identifier"
ref_output = (
"ID identifier\n"
"BF unknown\n"
"P0 a b c d\n"
"00001 0 0 0 0 X\n"
"00002 2 789 99999 89 X\n"
"00003 99999 8888 2 2 X\n"
"00004 99 0 999999 0 X\n"
"00005 0 0 0 0 X\n"
"XX\n"
"//"
)
output = PB.count_to_transfac(identifier, ref_input)
ref_output_lines = ref_output.split("\n")
output_lines = output.split("\n")
self.assertEqual(len(ref_output_lines), len(output_lines), "Not the right number of lines")
for ref_line, line in zip(ref_output_lines, output_lines):
print("ref:", ref_line)
print("out:", line)
self.assertEqual(ref_line, line)
# G. E. Crooks, G. Hon, J.-M. Chandonia, and S. E. Brenner.
# 'WebLogo: A Sequence Logo Generator.'
# Genome Research 14:1188–90 (2004)
# doi:10.1101/gr.849004.
# http://weblogo.threeplusone.com/
#-------------------------------------------------------------------------------
if options.logo:
# read count file
#-------------------------------------------------------------------------------
f_in = open(options.f, 'r')
count_content = f_in.readlines()
f_in.close()
# convert a table of PB frequencies into transfac format as required by weblogo
# http://meme.sdsc.edu/meme/doc/transfac-format.html
transfac_content = PB.count_to_transfac(options.f, count_content)
# write transfac file (debug only)
#-------------------------------------------------------------------------------
debug = False
if debug:
transfac_name = options.o + ".PB.transfac"
f_out = open(transfac_name, 'w')
f_out.write(transfac_content)
f_out.close()
print("wrote {0}".format(transfac_name))
# define output file name
#-------------------------------------------------------------------------------
logo_file_name = options.o + ".PB.logo.pdf"
if options.residue_min or options.residue_max:
# G. E. Crooks, G. Hon, J.-M. Chandonia, and S. E. Brenner.
# 'WebLogo: A Sequence Logo Generator.'
# Genome Research 14:1188–90 (2004)
# doi:10.1101/gr.849004.
# http://weblogo.threeplusone.com/
# -------------------------------------------------------------------------------
if options.logo:
# read count file
# -------------------------------------------------------------------------------
f_in = open(options.f, "r")
count_content = f_in.readlines()
f_in.close()
# convert a table of PB frequencies into transfac format as required by weblogo
# http://meme.sdsc.edu/meme/doc/transfac-format.html
transfac_content = PB.count_to_transfac(options.f, count_content)
# write transfac file (debug only)
# -------------------------------------------------------------------------------
debug = False
if debug:
transfac_name = options.o + ".PB.transfac"
f_out = open(transfac_name, "w")
f_out.write(transfac_content)
f_out.close()
print("wrote {0}".format(transfac_name))
# define output file name
# -------------------------------------------------------------------------------
logo_file_name = options.o + ".PB.logo.pdf"
if options.residue_min or options.residue_max:
