import collections
from scipy.stats import chisquare
seq = ''
count = 0

with open('sequence.fasta', 'r') as inp:
  line = inp.readline()
  while line:
    seq += line.strip()
    line = inp.readline()
  for i in range(len(seq)):
    if seq[i:i+6] == 'GAATTC':
      count += 1
  print(count)

nuc_counter = collections.Counter(seq)

cnt_G = nuc_counter['G']/len(seq)
cnt_C = nuc_counter['C']/len(seq)
cnt_A = nuc_counter['A']/len(seq)
cnt_T = nuc_counter['T']/len(seq)

exp_num = round((cnt_G * (cnt_A ** 2) * (cnt_T ** 2) * cnt_C) * len(seq))   
print(exp_num)
p_value = chisquare([count, exp_num], ddof=0)
print(p_value)