The file "question_HMM.nb" implemented a hidden markov model for
annotating CpG islands in a DNA sequence as you have seen in the
demonstration. It obtains the input sequence from file "seq.txt" and
then computes the probability of a given sequence by Forward-Backward
procedure and computes the most likely state sequence of the model by
the Viterbi algorithm.

The model contains 8 states {A+, C+, G+, T+, A-, C-, G-, T-}. Each state
X+ or X- generates only one symbol X (where X is A,C,G,T). You are given
the transition probability matrix. Your job is to choose the reasonable
parameters for making the model work.

You have to specify the following parameters in the Mathematica file:
(1) p: the probability for being in a CpG island.
(2) q: the probability for being in the ocean.
(3) The initial state distribution.

Explain how you choose these parameters. Try to run 5 settings of these
parameters and reports the number of CpG islands and the percentage of
the CpG islands in the given DNA sequence.