Gene Sequence Clustering Based on the Profile Hidden Markov Model with Differential Identifiability

Algorithm 1

Profile hidden Markov model with differential identifiability (DI-PHMM).

	Input: Training sequence data and differential identifiability parameter
	Output: A PHMM Calculate probability of transition from state to state . ;
(1)	//Generate privacy noise for matching state and insert state emission probabilities.
(2)	for in (‘A’, ‘G’, ‘T’, ‘C’)
(3)	,
(4)	;
(5)	//Calculate the emission probability of state to state .
(6)	for in (‘match’, ‘insert’)
(7)
(8)	//Validity test. Check whether the emission probability is valid. If not, go
(9)	back to step 2.
(10)
(11)	Regenerate privacy noise
(12)	Return