ABSTRACT
Hidden Markov models (HMMs) have been applied as statistical models in the area of speech recognition since the early 1970s. The use of HMMs for protein modeling was introduced in the early 1990s by Haussler et al. [1] and Krogh et al. [2]. The general HMM structure to model protein sequence families is known as profi le HMM. A possible way to construct (or learn) such profi le HMMs is by using a multiple sequence alignment (MSA) of proteins belonging to the same family as an input (which is explained in more details in Section 10.2).
