Hidden Markov Models for biological sequence analysis

This thesis aims at providing an overview on Hidden Markov Models (HMMs), a class of statistical models widely used for modeling situations where some events are "hidden". We start by introducing the basic concepts about HMMs and their three fundamental problems, namely, the Evaluation, Decoding and Training problems. Then, we illustrate the main HMM variants developed in the literature and some of their applications to biological sequence analysis. Finally, we propose the use of a weighed profile-HMM to model a family of peptides inhibitors of a specific target protein. The aim is to use the resulting model to discover new potential peptides candidates.