The Highest Expected Reward Decoding for HMMs with Application to Recombination Detection

TL;DR

This paper introduces HERD, a new decoding algorithm for HMMs that accounts for boundary uncertainty, demonstrated on viral genome recombination detection, improving over traditional Viterbi decoding.

Contribution

The paper presents HERD, an efficient decoding algorithm for HMMs that handles boundary uncertainty, with application to viral recombination detection.

Findings

HERD effectively models boundary uncertainty in HMM decoding.

Application to viral genomes shows improved recombination detection.

HERD outperforms traditional Viterbi decoding in relevant tasks.

Abstract

Hidden Markov models are traditionally decoded by the Viterbi algorithm which finds the highest probability state path in the model. In recent years, several limitations of the Viterbi decoding have been demonstrated, and new algorithms have been developed to address them \citep{Kall2005,Brejova2007,Gross2007,Brown2010}. In this paper, we propose a new efficient highest expected reward decoding algorithm (HERD) that allows for uncertainty in boundaries of individual sequence features. We demonstrate usefulness of our approach on jumping HMMs for recombination detection in viral genomes.