A hidden Markov model for estimating retrovirus activities from expressed sequence databases
Human endogenous retroviruses (HERVs) are remnants of ancient retrovirus infections and now reside within the human DNA. Recently HERV expression has been detected in both diseased patients and normal tissues. However, the expression levels of individual HERV sequences are mostly unknown. In this work we introduce a generative mixture model, based on Hidden Markov Models, for estimating the activities of the individual HERV sequences from databases of expressed sequences. We determined the relative expression levels of 91 HERVs; the majority of their activities were previously unknown.