Detecting virus-specific effects on post-infection temporal gene expression

Abstract

BackgroundDifferent types of viruses have different envelope proteins, and may have their shared or distinctive host-virus interactions which result in various post-infection effects in humans and animals. These effects often do not appear at once but take time to unfold. To characterize the virus-specific effects, we applied a Multivariate Polynomial Time-dependent Genetic Association (MPTGA) method, previously proposed for detecting differences in temporal gene expression traits, to test for the differences in mouse lung transcriptome response to infection of different subtypes of influenza A viruses.ResultsWe compared two methods: the Multivariate Polynomial Time-dependent Genetic Association (MPTGA) method, and the conventional modified t-test, to study the virus-specific effects on mouse lung gene expression. Both methods found H3N2 to be the most different virus among the three viruses tested, with the largest number of genes with H3N2-specific effects. However, the MPTGA method demonstrated much higher power of detection, and the detected genes with virus-specific effects showed better biological relevance.ConclusionsTranscriptome response to virus infection is dynamic. MPTGA which leverages temporal gene expression traits showed increased power in detecting biologically relevant virus-specific effects comparing with conventional t-test method.