(264h) Constructing Influenza‐induced Host Response Models from Microarray Data

The influenza virus remains a potent threat to public health. Recently, researchers have considered functions of the host immune response as potential therapeutic targets but limited data, strain‐specific responses, and high variation of the host response between typical cell and animal infection models has limited the development of host‐based drug therapies. In our group, we have focused on developing a technique referred to as systems inference microarray analysis (SIMA) for developing prediction‐capable, species‐specific models of the immune response using whole genome data. We exploit biological mechanisms such as gene co‐expression to reduce whole genome data down to clusters of genes which represent the activity of biological pathways or processes. A collection of nonlinear equations which are strictly a function of virus growth are then fit to the expression dynamics and used to predict the host gene response in new infections. Applying this technique to time‐course microarray data developed from mouse lung infected with three different influenza isolates, we have identified novel mechanisms regulating a key inflammatory event – the cytokine storm. Further, applying the approach to cell plate data, we have found that the dynamics of inflammatory and immune response pathways are highly variant. This suggests that in vitro data is a poor biological model of the in vivo inflammation response. We believe applying the SIMA approach to a more extensive data set will lead to a more complete understanding of infection‐induced inflammation and guide the development of future treatments.