Differential variation and expression analysis

Abstract

We propose an empirical Bayes approach using a three‐component mixture model, the L2N model, that may be applied to detect both differential (mean) expression and variation. It consists of two log‐normal components (L2) for differentially expressed (dispersed) features: one component for underexpressed (dispersed) features and the other for overexpressed (dispersed) features, and a single normal component (N) for nondifferentially expressed (dispersed) features. Simulation results show that L2N can capture asymmetries in the numbers of overexpressed and underexpressed (dispersed) features (e.g., genes) when they exist and can provide a better fit to data in which the mixture component distributions are not well separated while also performing well under symmetry and separation. The L2N model is implemented in an R‐driven, user‐friendly, graphical interface called DVX, for differential variation and expression analysis, which does not require the user to have R programming knowledge. The interface also includes an implementation of differential expression analysis via the limma package, and a differential variation and expression analysis using a three‐way normal mixture model. It offers a set of diagnostics plots, data transformation tools, and report generation in Microsoft Excel‐ and Word‐compatible formats. The interface is available on the web at https://haim-bar.uconn.edu/software/DVX/.