Figure 8.

Detecting changes in lymphocyte cell counts and the appropriateness of follow-up analyses using CTen. (A) In infected tissue, as B cells (green), T cells (black), and macrophages (dark red) migrate into the sample area, they bring with them their own unique quantities of RNA, resulting in conserved expression patterns proportional to the cell type's increased (or decreased) presence. Here, we illustrate the potential results from a clustering study using WGCNA [15], resulting in (B) 4 clusters with unique temporal expression profiles; 3 of which are highly correlated to the changes in the number of lymphocytes shown in (A). (C) Analyzing the genes present in each cluster, CTen can distinguish which clusters represent gene expression and which represent cell migration, and determine the cell type responsible for the observed gene regulation. Using a "weighted-ranking" strategy (see text), CTen produces a heatmap showing the most enriched cell types for each cluster. Based on the CTen result, only the orange cluster is appropriate for further analysis using traditional bioinformatic techniques while the green, black, and dark red clusters reflect the relative changes in the number of lymphocytes during the infection. (D) We propose a new analytical workflow strategy which ensures that continued analysis of in vivo microarray data properly identifies events which may be coordinated with (or coordinating) cell migration.

Shoemaker et al. BMC Genomics 2012 13:460   doi:10.1186/1471-2164-13-460
Download authors' original image