Expression data from C2C12 cells with or without Smad3 overexpression

The goal of this study was to identify genes in C2C12 myoblasts whose expression was altered by overexpression of Smad3. Four total samples were analyzed. Two biological replicates were prepared from C2C12 cells infected with retroviral vector encoding wildtype Smad3 and exposed for 24 hours to 100 pM TGF-β. Two biological replicates were prepared from C2C12 cells infected with a control retrovirus that did not encode Smad3 (empty vector, EV) and exposed for 24 hours to 100pM TGF-β. To identify genes with altered expression between empty vector controls and cells overexpressing wild-type Smad3, we first pre-processed the four arrays using robust multiarray averaging as implemented in the R software package ‘xps’, version 1.10.2 (http://www.bioconductor.org/packages/release/bioc/html/xps.html). This preprocessing corrects for background noise and array effects, and aggregates probe data to 28,836 genes. Duplicates were averaged on the base-2 logarithm scale, and then we took differences between the averages to obtain logarithm-scale fold changes. Genes were selected which showed at least a two-fold change (raw scale) between wild-type and empty vector expressing cells. Due to the limited sample size, we did not apply any statistical tests to estimate false-discovery rate. This resulted in 79 genes that were at least two-fold higher in the wild-type Smad3-expressing C2C12 cells compared to empty vector cells and 25 genes that were at least two-fold lower in the wild-type Smad3-expressing cells than the in the empty vector cells.