We did not detect significant interaction between alcoholic status and ethanol exposure. After correction of the interaction term for multiple testing, only one probe set for an unknown transcript had an FDR < 0.95. This may be an issue of power, given the relatively small number of genes detected as differentially expressed between the alcoholics and controls. There was substantial heterogeneity between LCL from different subjects, which reduces power to detect differences between alcoholics and controls but did not greatly interfere with detection of the effects of ethanol because of our paired design.
