Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is frequently used as loading control in immunoblot experiments (Hsu et al. 2014;Oldenburger et al. 2014;Wang et al. 2014).We have recently reported a study on the homologous desensitization of human β 3 -adrenoceptors stably expressed in human embryonic kidney cells (Michel-Reher and Michel 2013). As part of that study, we have used immunoblots to determine the effect of treatment with the β-adrenoceptor agonist isoprenaline (10 μM for 24 h) on the abundance of several G-protein α-subunits at the protein level. After analysis of these G-proteins, the blots were stripped and reprobed with a GAPDH antibody (mouse monoclonal IgG, Bioconnect MAB-374, 1:2,000,000). GAPDH immunoreactivity was quantified using a secondary antibody (Lumilight plus kit, Roche Applied Science) by an LAS3000 (Fujifilm, Düsseldorf, Germany) using AIDA software (version 3.52, Raytest, Straubenhardt, Germany). Data were obtained from ten blots, each with four pairs of samples from isoprenalineand vehicle-treated cells, resulting in 40 pairs of GAPDH quantification with and without isoprenaline treatment. During our initial analysis of the G-protein data, we noted that expression of the data relative to GADPH abundance tended to increase variance. As recently discussed (Motulsky 2014), normalization should only been performed for good reasons. As normalization for GADPH abundance did not improve the variance of G-protein expression, we had decided to report the non-normalized data (Michel-Reher and Michel 2013).We have now analyzed GADPH data from that study to explore whether isoprenaline treatment may have altered its expression at the protein level. This indicated a 32 % lower expression of GADPH protein in isoprenaline-as compared to vehicle-treated cells (173 vs. 254 arbitrary units; mean difference 80 (95 % confidence interval=35-127, t=3.541, df= 39, p<0.01 in a paired, two-tailed t test as determined by Prism 6.07, GraphPad Software, La Jolla, CA, USA; Fig. 1). As the standard deviation within each group was greater than the mean value, we also performed an analysis of the logtransformed data. This yielded a mean difference of 0.1904 log units (95 % confidence interval=0.1043-0.2765, t=4.473, df=39, p<0.0001 in a paired, two-tailed t test; Fig. 1).As the hypothesis of regulation of GADPH expression by isoprenaline treatment was only developed after we had seen the data, these p values must be considered descriptive rather than hypothesis testing (Motulsky 2014). Nevertheless, they seem sufficiently robust to allow detection with both the original and the log-normalized data. Moreover, a 30 % lower GADPH protein abundance under conditions of increased sympathetic drive has also been reported from a rat heart failure model (Brattelid et al.