Normalization

Introduction As it is difficult to accurately determine the amount of mRNA used for each array, normalization of the data obtained from different array experiments for a direct comparison is critical for adapting DNA microarray to the quantitative analysis. However, the commonly used normalization method has been developed based on the assumption that there is no net difference in overall mRNA levels between different sample (constant mRNA levels)(1). Although this default method has been proved to be successful for most DNA microarray experiments monitoring the relative mRNA levels of specific genes in two different samples, it is not applicable for polysomal profiling. This is because the overall mRNA level, as well as the mRNA levels of specific genes, is expected to differ a lot between different fractions of the gradient. To solve this problem, we developed a new normalization method by doping a pool of in vitro transcribed polyA tagged Bacillus subtilis mRNAs with known concentrations into both RNA samples and reference samples as an internal standard for quantitative normalization. References: 1) M. B. Eisen, P. O. Brown, Methods Enzymol 303, 179-205 (1999).

Methods and Results Normalization is critical in order to directly compare each transcript's abundance at different fractions measured by different arrays. For normalization, an internal standard was prepared using a pool of in vitro transcribed Bacillus subtilis mRNAs with artificial polyA tags. PCR products of five B. subtilis genes (ATCC#87482 (Trp), #87483 (Dap), #87484 (Lys), #87485 (Thr), and #87486 (Phe)) were printed onto yeast DNA microarrays (25-125 spots/each array in different locations). In vitro transcribed B. subtilis mRNAs (Ribomax large scale RNA production system-T3, Promega) were pooled as a cocktail at various concentrations of each B. subtilis mRNA. To determine the linear readout range of the spiked-in B. subtilis mRNAs in microarray hybridization, 15 mg total RNA sample were spiked with B. subtilis cocktail A (Trp - 20 pg, Dap - 60 pg, Lys - 200 pg, Thr - 500 pg, and Phe - 1000 pg) and fluorescently-labeled in red (Cy5-dUTP), while 15 mg of the same total RNA sample was spiked with B. subtilis cocktail B (Trp - 1000 pg, Dap - 500 pg, Lys - 200 pg, Thr - 60 pg, and Phe - 20 pg) and fluorescently-labeled in green (Cy3-dUTP). The cocktails A and B represent the minimum and maximum range of yeast mRNA levels, 20-1000 pg, but in a reversed order so that they can also cover a wide range of red to green ratios. The linear ranges of these spikes in both channels were measured by plotting the average intensity of each B. subtilis transcript as a function of its concentration. As shown bellow, both channels have linear readout range within 20-1000 pg.