RNA amplification results in reproducible microarray data with slight ratio bias

Abstract

Microarray expression analysis demands large amounts of RNA that are often not available. RNA amplification techniques have been developed to overcome this prcblem, but limited data are available regarding the reproducibility and maintenance of original transcript ratios. We optimized and validated two amplification techniques: a modified in vitro transcription for the linear amplification of 3 microg total RNA and a SMART PCR-based technique for the exponential amplification of 50 ng total RNA. To determine bias between transcript ratios, we compared the expression profiles in mouse testis versus spleen between the two amplification methods and a standard labeling protocol, using microarrays containing 4596 cDNAs spotted in duplicate. With each method, replicate hybridizations were highly reproducible. However, when comparing the amplification methods to standard labeling, correlation coefficients were lower. Twelve genes that exhibited inconsistent or contradictory expression ratios among the three methods were verified by quantitative RT-PCR. The amplification methods showed slightly more discrepancies in the expression ratios when compared to quantitative RT-PCR results but were more sensitive in terms of detecting expressed genes. In conclusion, although amplification methods introduce slight changes in the transcript ratios compared to standard labeling, they are highly reproducible. For small sample size, in vitro transcription is the preferred method, but one should never combine different labeling strategies within a single study.