RESUMEN
Hybridization-based and sequencing-based technologies have found a widespread application in gene expression profiling analysis but much ambiguity exists regarding their reliability. This study developed a framework based on three parameters: detection ability, repeatability, and accuracy to evaluate the reliability of gene expression profiling technologies. The fraction of coverage of detected transcript category, the degree of variance for the number of differentially expressed genes (DEGs), the consistency of DEG category, and suspected false discovery rate (sFDR) were first introduced as statistical indices. In order to validate the availability of these indices, based on the same RNA extract, the analysis was performed by comparing gene expression differences between wild-type and transgenic rice using deep sequencing and microarray. The results suggested that the parameters were available and showed advances in the determination of gene expression differences. Based on relative self-comparison design, suspected false positive genes were easily identified from all DEGs detected, which was difficult for quantitative real-time polymerase chain reaction (qRT-PCR) validation when the count of DEGs was enormous. In addition, sFDRs had advantages in the accuracy evaluation for previous datasets.