Current safety assessment for novel crops, including transgenic crops, uses a targeted approach, which relies on compositional analysis. The possibility that transgene expression could lead to unintended effects remains a debated issue. This study used transcriptome profiling as a nontargeted approach to evaluate overall molecular changes in transgenic soybean cultivars. Global gene expression was measured in the first trifoliate leaves of two transgenic and three conventional soybean cultivars using the soybean Affymetrix GeneChip. It was found that gene expression differs more between the two conventional cultivars than between the transgenics and their closest conventional cultivar investigated and that the magnitudes of differences measured in gene expression and genotype (determined by SSR analysis) do not necessarily correlate. A MySQL database coupled with a CGI Web interface was developed to store and present the results ( http://soyxpress.agrenv.mcgill.ca/). By integrating the microarray data with gene annotations and other soybean data, a comprehensive view of differences in gene expression is explored between cultivars.