Since the sequencing of the Drosophila genome, large-scale expression studies using DNA microarrays have been employed in diverse ways to reveal the genetic networks that drive development. Time-course experiments using DNA microarrays have identified subsets of genes expressed at specific developmental stages throughout Drosophila embryogenesis [1-5]. The first few hours of embryogenesis are marked by active changes in cell fate, cell shape and cell rearrangements that are accompanied by dynamic changes in gene expression. Recent studies have determined that more than 2,000 genes, or approximately 15% of the genome, display increased expression at these early stages [3-6]. These systematic genome-wide approaches provide a global perspective on the dynamics of gene expression during embryonic development.

A major challenge now posed by these studies is how to use the overwhelming amount of information from whole-genome studies to identify specific genes with key roles in embryonic morphogenesis. The ability to integrate data from stage-specific gene-expression studies, high-throughput spatial analysis of gene-expression patterns and bioinformatics tools for protein sequence analysis is critical to the process of candidate gene selection (Figure ​(Figure1).1). The paper published in BMC Biology by Cambiazo and colleagues (Zúñiga et al. [7]) describes an integrated genome-wide approach to finding genes that are specifically upregulated during early Drosophila development. The authors use differential gene-expression profiling, bioinformatics, in situ analysis and functional characterization to identify a novel gene that is required for embryonic morphogenesis, as well as a number of interesting candidate genes for future functional analysis. Here, we discuss how these and other genomic datasets have been coupled with targeted functional analysis to gain insight into the molecular mechanisms underlying morphogenesis.

Using suppression subtractive hybridization (SSH) and microarray analysis, Zúñiga et al. [7] isolated a set of transcripts that are differentially expressed between gastrulating and blastoderm Drosophila embryos. The SSH technique is able to circumvent the inability of most high-throughput gene-expression profiling techniques to isolate transcripts of low abundance [8]. Using this method, the authors identified 114 genes and 4 noncoding RNAs that are more highly expressed in gastrulating embryos compared to the earlier blastoderm stages. To validate their experimental approach, the authors compared their gene list with the approximately 2,000 genes found to be expressed in the early embryo in other genome-wide expression studies [1,4,5,9]. More than half of the genes identified were previously reported to be expressed during gastrulation. In addition, their list contains 55 genes that have not been functionally characterized. Two of these genes are not represented on commercially available gene-expression chips, indicating that this technique is able to isolate transcripts that were not previously predicted. Of note, 19 of the novel genes are predicted to encode secreted or transmembrane proteins that may play a direct role in cell-cell communication during morphogenesis. Using in situ analysis the authors were able to determine that 12 out of 15 of these genes are expressed in a spatially restricted pattern in regions undergoing distinct morphogenetic processes.

While microarray analysis and ChIP:chip data can provide insight into the temporal expression and transcriptional regulation of genes during development, these approaches do not offer clues as to the spatial distribution of gene expression. Combining in situ expression patterns with genomic data on expression levels is one way to identify sets of genes involved in related developmental processes. In Drosophila, two groups have determined the spatial and temporal expression pattern of around 25% of the genes expressed in the embryo [2,6]. Lecuyer et al. [6], used fluorescent in situ hybridization (FISH) to generate high-resolution images for around 2,500 mRNAs throughout early Drosophila embryogenesis. This approach, which allows for single-cell resolution, led to the surprising discovery that over 70% of detected mRNAs are subcellularly localized, suggesting that uncharacterized transcripts can be classified on the basis of their localization as well as their expression. Both groups have generated web-based databases that offer a range of search options, including stage, tissue, gene name and predicted function. Analysis of these rich datasets may identify genes involved in morphogenesis that have failed to be identified in genetic screens.