The virtues of signaling networks aside, minor differences in gene expression can potentially result in dramatically different developmental programs. To illustrate, the expression of single “selector genes,” such as Ubx or Tbx5, distinguishes whether a wing or haltere will develop from a Drosophila imaginal disc or whether a forelimb or hindlimb will outgrow from the chick body wall (Weatherbee and Carroll 1999xWeatherbee, S.D. and Carroll, S.B. Cell. 1999; 97: 283–286Abstract | Full Text | Full Text PDF | PubMed | Scopus (42)See all ReferencesWeatherbee and Carroll 1999). Elegant examples from Drosophila show that subtle differences in gene expression induced by a RTK can influence biological output. The Torso receptor signals through the Drosophila Grb2 ortholog and the Ras pathway to induce the expression of tailless (tll) and huckebein (hkb), which specify terminal cell fate. The recruitment of Csw (the ortholog of the tyrosine phosphatase Shp2) to Torso mediates the transmission of receptor signals, as demonstrated by a Tyr→Phe substitution in Torso, which blocks this association and yields a loss-of-function phenotype. In contrast, mutation of the Torso-binding site for RasGAP gives a gain-of-function phenotype, as might be expected if RasGAP inhibits Ras signaling. An important effect of these competing regulators is to define the spatial boundaries of tll/hkb expression. Interestingly, substitution of the RasGAP-binding site restores signaling to a Torso mutant that cannot bind Csw, apparently because a major substrate for the Csw phosphatase is the pTyr motif that recruits RasGAP (Cleghon et al. 1998xCleghon, V., Feldmann, P., Ghiglione, C., Copeland, T.D., Perrimon, N., Hughes, D.A., and Morrison, D.K. Mol. Cell. 1998; 2: 719–727Abstract | Full Text | Full Text PDF | PubMedSee all ReferencesCleghon et al. 1998). However, these compensatory signals are still required for proper refinement of the Torso signal, as survival to adulthood is markedly reduced in flies expressing a Torso mutant lacking both Csw and RasGAP-binding sites.Oligonucleotide and cDNA arrays are clearly changing the ways we think about cell biology. The paper by Fambrough et al. reveals both the power of this approach, and also our ignorance about cellular function. It will be fascinating in the future to extend the gene chip technology to a more comprehensive analysis of growth factor signaling. This approach may also be useful in resolving the role of other PDGFR binding partners, such as Src family kinases, in regulating the transcriptional response to PDGF. It may be particularly revealing to use inhibitors to proteins more proximal to the IEG promoters, such as components of the MAP kinase cassettes, to tease out whether there is increased specificity as one gets further down the signaling pathways. Regardless, Pandora's box has now been opened.