Jonathan M. at evolution news recently criticized the hypothesis of a monophyletic origin of the eyes.
The common evolutionary rationalization of this phenomenon is to posit that the gene in question had some kind of propensity for promoting the development of the respective structure. But this solution appears dubious, particularly in the case of the even more spectacular example of eye development…
Among these are the mammalian "Six" genes, and their analogue in Drosophila called sine oculis. These genes are deployed somewhat later in development than is Pax6. Even the convergently deployed genes Dach and dac (vertebrates and Drosophila respectively) are seemingly homologous in structure, but each possesses quite a restricted role and is utilized late in development. Moreover, the eyes of vertebrates develop from two embryonic tissues, namely the epithelium and optic vesicle, whereas the eyes of Drosophila develop from a single embryonic tissue: the imaginal disc.
The 3 subfamilies of Six genes are also conserved, 2 are involved in eye develpment , and 1 in muscle development both in mice and Drosophila. This shows that transcription factors can control any gene or developmental program as long as their targets have the appropriate cis-regulatory sequences. Therefore, their involvement in the eye or muscle program is due to evolutionary historical reasons, i.e. the latest common ancestor of mammals and insects already used them to control the respective developmental programs.
The structural genes located further downstream in the cascade are much less conserved. This means that with the same toolbox of transcriptional regulators very different eye-types can be generated.