This web page was produced as an assignment for an undergraduate course at Davidson College.

Review Paper Concerning:

Analysis of the eye developmental pathway in Drosophila using DNA microarrays

Lydia Michaut, Susanne Flister, Martin Neeb, Kevin P. White, Ulrich Certa, and Walter J. Gehring

PNAS (2003) Vol. 100, p. 4024-4029.

Summary of Results

Prior to this paper, research in the field of development had identified a class of transcription factors encoded by Pax-6 genes that activate the genes required for the construction of an eye. Expression of one of these transcription factors, eyeless (ey) in leg disks of Drosophila was shown to result in ectopic eye formation. The research reported in this article examines the genes expressed during this ectopic eye formation with DNA microarrays in an attempt to elucidate the signal cascade involved in eye development.

DNA microarrays are small chips containing many oligonucleotide sequences corresponding to the genes in an organism. The researchers monitored gene expression by obtaining labeled RNA from the desired tissue and hybridizing it to the microarray. Colored spots on the microarray thus indicated that the gene corresponding to that spot was expressed (Campbell, 2002). To determine which genes were induced by ey, thereby investigating the eye formation signal cascade, the researchers compared RNA from normal leg discs, normal eye discs, and leg discs with ectopic expression of ey. Genes expressed in both eye discs and ey leg discs, but not in normal leg discs, were considered good candidates for genes triggered by ey in the eye developmental pathway. To further ensure the reliability of their data, the researchers tested the RNA on two different microarrays in order to control for the artifacts produced by differences in microarray design, and considered those genes whose expression was detected with both microarrays.

With these methods, the researchers identified 371 genes whose expression was induced by ey, but only 55 of these genes were detected by both microarrays to be expressed in normal eye discs, not expressed in normal leg discs, and induced in leg discs by ectopic expression of ey (Fig.1A). The function of many of these genes was already known, and, in figure 1B the researchers noted that the majority of the genes induced were transcription factors, signal transducers, or enzymes. In table 1, the paper presents all 55 genes identified, their functions, if known, and their statistical significance on both microarrays. When they compared this set of induced genes to all those expressed in normal eye discs, the researchers found that some genes known to be necessary for eye development were not detected, indicating that the genes identified are more specifically involved in the retinal differentiation pathway rather than general morphogenesis.

Among the genes identified, 18 transcription factors were already known to be associated with eye development, and specifically with photoreceptor differentiation. Three other genes identified were known to encode proteins that function at the top of the eye development cascade. These results show that this method can indeed identify genes induced by ey and gives credibility to the other results. Some transcription factors and signal transducers identified, including RacGTPases, are known to play roles in other developmental pathways, but had not previously been associated with eye development. The potential or known roles of these genes in eye development are discussed in the text, and table 2 shows more detailed microarray results for the genes discussed. The expression of several of the genes (highlighted in tables 1 and 2) was confirmed by serial analysis of gene expression (SAGE) in which diagnostic tags are located among transcribed sequences from a cell and then sequenced to determine which genes are expressed (SAGE Molecular Genetics Laboratory, 2003).

Though known genes were detected, many of the genes identified (11 of the 55) were completely uncharacterized with no known molecular function (Table 1). Figure 2 shows the data from experiments intended to further confirm the expression of several of the detected genes and to show where in the eye imaginal disc the expression occurs. Quail protein was detected posterior to the morphogenetic furrow by immunofluorescence, and 13 other genes were tested by in situ hybridization with antisense RNA probes in the eye discs. The positional expression information provides a first clue to the function of some of the previously unidentified genes.

From the data, the researchers conclude that ey functions to activate several different signaling pathways that carry out the functions of photoreceptor specification, actin binding, cell adhesion, and cell division necessary for the development of an eye.


This paper shows that microarrays are very powerful tools for gene expression analysis. The microarrays used in this research took a task that would be unfathomable with a Northern Blot and found candidate genes in the expression pathway. However, there are dangers in experiments like this in which much analysis is done with computer programs and very little raw data is shown. The results of such experiments are highly dependent on the researcher’s choice of parameters required for a positive result and the design of the microarray itself. A misguided choice of parameters could lead to false data or data subtly obscured to fit expectations.

This paper addresses some of these concerns well by comparing the results of two differently designed microarrays and focusing on the gene expression detected by both. The researchers also clearly outline their requirements for significance in the paper so that other scientists may evaluate the validity of their results. By giving specific examples of genes and discussing the patterns of results on each microarray, the researchers further demonstrate the validity of their method.

However, the fact that the researchers identify and report 55 genes that were detected by both microarrays, but then proceed to analyze and draw conclusions from many genes that were only detected by one microarray raises some concern. The discussion of some of these genes may be validated by the fact that they are already known to be involved in eye development and their expression was confirmed by SAGE, but the decision to consider all the genes shown to be induced by either array seems tenuous right after a discussion of the variability of the microarrays.

To provide evidence that the microarray test is valid and that the genes for which it detects expression are actually being expressed, the authors perform in situ hybridization. This is another good way to evade the potential pitfalls of microarray data listed above and provides raw data that the readers can evaluate apart from the author’s interpretation. Unfortunately, the dark stains representing expression are often faint and the figure does not show a clear negative control or a positive control for immunofluorescence.

The comparison of ey transfected leg discs to wild type eye and leg discs provides a good standard for evaluation of the genes detected by the microarrays. These controls convince the reader that the genes identified are not just normally expressed in the leg and are not an artificial result of ectopic expression, but actually are expressed in the normal eye during development. One possible drawback to the elimination of the genes that are expressed both in ey and normal leg tissue is that some genes which are expressed in normal leg tissue may also play a role in eye development, but these genes are not detected.

Though the consideration of the genes outside the 55 detected by both microarrays may not be wise, the authors present convincing experiments that the use of microarrays can identify genes induced by the ey gene in eye development.

Future Directions

In this paper, the authors present convincing data that they have identified candidates for genes that are expressed when induced by the master control gene ey. However, at this point, the genes are only candidates and their specific roles must be further verified and determined. Next, the researchers might test genes that they expect to operate soon after ey in the signal cascade. The function of a particular gene identified could be determined by ectopic expression of the gene, using similar methods to those accomplishing ectopic expression of ey. If a partial eye were formed or some cell differentiation occurred, it would suggest that the gene tested was triggered by ey to begin a cascade in eye development and the specific function of that cascade could be identified.

To determine which genes of those identified are triggered specifically by ey rather than by another protein further down the cascade, future work could look for a consensus binding sequence or motif among the genes identified where the ey transcription factor could bind to the induced genes. Looking for protected regions with DNase footprinting could be used to determine whether the transcription factor ey bound to the different candidate genes, and then sequencing and computational methods could look for a consensus binding sequence among the identified genes. If the candidate genes are truly induced by ey as the microarrays suggest, at least some should have sites that will bind to ey.

Several papers discussing eye development in Drosophila, including this paper, express a desire to learn about the similarities between fly eye formation and mammalian and human eye development. Ectopic expression in flies of Pax-6 genes from other species has indicated that the master control gene is well conserved between species (Halder, et. al. 1995). There must be differences further down in the cascade that result in the differentiation of a compound eye from a camera-type eye. Future research could compare the set of eye development genes identified by this paper with expressed genes in developing mammalian eyes to evaluate further similarities and differences. Labeled single stranded portions of these identified fly eye development gene sequences could be used to probe the mammalian eye in an in situ hybridization process, or the mammalian RNA transcripts could be tested on a microarray and compared to the microarray results for Drosophila. Similar genes identified are likely involved in the very basic processes of morphogenesis. Fly genes without mammalian homologs are likely involved in processes specific to the development of the compound eye.


Campbell, Malcolm. 2002. Introduction to DNA Microarrays. <> Accessed 2003 Apr 23.

Halder, Georg et. al. 1995. Induction of Ectopic Eyes by Targeted Expression of the eyeless Gene in Drosophila. Science. 267: 1788-1792.

Michaut, Lydia et. al. 2003. Analysis of the eye developmental pathway in Drosophila using DNA microarrays. Proc Natl Acad Sci USA 100: 4024-4029.

SAGE Molecular Genetics Laboratory. 2003. SAGE Overview. <> Accessed 2003 May 1.


Questions or Comments? Please e-mail Rachel Patton McCord


Rachel Patton McCord's Molecular Homepage

Davidson College Molecular Biology

Davidson College Home