Modification of cis-regulatory sequences is an important means by which developmental processes can evolve and generate changes in patterning and morphology. Understanding how these cis-regulatory enhancer sequences control gene expression is critical not only for understanding the mechanistic basis of development, but also for understanding how the systems are modified or evolve over time.

In this work I describe the discovery and characterization of many new enhancers that regulate genes for which a similar enhancer driving a similar expression pattern was previously known; these are "shadow" enhancers. The genes examined are critical developmental control genes responsible for the patterning and subdivision of the early Drosophila embryo. At first glance, many of these enhancers seem redundant, but are found to be evolutionarily conserved, suggesting a role in fitness and specific developmental function.

Sufficiently controlled tests of shadow enhancer function required the development and use of techniques such as bacterial artificial chromosome (BAC) recombineering and quantitative confocal imaging. Recombineering allowed the precise modification of large genomic regions and an unprecedented level of control in evaluating enhancer function when combined with the use of reporters and genetic rescue experiments. The ability to accurately quantify gene expression was aided by the development of a reporter gene containing an intron, allowing clear visualization of sites of active transcription. These tools facilitated the evaluation of shadow enhancer function.

Evidence is presented which shows that in some cases one enhancer of the shadow enhancer pair can be removed without disrupting core function of the associated gene under normal developmental conditions. The system breaks down, however, under stress such as elevated temperatures during development, but only when one enhancer is missing. These results suggest that shadow enhancers represent a novel mechanism for the canalization of gene expression in varying environmental conditions and genetic backgrounds.

It is shown that in several cases shadow enhancers help ensure complete patterns of transcriptional activation across domains of gene expression. We present a mechanistic model in which increasing the number of enhancers for a given pattern helps ensure proper transcriptional state even allowing for stochastic interruptions of underlying molecular interactions such as enhancer/promoter looping or activator/repressor binding. It is suggested that shadow enhancers play a key role in ensuring robustness and reliability of gene expression patterns in development.

We propose a model whereby cryptic duplication events lead to the birth of shadow enhancers. Such enhancers can provide immediate value for fitness, for example, by ensuring robustness in response to environmental or genetic variations. The second enhancer might facilitate the evolution of novel patterns of gene expression, or become incorporated into the core developmental machinery to produce precise and rapid patterns of gene expression.

While some vertebrate species such as zebrafish have demonstrated a robust repair and regeneration ability, the extent of this regenerative capacity in mammals is limited. In humans, it is unknown whether this process even occurs. Vision loss resulting from retinal ganglion cell (RGC) death is a leading cause of irreversible blindness. There are several areas of active research aimed at restoring retinal ganglion cells. One of them is endogenous repair. However, for this approach to work, we need to gain a better control of the developmental mechanisms leading to RGC formation. In this study, we use ectopic overexpression of four specific transcription factors involved in ganglion cell specification and differentiation to bring about the direct conversion of ganglion cells from human pluripotent stem cells. If these newly differentiated neurons turn out to be truly bonafide ganglion cells, then the idea of generating new ganglion cells to improve the prognosis of retinal diseases will be better understood.

Retinal degeneration (RD) is a leading cause of vision impairment and blindness worldwide. Diseases which cause RD, including glaucoma, retinitis pigmentosa (RP) and age-related macular degeneration (AMD) are characterized by dysfunction and/or loss of key retinal neurons. Studying these human diseases in model organisms is useful but often approaches a limit at which the model organism does not sufficiently capture the human context. Many groups are utilizing in vitro methods to study the human retina, the most prominent being human induced-pluripotent stem cell (iPSC) derived neurons and iPSC-derived 3D retinal organoids (rORGs). A key first step to studying diseases like AMD and RP is establishing accurate in vitro models that can recapitulate the complexity of the retina and its many cell types. To address this need, I explored transcriptomic patterns in rORGs focusing on photoreceptors (PRs). I found that the rORG-derived PRs (O-PRs) exhibited a PR-like identity with both rod and cone markers detected. Importantly, I also compared these in vitro O-PRs with in vivo datasets, and found that they share similarities in their transcriptomic profiles. This work is intended to serve as a resource for scientists seeking to perform similar analyses on retinal datasets as well as cell types beyond the nervous system. Toward this goal, I have documented all the code used to generate results in a publicly available repository with guides on how to perform the analysis regardless of prior bioinformatics experience.