Abstract
Neurodevelopmental disorders like autism spectrum disorder affect millions globally, with substantial personal, medical, and societal consequences. While genome-wide association studies have implicated thousands of genetic variants in disease risk, the vast majority reside in non-coding regulatory regions where their functional mechanisms remain unclear. This challenge is exacerbated by the inability of existing massively parallel reporter assays (MPRAs) to distinguish transcriptional regulation from post-transcriptional control, and their poor performance in living tissue where cell-type-specific and developmental contexts are critical. This dissertation presents the Inferred Stability Optimized Massively Parallel Reporter Assay (ISOMPRA), a redesigned MPRA platform that independently quantifies nascent and mature RNA to disambiguate regulatory mechanisms while maintaining compatibility with in vivo applications. Chapter 1 establishes the conceptual foundation of this dissertation, demonstrating how non-coding regulatory variation in enhancers and untranslated regions converge on shared neurodevelopmental pathways despite remarkable genetic heterogeneity, followed by an investigation into MPRAs and their history. Chapter 2 validates ISOMPRA through systematic in vitro benchmarking, revealing that conventional approaches misclassify 60% of elements as transcriptionally functional enhancers or repressors when effects arise from altered RNA stability. Testing 466 autism-associated enhancer variants and 3,325 3′ UTR variants demonstrated that regulatory elements frequently exhibit dual functions—simultaneously modulating transcription and post-transcriptional processes. ISOMPRA achieved superior reproducibility with 8-fold reduced sequencing requirements compared to traditional barcoded methods. Chapter 3 implements ISOMPRA in developing mouse brain via AAV delivery, demonstrating substantially improved performance over conventional approaches in complex neural tissue. Using Cre-dependent designs in cortical excitatory neurons enabled cell-type-specific regulatory measurements from single tissue samples, revealing that most variants exhibit context-dependent effects, with regulatory function determined by cellular environment rather than being an intrinsic sequence property. Chapter 4 systematically investigates how promoter architecture shapes regulatory landscapes—an often-overlooked design parameter in MPRAs. Motivated by the reduced dynamic range and lower replicate correlation observed in vivo with minimal promoters (Chapter 3), I evaluated whether promoter identity could account for this variability and improve assay performance in neural contexts. Testing seven distinct promoters across human and mouse cell lines revealed that promoter choice alone explained nearly 40% of expression variance—far exceeding contributions from cell type or RNA processing. Strikingly, the majority of tested variants exhibited promoter-dependent regulatory effects, underscoring that variant interpretation cannot be divorced from promoter context. Moreover, neuronal-like cell types displayed a more restrictive, repression-dominated regulatory architecture relative to the more permissive expression profiles of non-neuronal cells. These findings establish promoter architecture as a critical determinant of regulatory activity and provide key design principles for optimizing in vivo MPRA performance. Chapter 5 translates these findings in vivo, demonstrating that strategic promoter selection substantially improves assay performance: optimized promoter choice increased correlation among biological replicates and enhanced detection sensitivity in brain tissue. Beyond these technical advances, systematic promoter comparison revealed that cell-type tropism determines which neural populations are functionally sampled. Only a small percentage of variants showed promoter-independent effects, with the majority exhibiting complex interactions between promoter choice and sex. Collectively, this dissertation establishes that regulatory variant function represents context-dependent phenomena shaped by promoter architecture, cellular environment, developmental timing, and biological sex. ISOMPRA provides essential infrastructure for translating genetic discoveries into mechanistic understanding of how non-coding variation disrupts neurodevelopmental gene expression programs, offering a framework for systematic variant interpretation in complex biological contexts.
Committee Chair
Joseph Dougherty
Committee Members
Jason Yi; Joshua Rubin; Michael White; Tychele Turner
Degree
Doctor of Philosophy (PhD)
Author's Department
Biology & Biomedical Sciences (Neurosciences)
Document Type
Dissertation
Date of Award
12-16-2025
Language
English (en)
DOI
https://doi.org/10.7936/8v6v-9954
Author's ORCID
0000-0002-5425-1476
Recommended Citation
Selmanovic, Din, "Optimizing Massively Parallel Reporter Assays for In Vivo Neurogenomics: Tools to Decode Enhancer Function in the Brain" (2025). Arts & Sciences Theses and Dissertations. 3691.
The definitive version is available at https://doi.org/10.7936/8v6v-9954