Though numerous treatments for breast cancer have been developed, recurrence still exists, biomarkers and risk assessment tools are thus desired by the field. Cell-free RNA in blood or serum is now receiving more and more interest as a pool of biomarkers and RNA sequencing was believed to be a powerful tool for its analysis. However, development of the field was hindered by large volume of serum required for RNA extraction.

In this study, by circumventing RNA extraction and with the inspiration from single-cell RNA sequencing, we developed a technology being able to construct RNA sequencing libraries in large scale with microliters of direct unprocessed serum input. Optimizations for sequencing and mapping qualities, library complexity and library construction efficiency were conducted by varying serum input volume, adding custom sequencing primer and applying liquid handling instrument epMotion 5075. The optimized strategy “construct 96 different libraries in one single automated batch with 7ul serum input” was proposed accordingly. A huge diversity of genes could be found in serum with this technology. The developed technology was then applied to 96 breast cancer patients’ serums for its performance and potential in clinical cases. Based on the sequencing data: 465, 601 and 1259 genes were identified by three methods respectively to behave differently in recurrence and non-recurrence patients; The two populations could be separated by principle component analysis with all three sets of genes mentioned above; Preliminary recurrence risk assessment was conducted successfully with the classifier random forest.

Epigenetic modifications play a pivotal role in gene regulations and thus heavily influence phenotypic outcomes. Comparative epigenomics, which incorporates both epigenome and genome into interspecies comparison, has become a powerful tool for revealing regulatory features of the genome and evolutionary properties of the epigenome. With the rapid growth of the high-throughput sequencing technologies, a tremendous amount of epigenomic datasets has been generated in various species, urging the development of systematic and quantitative approaches to the integrative comparison of the genome and the epigenome.

In this dissertation, I first presented a likelihood approach to testing hypotheses on the co-evolution of genome and epigenome. By converting evolutionary biology hypotheses into explicit probabilistic forms, I was able to establish a class of evolutionary models to quantify the dependence of interspecies epigenomic variations on underlying sequence variations between orthologous regions. To better facilitate functional annotation of the genome, I further developed an algorithm named EpiAlignment based on the evolutionary models, which incorporates both sequence and epigenomic data into alignment. A web service was also implemented to make the algorithm accessible to the research community at large. Using EpiAlignment, I searched for genomic region with similarities in both sequences and epigenomic modification patterns between human and mouse, revealing thousands of pairs of cis-regulatory elements with potential functional correspondence. In summary, this dissertation work brings mathematical rigor to comparative epigenomic studies, laying foundation to functional genomic studies with evolution-based approaches.

SMARCAD1 is a matrix-associated actin-dependent regulator of chromatin that encodes SWI/SNF subfamily of helicase proteins. It has been shown that the functions of SMARCAD1 are linked to histone 3 citrullination, a specific type of histone post-translational modifications. Histone citrullination can lead to alterations in protein functions, and affect gene expressions. This peptidylarginine deiminases (PADIs) catalyzed post-translational modification also increases in the

progression of cancer. Interestingly, human renal cancer cells exhibit low to non-detection of SMARCAD1 protein expression. It is anticipated that an association between SMARCAD1 and histone 3 citrullination function critically in renal cancer and tumor formation. Western blot was performed on mouse cells fist followed by multiple renal cancer and non-cancer human cell lines, to determine SMARCAD1 and histone 3 citrullination protein expression levels and their potential correlation. Multiple cell treatments were introduced and CRISPR/Cas9 gene editing system was used to create SMARCAD1 gene knocking out condition. A negative correlation was determined between SMARCAD1 and citrullination expression level, with a potential positive feedback loop in between SMARCAD1 and histone 3 citrulline enzyme PADI4 in certain cell lines. Moreover, this negative correlation between two protein levels serves as an indicator of renal cancer existence; decrement of SMARCAD1 protein level holds true in renal cancer cells, in parallel with an increment of the histone 3 citrulline protein expression.

Ribonucleic acids (RNAs) serve as one of the main building blocks of organic life and with the advent of new understanding beyond an intermediary for protein coding, understanding it's function becomes imperative. Previous tools study RNA expression from cell populations such as tissue or cell lines, however, an expansive area of research has developed for the individual cells. The recognition of cell heterogeneity in normal and diseased tissues further highlighted the importance of accurate RNA detection in single cells. Single molecule RNA fluorescence in situ hybridization (smRNA-FISH) has revolutionized RNA detection and quantification by visualizing single RNA transcripts.

The work in this dissertation improves upon smRNA-FISH by developing quantum dot based fluorescent in situ technologies that uses hybridization of quantum dot-labeled single stranded DNA oligonucleotides to the RNA targets for visualization and quantification. In chapter 2, we demonstrate our proof-of-concept quantum dot single molecule RNA-FISH (qdot-smRNA-FISH) by comparing the sensitivity and specificity of qdot-smRNA-FISH against commercialized smRNA-FISH. In chapter 3, we describe qdot-smRNA-FISH for detection of isoforms (isoFISH), where different from smFISH that requires dozens of hybridization probes per gene, isoFISH requires only two hybridization probes per splicing variant. In chapter 4, we used the co-localization of multi-color quantum dot FISH (CoQ-FISH) to observe for the first time a fusion transcript formation of Eml4-Alk in non-small cell lung cancer biopsy samples from trans-splicing and not chromosomal rearrangement. In chapter 5, we develop cell surface qDot RNA-FISH (Surface-FISH), a method for visualization of RNAs on the cell surface. We identified RNA targets using RNA-seq on membrane coated nanoparticles from EL4 and applied Surface-FISH to observe regions of Malat1 and Neat1. In chapter 6, we developed Surface-PromFISH that uses randomized antisense probes to promiscuously label as many csRNAs as possible and made it compatible with co-staining of cell membrane proteins. We subjected the hybridization targets to RNA sequencing (Surface-FISHseq), which led to 51 putative PBMC csRNAs and we identified a potential function where antisense blocking of FNDC3B and CTSS lowers monocyte attachment.

While it is widely understood that proteins are the functional tools of a cell, there are still no methods that efficiently illuminate the cell-wide networks of protein-protein and protein-RNA interactions through which proteins can act. Existing methods of detecting protein interactions primarily take one of two forms, they are either highly parallelized ‘one-by-one’ assays or several ‘one-by-many’ assays. However, these efforts are expensive and time intensive when attempting to address a cell-wide network.

This dissertation describes the development and validation of three technologies, a protein barcoding technology which generates a library of proteins labeled with specific nucleic acid sequences (SMART-Display), a high throughput proximity ligation technology that elucidates protein-protein interactions (PROPER-Seq), and a high throughput proximity ligation technology that elucidates protein-RNA interactions (PRIM). Leveraging these technologies, protein-protein and protein-RNA interactions can be assessed for a given cell type with an “all-vs-all” approach. The PRIM and PROPER-Seq workflows are characterized by a drastic increase in the number of interactions assayed in a single experiment relative to existing techniques, minimal labor, cost, and time, suitability for automation, and accessibility to any benchtop scientist, as they are devoid of the need for specialized technology or equipment.

SMART-Display produces highly complex protein libraries that closely reflect the mRNA population of the cell-type they are derived from, and which contain protein-specific DNA barcodes that can be used for proxy identification of the proteins themselves. This enables the application of PRIM and PROPER-Seq with libraries containing tens of thousands of uniquely labeled protein interactors. Precision and sensitivity analyses indicate that PROPER-Seq technique demonstrates a detection of interactions that is similar to, or better than, the levels demonstrated by gold-standard techniques such as yeast-2-hybrid. Several novel PROPER-Seq interactions were validated in vivo. Preliminary PRIM data indicates sufficient quality for precision and sensitivity analysis and both literature and in vivo validation.

The spatial interactions between chromatin, RNA, and proteins inside the cell nucleus underpin the exquisite execution of cellular functions. RNA, one of the most versatile molecules, plays crucial roles in forming membrane less nuclear bodies, modulating chromatin architecture, and regulating gene expression. Despite a few well-studied ncRNAs, the roles of many other RNAs in gene regulation and chromatin organization remain unclear. This thesis aims to provide a systematic approach to delineate functionally important chromatin-associated RNAs (caRNAs).Chapter 1 establishes a baseline model of caRNA localization relative to chromatin structure and the influence of caRNAs on chromatin organization. By jointly comparing chromatin structure with genome-wide RNA-chromatin interactions in human embryonic stem cells, I found that caRNA localization is largely confined and shaped by local chromatin structures such as compartments, topologically associated domains (TADs), and chromatin loops. Chromatin loop anchors are hotspots for caRNA interactions, and depletion of loop anchor caRNAs leads to the formation of new chromatin loops. Chapter 2 addresses the challenge of identifying functional caRNA-chromatin interactions, where nascent transcripts significantly contribute to RNA-chromatin interactions, complicating the identification of functional caRNAs. I proposed a modified version of the iMARGI technique that introduces RNase A treatment before crosslinking. This method effectively removes non-protected caRNAs, such as nascent transcripts, leaving only stringent and protected interactions. We found that RNase-inaccessible caRNAs exhibit higher sequence conservation and are enriched at transcription factor binding sites, histone modifications, and other functional genomic regions. Additionally, we identified specific RNA species associated with transcription factor and histone modification sites. Chapter 3 presents MUSIC, a novel technology I co-developed for simultaneous profiling of DNA-DNA, RNA-DNA, and RNA-RNA interactions along with the transcriptome at single-cell resolution. Applying MUSIC to brain tissue from control and Alzheimer's disease (AD) patients, I find that transcriptomically aged cells have fewer local chromatin contacts and are more prevalent in AD patients. The research also demonstrates that the cell type exhibiting chromatin contacts between cis eQTL and a promoter tends to be that in which these cis eQTL specifically affect the expression of their target gene. Furthermore, I reveal that female cortical cells exhibit highly heterogeneous interactions between XIST non-coding RNA and chromosome X, with decreased XIST-chrX enrichment linked to increased chrX gene expression. In conclusion, this thesis provides valuable tools and insights into the functional roles of caRNAs in nuclear organization and gene regulation, offering new technologies for studying the RNA-DNA interplay in both normal and disease cellular states.

At any time hundreds of thousands of macromolecular interactions occur in a cell, mediating functions that maintain normal cellular activities. A fundamental problem in molecular biology is to catalog these interactions and to decipher their functional consequences. High throughput sequencing has made it possible to characterize some of these interactions rapidly, at high-resolution, and in vivo (e.g., protein-DNA binding via ChIP-Seq and protein-RNA binding via CLIP-Seq). But many interactions are not susceptible to these methods (e.g., RNA- RNA complexes, ncRNA-DNA binding, protein-protein interactions).

This thesis aims to address this gap by coupling high-throughput sequencing with proximity-ligation-based methods. In proximity ligation, spatially proximate nucleic acids ligate to one another, forming a chimeric oligonucleotide. Observation of a chimera composed of X and Y suggests that X and Y must have been near one another in the original sample. As a result, questions about spatial arrangement become questions about sequence composition, making it possible to take advantage of high-throughput sequencing. Using this general approach, we developed high-throughput technologies to study RNA interactions with different types of molecular partners: RNA, chromatin, and lipid.

In Chapter 1, I review and discuss many different high-throughput techniques to map RNA structure and RNA-RNA as well as RNA-chromatin interactions. I also provide the biological insight that can be gained from the type of data generated by the new technologies.

In Chapter 2, I describe the development of MARIO, a technology to map RNA-RNA interactions. This method produces a global map of RNA-RNA interactome and RNA structures in vivo. The information will provide roadmaps to systems level understanding of cellular regulations through RNA-RNA interactions.

In Chapter 3, I describe the development of MARGI, a technology to map RNA-chromatin interactions. Mapping RNA-chromatin interactions identify RNAs that bind to the chromatin and shed light on the functional roles of chromatin associated RNAs in gene regulations.

In Chapter 4, I describe the discovery of a new class of RNA, named cell surface RNA. These are RNAs that are expressed outside of the cell membrane. I also develop methods to isolate and identify these surface RNAs and protein binding partners of the RNAs

RNA has been proved to interact with lipid bilayers and various proteins near the membrane, however; the existence of cell surface RNA was less explored. Using CLICK reaction and fluorescence visualization, preliminary research in Zhong Lab had suggested the existence of cell surface RNAs(csRNA) that attached firmly on the outer membrane of mouse and human cells. For the first step in this study, functionality assay of cytotoxicity, which measured by LDH release of human immune Natural Killer cell line NK92, showed that the global csRNA perturbation could significantly impact NK92’s cell killing activity. To further capture and characterize csRNAs using sequencing data, we have developed a method to pull-down lipid-associated RNAs using lipid-coated beads followed by RNA-sequencing. csRNAs were identified based on differential analysis of RNAs bound to 8 types of membrane-associated lipid against the control bead. Candidate RNAs discovered were validated using two sequencing techniques developed by the Zhong Lab, i.e. SurfaceClick and SurfaceSeq, to further remove background noise. Functional and structural analysis of validated csRNAs showed significant enrichment on immune and cancer-related functions as miRNAs. Furthermore, the analysis result further indicated potential cellular functions such as cell-cell recognition, structural support, and signaling regulation. In a nutshell, this study indicated the existence of RNA on cell’s outer membrane as functional and structural groups for cellular functions such as anti-tumor cytotoxicity and immune response.