Lung cancer is one of the most aggressive cancers and the leading cause of cancer mortality in the US, mainly due to the lack of early detection. Meanwhile, gene expression profiling can identify molecular responses to carcinogen exposure and tumorigenesis. We have previously identified lung cancer-associated gene expression alterations in the normal bronchial airway epithelium of ever smokers with and without lung cancer. These alterations are the basis of a diagnostic test that is useful in clinical decision-making in patients with suspect lung cancer. Despite this success, further improvements in early lung cancer diagnosis are needed, along with a better understanding of airway biology during the initiation and development of lung cancer. Towards these goals, for the first aim of my thesis, I explored whether normal-appearing bronchial airway gene expression reflects lung cancer histologic subtypes. Genes differentially expressed in the bronchial airway between patients with lung squamous cell carcinoma and lung adenocarcinoma were identified and confirmed in independent data. Using a method developed based on independent component analysis (ICA), cell type-specific gene modules were derived from airway single-cell RNA-sequencing data and shown to be altered between lung cancer subtypes. Secondly, I sought to investigate whether integrating the bronchial airway molecular biomarker with radiomic features (i.e., quantitative features derived from radiographic images) could yield a better diagnosis for lung cancer screening. Using clinical variables, molecular signatures, and radiomic imaging features, I built and tested an integrated biomarker to improve discrimination between malignant and benign Indeterminate Pulmonary Nodules (IPNs). Finally, as COVID-19 became a pandemic during my thesis work, I sought to utilize large-scale genomic data from multiple cohorts to investigate possible clinical risk factors related to SARS-CoV-2 entry and disease severity. My analysis showed that smoking affects the expression of host genes involved in SARS-CoV-2 entry differently in the nasal and bronchial airways. The work has implications about how smoking might modulate SARS-CoV-2 infection and COVID-19 disease severity. Collectively, this work leverages computational approaches to identify airway biology associated with lung cancer subtypes, improve the diagnosis of lung cancer in patients with IPNs, and reveal relationship between smoking and SARS-CoV-2 infection. / 2024-02-18T00:00:00Z

Bovine respiratory disease (BRD) is a highly dynamic disease complex that results from host, microbial agent, and environmental interactions. Despite nearly a century of targeted research, BRD remains the most economically damaging disease in beef cattle production and appears to be increasing in global incidence. While modern modalities for BRD detection exist, clinical diagnosis and management decisions largely depend upon clinical observations and their associated risk of disease. Though these approaches lack precision, they remain in use for many reasons, including fiscal and time constraints within beef production systems. Advancements in high-throughput sequencing have demonstrated the ability to provide insight into complex biological disorders, leading to the development of predictive biomarkers and individualized therapy. Through the use of observational research methods and previously published data, transcriptome analyses were used to capture biological information related to the host-disease or host-pathogen relationship. These studies independently elaborated findings related to host management of inflammation, ultimately being associated with both acquisition and severity of BRD. Through advances in sequencing technology and data analysis methodology, novel components related to host inflammatory mitigation and antimicrobial defense are described for clinical BRD. Factors related to increased alternative complement activation, decreased specialized proresolving lipid mediator biosynthesis, decreased antimicrobial peptide production, and increased type I interferon stimulation were associated with severe clinical BRD. These findings define molecular networks, mechanisms, and pathways that are associated with BRD outcome, and may serve as a foundation for precision medicine in beef cattle.

With the rapid progress of next-generation sequencing (NGS) technologies, new tools and methods have emerged to investigate the transcriptomics of various organisms. RNA sequencing (RNA-seq) employs NGS to evaluate the presence and abundance of RNA transcripts in biological samples. This technique offers a comprehensive snapshot of the RNA dynamics within cells. With the ability to profile the entire transcriptome of organisms rapidly and accurately, RNA-seq has become the state-of-the-art method for transcriptome profiling, surpassing the traditional microarray approach. Single-cell RNA sequencing (scRNA-seq) was introduced in 2009 to profile the single-cell gene expression in highly heterogeneous samples such as brain tissue and tumors. The advancement of scRNA-seq technologies enables the in-depth transcriptomic study in each cell subtype. When selecting an scRNA-seq method, researchers must weigh the trade-off between profiling more single cells versus obtaining more comprehensive transcripts per cell, while considering the overall costs. The throughput of full-length scRNA-seq methods is usually lower, as each single cell needs to be processed separately to produce scRNA-seq libraries. However, full-length methods enable the researchers to investigate the splicing variants and allele-specific expression. Non-full-length methods only capture the 3' or 5' ends of transcripts, which limits their application in isoform detection, but as cells are pooled after barcoding for cDNA synthesis, the throughput is 2–3 orders of magnitude higher than full-length methods. We developed a droplet-based platform for full-length single-cell RNA-seq, which enabled the efficient recovery of full-length mRNA from individual cells in a high-throughput manner. The developed platform can process ~8,000 single cells within 2 days and detect ~20% more genes compared to Drop-seq. Besides scRNA-seq technology development, we also applied a low-input RNA-seq method to study the transcriptomics in different biological samples. When handling precious biological samples, a low-input method is necessary to profile the transcriptome of homogeneous cell populations. We first studied the epigenomic and transcriptomic regulations in colorectal cancer (CRC) using MOWChIP-seq, a low-input high-throughput method, in conjunction with our low-input RNA-seq approach. Fusobacterium nucleatum (Fnn) is closely related to the progression of cancers like CRC and pancreatic cancer. However, the molecular mechanisms of how Fnn adjusts the tumor microenvironment (TME) and leads to poor clinical outcomes are still unclear. In this in-vitro study, we characterized how hypoxia, an important TME ignored by previous research, facilitates Fnn infection of CRC and corresponding alterations of global epigenome and transcriptome. We infer that hypoxia has similar effects as Fnn infection alone on the CRC cells. The Fnn infection under hypoxia further boosts the proliferation and progression of CRC. We then applied our low-input RNA-seq method to study brain neuroscience and immunology. Psychedelics like DOI show promising clinical efficacy in patients with psychiatric conditions. Although psychedelics exhibit rapid antidepression action and long-lasting effectiveness compared to conventional treatment, their acute psychotic symptoms and potential for drug abuse discourage their application in clinical practice. In this case, it is important to comprehend the molecular mechanisms responsible for psychedelics' clinical efficacy. This understanding can pave the way for the development of improved treatments that do not rely on psychedelics. After profiling the transcriptome of mouse brain samples exposed to psychedelics with different post-exposure times, we concluded that the psychedelic-induced transcriptomic variations are more transient than epigenomic changes. In the second brain neuroscience project, we first applied 3-color FACS sorting to differentiate four neuron and non-neuron subtypes in human postmortem prefrontal cortex tissues. Then we profiled the gene expression of the four subtypes and validated the FACS sorting by examining the expression of marker genes. Differentially expressed genes between each subtype and the others were extracted and proceeded to gene ontology analysis. We identified unique altered biological pathways related to each subtype. The immunology research focuses on revealing the difference between low-grade inflammation and monocyte exhaustion, as well as the unique biological pathways they regulate. Therefore, we profiled the transcriptome of bone marrow-derived monocytes stimulated by PBS control, a low- or high-dose LPS. In addition to wild-type mice, we also included TRAM-deficient and IRAK-M-deficient mice. We concluded that low-dose LPS specifically regulates the TRAM-dependent pathway of TLR4 signaling, and high-dose LPS exclusively upregulates exhaustion markers by impacting metabolic and proliferative pathways. / Doctor of Philosophy / Transcriptomics is the comprehensive study of RNA transcripts derived from an organism's genome. RNA plays a vital role in maintaining the fundamental functions of cells and organisms. In eukaryotes, the genetic information stored in the DNA of cells is transferred to messenger RNA (mRNA) molecules through a process called transcription. These mRNA molecules serve as a bridge between DNA and proteins, as they carry the instructions encoded in genes to ribosomes for protein synthesis. Studying mRNA transcripts reveals various cellular mechanisms and their impact on overall organism function, gene regulation, and disease pathways. With the aid of next-generation sequencing, various RNA-seq approaches have been developed to study mRNA transcripts quantitatively in the past decades. To better understand the gene expression regulations in biological samples, we first applied bulk RNA-seq to profile the transcriptome of various samples under different conditions. Our in-house bulk RNA-seq protocol has been proven to be both high-performance and cost-effective compared to commercial kits. To better understand cellular diversity and uncover rare cell types in heterogeneous biological samples, we developed a droplet-based scRNA-seq platform that can recover full-length mRNA transcripts in a high throughput manner. It can profile the transcriptome of thousands of single cells within two days. It combines the advantages of the droplet-based scRNA-seq method (high throughput) and the well plate-based scRNA-seq method (full-length mRNA recovery).

Human immunodeficiency virus type 1 (HIV-1) is the causative virus for acquired immunodeficiency syndrome (AIDS). AIDS is characterized by chronic inflammation and reduction of CD4+ T-cells in the blood. This leads to the patient becoming immunocompromised and much more susceptible to disease in general. Different phenotypes in the progression of AIDS have been observed in patients in either progressing to AIDS faster as a Rapid Progressor (RP), or slower as a Long-Term Non-Progressor (LTNP). Researching elements that result in the LTNP phenotype is of interest as it has the potential to offer alternative treatments and therapies to those suffering from HIV and improve their quality of life. A separate genome wide association study into a population of LTNP patients had associated the R77Q mutation of viral protein R with the LTNP phenotype. Although this association has been controversial, recent work has shown that the R77Q mutation promotes apoptosis in a variety of cell lines compared to unmutated virus. However, the mechanisms behind the increase in apoptosis remain a place for further research. In this thesis, we attempted to elucidate some of the exact changes in gene expression between cells infected with the R77Q mutation and those without in the induction of apoptosis. We observed that apoptosis could be detected approximately 24 hours after infection via Annexin V staining, but there were no significant differences in the expression of genes within the first 24 hours. Furthermore, we observed 289 genes were then differentially expressed at 72 hours post infection. Analysis through SPIA revealed that the c-myc transcription factor pathway was activated in the R77Q infected cells and further analysis of the individual genes suggested less inflammatory signals in R77Q populations as well as an overall increase in antiapoptotic genes in WT infected cells. Exploration into the ANT1, Bax, and B-cl2 genes revealed that B-cl2 was upregulated in WT HIV infections and provided a degree of protection from apoptotic processes. Differences found in the expression of genes between R77Q and WT infected cells support the observation of increased apoptosis in R77Q infections and identify several targets for further research into the LTNP phenotype.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.