May 8

Marlene Tejeda
Advisor: Lindsay Farrer
Title: Multiple Viruses Detected in Human DNA are Associated with Alzheimer Disease Risk

Abstract:
Multiple infectious agents have been identified as risk factors for Alzheimer’s disease (AD). Viruses can cause inflammatory damage and potentially increase the formation of the AD hallmark proteins amyloid-β protein and hyperphosphorylated tau. Here we explore the role of herpes viruses in the pathogenesis of AD and highlight the importance of genetic factors in modulating these associations.

 
March 20

Helen Scott
Advisor: Daniel Segrè
Title: Metabolic Flux Modeling in Marine Ecosystems

Abstract:
Ocean metabolism constitutes a complex, multiscale ensemble of biochemical reaction networks, harbored within and between the boundaries of a myriad of microbial cells. Gaining a quantitative understanding of how these networks operate requires mathematical tools capable of solving in silico the resource allocation problem that each cell faces in real life. Towards this goal, stoichiometric modeling of metabolism, such as flux balance analysis, has emerged as a powerful computational tool for unraveling the intricacies of metabolic processes in microbes, microbial communities, and multicellular organisms. Here we provide an overview of this approach and its applications, future prospects, and practical considerations in the context of marine sciences. We explore how FBA has been employed to model marine organisms, help elucidate nutrient cycling, and predict metabolic capabilities within diverse marine environments, and highlight future prospects for this field in advancing our knowledge of marine ecosystems and their sustainability.

 
March 6

Andrew Chen
Advisor: Stefano Monti
Title: Network-Based Multi-Domain Recontextualization of Omics Signatures

Abstract:
Gene and protein expression signatures (omics signatures) provide a summarized view of the distinctive biological processes occurring in a cell. These omics signatures are central to our understanding of the molecular changes that occur in cells in response to different perturbations, e.g., gene knockdown, small molecule treatment, or disease phenotype. However, these omics signatures are biased by our use of cancer cell lines and animal models that are more feasible to generate in experimental settings. To address this bias, we are working on developing a page-rank based network propagation method that leverages gene regulatory networks to prioritize portions of the gene signature that are more relevant to the biological context of choice.

 
February 21

Yusuke Koga
Advisors: Joshua Campbell
Title: Validation of the bronchial airway-derived signature for lung function decline profiled by RNA-seq

Abstract:
Previous studies have identified microarray-based bronchial gene expression signatures associated with COPD and the rate of FEV1 decline. Validation of the phenotypic association of these signatures was performed on an independent airway brushings cohort profiled with total RNA sequencing, in which concordant enrichment of the signatures based on clinical phenotypes was ascertained. (FDR q < 0.05) Furthermore, in a single cell cohort, genes displaying increased and decreased expression with rapid decrease of FEV1 were enriched in the goblet and ciliated cell populations, respectively. These results demonstrate the robustness of the gene expression signatures despite platform differences (Affymetrix microarrays vs. RNA-seq).

 
January 24

Regan Conrad
Advisor: Jennifer Beane
Title: Single cell gene expression analysis of the field of precancerous LUSC

Abstract:
Prior work has described transcriptional alterations in the airway field of patients with lung cancer that are associated with a decrease in immune-related activity, shifts in secretory cell type populations, and enrichment of gene pathways associated with proliferation and cell cycle. However, much of this knowledge comes from using bulk RNA-seq to investigate changes in the airway field of injury; there have not yet been any such studies conducted using scRNA-seq. Here, we leverage the increased resolution of single-cell sequencing methods to examine airway brushes obtained from patients with lung squamous premalignant lesions and lung cancer to understand key changes in cell states associated with lesion histology and smoke exposure.

Bronchial brushes, nasal brushes, and endobronchial biopsies were collected from the airways of 46 patients undergoing bronchoscopy to follow lung squamous premalignant lesions. The cells were sorted into 96-well plates and gated on CD45-, EPCAM+, and Live sorting for the Cel-Seq2 sequencing pipeline. The celda (Cellular Latent Dirichlet Allocation) package was used to cluster the cells and find modules of co-expressed genes.

After separate quality control of cells from bronchial brushes and endobronchial biopsies, 14,187 cells remain for clustering and analysis. Using canonical markers, we can identify cell types expected from airway samples, and observe a shift in proportions of secretory cell types associated with smoke exposure. We have also identified a population of basal cells in both bronchial brushes and biopsies that originates from patients with high-grade lesions; this population clusters separately from the basal cells that originate from patients with lesions of a lower grade. The gene module upregulated in this high-grade basal cell population also associates with samples taken from higher-grade, progressive lesions in other datasets.

Our results support prior studies conducted with samples profiled with RNA-seq that have shown field effects due to tumor presence and smoke exposure. Additionally, a high-grade basal cell population observed in both endobronchial biopsies and brushes and a set of co-expressed genes upregulated in this cell population suggests a cell type shift associated with lesion histology and could be the basis for a high-grade premalignant lesion signature in the accessible upper airway.

 
January 10

Neal Kewalramani
Advisors: Mark Crovella & Andrew Emili
Title: Identifying Key Cancer Protein-Protein Interactions in Endothelial-Mesenchymal Transition Using Co-Fractionation Mass Spectrometry and Neural Networks

Abstract:
The tumor microenvironment allows for the proper conditions for tumors to metastasize. Endothelial cells (ECs) in the tumor microenvironment allow for the formation of new blood vessels by undergoing a partial phenotypical change to mesenchymal stem cells. Therefore, to gain an understanding of tumor metastasis and identify druggable targets to prevent it, knowledge of endothelial-mesenchymal transition (EMT) is needed. In this study, we use co-fractionation mass spectrometry (CFMS) data to identify the key proteins involved in EMT for tumor growth. The CFMS experiment consists of relative protein abundances across eight timepoints. We identify key changing PPIs by using a novel application of a neural network architecture. The results indicate that we are able to identify key proteins that are involved in EMT or cancer from the literature, and our methodology can be used to suggest new previously unidentified proteins for future experimentation and validation.

 
Nate Borders
Advisor: Mo Khalil
Title: Accessible and Improved Phage-Assisted Continuous Evolution (PACE) with the Development of a Miniature Bioreactor: the min-eVOLVER

Abstract:
Phage-assisted continuous evolution (PACE) is a transformative method for the directed evolution of macromolecules, yet its widespread application has been hindered by the need for bulky, expensive hardware and a susceptibility to various failure modes. To overcome these obstacles, we present the min-eVOLVER: a cost-effective and Python programmable bioreactor system that significantly enhances the reliability and accessibility of PACE. We apply the programmability of min-eVOLVER to make improvements to the PACE algorithm, including control of bacterial host cell density and chemical induction of evolutionary drift. We hypothesize that this will vastly increase the evolutionary distance that can be traveled in a single campaign, decrease probability of phage washout, and effectively obviate the commonly used but slow and failure prone onramp to PACE, PANCE (Phage-assisted non-continuous evolution). These enhancements not only streamline the PACE process but also make the min-eVOLVER a preferred tool for both prototyping and conducting PACE campaigns. Consequently, the min-eVOLVER represents a pivotal step forward in the field of molecular evolution, offering a more efficient, reliable, and accessible platform for scientific exploration and discovery.