Fall 2024 Student Seminars

December 18

Katie Atherton
Advisors: Jenny Bhatnagar and Daniel Segrè
Title: Dysbiosis in Urban Trees

Abstract: The tree microbiome is a critical determinant of tree and ecosystem functioning, but human disturbances can disrupt natural microbial-tree relationships. Here, we report strong patterns of dysbiosis in the urban tree microbiome across a model urbanization gradient, where the most urbanized trees have low mutualistic root symbiont and potential leaf symbiont abundances, which were replaced with a diversity of root decomposers and plant, animal, and human pathogens. We also found dysbiosis in the biogeochemical cycling capabilities of the urban tree microbiome, with high potential for nitrogen loss and reduced capacity for methane consumption. City street trees had distinct, but less diverse microbiomes than trees in rural forests. Our results have implications for Urban One Health, suggesting that environmental health may be impacted by urbanization and management efforts to resolve urban tree dysbiosis might improve environmental resillience and human health

 
Regan Conrad
Advisor: Dr.Jennifer Beane
Title: Single-cell Characterization of Airway Field Alterations in Patients with Lung Squamous Premalignant Lesions

Abstract: Lung cancer remains the leading cause of cancer-related deaths globally, highlighting the urgent need for improved early detection methods. Low-dose computed tomography (LDCT) screening in high-risk populations reduces mortality by identifying lung cancers at earlier stages. Gene expression profiling from bronchial and nasal brushings has been shown to distinguish benign from malignant lung nodules detected via LDCT. Lung premalignant lesions (PMLs), which precede invasive cancer, offer an opportunity to further reduce mortality; however, effective detection strategies for PMLs are still limited. In this study, we employ single-cell RNA sequencing (scRNA-seq) to characterize molecular and cellular changes within the airway field of patients with PMLs.

We collected endobronchial biopsies from suspect PML sites, along with bronchial and nasal brushings from normal epithelium in high-risk individuals. Epithelial and immune cells were profiled using scRNA-seq via the CEL-Seq2 plate-based protocol, and cell/gene clustering was performed using Celda. Differential abundance analysis and linear mixed models were used to identify cell-type shifts and gene module expression differences across sample type, smoking status, and lesion histology. Findings were validated using independent bulk RNA-seq data from bronchial brushings.

Analysis of epithelial cells revealed nasal brushes were enriched in 9 cell clusters with high expression of gene modules involved in detoxification and mucus production. Across all sample types, 9 gene modules linked to xenobiotic metabolism were significantly associated with smoking status, with the strongest effects observed in bronchial biopsies, intermediate effects in bronchial brushes, and the weakest in nasal brushes, demonstrating a stepwise trend. We identified 5 basal cell clusters enriched for carcinoma in situ and tumor cells (high-grade basal cells), which expressed 7 gene modules associated with keratinization and inflammation. Notably, one bronchial brush sample contained cells clustering with high-grade basal cells. A gene module upregulated in these high-grade cells was associated with the most severe lesion histology observed during the procedure, as validated in bulk RNA-seq of bronchial brushings.

This work uncovers molecular and cellular alterations linked to smoking and lesion severity across the airway field in patients with PMLs. Our findings indicate that airway brushing samples, particularly from bronchial regions, may help detect high-grade PMLs and inform early intervention strategies.

 
December 4

Nathan Sahelijo
Advisor: Gyungah Jun
Title: Brain Cell-Type Specific Genetic Subtyping and Targeted Drug Repurposing for Alzheimer’s Disease

Abstract:
Alzheimer’s Disease is characterized by a complex and heterogeneous etiology and gradual progression, leading to high drug failure rates in late-stage clinical trials. To improve patient stratification and develop targeted therapies for precision medicine, we introduced an innovative procedure utilizing cell-based co-regulated gene networks and polygenic risk scores. By defining genetic subtypes based on the extremes of these distributions, we evaluated correlations with previously defined AD subtypes characterized by domain-specific cognitive functioning and neuroimaging biomarkers. Priority gene targets for these genetic subtypes were identified using a PageRank algorithm. Existing drugs targeting these genes revealed associations with tau pathology and neuroinflammation, highlighting candidates currently used for hormone therapy, diabetes, hypertension, and epilepsy as potential repurposing options for AD. Experimental validation with human induced pluripotent stem cell (hiPSC)-derived astrocytes demonstrated that estradiol, levetiracetam, and pioglitazone reduced the expression of APOE and complement C4 genes, indicating their potential for repurposing in AD therapy.

 
Yusuke Koga
Advisor: Josh Campbell
Abstract: Immune microenvironments (IM) in non-small cell lung cancer (NSCLC) and the surrounding lymph nodes (LNs) are associated with clinical stage and outcomes. However, the IM in Stage I NSCLC, particularly in the associated metastasis-negative LNs, has not been fully described. Using single-cell and spatial profiling, we have begun identifying immune cell subtypes and their spatial interactions in LNs, aiming to propose how various immune niches may relate to NSCLC severity.

 
November 20

Neal Kewalramani
Advisors: Mark Crovella & Ruben Dries
Title: Characterizing the epithelial-mesenchymal transition landscape using deep learning features

Abstract:
Epithelial-mesenchymal transition (EMT) is a cellular transition process that occurs in many different biological processes. The tumor microenvironment is one such biological area where the study of EMT is important because the transition to mesenchymal cells drives metastasis. However, little is known about the intermediate states within EMT and how these cells are driven to the transition process. Deep learning has been a useful tool to extract important features from input data, and these features have been shown to provide biological insight. Here, we use deep learning features to analyze co-fractionation mass spectometry data and single-cell RNA data to identify key proteins in these intermediate states.

 
Zainab Khurshid
Advisors: Lindsay Farrer & Xiaoling Zhang
Title: Exploring the relationship between Telomere length and Alzheimer disease

Abstract:
Alzheimer’s Disease (AD) is a neurodegenerative disorder that affects millions of people worldwide, particularly those over the age of 60. Telomere length (TL), a biological marker of aging, is known to decrease with age. While numerous studies have explored the relationship between AD and TL, the exact mechanisms and pathways linking the two remain unclear. To investigate this potential relationship, we conducted various analyses using multiple datasets, including the Framingham Heart Study (FHS), the Alzheimer’s Disease Neuroimaging Initiative (ADNI), and the Alzheimer’s Disease Sequencing Project (ADSP). These analyses included a genome-wide association study (GWAS) on AD and its interaction with TL, as well as a survival analysis to examine the relationship between the two while also adjusting for other covariates such as sex and APOE alleles.

 
November 6

Lina Kroehling
Advisor: Stefano Monti
Title: High-resolution Characterization of Age-specific Changes in HPV-negative HNSCC Tumors

Abstract:
It’s been shown that elderly patients with head and neck squamous cell carcinomas (HNSCC) exhibit diminished survival outcomes compared to their younger counterparts. While the convergence of aging hallmarks and cancer hallmarks offers valuable insights, further work is needed to elucidate the age-specific mechanisms influencing HNSCC beyond the shared characteristics of these biological processes. To this end, we have assembled a hiqh-quality human single-cell RNA-sequencing HNSCC atlas profiling more than 230,000 cells across more than 50 patients, with ages ranging between 18 and 89, which provides a unique resource to investigate age-associated changes in the disease’s heterogeneity.  To create the atlas, we integrated six publicly available single-cell RNAseq datasets from 54 HPV-negative patients. Cells were clustered, classified, and characterized by gene set enrichment analysis, both in the epithelial cell compartment and in the tumor microenvironment (TME). Differential cell type proportion analysis was performed to identify cell type compositional changes associated with age.  CNV analyses were performed to identify cancer subclones and assess level of copy number variation across tumors and age. Interestingly, we identify programs that associate with age, such as the partial-EMT signature, extracellular matrix-related processes, and dysfunction in epithelial cells, fibroblasts, and T cells respectively.  We also identify distinct cell populations, such as vascular endothelial cells, that are more prevalent in older patients.  Further analyses are ongoing, and we plan to functionally validate the hypotheses generated, specifically the presence of differentially abundant cell populations, and age-specific ligand-receptor signaling events that lead to tumor growth.

 
Nate Borders
Advisor: Mo Khalil & Kirill Korolev
Title: Hardware and control algorithms for continuous directed evolution

Abstract:
Continuous directed evolution is a powerful strategy for rapidly evolving biomolecules and organisms, with applications in medicine and metabolic engineering. This presentation explores the development of hardware and control algorithms that optimize two distinct continuous evolution systems: Phage-Assisted Continuous Evolution (PACE) and a novel bioreactor for evolving phototrophic microbes. PACE is a well-established method that harnesses bacteriophage replication on E coli. hosts to drive the rapid evolution of proteins. Despite its efficiency, PACE often faces challenges with (among others) hardware reliability, E coli host cells dysregulation, and phage failure to adapt to the rigors of PACE, highlighting the need for control over selection pressure, phage population dynamics, and evolutionary trajectory. We present a minimal PACE hardware which enables the use of a suite of algorithms designed to dynamically induce genetic drift to enable phage adaptation and flow rates to toggle selection pressures, which stabilizes PACE over extended runs while maximizing evolutionary outcomes.

Parallel to PACE, we have developed a bioreactor that supports the continuous evolution of phototrophic organisms, such as cyanobacteria. Unlike traditional systems, our bioreactor implements a selection algorithm that controls light exposure, temperature, or chemical concentration providing a tailored environment that maintains selective pressure for desired traits. By integrating these control algorithms, we demonstrate the potential to improve not only the speed and efficiency of continuous directed evolution but also the scalability of evolving diverse organismal functions across different biological platforms. This work paves the way for more versatile continuous evolution systems capable of addressing a wide range of biotechnological challenges.

 
October 23

Tong Tong
Advisors: Lindsay Farrer & Xiaoling Zhang
Title: Multi-ancestry genome-wide gene×age interaction study identifies novel loci associated with late-onset Alzheimer’s Disease

Abstract:
Background: Age is the largest risk factor for late-onset Alzheimer’s Disease (LOAD). Although >80 genetic loci have been associated with LOAD, little is known about the age dependencies of these associations except the APOE region.

Method: We performed cross-ancestry and ancestry-specific genome-wide gene-age interaction and age-stratified association study using TOPMed-imputed genome-wide association study (GWAS) data from Alzheimer’s Disease Genetics Consortium (ADGC) including 34,833 non-Hispanic Whites (NHW), 7,264 African Americans (AA), 3,232 East Asians (EA), and 2,024 Caribbean Hispanics (CH) aged 60 years and older. We chose an approximate median age-at-onset threshold of 75 to dichotomize participants into younger (Ncase:Ncontrol=10, 905:11,624) and older (Ncase:Ncontrol=9,982:14,842) groups. We first evaluated the association of LOAD with the interaction of SNP×age within each cohort using a linear mixed-effect model including a binary term for the age group, numeric age, sex, and the first 5 principal components (PCs) of ancestry. Results from each cohort were combined using a fixed-effect model to jointly estimate the regression coefficients of SNP and SNP×age terms, both within-ancestry and cross-ancestry analyses. We also performed an age-stratified analysis that included all controls regardless of age to increase power and using the same linear model without the interaction term. Results across cohorts were combined by meta-analysis using a fixed-effect model within-ancestry and a modified random-effect model across ancestry.

Result: In cross-ancestry joint meta-analysis, we identified 6 genome-wide significant (GWS) loci, 4 of which showed nominal evidence of age-dependent association with LOAD, including CD2AP (rs5876027, PJoint=2.33×10-8, PSNP×age=0.043), PICALM (rs583162, PJoint=2.34×10-8, PSNP×age=7.60×10-3), APOE (rs429358, PJoint=0, PSNP×age=1.58×10-7) and LILRA5 (rs1761461, PJoint=4.86×10-8, PSNP×age=0.032), for the first time. Within-ancestry joint meta-analysis identified novel associations with VXN (rs146432976, PJoint=2.29×10-8, PSNP×age=3.08×10-5) in AAs and PALM2-AKAP2 (rs67191673, PJoint=3.46×10-9, PSNP×age=0.013) in EAs. CR1 (rs6661489, PJoint=3.54×10-8) and TREM2 (rs75932628, PJoint=8.61×10-10) were associated with LOAD in NHWs but showed no evidence of age-dependent effects. In cross-ancestry age-stratified GWAS, BIN1, PICALM, and APOE region genes reached GWS in both strata, while MS4A6A was only significant in older group (rs7232, OR=0.88, P=4.23×10-8). A GWS association with PALM2-AKAP2 (rs67191673, OR=1.94, P=3.64×10-9) was identified in the EA younger group.

Conclusion: We identified several ancestry-specific age-dependent association with previously established and novel loci.

 
October 9

Zhaorong Li
Advisors: Juan Fuxman Bass and Adam Labadorf
Title: Decipher Cell Type Specific and Prognosis Related Gene Signatures from Tumor RNA-Seq Datasets

Abstract:
The study of tumor micro-environment focuses on the behaviors of different cell types in the tumor tissue. While single cell level data allows us to study the transcriptomic and epigenetic landscapes of different cell types in tumor micro-environment, we do not have enough of them to associate the certain transcriptomic/epigenetic profiles with prognosis and phenotypes. Here we propose to use the BayesPrism algorithm (Chu et a., 2022) to perform de-convolution analysis and obtain the cell type specific expression profiles, and analyze them using independent component analysis to explore cell type specific gene expression signatures that are associated with prognosis and phenotypes.

 
September 25

Kelley Anderson
Advisors: Jennifer Beane and Marc Lenburg

Abstract:
Lung cancer is the leading cause of cancer-related death. Lung adenocarcinoma (LUAD) is the most common form of lung cancer, and is a heterogeneous  disease with significant variability in clinical outcomes. Features of LUAD premalignant lesions (PML) leading to aggressive disease are poorly characterized. We hypothesized that transcriptomic changes in PML are linked to distinct clinicopathologic features that lead to malignant disease. We performed bulk RNA and exome sequencing of tissue from tumor resections that included PML, tumor, and normal tissues. We utilized archetypal analysis of transcriptomic data to capture phenotypic heterogeneity in PML. These archetypes we identified correlated with clinical and molecular phenotypes. Each archetype was distinctly enriched for different biological pathways, including immune and cell cycle pathways. Changes in alveolar intermediate cell types and enrichment of driver mutations also differentiated archetypes. By projecting these archetypes into early stage LUAD sequencing data, we found that different archetypes were also significantly associated with prognosis. Molecular signatures measured in PMLs may enhance our understanding of diverging patterns of dysregulation that occur during LUAD carcinogenesis, and implicate immunotherapeutic strategies to prevent their progression to cancer.

 
Nick O’Neill
Advisors: Xiaoling Zhang and Lindsay Farrer
Title: Characterizing Cell-type and Neuron Subtype Activity and Abundance in Alzheimer’s Disease

Abstract:
Alzheimer’s disease (AD) is a complex, neurodegenerative disease whose primary progression is characterized by increasing amyloid-β (Aβ) plaque burden followed by increasing neurofibrillary tau tangles (NFT) and cognitive decline. Many subjects fall outside of this progression, exhibiting NFTs without Aβ or maintaining cognitive performance in the face of AD pathology. AD progression, including exceptions to traditional progression, are characterized by changes in cell-type abundance and cell-type specific activity. Our work investigates these topics by utilizing single-nuclei datasets in combination with large-scale bulk RNA-seq and whole genome sequencing datasets.  We deconvolute AD brain bulk RNA-seq to assess the relationship between cell-type abundance and AD endophenotypes, including cognitive resilience to AD pathology, in brain regions vulnerable and spared in the disease. We also identify and discuss genetic drivers of cell-type abundance changes. Finally, we generate highly cell-type specific AD polygenic risk scores (ct-ADPRS) to investigate the relationship between cell-type activity and AD progression.

 
September 11

Michael Silverstein
Advisors: Daniel Segrè and Jennifer Bhatnagar
Title: Microbiome structure and implications for climate change mitigation

Abstract:
Climate change continues to threaten the stability of the biosphere, increasing the demand for mitigation strategies. One exciting opportunity is soil microbiome engineering, i.e., the use of a microbial inoculum to induce enduring, stable modifications to a natural soil microbial community and the ecosystem functions it regulates. While environmental microbiome engineering has existed for at least a century, the properties underlying this process and strategies for maximizing its efficacy are yet to be uncovered. Here, I will discuss a recent experiment (https://www.nature.com/articles/s41559-024-02440-6) which suggests how more complex environments, like forests, may be more susceptible to microbiome engineering than simple ones and my plan for upcoming related work.