15064 research outputs found
Sort by
Deciphering Mechanisms of CAR T Cell Differentiation and Efficacy with Multimodal Sequencing Analysis
No full textChimeric antigen receptor (CAR) T-cell therapies have transformed the therapeutic landscape for hematological malignancies by providing patients with relapsed or refractory disease a late-line option. While many patients can achieve long-term remission, the clinical efficacy of CAR T cells has been limited by high rates of non-response and early relapse. Thus, new strategies to improve existing CAR T-cell products and develop novel CAR T-cell therapies are necessary. In this dissertation, I leverage multimodal sequencing analysis of patient-derived CAR T cells to investigate three central questions in the field: 1) How do CAR T cells differentiate following infusion? 2) What immunological factors enable successful salvage CAR T-cell therapy? 3) Can durable patient responses be predicted based on early CAR T-cell signatures? In Chapter 1, I provide an overview of CAR T-cell therapy and current discoveries of in vivo CAR T-cell differentiation and functionality enabled by single-cell sequencing. In Chapter 2, I describe how post-infusion CD19 CAR T cells follow a two-stage differentiation model governed by distinct phenotypic and clonotypic profiles. In Chapter 3, I report that a T resident memory-like BCMA CAR T-cell phenotype plays an important role in enabling successful salvage therapy after failed BCMA CAR T-cell therapy. In Chapter 4, I provide an in-depth characterization of the early drivers of CAR T-cell stemness that distinguish durable versus nondurable response in patients receiving BCMA CAR T-cell therapy. Altogether, these findings provide a mechanistic foundation that enable the development of next-generation CAR T-cell therapies
The Unsettled Queer Body in Spring Fever (2009): Politics of Looking, Haptic Voyeurism, and A Pornographic Discourse
Get PDFThis thesis examines queer embodiment and spectatorship in Lou Ye’s Spring Fever (2009), a Chinese underground film that centers on Jiang Cheng, a closeted gay man navigating romantic entanglements and social marginality in 2000s Nanjing city. By analyzing the film’s use of voyeuristic perspectives, haptic visuality, and pornographic discourse, I argue that Lou crafts a complex cinematic language that reveals how queer life is simultaneously visible and vulnerable, legible and erased, desired and disciplined. Through techniques such as handheld camerawork, jump cuts, intimate close-ups, and perspectival alignment with a character’s POV, Lou constructs a haptic aesthetic that not only renders queer bodily desire as something continuously negotiated, but also implicates the viewer in both the quasi-surveillant perspective and affective disorientation experienced by the protagonist. In particular, the thesis analyzes the film’s sex scenes as moments of ambivalent intensity. Rather than functioning as pornographic spectacle or narrative climax, these scenes foreground the condition in which queer intimacy remains inescapably surveilled and persistently unsettled—even within acts and spaces of supposed privacy and closeness. In this context, the queer body becomes not a symbol of liberation, but a surface of inscription—marked by historical trauma, state power, and personal compromise. By situating Spring Fever within broader discourses of queer theory, Sinophone cinema, and Chinese modernization, this thesis contributes to conversations on how queerness is represented and felt across cultural boundaries. It challenges Euro-American paradigms of queer visibility by attending to China’s “nonconfrontational” strategies of resistance and the cinematic techniques used to express them. Ultimately, Lou Ye’s film offers not a celebration of queer identity, but a meditation on its precarious endurance—one that invites viewers to reckon with their own role in the politics of looking, feeling, and knowing
Polyploidy, Phylogeny, and Morphology of <i>Rhododendron's</i> (Ericaceae) 'Hardy Mountaineers' from the Hengduan-Himalayan Mountain Region
Get PDFMountain systems have long served as hotspots of plant diversification, where complex evolutionary processes can obscure species boundaries and phylogenetic relationships. This dissertation investigated the evolutionary history, polyploidy, and morphological diversification of Rhododendron (Ericaceae) subsection Lapponica and closely related taxa in section Rhododendron, a group that dominates alpine ecosystems in the Himalaya and Hengduan Mountains. By integrating chloroplast phylogenomics, nuclear target-capture sequencing, and quantitative morphological analyses, I examined how geography, ploidy, and morphology relate to species' relationships in this taxonomically challenging group. In Chapter 2, plastome phylogenies from across section Rhododendron revealed frequent discordance with traditional taxonomy. At broad phylogenetic scales, plastid similarity correlated significantly with both geography and ploidy—marking the first reported association between plastome variation and ploidy level in Rhododendron. At finer phylogenetic scales, geographic proximity more often predicted chloroplast similarity than morphological species identity, suggesting geographically structured chloroplast capture via hybridization. Although many species’ relationships within section Rhododendron remained unresolved, a set of diploid taxa formed strongly supported monophyletic clades, offering key reference points for interpreting maternal lineage structure and reproductive isolation. Chapter 3 employed nuclear target-capture sequencing to infer phylogenomic relationships within subsection Lapponica. The clade was resolved as monophyletic, and several diploid species received strong support. However, polyploid taxa frequently clustered by ploidy level rather than by taxonomic species identity, potentially reflecting widespread gene flow, hybridization, and independent origins of polyploid lineages. Cytonuclear discordance was particularly pronounced in polyploid clusters, providing additional evidence for geographic chloroplast capture and reticulate evolutionary histories. Chapter 4 examined whether floral and vegetative morphology corresponded to species boundaries, and similarly, whether clades supported by the nuclear phylogeny in Lapponica aligned with morphological groupings. Morphometric analyses showed substantial overlap in trait values across taxa, with most floral traits offering little taxonomic resolution or phylogenetic signal. Stamen length was the only trait to exhibit consistent and statistically significant phylogenetic conservatism, likely reflecting constraints tied to pollination function. In contrast, floral shape factor traits appeared evolutionarily labile and were not strongly structured by phylogeny. Multilevel linear models showed that ploidy and elevation had minimal influence on floral trait variation. Together, these findings demonstrated that evolutionary patterns in Rhododendron, especially subsection Lapponica, were shaped by a complex interplay of polyploidy, geography, and hybridization, while traditional morphological characters contributed limited resolution for taxonomic or phylogenetic inference. This work uncovered new polyploid information for previously unexamined taxa, clarified some species-level relationships within a dominant alpine shrub lineage, and highlighted the importance of genomic and quantitative approaches in understanding recent and reticulate plant radiations in montane biodiversity hotspots
A Secure Hub for Access, Reliability, and Exchange of Data (SHARED)
Get PDFThe University of Chicago is developing SHARED (Secure Hub for Access, Reliability, and Exchange of Data) as a comprehensive resource for data-driven research and an integrated data management platform. Funded by the NSF, SHARED offers federated data storage across disciplines to promote collaboration and exemplary data management practices. This infrastructure extends beyond managing 'active' data by integrating with UChicago's institutional repository for data sharing in line with FAIR principles. The initiative is spearheaded by the University Library and Research Computing Center, alongside several university departments. SHARED aligns with the university's data lifecycle strategy, ensuring data access, analysis, publication, distribution, and long-term archiving. It supports diverse scientific applications, from cosmological studies to linguistic research, with a four petabyte ceph-based storage platform. Projects include dark matter searches, simulations of cosmic reionization, and cognitive process studies. SHARED promotes interdisciplinary research and offers educational opportunities through collaborations with minority-serving institutions and K-12 student engagement. Initiatives like the Data Science Preceptorship program with Chicago City Colleges enhance workforce diversity and data-related education. This poster will showcase SHARED's approach to integrating active storage with long-term data sharing and preservation through repositories.</p
Reworking Work: Reclaiming Balance and Recentering in Christianity and Away From the Idolatry of Work
No full textNo abstrac
Machine Learning for Complex Physical Systems: Applications to Climate Forecasting and Dynamical Systems Simulation
Get PDFThis dissertation investigates machine learning (ML) approaches for modeling complex physical systems across multiple domains. We examine three interconnected research areas: subseasonal climate forecasting, chaotic dynamical systems, and quantum mechanical simulations. Across these diverse domains, we address common challenges in the application of ML to physical systems: effectively encoding information to capture important features of spatiotemporal dynamics, developing robust emulators of physical systems that preserve critical invariant properties such as chaotic attractors, and systematic incorporation of known physical constraints into neural architectures and training schemes to simultaneously enhance prediction accuracy and computational efficiency. These methodological considerations form the conceptual framework for our various applications. For subseasonal climate forecasting, we develop frameworks that use lagged ensemble members and observational data to significantly improve temperature and precipitation forecasts. In chaotic dynamical systems, we propose neural operator training approaches that preserve invariant measures through optimal transport distance minimization and contrastive learning, maintaining statistical fidelity in long-term simulations. For quantum mechanics, we introduce Deep Stochastic Mechanics (DSM), a framework inspired by stochastic mechanics and generative diffusion models that may have far lower computational complexity in higher dimensions compared to traditional numerical methods by exploiting wave function latent structure. Our DSM simulations of bosonic systems outperform conventional approaches in both accuracy and computational efficiency. Furthermore, we extend DSM to effectively model fermionic systems, demonstrating its capabilities through a hydrogen molecule time dynamics simulation. Throughout these applications, we integrate domain-specific physics with advanced learning techniques to enable more accurate, efficient, and physically consistent simulations
Dire et ne pas dire : la Conversion d'un chevalier et la Confession contournée de Montaigne — deux études sur le non-dit dans Le Chevalier au lion et Les Essais
Get PDFThis dissertation deals with two different works The Knight of the lion of Chrétien de Troyes (circa 1170-1180) and The Essays of Montaigne through the theoretical perspective of the unspoken borrowed to Nietzsche’s Hinterfragen taken up by Pierre Macherey’s A Case of literary production [1st edition, 1966]. In a first part, this work advocates for a fresh reading of Yvain’s so-called « madness ». It argues that it is rather a case of spiritual conversion : a Christian Metanoia explained via Foucault’s posthumous book Confessions of the Flesh (2018) which tackles texts belonging to early Christianity. Overall, this dissertation pleads in favor of a new analysis of this novel of Chrétien by insisting on the fact that the genre of the novel encompasses theological debates of its time, such as that on penance, while also borrowing other cultural matters other than the sole « matière de Bretagne ». We show that the novel has a transgeneric quality from a literary standpoint but also from a social point of view and from a theological one. It also challenges the divide between human and non-human, while it triggers a debate about knighthood through a reflection involving a gender component. That we are dealing with a « courtly » novel is thereby fundamentally questioned. In a second part, this dissertation deals with the unspoken of Augustine’s Confessions within The Essays. It argues, in line with the work of Erich Auerbach, Gisèle Mathieu-Castellani or more recently that of Takeshi Kubota that Augustine’s Confessions are indeed present in The Essays albeit non explicitly quoted in Montaigne’s text. After having tracked the name of Augustine in The Essays, it turns out Montaigne plays with the genre of confession, diverts it and subverts it, offering an anarchic version of it on the verge of being mystical. This work thus all the way through questions the grounds for this augustinian unspoken while reflecting at the same time upon its implications in terms of genre on this hypertext (Gérard Génette) that are The Essays. Overall, through a philosophy of literature approach defined by Pierre Macherey, this work proposes a fresh perspective on the History of French Literature in a revised epistemological frame which takes seriously the religious literature to engage both those works. The unspoken theoretical angle allows to study them by an in-depth questioning of the silence that has allowed their existence
Shadows in the Detector: Searching for Axino Dark Matter as Missing Energy in ATLAS and Commissioning MET and Jet Triggers
Get PDFThis thesis presents a search for long-lived particles decaying to axinos using 137 fb of proton-proton collision data collected by the ATLAS experiment at TeV. The analysis focuses on events featuring displaced vertices with high invariant mass and track multiplicity, accompanied by substantial missing transverse energy. Results are primarily interpreted in the context of DFSZ axino models, where the axino serves as the lightest supersymmetric particle and a potential dark matter candidate. Background estimation is performed using data-driven techniques, with systematic uncertainties carefully evaluated through multiple validation regions. No significant excess over the estimated background is observed, allowing exclusion limits to be set on the model parameters at 95% confidence level. This work also details contributions to the gFEX system — a component of the ATLAS Level-1 trigger upgrade that enhances detection capabilities for missing energy and jet signatures. These developments strengthen future searches for axinos and other phenomena involving missing transverse energy, advancing our understanding of dark matter and beyond Standard Model physics
Positional distribution of transcription factor binding sites in the human genome
Get PDFAs transcription factors (TFs) play a major role in gene regulation, we studied their binding motifs (positional weight matrices, PWMs) and binding sites (TFBSs) in the human genome, and how TFs bind DNA motifs, including the involvement of binding co-factors. Using the chromatin immunoprecipitation sequencing data recently released by ENCODE (Encyclopedia of DNA Elements), we obtained new PWMs for 196 TFs and revised PWMs for 119 TFs. From these and the PWMs previously obtained for 235 TFs, we inferred the canonical PWMs for 500 TFs, including 243 new PWMs. Analysis revealed that most TFBSs are in introns (42.6%) and intergenic regions (31.6%), with only 11.3% in promoters. However, the TFBS density is considerably higher in promoters, showing a bell-shaped distribution of TFBSs with a peak at the transcription start site. Many TFBSs lie close to CTCF (CCCTC-binding factor) binding sites. Tethered binding is far more frequent than co-binding, with the latter often requiring co-factors