Alliance One Tobacco (Malawi)
Academic Research Repository at the Institute of Developing EconomiesNot a member yet
52305 research outputs found
Sort by
Regularization in Reinforcement Learning: Equivalences and Novel Methods
No full textReinforcement learning (RL) is a powerful framework for sequential decision-making, with applications ranging from robotics to healthcare. However, in real-world settings, such as mobile health (mHealth), RL faces challenges due to limited data and the need for generalization beyond observed experiences. Regularization -- a set of techniques that constrain model complexity to prevent overfitting and promote generalization-- plays a crucial role in overcoming these challenges. This dissertation critically examines existing RL regularization methods, uncovers novel connections between them, and introduces new approaches inspired by the challenges of mobile health studies.
One focus of this work is establishing theoretical connections between existing regularization methods. We prove that discount regularization produces the same optimal policy as a Bayesian prior on the transition function and a penalized Q-function, and is also equivalent to a truncated lambda return. These relationships reveal underlying assumptions and limitations of discount regularization.
This work also focuses on introducing novel regularization methods. First we introduce a state-action-specific regularization method that mitigates the limitations of discount regularization uncovered in our analysis. We also propose a novel Bayesian hypothesis testing-based regularization approach that leverages prior study data to improve learning while adapting to differences between the environments of the prior and current studies. This is particularly useful in mobile health applications where feedback is sparse and exploration is limited.
Through theoretical analysis and empirical validation, this dissertation advances the understanding of RL regularization methods and introduces new techniques that enhance generalization in data-constrained environments. These contributions provide a principled foundation for improving RL applications in healthcare and beyond.Engineering and Applied Sciences - Applied Mat
Programmable protein expression with RNA sensors
No full textRNA sequencing continues to reveal new cell type markers and gene expression profiles1. Biological tools that directly couple gene expression information to transgene expression are desirable for cell-type and state specific transgene expression. We developed Reprogrammable ADAR Sensors (RADARS), a technology that uses base-pairing and RNA editing by adenosine deaminases acting on RNA (ADAR) to translate a transgene cargo only in cells with target RNA expression. RADARS is reprogrammable based on RNA target sequence, which positions RADARS to deliver therapeutic cargoes to diseased cell-types or tissues that exhibit unique RNA expression, such as cancer. Many cancer types have unique fusion RNAs as consequence of gene fusions, which arise through somatic structural variation and often drive cancer progression. Given their cancer-specificity, targeting fusions has precedent in oncology–yet many cancers driven by fusions do not have targeted therapies. We hypothesize that RADARS can leverage cancer specific RNA fusions as therapeutic vulnerabilities. We tested the selective activation of cell killing payloads using RADARS against three clinically relevant fusion RNA classes: (1) a driver fusion in fibrolamellar hepatocellular carcinoma; (2) gene fusions found in therapy-resistant metastatic breast cancer; (3) endogenous fusion RNAs in breast cancer lines. Preliminary data suggest that fusion RNAs may serve as conditional biomolecules for delivering cancer-killing protein payloads.Biology, Molecular and Cellula
Investigation of the impact of αCD45 cellular backpacks on the mechanobiology of T cell activation and tumor cell cross-talk
No full textImmune checkpoint blockade-resistant solid tumors present a major therapeutic challenge. Although adoptive T cell therapies, such as CAR T cells, have shown remarkable success in treating hematological malignancies like leukemia, they have not yet achieved comparable efficacy in solid tumors. In the majority of patients with solid tumors, transferred T cells rapidly lose their functional phenotype following adoptive transfer, leaving them vulnerable to the immunosuppressive tumor microenvironment (TME) [#zhao2019TcelltherapyReview]. Considering the significant promise of adoptive T cell therapies, there is intense interest in developing strategies to enhance their persistence and function within solid tumors. Recent work under review for publishing in the Mitragotri lab has shown that equipping primed polyclonal murine CD8+ T cells with micropatches, namely "cellular backpacks (BPs)", can enhance their anti-tumor response against aggressive tumors by providing localized stimulation. These BPs are 6 μm polymeric poly(lactic-co-glycolic acid) (PLGA) microparticle disks functionalized with anti-CD45 antibodies[#Fukuta_Mitragotri_neutrophilBPBrain], which can attach to the cellular membrane of T cells. BPs show strong clinical promise as a companion therapy to extend the persistence of T cell therapies. Therefore, to fully realize their potential, it is essential to understand the diverse ways in which BPs influence T cell behavior, in order to improve the technology and optimize for its clinical application. Previous studies have investigated the biological effects of BPs on immune cells through receptor clustering[#Prakash_2023_BP_NKcells, #Prakash_2024_BP_Bcells] and biophysical interactions[#Ninad_2024_neutrophilBP], as well as their chemical effects for drug and cytokine release[#Kapate_2023_BPMacrophageforTBI, #Kapate_2023_BPMyeloidforMS, #Shields_2020_BPMacrophage]. However, the mechanical impact of BPs on the cytoskeleton and plasma membrane has not been thoroughly explored. Preliminary observations have shown that BPs induce morphological changes and cytoskeletal remodeling in immune cells while still permitting migration, suggesting a unique form of mechanical modulation. The consequences and depth of these effects, however, have remained largely uncharacterized. This thesis specifically has addressed this previously unexplored topic by investigating the impact of BPs on the mechanobiology of T cells. Although interest in T cell mechanosensation has grown in recent years, most studies have focused on the formation of the immunological synapse (IS) and the influence of substrate stiffness, leaving other mechanical aspects of T cell regulation underexamined. BPs provide a unique platform to investigate these questions, as they induce cytoskeletal and morphological changes while maintaining T cell motility and functionality.
This work has identified and characterized two well defined aspects of mechanobiology to shed light on how BPs influence T cell mechanobiology, contributing new insights with potential implications for optimizing BP-based immunotherapies. First, the project explored whether BPs could alter the mechanics of physical tumor cell cross-talk with T cells. A 2022 study revealed a new mechanism of immune evasion whereby cancer cells extend tunneling nanotubes (TNTs) structures to pump out the mitochondria of immune cells. This mechanism has been shown to promote premature exhaustion of T cells in the TME, contributing to their lack of efficacy[#saha2022nanotubes]. Exocyst complex proteins of the Sec family and the Rho and Ras GTPase family are known to be involved in actin remodeling during TNT formation and mitochondrial trafficking. The same protein complex is involved in cytoskeletal remodeling due to a mechanical stressor. Hence, this project hypothesized that modifying T cells with cellular BPs would alter this complex from allowing mitochondrial transfer and protect them from mitochondrial theft by cancer cells, thereby prolonging their persistence. We investigated whether BP attachment on T cells would hinder mitochondria transfer by co-culturing them with tumor cells. Although BP attachment to Jurkat T cells was achieved with high efficiency (~80%), it did not impair mitochondrial transfer, suggesting that BP-induced cytoskeletal remodeling may not be sufficient to disrupt TNT formation or that alternative mechanisms of transfer may exist. These results suggest that BPs do not interfere with beneficial intercellular mitochondrial exchange, which could allow T cells to remain metabolically supported in the tumor microenvironment.
The second focus investigated whether a biomechanical effect participates in the non-specific activation of T cells caused by BP attachment. Shear stress alone has been shown to activate T cells via the Piezo1 stretch-activated calcium (Ca2+) channel[#sarna2024_ShearStressTcells], and Ca2+ influx is a critical downstream event in T cell receptor (TCR)-mediated signaling. Thus, we hypothesized that the mechanical interaction of the T cell membrane with a BP could trigger Piezo1-mediated Ca2+ entry, thereby promoting an activated T cell phenotype. To test this hypothesis, a pharmacological inhibitor of Piezo1 was applied to T cells, and intracellular Ca2+ levels were monitored. As expected, BP attachment led to elevated intracellular Ca2+ levels, an early marker of activation. However, pharmacological inhibition of Piezo1 did not significantly reduce this Ca2+ influx. Interestingly, longer-term studies revealed that Piezo1 inhibition suppressed the upregulation of the activation marker CD25 and also reduced BP retention on the T cell surface, suggesting that Piezo1 contributes to sustained activation and cytoskeletal stabilization in the context of BP attachment. This indicates that Piezo1 may not initiate the Ca2+ influx but instead plays a role in maintaining T cell responsiveness to sustained mechanical stimuli.
In summary, this thesis identified BPs as a unique platform for modulating T cell mechanobiology, revealing both their compatibility with intercellular communication and a potential role for Piezo1 in sustaining mechanical activation of T cells by BPs. These insights lay the groundwork for refining BP-based immunotherapies and uncovering new strategies to support T cell function in solid tumors.Graduate Educatio
Dear Father: A Memoir on Love, Family, and a Complicated Paternity
No full textThe epistolary form provides a unique way for the author to shorten the narrative distance by creating a level of intimacy with the reader that one can’t quite achieve with other forms. By using this form, the author allows the audience to peer into the world they are painting in a such a way that removes the veil of distance and invites the reader to fully submerse themselves within the text, the subtext, and the story itself. Dear Father addresses a complex relationship between a daughter and her adoptive father, while using the epistolary form to create a certain intimacy that would be difficult to achieve otherwise.Extension Studie
A Causal Inference Framework for Identifying Critical Windows of Time-Varying Exposures
No full textThere has been great clinical interest in the concept of ‘critical windows’ or ‘sensitive periods’ of environmental exposures. The concept of a critical window, defined formally in this dissertation, refers to a specific time period during which an individual is more susceptible to developing an outcome in response to a particular exposure than at other times. The statistical methods used to identify these critical windows have several limitations, and applied researchers are often limited to using methods developed for different research questions which can lead to bias, inflated Type I error rates, and low power.
Studies on environmental exposures are almost always observational by necessity, and it remains an ongoing challenge to interpret results as the causal effect of intervening on the exposure rather than merely as an association between the exposure and the outcome. Because the methods currently in use have not been previously examined through a causal inference lens, results across studies are difficult to compare, even if the same covariates are used.
This dissertation seeks to combine these two areas of interest by proposing a framework for the identification of critical windows from a causal inference perspective. Throughout this work, we demonstrate how different methods should be employed to answer subtly different research questions and compare our methods to existing approaches through simulations where appropriate.
In Chapter 1, we introduce our novel flexible CAusaL Identification of Critical windOws - Modified Treatment Policy (CALICO-MTP) framework, extending previous work on using a dose modification scheme to estimate the causal effect of continuous exposures. We propose dividing the concept of critical window identification into three distinct research questions, each addressed with different approaches. These questions are: 1) Curve estimation: what does the exposure-outcome relationship look like over time? 2) Hypothesis testing: is there any time window during which there is an effect of intervening on the exposure? and 3) Window selection: after determining that there is a causal relationship, what is the critical window for that exposure? For the first question, we propose a curve estimation strategy to yield results similar to those of the commonly used distributed lag model (DLM). For the second, we propose estimating the effect of intervening on
all biologically plausible windows and combining the p-values using the Aggregated Cauchy Association Test (ACAT), a p-value combination method that accounts for strong correlations between test statistics. For the third, we discuss strategies for selecting the window once the global null has been rejected. We apply these methods to a dataset from Beth Israel Deaconess Medical Center (BIDMC) and compare them to previous results regarding the effect of Nitrogen Dioxide (NO2) exposure on the 32-40 week fetal head circumference as measured by ultrasound45, and we present a novel visualization for the causal effect of intervening on time intervals.
In Chapter 2, we present a variant of this framework, CALICO-ADRF, that explicitly models nonlinear dose-response relationships by estimating the Average Dose Response Function (ADRF) for each time window. This nonlinear relationship is particularly relevant for environmental exposures such as metals, where some are necessary minerals at low exposures but act as toxins at high exposure levels, and temperature, which may exhibit a thresholding effect for certain outcomes. We use a scalar test statistic that is the integrated squared derivative of the estimated ADRF to perform global hypothesis testing with Type I error control and improved power compared to the methods of Chapter 1 for biologically-plausible nonlinear dose-response curves. We demonstrate these results looking at the effect of maternal prenatal temperature exposure and birthweight for full-term deliveries in the same BIDMC cohort.
In Chapter 3, we present a discussion of causal inference concepts specifically tailored to the methods most commonly used for time-varying environmental exposures, offering a novel perspec- tive for researchers. We present a framework through which the target estimand of different mod- eling approaches can be compared, improving the ability to draw meaningful and comparable conclusions across studies. We explore the different estimands that these methods target and illustrate when these estimands align or diverge depending on the underlying causal structure of the exposure. Finally, we provide guidance for researchers on how to appropriately align their methodological choices with their research questions.Biostatistic
