71446 research outputs found
Sort by
Systemic Treatment, Subjective Cognitive Function, and Survival in Prostate and Breast Cancer: Evidence from Patient Reported Outcomes and Real-World Data
No full textProstate and Breast cancer are the most commonly diagnosed malignancies and the
second leading causes of cancer-related death for men and women, respectively. In the United
States, there are more than 3.5 million prostate cancer survivors, and more than 4 million breast
cancer survivors, with these populations projected to continue to grow in the coming years.
Improvements in screening practices, earlier detection, and advancements in treatment have
prolonged survival among those diagnosed across both cancer sites. For cancer survivors with
metastatic prostate cancer and those with hormone receptor-positive (HR+) breast cancer,
systemic hormone therapies play a critical role in reducing mortality by targeting distinct
biological pathways. Despite these therapeutic advancements, there remains a limited
understanding of the long-term role hormone therapies play in cancer survivorship, particularly
with respect to sustained use. In addition, methods for measuring treatment adherence using
real-world data are not well established. This dissertation aims to bridge these research gaps by
investigating various aspects of cancer survivorship, with a focus on the long-term use of
systemic hormone therapies and on measuring adherence in real-world settings.
In Chapter 1, we utilized data from the International Registry for Men with Advanced
Prostate Cancer (IRONMAN) to examine the association between first-line systemic hormonal
treatment and subjective cognitive function trajectories among men with metastatic hormonesensitive
prostate cancer (mHSPC). We fit joint longitudinal and survival models for two
subjective cognitive function scales administered to IRONMAN patients at enrollment and during
follow-up for up to 2 years. We found that there was no evidence of worse cognitive decline over
time for individuals treated with androgen deprivation therapy (ADT) plus androgen receptor
pathway inhibitors (ARPIs), nor ADT + chemotherapy, when compared to individuals treated
with ADT monotherapy. Our study suggests that additional systemic hormonal therapies used to
treat mHSPC survivors do not negatively impact subjective cognitive function.
In Chapter 2, we further examined subjective cognitive function near diagnosis and
longitudinal trajectories and their association with mortality in the IRONMAN registry, among
those with metastatic prostate cancer. This association was evaluated by fitting joint longitudinal
and survival models for a follow-up period of up to 4 years. The findings from this study
revealed that poor subjective cognitive function near diagnosis and longitudinally was
associated with higher mortality. Our study highlights that subjective cognitive function can
provide an additional point of intervention to improve overall survival in this patient population.
In Chapter 3, we compared measurements of adherence for endocrine therapy when
using electronic health records (EHR) and health claims data among a population of 79
hormone receptor-positive breast cancer survivors. We investigated whether there were
differences across race and ethnicity, primary language spoken, and insurance status. We
found that when adherence is measured using only EHR data, adherence estimates are
consistently higher than when measured from health insurance claims data. Our study highlights
the importance of data source selection when assessing adherence to endocrine therapy and
that EHR and claims data offer complementary insights.
In conclusion, investigating research questions that extend beyond survival to
encompass life after diagnosis, while also highlighting the need to refine approaches to
measuring treatment adherence, is essential for informing care for cancer survivors. This work
provides a foundation to facilitate research and interventions aimed at enhancing the well-being
and quality of life of the continuously growing population of prostate and breast cancer
survivors.Population Health Science
Decoding Solar System Volatile Compositions: Laboratory Investigations of Ice Entrapment and Deuterium Fractionation
No full textStars and planets form within dense molecular clouds composed of gas and microscopic dust grains. On the cold surfaces of these grains, volatile molecules such as H2O, CO2, CH4, and N2 condense to form icy mantles that serve as active chemical reactors. Within these ices, physical trapping, isotopic exchange, and radical chemistry govern how volatile elements are stored, transformed, and ultimately delivered to planets. Understanding these processes is key to tracing the molecular history that connects interstellar clouds to comets, protoplanetary disks, and potentially habitable worlds.
In protoplanetary disks, and in planet formation more broadly, the distribution of volatiles between gas and ice is critical for setting the compositions of the solids that become incorporated into planets, moons, and comets. If ices were pure, this distribution could be described by a sequence of snowlines. In reality, astrophysical ices are complex mixtures, and their compositions at different disk temperatures depend on how different volatiles interact within the ice matrix. Physical trapping, or entrapment, of hyper-volatile molecules such as CO, CH4, N2, and Ar within less volatile ices like H2O or CO2 can delay sublimation to temperatures above their pure desorption points, altering the inventory of gases released into the disk and of the solids incorporated into planetary bodies. Laboratory desorption experiments conducted under astrophysically relevant conditions show that entrapment efficiency is primarily governed by the physical structure of the ice rather than by chemical interactions between species, indicating that entrapment is a mechanical trapping process. These results provide quantitative constraints on how hyper-volatiles are retained or released, influencing the volatile distribution inherited by disks and icy bodies.
Icy grains are not simply passive reservoirs of interstellar species; photochemical processing actively introduces new chemical and isotopic complexity. In the second part of this thesis, ultraviolet irradiation of mixed H2O:CD4 and D2O:CH4 ices was used to investigate solid-state hydrogen–deuterium exchange. These experiments reveal that photodissociation of both water and methane produces reactive OH and OD radicals that drive abstraction and recombination reactions, leading to the formation of HDO and CD3H/CH3D. The efficiency and directionality of this exchange differ between the two isotopic systems, indicating that isotopic equilibration is incomplete. The results suggest that local irradiation conditions can partially reshape primordial D/H ratios, providing a mechanism to explain the diversity of isotopic signatures measured in cometary volatiles and interstellar ices.
The final part of this work explores oxygen-atom insertion as a pathway to form methanol and its deuterated analogs under cold, barrierless conditions. Experiments using mixed CH4 and CD4 ices with 18O2 demonstrate that excited O(1D) atoms readily insert into C–H and C–D bonds to form CH3OH and CD3OD. A subtle isotopic bias toward the formation of deuterated methanol indicates that zero-point energy differences and bond energetics influence the insertion outcome. These findings identify oxygen insertion as a new solid-state pathway for deuterium fractionation in interstellar organics, offering an explanation for otherwise puzzling isotopic enrichments observed in methanol and related molecules.
Together, these studies provide a unified view of how the physical structure and photochemistry of ices regulate volatile retention, isotopic fractionation, and the emergence of molecular complexity in star- and planet-forming regions. By combining systematic laboratory experiments on entrapment, H/D exchange, and oxygen insertion, this thesis connects the chemical evolution of interstellar ices to the processes shaping planetary compositions. The results reveal how cold solid-state chemistry governs the inheritance of volatiles from molecular clouds to planetary systems, bringing us closer to understanding the molecular origins of habitability.Astronom
Evaluating the Pharmacokinetics of Topically Applied Small Molecule Drugs in Skin using Stimulated Raman Scattering Microscopy
No full textUnderstanding the pharmacokinetics of therapeutic compounds is crucial to the development of safe and effective therapeutics, including topical compounds, which are applied to the skin and other exterior sites of the body. Topical drugs are used to treat numerous dermatological conditions, ranging from cosmetic concerns to autoimmune conditions. Topical drug efficacy depends on topical drug pharmacokinetics in the skin, but cutaneous pharmacokinetics can be especially challenging to evaluate due to the barrier function and complex structure of the skin. Because of this, several aspects of cutaneous pharmacokinetics are not well understood, such as drug uptake to specific skin regions over time. While a variety of methods can be used to measure or visualize topical drug uptake in skin, many of these lack sufficient spatiotemporal resolution to evaluate cutaneous pharmacokinetics over time, especially in specific skin regions. Previous works have shown the value of using stimulated Raman scattering (SRS) microscopy as a tool to study topical drug uptake, since SRS imaging is rapid, chemically-specific, non-invasive, and SRS signal is proportional to drug concentration.
This work describes the development of SRS imaging methods to evaluate topical drug pharmacokinetics, helping reveal previously poorly understood aspects of topical drug uptake and cutaneous pharmacokinetics, such as the effects of perfusion on cutaneous pharmacokinetics, how pharmacokinetics in the stratum corneum relate to pharmacokinetics in deeper skin layers, and the temporal dynamics of topical drug delivery to the sebaceous glands. Measuring uptake of tazarotene with SRS in in vivo and ex vivo mouse ears in a paired experiment revealed that differences in pharmacokinetics were observed in the presence of perfusion. Using adaptive optics to correct for wavefront aberrations and overcome SRS imaging depth limitations allowed the capture of ruxolitinib uptake data in the stratum corneum, viable epidermis, and dermis. Finally, developing an SRS imaging method to detect drug uptake in sebaceous glands and performing corrections to account for light attenuation variations and lens effects enabled the measurement of tazarotene uptake in sebaceous glands, representing a novel approach of assessing topical drug delivery over time to structures located deep in the skin. The methods developed here present promising avenues for future studies and further illumination of cutaneous pharmacokinetics and topical drug uptake.Biology, Molecular and Cellula
Essays on Applied Political Methodology for Political Campaigns
No full textQuantitative social science techniques are increasingly used and misused by American political campaigns. In their efforts to influence a nationalized and polarized electorate, practitioners collect ever-larger amounts of data and conduct experiments to understand the position and malleability of the electorate. This rise in quantitative techniques has led to efficiency---increasing the concentration of spending and attention on the most persuadable voters in the most pivotal states---as well as stagnation. We now confront a political landscape crowded with analysis that can be critically informative but is also often incomplete, biased, and reliant on largely private data. A critical issue is the frequent use of quantitative social science and machine learning tools for inference without adequate attention paid to uncertainty and generalizability. More transparency and emphasis on the fundamentals of inference would allow practitioners to better understand the electorate, how and when to take action, and to appropriately consider uncertainty and the range of possible outcomes in political scenarios. To do this, both applied researchers and practitioners need to move toward an iterative use of research to build knowledge over time, rather than relying on single studies to answer big questions across broad contexts. In this work, I contribute to this agenda with new methods campaigns can use to understand the electorate and offer an overall approach to iterative research that helps resource-constrained actors optimize their influence.Governmen
Characterization of Thymocytes and Splenocytes in Gdf8^Gdf11MD mice
No full textImmunosenescence refers to an age-related decline in immune function, primarily characterized by the gradual degeneration of the thymus, also known as thymic involution. Transforming growth factor-β (TGF-β) signaling is integral to the formation and normal function of the thymus. As a member of the TGF-β superfamily, Growth differentiation factor 11 (GDF11) and its receptors are expressed during thymocyte development. Moreover, GDF11 plays a critical role in the immune system by reducing inflammation through the regulation of specific pathways and immune cell functions. More importantly, GDF11 has been shown to have potential rejuvenating effects on various organs, including the heart and brain. However, its impact on thymic rejuvenation remains unknown. To study the effect of GDF11 on thymic involution, here we utilized the Gdf8^Gdf11MD mice, which have increased levels of circulating GDF11, and analyzed changes in their thymocyte and splenocyte populations. Our results showed that young adult female (4-month-old) and male (3-month-old) Gdf8^Gdf11MD mice exhibited a trend toward accelerated thymic involution compared to controls. Although this difference diminished with age, the reduction in thymic cellularity persisted into late adulthood (8-month-old) in male mutants. Notably, significant changes in the absolute number of multiple thymic subsets were observed in young (3-month-old) and pre-middle-aged (8-month-old) Gdf8^Gdf11MD males, indicating potential alterations in thymocyte development. Using flow cytometry, distinct immune cell subsets within the spleen were analyzed to further explore the effects of accelerated thymic involution on peripheral T cells in mutant mice. We found that, despite a significant increase in spleen weight in 12-month-old female and 8-month-old male mutants, there were no differences in the frequency or absolute number of splenocytes compared to controls. In addition, 4-month-old female Gdf8^Gdf11MD mutants exhibited an increased spleen-to-body weight ratio, accompanied by a decreased proportion but increased number of CD4⁺ central memory T cells (TCM). These findings demonstrate a possible correlation between GDF11 expression levels and thymic involution, highlighting the important role of GDF11 in immunosenescence.Graduate Educatio
Mythologies and Sound: A Composition Portfolio
Get PDFBefore starting my doctoral studies at Harvard, I was primarily focused on both
the improvisatory nature of music making and conceptualisations of sound as image.
Over the six years I spent at Harvard developing my musical thinking, the
improvisatory approach toward music making remained at my core, while the mystery
of sound-as-image was the focus of my conceptual development.
A leap was made in my music when I began combining my drawings with my
music in the form of experimental stop animation. This approach, first attempted in
Animal (2022), developed significantly, both technically and aesthetically, through the
works Fabric of Sorrow (2023), Mom (2023), root (2023) and most recently in bluer
womb (2025) where a complex polyphony between sound and image is explored. Yet,
while literal images found a place in my music in 2022, for me, sound itself has always
been an image –– sound can be experienced as a landscape, as a color, or simply as a
line. While in the past this was wholly conceptual or personally experiential, in recent
works, this musical thinking has emerged as a tangible quality of my music.
While there is the literal image on a screen in many of my recent works, my
conceptualisation of sound-as-image drove my compositional language in a new
direction during this degree –– wide polyphonies of musical material exist in my recent
works; compositional systems of material interrelation in which the form of the work is
‘opened’ as a space and this space can become a landscape which the materials
themselves sustain and move through. This can be seen in my works with animation,
but also in the instrumental works without visuals included in this portfolio: children's
games (2023), flocks of birds flying out of her belly (2024), 10am is when you come to
me (2025).
Working on these pieces, at some point it occurred to me that my works have
always been an expression of my nostalgic relationship with childhood. This nostalgia
exists as sensory fragmentations: stories, memories, spaces, people, landscapes, sounds.
As a child I would create stories, fantasies as a way to escape. Now, as an adult I create
images of nostalgia to reconnect with what I was trying to escape from. In this way, the
music I make is nostalgic. It is a translation of childhood nostalgias, a sonification of
these sensory fragments, these images, these flashes of feeling.
For me, my nostalgia runs deeper than a simple melancholy for ‘what was
before’. Nostalgia is in a sense mystical. While being deeply personal, my nostalgia is
also beyond me. It is a feeling that connects us to something larger than ourselves; it is a
connection to a shared, generational, foundational something. It’s in this way that I
believe nostalgia is mythological in a true sense –– nostalgia is a foundational,
somehow shared story that we appeal to for our sense of identity, and, in my work,
nostalgia presents itself as a work of art.
By engaging with the individual and their myth, the personal and shared, my
works become political –– as a myth and as a space, it’s my hope that my music is
something an audience could take refuge in and in which their nostalgia (their myths)
can be reactivated and, potentially, reshaped.Musi
Epistemic Limits of Trustworthy Machine Learning
Get PDFTheoretical understanding of a system’s limits has long driven technological breakthroughs. Carnot delineated the fundamental limits of heat engine efficiency, paving the way for the design of modern state-of-the-art engines. More than a century later, Claude Shannon unraveled the fundamental limit of communication, known as channel capacity. This insight revolutionized communication systems, enabling continual improvements that ultimately led to wireless communication as we know it today.
This thesis discusses the epistemic limits of machine learning (ML) and leverages them to improve the trustworthiness of ML systems. ML models have an epistemic limit when proving one of their properties is impossible. Epistemic refers to the impossibility of providing theoretical guarantees (knowledge) about a model's property. Epistemic limits are information-theoretic converse results on the hypothesis test that checks a model's property.
First, we prove a limit on how much information personalized models can use while ensuring reliable test for performance gains across all users -- epistemic limits of personalization. We leverage this limit to develop a tool to help with feature selection. Second, we show a limit for reliably testing if model performance is equitable across multiple demographic groups --epistemic limit of fairness testing. We exploit this limit to design a metric for efficient algorithmic bias detection.
Third, we prove a limit for testing if one model outperforms another on average -- epistemic limit of model selection.
We use this result to delineate the set of indistinguishably good models --Rashomon set. Finally, we argue that the epistemic limits in model selection imply that explaining the predictions of ML models is necessary. Then, we develop efficient methods for explaining the content produced by large language models.Engineering and Applied Sciences - Applied Mat
Agrarian Developmentalism: The Politics of Development Strategies in Latin America
No full textThe conventional wisdom holds that governments extract from agriculture to promote industrialization. However, development strategies are not uniform. While some governments did extract from agriculture, others actively supported the sector in what I call agrarian developmentalism. Why do some governments support agriculture while others extract from the sector during industrialization?
I argue that this variation is explained by the structure of the party system and rural producers’ legislative strength. Integrative party systems—where parties rely on mixed rural-urban constituencies for electoral support—are more favorable to agriculture. In these systems, politicians from different parties are more likely to endorse pro-rural programs, include rural producers in governing coalitions, and craft intersectoral bargains to solve urban-rural distributional conflicts. In contrast, segmented party systems—where politicians specialize in representing either urban or rural constituencies—foster intersectoral conflict, making generous government support for agriculture less likely. Additionally, rural producers’ legislative strength determines the level of extraction. This formal power allows them to systematically constrain extraction from agriculture. The combination of an integrative party system and rural legislative strength leads to agrarian developmentalism.
I develop the argument through a comparative historical analysis of two countries with divergent development strategies crafted during World War II: Colombia and Chile. In Colombia, the government supported agriculture generously through price support, tariff protection, and public spending. By contrast, the Chilean government sidelined agriculture to promote industrialization. Drawing on original data collected over twelve months of fieldwork from a range of historical sources, I show how variation in the type of party system and rural producers’ legislative strength shaped agricultural policy. I also demonstrate that alternative approaches focusing on the structural or instrumental power of rural producers have limited leverage to explain whether development strategies favor agriculture.
The main contribution of this study is to show that patterns of political representation through the party system matter for economic policy outcomes. Against the literature portraying landowners as very powerful actors, my findings reveal that they were mainly reacting to rather than shaping economic policy. Second, I offer a new perspective on Latin American party systems. Parties in integrative party systems are often dismissed as “catch-all” clientelistic machines that lack programmatic commitments. I demonstrate that these parties do have programmatic agendas that influence how politicians govern and how development strategies unfold. Third, my work joins recent scholarship that challenges the alignment between landowners and authoritarianism. I show that independent legislatures provide landowners with a unique institutional arena to constrain extraction, an arena absent under authoritarian regimes. Thus, I argue, democracies with broad-based coalitional parties are conducive to a pro-agricultural development strategy.Governmen
Target identification by chemical proteomics and bioinformatic investigations with small molecule ligands
No full textUnderstanding the protein targets and mechanisms of action of bioactive small molecules remains a central goal in chemical biology and drug discovery. Unbiased approaches, particularly chemical proteomics and systems-level bioinformatics, have emerged as powerful strategies to uncover the molecular interactions and biological consequences of small molecule engagement in cells. In this thesis, I apply these complementary methodologies—activity-based protein profiling (ABPP), photo-affinity labeling (PAL), and bioinformatic analyses—to investigate two distinct but mechanistically rich chemical spaces: opioid analgesics such as heroin and morphine, and cereblon (CRBN)-binding immunomodulatory drugs (IMiDs) that induce targeted protein degradation.
Chapter 1 introduces the conceptual and methodological foundation for this work, beginning with a background on ABPP and PAL, highlighting their ability to turn small molecules into molecular probes that illuminate their target landscapes through quantitative mass spectrometry. The chapter also introduces gene ontology (GO) annotations and the database for annotation, visualization, and integrated discovery (DAVID) bioinformatics suite, which are essential for contextualizing proteomic hit lists. Further, I provide an overview of machine learning principles and illustrate how these computational tools can accelerate target identification and therapeutic discovery. The chapter concludes with two thematic deep dives: a historical and pharmacological overview of opioids and their receptors and the development of chemical probes to study them, and a survey of CRBN biology, including the evolution of IMiD-based molecular glues and the discovery of the endogenous CRBN degron—the cyclic imide degron—formed either spontaneously from protein damage or enzymatically via Protein-L-isoaspartate O-methyltransferase (PCMT1).
Chapter 2 describes the design, synthesis, and characterization of novel chemical probes for morphine and heroin, including photo-click morphine (PCM-1, PCM-2) and the acyl-donating probe Di-Alkynyl-Acyl-Morphine (DAAM). These tools were validated for their ability to engage opioid receptors and induce G-protein signaling. Confocal imaging revealed receptor-independent localization of these probes to lysosomes across diverse cell lines. Chemoproteomic analysis uncovered voltage-dependent anion channel 1 (VDAC1) as a shared target across all probes and identified lysine 234 of solute carrier family 25 member 3 (SLC25A3) as a selective site of acylation by DAAM. This residue was later confirmed to be directly acetylated by heroin itself, representing one of the first demonstrations of small-molecule-mediated post-translational modification by heroin. These findings suggest potential mechanisms for the mitochondrial and metabolic dysfunctions associated with chronic heroin use.
Chapter 3 focuses on the proteomic and computational reanalysis of the Broad Institute’s PRISM screen using the CRBN-dependent degrader DEG-35. By separating wild-type from mutant populations and refining statistical comparisons, I report biologically meaningful indicators of sensitivity, including intact p53 signaling, protein Mdm4 (MDM4) dependency, and alterations in the switch/sucrose non-fermentable (SWI/SNF) complex. I report the development of a multilayer perceptron (MLP) model trained on the Profiling Relative Inhibition Simultaneously in Mixtures (PRISM) viability data to predict DEG-35 sensitivity across the extended DepMap cell line repository. The model achieved strong performance (Area under the Receiver Operative Characteristic Curve (AUROC) = 0.98) and accurately identified sensitive cell lines, including SCCOHT-1, which is actively being experimentally validated. This work highlights the potential of integrating chemogenomics data with machine learning to stratify responders and uncover mechanistic biomarkers.
Chapter 4 presents a novel inversion of traditional GO analysis using the DAVID tool: instead of applying GO terms post hoc to hit lists, I used it as a filtering method for hypothesis generation. Starting with all human proteins ending in C-terminal asparagine or glutamine—potential substrates for PCMT1-mediated cyclic imide formation—I performed annotation clustering and literature-driven prioritization. From this list, Hairy and Enhancer of Split 1 (HES1) emerged as a top candidate. In vitro assays confirmed that PCMT1 catalyzes the formation of the cyclic imide degron on the HES1 C-terminus, enabling CRBN binding. Full-length HES1 showed PCMT1-dependent CRBN engagement, and follow-up experiments in SH-SY5Y cells demonstrated that knockout of either PCMT1 or CRBN led to impaired neuronal differentiation, suggesting a role for this degradation pathway in developmental timing. These findings advance our understanding of CRBN’s endogenous substrates and open new directions for exploring its role in neurodevelopment.
Together, this thesis demonstrates how integrative chemical biology—uniting chemical proteomics, computational modeling, and bioinformatics—can illuminate the complex interactions between small molecules and the proteome, uncover new biological mechanisms, and inspire new directions for therapeutic or biological discovery.Chemistry and Chemical Biolog
House of Five
No full textThis thesis presents the opening chapters of my novel, House of Five. Set in Fig
Town, a fictional port on the Aegean Sea, the four chapters move between an empire’s
final years in the 1910s and the era of its modern republic. Irene arrives at the Lyre
House, which is perched above the harbor, blaming herself for her mother’s death and
certain she doesn’t belong. Half orphanage, half choir school, the Lyre House is unlike
any other place to grow up. Though reluctant at first, Irene finds purpose there in singing
alongside girls drawn from across the fading empire—Aysel, Francesca, Nara, and
Rivka—under exacting foreign instructors. She has always sensed the untamed power in
her voice, and the Lyre House shows her what it can do when interwoven with theirs. As
the empire fractures and the patrons come to prize solos and favors, the girls are tempted
into rivalry, threatening the fragile chorus. Over a century later, the Lyre House stands in
ruins. Developers circle the crumbling building, one intent on demolition, the other
attuned to what lives on in the walls. Narrated by Irene’s lingering spirit, now fused with
the house itself, House of Five is a novel about plural belonging in fractured time and the
architecture of memory: how a building preserves what people forget.Extension Studie