Stanford University

NSF Awards · FY 2026 · 2026-09

The rapid advancement of artificial intelligence (AI) has produced remarkable achievements, from generating realistic text and images to designing new proteins and supporting healthcare delivery. However, the opaque nature of these powerful technologies poses significant challenges to scientific progress and public trust. When AI systems make predictions in sensitive domains such as medicine, education, or public policy, we need to understand which factors drive their decisions and whether their discoveries are reliable and reproducible. This project develops methods to extract interpretable, verifiable and trustworthy insights from sophisticated AI algorithms. By doing so, it will accelerate scientific discovery across disciplines, while strengthening public confidence in data-driven research. Beyond methodological innovation, the project will contribute to training the next generation of AI researchers and data scientists. This project will develop novel statistical methods for testing large number of hypotheses about variable importance in complex predictive models, with rigorous control of false discovery rates. The research will focus on context-dependent variable importance, studying how explanatory variables influence outcomes under different conditions and through non-additive interactions, leveraging sophisticated machine learning architectures. The methodological framework will integrate recent advances including knockoff inference, e-values, conditional randomization tests, and explainable AI techniques. A key innovation will be designing inference procedures robust to multiple data passes, avoiding selection bias and circular reasoning, while adapting to the signal in the data. Applications will focus on genomic data analysis, including inference on gene-gene interactions. The project will contribute to training one graduate student. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-09

Modern artificial intelligence (AI) systems are increasingly limited not by arithmetic, but by memory. As frontier AI models become more capable, they require far more data to be moved, stored, and accessed efficiently. These workloads systematically generate large volumes of short-lived data that are written in memory, consumed, and quickly discarded, as well as long-lived data that must be retained reliably across much longer time scales. Conventional memory systems are poorly optimized to this behavior, as they are typically designed as one-size-fits-all storage, resulting in excessive energy consumption and increasingly limited density scaling. This project addresses that mismatch by developing a computing infrastructure that treats data persistence as a central design consideration by statically and dynamically matching short-lived and long-lived data to differentiated memory architectures and technologies, each optimized for the appropriate retention window. The result is a more efficient and sustainable foundation for accelerating large-scale AI systems in both datacenter and edge settings. The project also supports education and workforce development through new course materials, industry engagement, and open-source resources that expand participation in next-generation AI hardware accelerator design and computer engineering. The project develops a retention-aware computing stack that aligns AI application data lifetimes with heterogeneous memory architectures and technologies offering different retention times and densities. The research is organized around four integrated activities: (1) building a profiling framework to characterize how long different data values remain useful and to map those lifetimes onto suitable memory tiers; (2) developing algorithmic techniques that restructure computation and data movement to better satisfy retention constraints; (3) designing compilation and scheduling methods for retention-aware data placement across differentiated memories; and (4) validating the resulting hardware-software co-design through the agile design and tapeout of memory-centric AI hardware accelerator chips. Together, these efforts will establish a cross-layer framework for rethinking memory-system design around data persistence and creating a foundation for more energy-efficient, high-performance hardware systems for the AI era. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-07

This award supports the development and applications of methods in the field of microlocal analysis. Roughly speaking, this field keeps track of the position and frequency, or momentum, of waves (or more generally, functions, such as the amplitudes and phases of waves) simultaneously. The planned applications, such as the analysis of long-time or far field behavior, are to wave propagation and other related phenomena such as quantum fields on curved spaces; curvature of the space itself is a feature of mathematical general relativity. Although the project concerns their mathematical theory, these problems are closely connected to the physical world. Wave propagation is ubiquitous in nature, with light and gravitational waves being important examples, and the latter (gravitation) giving rise to curved spacetimes. Scattering theory of quantum particles is another subject governed by microlocal analysis: these aspects enter into the description of quantum waves at large distances. Many of the projects are suitable for research by doctoral students, and the Principal Investigator (PI) strives to contribute to the education of a new generation of mathematicians and scientists. Some of the projects describe the long-time or far field behavior, including existence, of waves, such as electromagnetic or gravitational waves, on curved spacetimes. The microlocal approach to analysis on these spaces has made breakthroughs possible in the PI's (in part collaborative) work on linear and non-linear problems on asymptotically hyperbolic (AH) spaces as well as Kerr-de Sitter (KdS) space (rotating black holes in a cosmological spacetime), culminating in the proof of the stability (as solutions of Einstein's equation) of slowly rotating KdS spaces with Hintz. More recently, with Hafner and Hintz, the PI extended some of these tools to the vanishing cosmological constant case, namely Minkowski and Kerr spaces - so far only to the linearized result in the Kerr case but including fast (subextremal) rotation. With Hintz and Petersen, the PI also extended the nonlinear stability result to fast rotating (subextremal) KdS spacetimes, though in a conditional manner. The projects here aim to extend these tools to further spaces, such as perturbations of subextremal Kerr spacetimes. Other projects study basic objects in quantum field theory, in particular the Feynman propagator, on curved spacetimes, and their uses for spectral theory, including for a spectral action principle. Two particular area of focus for this are radiative spacetimes that arise by solving Einstein's equation in general relativity with perturbed Minkowski initial data (which gives rise to gravitational radiation as the spacetime settles down to Minkowski space) and perturbations of de Sitter spacetimes, which are the underlying spacetimes for positive cosmological constant problems. The microlocal techniques involved in understanding the latter also relate to wave diffraction from boundaries or edges, the quintessential near field phenomena on which the PI had worked extensively; a longer-term project is to use the new insights to gain a better understanding of these phenomena. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-07

Children's brains grow and change in ways that help them understand what they see. Over development, babies and children learn to recognize familiar faces and places, make sense of words on a page, and make visual decisions such as "Is this my mother?" or "What word am I reading?" This project explains how children develop visual intelligence by studying how brain structure and brain function change together from infancy to young adulthood. The research may help scientists create earlier ways to detect when development is going off track, so children and families can receive earlier support and have better long-term outcomes. The project may also inspire new kinds of biologically inspired artificial intelligence (AI). Most AI systems have a fixed design, but children's brain tissue grows as visual abilities develop. Understanding how changes in brain tissue and visual abilities are linked may help engineers build biologically inspired AI systems with adaptable designs that can grow and become more efficient over time. The project studies how brain activity and brain tissue change together in the ventral visual stream, a major brain pathway that supports visual understanding, including face recognition, place recognition, reading, and visual decision-making. The research team uses functional magnetic resonance imaging (fMRI) to measure brain responses to visual inputs and quantitative MRI (qMRI) to measure brain tissue properties in the same children. The project tests when and how brain responses and tissue develop across infancy and childhood, and whether these developmental changes are linked. This work will provide brain charts of the development of the visual system. Because qMRI reflects brain tissue properties indirectly, the project also combines qMRI with direct measurements of biological features from rare pediatric brain tissue samples, as well as with MRI phantoms. MRI phantoms are lab-designed objects that mimic brain tissue and allow researchers to precisely control specific features, such as the amount of myelin or iron. The project will test biological features that change during development, including myelin, a fatty membrane that wraps around nerve fibers and helps brain signals travel efficiently, and iron, which supports brain function. Computational modeling will connect qMRI measures to these biological features, which will make it possible to use qMRI data from infants and children to infer which features of brain tissue are developing. Understanding how brain responses and brain tissue develop together may support earlier detection of atypical neurodevelopment and inform new types of AI systems whose architecture can adapt over time and become more efficient. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NIH Research Projects · FY 2026 · 2026-06

Project Summary/Abstract All living organisms, from the simplest microbes to complex humans, produce lipids. Lipids are a broad class of organic molecules that include, among others, fatty acids, cholesterol, and isoprenoids. Lipids play critical roles in membrane structure, cell signaling, and cell development and have been of interest to a broad range of scientists. Biochemists have elucidated the elegant biochemical reactions and pathways required to synthesize them, physiologist have uncovered various structural and functional roles of lipids, and geologist have shown that the recalcitrant nature of lipids makes them excellent molecular fossils that can inform our understanding of the ancient Earth’s ecosystems. In my research group, we have focused on understanding the biosynthesis of lipids in microbes with the primary goal of revealing novel lipid producers, new lipid structures, and distinct biosynthetic pathways. Through our work, we have made significant discoveries in lipid biosynthesis – from revealing that archaea utilize unique radical chemistry to modify their distinct isoprenoid membranes to discovering de novo synthesis of cholesterol in bacterial species. Over the next five years, we would like to build on the foundation of these insights to move beyond studying lipid biosynthesis to lipid function, regulation, and transport in these microbes with a particular focus on sterols in bacteria. Sterol lipids are essential and ubiquitous components of all eukaryotic cells. The most well-studied sterol is cholesterol, the primary sterol produced by vertebrates and a key lipid in human health. High levels of cholesterol are linked to increased risks of cardiovascular disease and impaired cholesterol trafficking has been implicated in a variety of defects including lysosomal storage diseases. Further, bacterial-cholesterol interactions within the human host have been proposed to affect human lipid metabolism, through the gut microbiome, and are critical for infection by a variety of intracellular bacterial pathogens. However, the molecular mechanisms involved in bacterial-cholesterol interactions, particularly in the context of human health, are not fully understood. Interestingly, while certain bacteria can metabolize human host cholesterol, these bacteria cannot produce cholesterol themselves. Indeed, no bacterium, whether host-associated or free-living, was thought to produce cholesterol de novo. My research group recently demonstrated the first instance of de novo cholesterol biosynthesis in bacteria, and we have identified potential molecular links in cholesterol-producing bacteria to cholesterol metabolism by human-associated gut bacteria. We now seek to explore the molecular mechanisms driving bacterial-cholesterol interactions in bacteria that produce cholesterol and bacteria that metabolize cholesterol - two distinct bacterial groups that could provide broad insight into unknown biological pathways impacted by cholesterol as well as potentially uncovering novel sterol regulatory and transport mechanisms. Ultimately, we aim to further explore cholesterol metabolism in more complex microbiome communities with the goal of developing strategies to reshape microbiomes for beneficial outcomes.

NIH Research Projects · FY 2026 · 2026-06

PROJECT SUMMARY A major challenge in modern biological data analysis is integrating and reasoning over the vast volumes of unstructured, multimodal data now available, such as images and text. While each modality offers complementary biological insight, they are not encoded in a shared format or language, making it difficult to align them or reason across them computationally. A core unmet need is the development of AI systems that can bridge this divide by learning shared representations across data types. This project targets building AI systems for aligning and reasoning jointly over biological image and text data. We focus on microscopy and the challenge of interpreting microscopy images often requiring integration with broader biological context—relating observed phenotypes to those seen in other experiments, identifying plausible mechanisms, and connecting to relevant prior studies. This knowledge is frequently buried in unstructured images and text scattered across publications, databases, and annotations. Conventional AI systems rely on supervised learning, which demands large amounts of manually annotated data and cannot scale to the complexity or breadth of modern biology. Training AI models using weak supervision offers a promising alternative: by learning from loosely aligned image-text pairs, models can capture cross-modal associations from noisy but abundant sources. Vision-language models (VLMs) built on this principle embed images and text into a shared semantic space and support flexible reasoning tasks such as retrieval and question answering. However, current models often suffer from “blurry vision”—they can identify broad semantic matches between images and text but fail to resolve the fine-grained visual distinctions essential for biological interpretation. The goal of our project is to overcome this limitation by advancing weak supervision methods that enable fine-grained alignment between biological images and text, with a focus on microscopy. We will curate a large and diverse dataset of fine-grained image-text pairs and train a visual encoder using multi-scale contrastive learning to integrate both global and local alignment signals. This encoder will power an agentic AI system for query-based interpretation of microscopy images, that can retrieve relevant biological evidence and generate natural-language interpretations of microscopy images in response to researcher queries. We will validate the system in expert-driven use cases spanning single-cell perturbation and tissue-level pathology, and disseminate it through integration into widely used imaging workflows. By building AI tools that help researchers connect microscopy image content to pathways, phenotypes, and prior studies, we aim to support flexible, biologically grounded exploration and accelerate data-driven discovery. By open-sourcing our datasets, methods, and trained models for fine-grained image-text alignment, we also aim to advance the broader capabilities of multimodal AI for biological data analysis.

NSF Awards · FY 2026 · 2026-06

A key challenge for the success of underground carbon storage is development of technology to ensure that the gas will not slowly leak back to the surface over hundreds or thousands of years. This research focuses on understanding how carbon dioxide interacts with underground rocks and fluids, potentially changing their structure and allowing leakage pathways to form. By improving this understanding, the project aims to make underground carbon storage more reliable and effective. The project is also creating elementary education materials, student training opportunities, and publicly accessible open-source software. This project will investigate how fluid flow, chemical reactions, and rock structure and composition interact at very small scales to influence large-scale properties such as permeability and porosity in carbonate rocks. The project is built around a tightly integrated experimental, numerical, and AI-enabled mathematical framework. Microfluidic experiments on real rock samples will be used to identify key processes controlling fluid-rock interactions for different flow conditions and rock compositions. High-resolution numerical simulations will be validated against experiments and deployed to quantify the coupling between different processes. An AI-enabled computation framework will translate small-scale processes into larger-scale models and capture the temporal evolution of rock permeability during reactions. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-06

Running artificial intelligence workloads requires vast amounts of memory and energy, motivating a new generation of memory and transistor technologies that pack more storage closer to compute and operate at lower power. These emerging devices are promising candidates for accelerating AI, but they experience data corruption at rates far higher than conventional hardware, making them unreliable without costly error-correction techniques that erase their efficiency advantage. This project develops software tools and mathematical foundations that allow AI algorithms to be written so that they are naturally tolerant of such hardware errors, making these devices viable for real-world use without requiring error correction. The educational plan trains the next generation of students from computer science and electrical engineering to collaborate across the hardware/software boundary, addressing a critical workforce gap for the design of next-generation AI computing systems. The technical objective of this project is to establish exchangeability - a statistical symmetry property present in hyperdimensional computing, a naturally error-resilient computational model - as a principled foundation for error-resilient algorithm design across a broader class of computations, including machine learning inference and optimization. The central hypothesis is that exchangeability is more widespread than previously recognized, and that programs possessing this property can tolerate hardware-induced data corruption without sacrificing correctness. This project pursues two research thrusts. In the first, a programming language is developed for expressing and analyzing exchangeable programs, characterizing which AI and optimization algorithms possess this property. In the second, a compiler is developed that automatically optimizes such programs for performance and error resilience, evaluated in simulation across realistic hardware error models. The educational plan is integrated with these thrusts through "Hardware Anywhere" open-source projects that engage computer science and mathematics students in hardware design problems, and through coursework that gives students hands-on experience developing software optimizations for emerging hardware platforms. This work opens new connections between statistics, programming languages, and AI systems, and will be among the first to establish exchangeability as an exploitable computational property for program optimization. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NIH Research Projects · FY 2026 · 2026-06

Project Summary/Abstract This research program explores how the ubiquitin-fold modifier 1 (UFM1) conjugation system orchestrates ribosome quality control (RQC) in mammalian cells. Our recent work has demonstrated that UFMylation spatially and temporally coordinates the disassembly and clearance of stalled ribosomes at the endoplasmic reticulum (ER), a process critical for maintaining translational fidelity and proteostasis. The proposed research will define the molecular basis by which UFMylation regulates ribosome recycling after normal termination and in response to collision-induced translational stalling. Using a multidisciplinary approach that includes genetics, biochemical reconstitution, cryo-electron microscopy, and ribosome profiling, we aim to delineate how UFMylation interfaces with canonical RQC pathways and adapts to cellular stress. The R35 mechanism will provide the flexibility to expand these foundational discoveries into new areas, including how dysregulation of this pathway may impact cell viability and stress resilience.

NIH Research Projects · FY 2026 · 2026-06

PROJECT SUMMARY Chronic pain is a major healthcare problem, affecting 20% of the population and costing society approximately $600 billion dollars per year. Intracranial neuromodulation therapies such as deep brain stimulation have great therapeutic potential, but their development is hindered by a poor understanding of brain circuits underlying pain. Identifying pain-specific brain circuits is challenging because of the confound of salience. Salient stimuli, such as a surprising loud sound, are defined as those that are highly distinct from the environment. They typically evoke several reflexive processes including increased attention, preparatory motor activity, and sympathetic arousal, such as pupil dilation. Painful stimuli are generally highly salient and evoke widespread activation in neural circuits that process salient stimuli (including non- painful stimuli, e.g., a surprising, loud sound). Thus, there is a need to identify neural circuits that are specific for pain, as distinguished from neural circuits for salience, to inform future targeted brain stimulation protocols for the treatment of chronic pain. Our central hypothesis is that nociceptive stimuli evoke pain- specific neural activity across widespread brain regions that is distinct from salience-related neural activity. Our approach will be to obtain intracranial electroencephalography (iEEG) measurements in 20 neurosurgical patients as they perform thermal pain and auditory salience tasks. iEEG measurements are possible in patients who are admitted to the hospital for evaluation of medically refractory epilepsy (who typically do not have chronic pain). Patients are implanted with intraparenchymal depth electrodes in widespread brain regions (based on clinical criteria alone) and given the option to participate in research studies within the safeguards of an established Institutional Review Board (IRB) protocol. These recordings provide the rare opportunity to measure neural activity in the human brain with high spatiotemporal resolution and broad anatomical sampling and can complement prior non-invasive neuroimaging studies. Our preliminary studies in two patients show that a) iEEG can identify neural populations that encode thermosensory intensity, pain ratings, and salience, and b) that we can measure pupil dilation following salient stimuli as an independent measure of sympathetic arousal in the clinical setting. We will build on these preliminary results to test our hypothesis via the following complementary Specific Aims: 1) To disentangle pain and salience neural circuits using thermosensory and auditory stimuli, and 2) To disentangle pain and salience neural circuits using concurrent pupil measurements. Regardless of the outcome, the proposed research will improve our understanding of human brain circuits that process pain and how they relate to those for salience. In the long-term, our studies can inform intracranial brain stimulation therapies to target pain-specific neural circuits in treatment-refractory chronic pain.

NIH Research Projects · FY 2026 · 2026-06

Summary Pediatric Crohn’s disease (CD) is a rare and chronic immune-mediated condition that is managed, but not cured, with continuous anti-inflammatory/immunosuppressive drugs. The condition leads to impaired growth, poor quality of life, and irreversible organ damage that often necessitates surgery, although surgery does not usually fully resolve or control the disease. Gut epithelial damage is critical to disease pathogenesis, but no current interventions aim to promote repair of the gut epithelium. Independent of the pathogenesis, regulatory T cells (Tregs) as CD therapeutics, have gained much attention for their potential to be impactful across many pathogenic CD manifestations. Indeed, Tregs can exert dual functions dampening the dysfunctional immune system in CD and contribute to tissue healing and regeneration. However, the current approaches to obtain Treg cells for clinical use have many hurdles and their use in CD has been difficult to explore. Our group has established a process to efficiently generate Treg-like cells, converting effector T cells by lentiviral transfer of the FOXP3 coding region, the essential transcription factor for Treg development and function. We have extensively characterized these engineered Tregs, called CD4LVFOXP3 cells, and we are gaining knowledge of their safety and efficacy in vivo in patients with IPEX syndrome. IPEX syndrome is considered a very early onset (VEO)-IBD, resembling pediatric CD, and manifesting with severe enteropathy, especially affecting the small intestine, and other autoimmune manifestations. CD4LVFOXP3 Treg-like cell product is currently in Phase 1 clinical trial at Stanford as a treatment for IPEX syndrome (NCT05241444). We propose that the clinical applicability of CD4LVFOXP3 goes beyond IPEX syndrome whereby administration could enhance gut epithelial cell regeneration and promote immune-modulation in pediatric CD. Here, we propose to test the efficacy of CD4LVFOXP3 in pediatric CD using 2 complementary models: 1. An organoid culture system derived from crypts from ileum tissues (called enteroids) obtained from pediatric CD patients that recapitulate the gut epithelial structure and function, including mucus production, regeneration and differentiation capacity, and epithelial barrier functions; and 2. a humanized mouse model of colitis, as the best available for pediatric CD, that shows impaired survival, inflammation and infiltration of T cells in the gut. In these mice, CD-like pathologies can be induced using the well-established trigger, 2,4,6-trinitrobenzene sulfonic acid (TNBS). Using these systems, we will test the hypothesis that CD4LVFOXP3 can provide a therapeutic modality for patients with pediatric CD by a) promoting tissue repair and regeneration and b) exerting immune suppression towards pathogenic effector T cells. Our goal is to use these systems to perform the required key pre-clinical studies to advance CD4LVFOXP3 Treg-like cells as a therapeutic for pediatric CD patients.

NIH Research Projects · FY 2026 · 2026-06

With over 107,000 drug overdose deaths in 2023, the U.S. opioid epidemic has become an unprecedented public health crisis that is driving the epidemiology of transmission of HIV, HCV, and related diseases. At the same time, homelessness in the U.S. has reached epidemic proportions, with an estimated 770,000 people experiencing homelessness (PEH) on any given night. In addition to poor health outcomes associated with being unhoused, PEH have been significantly impacted by the opioid crisis. Our proposed work focuses on modeling the opioid epidemic and homelessness, including epidemiological links to HIV, HCV, and other communicable diseases, and modeling the health and economic impacts of housing provision, opioid use disorder (OUD) treatment, and other interventions for PEH with OUD. Importantly, we will consider the impact of barriers and enablers for such strategies. Our proposed work is highly significant and uses novel model-based approaches to assess the epidemiologic impact of housing status and opioid use on HIV, HCV, and related diseases and to assess how prevention and mitigation strategies such as housing provision and medication-assisted treatment can effectively improve health outcomes. Our economic analyses will capture the broad societal impacts, costs, and savings of interventions. Our work will provide clinicians, policymakers, and community organizations with critically needed epidemiological and policy guidance about how combinations of strategies can efficiently mitigate the consequences of homelessness and opioid use and how such interventions can be integrated to reduce incidence of HIV, HCV, and related diseases and to improve quality of life for PEH with opioid use disorder. Efforts are urgently needed to improve the health of PEH with OUD. The development of effective housing and OUD treatment options for this large population should be an urgent national priority.

NIH Research Projects · FY 2026 · 2026-06

Crabtree, Kay, Gray and Woyach REWIRING ONCOGENIC PATHWAYS TO INDUCE CELL DEATH IN CHRONIC LYMPHOCYTIC LEUKEMIA ASSIST ID1959504 Summary Chronic lymphocytic leukemia (CLL) is the most common chronic leukemia in the US. CLL remains incurable with ~4,500 death per year. The most urgent need for new therapies involves patients who have progressed after targeted therapies that now define modern care - BTK inhibitors and BCL-2 antagonists. The second major unmet need concerns patients with Richter transformation (RT), particularly because the incidence of RT climbs to ~ 15 % after six years of BTK-inhibitor therapy. Once CLL evolves into this aggressive lymphoma, median overall survival falls below 1 year. Taken together, progressive or relapsed CLL, particularly BTK/BCL- 2 double-refractory disease and RT, is a critical unmet medical need that demands new and innovative therapeutic strategies. The Crabtree and Gray laboratories have produced a new class of small molecules that rewire mutated cancer drivers to directly activate precise and powerful pathways of programmed cell death using chemically induced proximity (CIP). This class of molecules, called TCIPs (Transcriptional/Epigenetic Chemical Inducers of Proximity) uses a unique dominant gain-of-function mechanism designed to eliminate cancer cells containing the driving mutation in the face of secondary driver mutations or alternate oncogenic pathways (Gourisankar et al Nature 2023; Sarott et al Science, 2024). Indeed, in DLBCL (Diffuse Large B Cell Lymphoma), TCIP1 eliminates triple-hit PDX models as a single agent in only 21 days without recurrence or evidence of toxicity in mice. These molecules use chemically induced proximity (CIP) to bring epigenetic or transcriptional activators to the promoters of cell death genes normally bound and repressed by the BCL6 oncogenic transcription factor. By this means the BCL6 protein is rewired to be an activator rather than a repressor of proapoptotic genes, stimulating their expression and specifically killing the malignant cells. Since the original submission of this application, we have surveyed our combinatorial library of TCIP molecules and have found that TCIP2 (BCL6-CDK9, BAK-04-232) is highly effective (IC50 1 -20 nM) at selectively killing leukemic CLL cells. Recent studies, that showed BCL6 to be the most frequently mutated non-Immunoglobulin (Ig) gene in CLL B cells, suggest that non-coding mutations in the BCL6 gene are positively selected and that BCL6 contributes to the pathogenesis of CLL. We will define the mechanism underlying killing of CLL cells and explore the way that antiapoptotic epigenetic states are rewired by TCIP2 to become proapoptotic. To understand the potential role of TCIPs in CLL therapy, we will define the molecular features of cells sensitive to TCIP2 with the goal of identifying patient selection criteria that predict responses. For further mechanistic understanding of TCIP killing, we will explore the role of somatic hypermutation and negative autofeedback at the BCL6 locus of CLL B-cells in the pathogenesis of CLL and in producing sensitivity to TCIPs. Finally, we will use robust murine models to test the efficacy of TCIP2 in vivo. At the conclusion of our studies, we expect to have laid the foundation for clinical trials that will define the potential use of TCIPs in the treatment of CLL.

NIH Research Projects · FY 2026 · 2026-05

Abstract Lung cancer is the most frequent cause of cancer-related deaths in the world. Metastasis is responsible for the overwhelming majority of these deaths, yet the mechanisms underlying metastasis remain poorly understood. This is due in large part to the great difficulty in modeling metastasis for high-throughput functional analyses, as well as recapitulating complex tumor genotypes observed in humans. Using CRISPR-Cas9 somatic genome editing and tumor barcoding, I have developed a novel metastasis assay in genetically engineered mice that can be used to explore functional genetic effects on metastasis and immune evasion. With this approach, I found that inactivation of Cdkn2a promotes metastatic colonization and growth of oncogenic KRAS-driven lung adenocarcinoma. Furthermore, the simultaneous inactivation of several genes along with Cdkn2a dramatically increases metastatic spread, allowing spontaneous metastasis to the liver, bone and brain. In this proposed study, I hypothesize that specific tumor genotypes involving inactivation of Cdkn2a can enable potent metastatic ability. I will explore the ability of Cdkn2a knockout to promote metastasis by profiling the metastatic cell state of Cdkn2a-deficient cells and dissecting the specific impact of deleting the constituent genes Ink4a and Arf (Aim 1). I will investigate the ability of defined, Cdkn2a-related complex genotypes to dramatically promote metastasis, and I will map metastatic phenotypes to specific gene functions (Aim 2). I will produce tumors with these complex genotypes through the use of Cas12a. I have generated a Cas12a transgenic mouse to take advantage of the unique features of the enzyme, which include the simplicity of targeting multiple loci using a single CRISPR array. Finally, I will explore the metastatic role of the Mtap/Cdkn2a/Cdkn2b locus, which is frequently deleted in its entirety, and test how targeted therapy of Mtap-deficient tumors affects metastatic spread (Aim 3). This work will uncover functional genetic mechanisms by which lung cancer can metastasize, establish critical models for relevant in vivo investigation of metastasis, and identify new therapeutic strategies for treating metastatic lung cancer.

NIH Research Projects · FY 2026 · 2026-05

Project Summary Human neuroscience does not currently have a noninvasive tool to focally modulate activity anywhere in the brain. This is a critical unmet need: without a focal brain-wide neuromodulation technique, hypotheses on neural circuit function generated from neuroimaging and behavioral studies remain correlational, and circuit-based treatments for central nervous system disease are limited by lack of depth and anatomical specificity. Transcranial Ultrasound Stimulation (TUS) is rapidly emerging a powerful tool for neuroscience and clinical applications. However, the mechanisms, targeting performance, and efficacy are currently not well understood. Mice are the ideal model system to study the biophysics of these processes, using well developed tools based on genetically encoded calcium indicators (GCaMPs). The goal of this research is to use an innovative ultrasound transducer that provides an order of magnitude smaller focal spots at frequencies relevant to humans and therefore enabling technology for research in mice. We then use this transducer to obtain preliminary data with a novel paradigm, based on hippocampal slice data, to probe whether there is a state dependency of TUS to the visual cortex.

NIH Research Projects · FY 2026 · 2026-05

Abstract Alzheimer’s disease (AD) is the most prevalent form of dementia and is characterized by the progressive degeneration of neurons, leading to cognitive decline. Cumulatively, over 10% of people over the age of 65 have AD, higlighting an urgent need for innovative therapeutic targets, particularly as the rate of AD continues to rise globally. Previous research has focused on key hallmarks of AD including neurofibril tau tangles and amyloid- beta (Aβ) plaques primarily in neuronal cells. However, emerging evidence suggests that non-neuronal cells, including microglia and vascular cells, play a critical role in both the initiation and progression of AD. This proposal aims to leverage a groundbreaking air-liquid interface (ALI) adult human brain organoid culture system, which co-preserves diverse cell types including neurons, oligodendrocytes, astrocytes, endothelial cells, pericytes, and microglia, to comprehensively explore the underlying mechanisms of AD focusing on tau and/or Aβ as disease drivers. These aims determine the cell types responsible for tau and Aβ seeding and spread, characterizing the influence of microglia and endothelial cells on Aβ aggregation and tau spread, and elucidating the sequential dynamics of tau and Aβ in disease phenotype development. Specifically, Aim 1 identifies cell types responsible for primary Ab and tau seeding and secondary spread of “prion-like” aggregates phenotypes. Aim 2 determines the role of microglia in Ab and tau deposition and spread. Aim 3 determines the role of the vasculature in AD spread and identifies biomarkers for early AD. Through these studies, the goal is to provide critical insights into the contributions of the neurovascular microenvironment inclusive of microglia on AD initiation and spread, with potential implications for identifying novel therapeutic strategies aimed at mitigating disease progression. The proposed experiments will span the entire award period, with the technical training Dr. Rada receives during the mentored phase laying the groundwork for her independent research. A dedicated team of expert mentors and mentoring committee will provide her with instruction in essential methods that are crucial for the success of her studies on AD. In addition to hands-on training, Dr. Rada will actively participate in local and national meetings and scientific conferences to expand her network, enhance the visibility of her work, and stay current with advancements in the AD field. She will also attend career planning courses and meet regularly with her mentors and advisory committee to discuss her scientific progress, as well as to strategically prepare for job applications and interviews. Both of Dr. Rada’s co-mentors are fully committed to her success, ensuring she is well-equipped to carry her innovative research plan forward as she transitions to establishing an independent academic research lab.

NIH Research Projects · FY 2026 · 2026-05

PROJECT SUMMARY/ABSTRACT Post-hemorrhagic hydrocephalus following severe intraventricular hemorrhage (IVH) is the most severe consequence of premature birth. There are no treatments available to prevent hydrocephalus after IVH has occurred. The long-term objective of this research is to advance therapies for the prevention of hydrocephalus in preterm severe IVH. To address this unmet need, we need a better understanding of the cellular mechanisms leading to hydrocephalus following severe IVH. We propose that ferroptosis-driven ependymal and choroid plexus epithelial cell death is a critical mechanism for developing post-hemorrhagic hydrocephalus in infants with severe neonatal IVH. This work is novel both conceptually and in its approach. We will utilize a novel, highly specific biomarker of ferroptosis, 3R dimer of the transferrin receptor 1, to characterize ferroptosis. We will employ our novel rat pup model of IVH that recapitulates the severity of the human hydrocephalus disease state more realistically that prior animal models. Finally, we will take a novel approach of utilizing a genetic defect in DMT1 in the Belgrade (b/b) rat to manipulate an intrinsic disease mechanism to understand its role in an extrinsic etiology (IVH). To test the hypothesis that ferroptosis is the mechanism of cell death in the ventricular ependyma and choroid plexus following severe IVH and leading to hydrocephalus, our experiments will address the following AIMS: (1) Test if ferroptosis mediates ependymal and choroid plexus cell death after severe neonatal IVH in post-hemorrhagic hydrocephalus. We will test if 3R dimer formation is higher in [1.1] rat pups with severe IVH and hydrocephalus than in controls, and [1.2] CSF of human infants with hydrocephalus following severe IVH, compared to control (non-IVH) and resolved IVH (without hydrocephalus). (2) Test if divalent metal cation-1 (DMT-1) mediates ependymal and choroidal cell iron uptake following severe IVH, thereby increasing ferroptosis and post-hemorrhagic hydrocephalus. We will test if [2.1] ependymal and choroid plexus 3R dimer formation and [2.2] rates of hydrocephalus in DMT-1 deficient Belgrade (b/b) rats are reduced compared to wildtype rats following IVH induction, reflecting decreased ferroptosis in the DMT-1 deficient rat. (3) Test if intracerebroventricular (ICV) iron chelation with deferiprone prevents post- hemorrhagic hydrocephalus by inhibiting cellular iron uptake and ferroptosis-mediated ependymal and choroid plexus cell death. We will test if ICV deferiprone decreases [3.1] ependymal and choroid plexus 3R dimer formation and [3.2] hydrocephalus in IVH rat pups compared to controls.

NIH Research Projects · FY 2026 · 2026-05

Project Summary Opioid overdose, particularly from fentanyl, is the leading cause of drug-related deaths in the U.S., largely due to opioid-induced respiratory depression (OIRD). Tolerance to opioids' euphoric and analgesic effects develop faster than tolerance to their respiratory effects, increasing the risk of fatal overdose. Naloxone reverses respiratory depression but also blocks pain relief, highlighting the need for alternative therapies. This project describes the first reports of rhythmic dopamine release, produced after repeated high-dose fentanyl exposure, that may underlie the generation of agonal breathing responses. These “dopamine rhythms” originate in the ventral tegmental area (VTA) with each dopamine peak preceding inspiration. Mice exhibit a concomitant increase in both dopamine peak amplitude and agonal breathing response magnitude with each drug exposure. This proposal investigates the origin, function and molecular underpinnings of this possible “safety signal.” My preliminary data demonstrate that opioid-induced rhythmic dopamine release produces robust stereotyped peaks and is remarkably reproducible. To understand the role of these rhythms, I will employ optogenetic manipulations, patch clamp electrophysiology and single-cell RNA sequencing (scRNAseq) strategies. Based on my data thus far, I hypothesize that these dopamine rhythms have evolved to help maintain breathing under hypoxic stress. In Aim 1 (K99), I will delineate the causal relationship between the agonal breathing responses and the dopamine rhythms by optogenetically activating and inactivating the rhythmic signals. In Aim 2 (K99), I will use patch clamp electrophysiology to determine the molecular underpinnings in VTA dopamine and GABAergic neurons that may underlie the generation of these rhythms. Aim 3 (K99) will use scRNAseq to delineate cellular targets that may contribute to OIRD or opioid-use disorder generally. Finally, Aim 4 (R00) will use function-oriented synthesis to create novel chemical compounds to test the causal role of HCN channels in rhythm generation. Together, this proposal will define the role of these novel dopamine rhythms with breathing responses after repeated fentanyl administration. This project is conceptually novel, as it may represent a new function of the mesolimbic dopamine system in survival. In the K99 phase, I will be mentored by Dr. Robert Malenka, co-mentored by Dr. Karl Deisseroth, and will be advised by an exceptional advisory team composed of Dr. Boris Heifets, Dr. Ali Mohebi and Dr. Longzhi Tan. With their support and the tremendous scientific environment at Stanford University, I will gain technical training on optogenetics and scRNAseq, as well as refine my skills with patch clamp electrophysiology. This training will prepare me for my long-term goal of leading my own research laboratory where I will utilize advanced experimental techniques and a circuit-level approach complexed with my background in drug drug discovery to identify novel therapeutic targets for OIRD.

NIH Research Projects · FY 2026 · 2026-05

PROJECT SUMMARY/ABSTRACT Efficient cell-specific delivery is the key bottleneck of applying macromolecule therapies to various diseases. Traditional delivery vehicles each have their own limitations. For instance, viral vectors and nanoparticles are limited by poor specificity, small cargo capacity, and low programmability to different tissue/cell types. As cell- based therapies gain clinical momentum, we propose a paradigm shift by using autologous or allogeneic cells as vehicles to deliver therapeutic biomolecules, enabling precise, context-dependent delivery to various cells. Inspired by the natural phenomenon of trogocytosis, where one cell acquires a patch of membrane and associated molecules from another cell through direct contact, we engineered donor cells with customized receptors and a pH-sensitive fusion protein (termed fusogen) to deliver large cargos such as CRISPR-Cas9 and base editors into recipient cells. Upon trogocytosis, the fusogen facilitates cargo endosomal escape, while a release domain preferentially cleaves at the acidic pH frees the cargo within the recipient’s cytosol or nucleus for downstream function (e.g., gene editing in the nucleus or triggering apoptosis in the cytosol). We termed this cell-based delivery system “Trogocytosis-based tRANSfer and Functional Effector Release” (TRANSFER). In the proposed studies, we aim to develop TRANSFER as a therapeutically relevant system and demonstrate its applications for disease treatment in vivo. In Aim 1, we will improve TRANSFER using primary cells as the donors (e.g., T cells and retinal pigment epithelial cells [RPE]) for cell-type-specific delivery to various primary recipient cells by refining key components, including transmembrane domains, fusogens, and release domains, developing targeting strategies for primary recipient cells, and expanding the range of cargos delivered. In Aim 2, leveraging our expertise on retinal biology and diseases, we will evaluate TRANSFER in simultaneous prime editor and neurotrophic factor delivery in a preclinical retinal degeneration model, providing a proof-of-concept that could extend beyond neurodegenerative diseases. Overall, this project lays the groundwork for a novel, powerful, patient-friendly cell-based system that addresses major hurdles in delivering macromolecular therapeutics (including editing enzymes and biologics). By synergizing the safety and targeted potential of cell-based therapies with the precision of engineered release mechanisms, TRANSFER could significantly expand treatment options for neurodegenerative, ocular, and other diseases where localized, large-cargo delivery is critical. If successful, it can greatly expand the methods for efficient therapeutic cargo delivery in vivo, fueling the development of next generation cell therapy and gene therapy toward disease such as retina disorder, muscular disease, and neurological diseases.

NIH Research Projects · FY 2026 · 2026-05

PROJECT SUMMARY/ABSTRACT In this R01 project, the researchers will use an innovative technology pipeline developed in the PIs’ laboratories to characterize Epstein-Barr virus (EBV)-infected B cells in patients with multiple sclerosis (MS). Infection with EBV is highly associated with MS, as almost 100% of MS patients contract EBV before disease onset. Interestingly, the virus is also linked to other autoimmune diseases, including systemic lupus erythematosus (SLE), despite differing disease presentations and underlying pathologies. EBV establishes latency in B cells and viral dysregulation of B cells is likely an important mechanism that contributes to autoimmunity. In addition, EBV-infected B cells might be specific to self-antigens of the central nervous system (CNS), and EBV might perpetuate their survival and cause their escape from tolerance. Capturing and characterizing latently EBV-infected B cells in healthy individuals and autoimmune patients has been a major challenge in the field, as only 1 in 104 – 106 B cells are infected with EBV, and there are no cell surface markers that would facilitate their enrichment. The researchers have developed a novel method, named ELUCIDATE (EBER-Labeled Unique Cell Identification and Deep Analysis of Transcriptomes, Immune Repertoires, and EBV-genes). This method enables the isolation of EBV-infected B cells by flow cytometry, based on the expression of two EBV-encoded small non-coding RNAs (sncRNAs, EBER1 and 2), and subsequent single-cell capture and sequencing to describe their transcriptomes, immune repertoires, viral genes and sncRNAs. This proposal builds on prior research the labs have undertaken in SLE, and it seeks to address critical questions regarding the effects of EBV-infection on B cells and its impact on MS. The research will identify specific characteristics of EBV+ B cells in MS compared to SLE and healthy individuals, investigate inflammatory pathways activated by EBV in MS EBV+ B cells, assess whether B cell receptors (BCRs) of EBV+ B cells recognize the patients’ CNS antigens, and evaluate potential transformations of B cells in cerebrospinal fluid (CSF). By uncovering the key dysregulated pathways influenced by EBV, this study aims to enhance our understanding of their roles in MS pathology.

NIH Research Projects · FY 2026 · 2026-05

Project Summary/Abstract Pancreatic ductal adenocarcinoma (PDAC) is an exceptionally deadly cancer with few available treatment options. The lack of treatment options and poor treatment outcomes is associated with its extensive fibrosis, which contributes to therapeutic resistance. Cancer-associated fibroblasts (CAFs) mediate local tissue remodeling and are implicated in PDAC’s extensive fibrosis. However, therapies targeting CAF development and activation have paradoxically led to worsening tumor growth and metastasis in PDAC. Recent analyses at the single cell resolution have identified a unique subpopulation of inflammatory CAFs in PDAC. Our preliminary work has focused on understanding CAF phenotypes, determining the factors that promote differentiation to inflammatory CAFs, and the influence of CAF phenotypes on the tumor microenvironment. Using primary human cell models, we have demonstrated that inflammatory CAF differentiation occurs in response to exposure to interleukin 1 (IL-1), lipopolysaccharide (LPS), or PDAC cell conditioned media and is dependent on intact fibroblast myeloid differentiation factor 88 (MyD88) signaling. Furthermore, inflammatory CAFs generated in this manner are potent suppressors of T cell activation, interferon-γ production, and cytotoxic T cell function. This revelatory preliminary work suggests that aberrant MyD88 signaling is responsible for tumor progression, which explains the poor outcomes associated with previous therapies. MyD88 presents a promising new therapeutic target for the treatment of PDAC. We will first assess whether inflammatory CAF differentiation can be suppressed by MyD88 inhibition in mouse models of PDAC. This work will elucidate the factors that control the inflammatory CAF phenotype in PDAC, the influence of the inflammatory CAF phenotype on the tumor microenvironment, and the resultant impact on tumor growth and disease progression. Next, we will determine the mechanism of T cell suppression in PDAC through identification of the secreted factors from inflammatory CAFs responsible for this activity using primary human cell models of patient-derived CAFs. These secreted factors present valuable targets for treatments that specifically prevent T cell suppression from inflammatory CAFs, likely to be more valuable for treatment than global suppression of MyD88 signaling. Third, we will explore a novel treatment regimen which leverages MyD88 pathway inhibition in PDAC in combination with immune checkpoint blockade. This work will utilize a currently available MyD88 pathway inhibitor which has not previously been investigated for efficacy in combination with immune checkpoint blockade. These experiments will solidify a novel relationship between inflammatory CAFs, MyD88 signaling, and T cell suppression. In addition to providing mechanistic insights into CAF differentiation in the development of PDAC, we propose to complete the preclinical work necessary to combine immune checkpoint blockade and MyD88 pathway inhibition in a clinical trial for PDAC.

NIH Research Projects · FY 2026 · 2026-05

PROJECT SUMMARY/ABSTRACT Depression affects millions of individuals worldwide, yet the treatments including transcranial magnetic stimulation (TMS), achieve only moderate success. While the exact mechanisms underlying the therapeutic effects of TMS remain unclear, they are in part attributed to the induction of neural plasticity. Notably, plasticity is most pronounced during non-rapid eye movement (NREM) sleep, particularly in slow-wave sleep (SWS), when synchronized oscillations between the cortex and thalamus may optimize neuroplastic changes. However, currently, TMS is administered in wake-state only. This research gap suggests a new frontier for TMS in depression - stimulating the brain during sleep rather than while awake. By timing TMS to coincide with critical neural events that promote plasticity and systems-level consolidation in NREM sleep, such as thalamocortical sleep spindles, we may enhance TMS efficacy to induce prefrontal plasticity for treating depression. I have developed an approach to deliver TMS during sleep, using automated systems for real-time detection of sleep stages, slow oscillations, and sleep spindles. My preliminary work has shown that intermittent theta burst stimulation (iTBS) of the primary motor cortex (M1) during NREM sleep produces more robust cortical changes than when applied during wakefulness and that spindle-guided iTBS further amplifies these plasticity effects. Building on this, I hypothesize that targeting the dorsolateral prefrontal cortex (dlPFC) with iTBS during NREM sleep will result in superior prefrontal plasticity, improving both brain function and behavior in depression. I will first test the effects of intracranial electrical iTBS of dlPFC during NREM sleep on brain activity in neurosurgical patients, using intracranial EEG (iEEG) to measure evoked responses and index plasticity (Aim 1). Next, I will focus on TMS-delivered dlPFC iTBS in depressed patients, using simultaneous TMS-EEG to measure noninvasive brain responses during NREM sleep and pre-sleep wakefulness conditions (Aim 2). Finally, I will investigate real-time sleep-spindle guided dlPFC iTBS in healthy individuals, hypothesizing that targeting events of spindles during SWS will induce even superior prefrontal plasticity and improvements in working memory (Aim 3). For training and professional development while pursuing these aims, I will rely on a mentoring team of world-class experts in invasive and noninvasive brain stimulation, depression, sleep and neural oscillations: Drs. Corey Keller, Josef Parvizi, and Andrea Goldstein-Piekarski, with Drs. György Buzsáki, and Manish Saggar as advisors. This work will deepen our understanding of the neural effects of TMS in the prefrontal cortex and its potential to enhance treatment outcomes in depression. Through this project, I will gain critical expertise in intracranial EEG, direct brain stimulation, pathophysiology of depression and learn how to design, recruit, and execute an independent clinical trial, all of which will prepare me to lead a future independent academic career in sleep-augmented brain stimulation therapies for psychiatric disorders.

NIH Research Projects · FY 2026 · 2026-05

ABSTRACT Tumor suppressor pathways are critical regulators of many different aspects of cancer development, including tumor initiation and growth. Identifying tumor suppressor genes, the pathways they control, and their interactions with other genetic alterations is essential for understanding tumorigenesis. While cancer genomics has revealed important tumor suppressor genes, many remain uncharacterized and there is a growing appreciation that genes and pathways that are not highlighted in cancer genomic studies can be fundamentally important in cancer. The function of putative tumor suppressor genes on cancer growth has been investigated mostly using cell lines and mouse models. Cell lines lack a natural microenvironment and have near optimal growth, while genetically engineered mouse model recapitulate human tumors but are limited in scale. Thus, a comprehensive approach to assess tumor suppressor gene function is lacking. To overcome these limitations, we integrated genetically engineered mouse model, multiplexed CRISPR-based genome editing, and tumor barcoding to assess multiple genotypes on lung tumor initiation and growth in parallel. We hypothesize that many uncharacterized tumor suppressor genes exist, and their effects are highly dependent on genomic context. Our preliminary data and novel in vivo models make us well-positioned to achieve these goals. In Aim 1 we will use CRISPR/Cas9 genome editing and tumor barcoding in a KRAS-driven mouse model to inactivate ~10,000 genes and quantify their effects on lung tumor initiation and growth. We will identify key genetic determinants of lung carcinogenesis and generate novel insights into the molecular mechanisms underlying tumor initiation and growth. In Aim 2, we will determine how tumor suppressor genes function across various oncogenic contexts, using multiplexed in vivo genetic epistasis experiments in EGFR-, KRAS-, and BRAF-driven lung tumors. Finally, in Aim 3, we will use somatic genome editing, diverse oncogene-driven lung cancer models, and in vivo Perturb- seq to study the molecular effects of single and combined tumor suppressor gene inactivation. This will provide a molecular framework to understand the effects of novel tumor suppressor genes and begin to uncover the molecular logic that drives the pattern of genomic alterations in human cancer. The unprecedented scale and resolution of these data will have a broad impact on our understanding of lung tumorigenesis with potential implications for tumor suppression across other cancer types. These innovative, multidisciplinary, and highly quantitative approaches will fundamentally improve our understanding of the pathways and biological processes that drive lung tumorigenesis in the physiologic in vivo setting and begin the systematic deconvolution of gene function during tumorigenesis.

NIH Research Projects · FY 2026 · 2026-05

PROJECT SUMMARY An estimated 10-25% of human embryos contain incorrect chromosome numbers (aneuploidy), the leading cause of pregnancy loss and birth defects. Most critically, 25% of human oocytes are aneuploid, with this rate dramatically increasing after age 35, contributing to age-related fertility decline. The disproportionately high error rates in female versus male meiosis highlight fundamental gaps in understanding oocyte chromosome segregation. A critical knowledge gap exists in how oocytes assemble functional spindles without centrosomes—termed acentrosomal meiosis. While centrosomes organize microtubules in sperm meiosis, oocytes employ poorly understood alternative mechanisms. Additionally, although microtubules comprise diverse tubulin isotypes, their mechanistic significance in oocyte meiosis remains unexplored. Recent discoveries linking oocyte-specific tubulin TUBB8 mutations to human infertility disorders underscore the importance of understanding isotype-specific contributions to chromosome segregation. My goal is to elucidate the genetic and molecular mechanisms critical for acentrosomal meiosis and how tubulin isotype composition impacts chromosome segregation fidelity. My preliminary data using C. elegans β-tubulin isotypes TBB-1 and TBB-2 reveal these isotypes differentially affect spindle length, structural integrity of metaphase arrested spindles, microtubule motor sensitivity, and anaphase segregation velocity. I have remodeled existing tools, including a strain co- expressing endogenously tagged subunit of microtubule severing enzyme katanin (GFP::MEI-1) and mCherry::H2B, and single-β-tubulin substitution strains that also express GFP-tagged α-tubulin (GFP::TBA-2), providing unprecedented opportunities to dissect acentrosomal spindle assembly. I hypothesize that acentrosomal spindle assembly requires coordinated interactions between microtubule-interacting proteins and that such interactions are regulated by tubulin isotype composition. Using quantitative proteomics, forward genetic screens, in vitro biochemical analysis and high- resolution microscopy: Aim 1 investigates the molecular basis of differential β-tubulin isotype contributions through isotype-specific proteomics and biochemical characterization; Aim 2 identifies katanin-mediated mechanisms essential for acentrosomal spindle assembly using proximity labeling and genetic screening. This research aligns with my career objectives to become an independent investigator studying microtubule-dependent chromosome inheritance. The K99 phase will provide essential training in proteomics, biochemical reconstitution, and genetic screening—techniques underutilized in oocyte research. My mentoring team, led by Dr. Anne Villeneuve at Stanford, combines expertise in such essential techniques positioning me to master approaches that differentiate my research program. This project will establish the first comprehensive analysis of tubulin isotype-specific roles in any system and develop mechanisms enabling accurate chromosome segregation without centrosomes. The innovative methodological approaches will provide robust foundations for R01 applications, ultimately advancing diagnosis for reproductive health challenges affecting millions of families worldwide.

NSF Awards · FY 2026 · 2026-05

This project focuses on developing the next generation of network scanning tools and methodologies for more efficiently finding all Internet devices and software, safely uncovering vulnerabilities and misconfigurations in them, and identifying device owners so that they can be notified before they are attacked. This project will introduce fundamentally new methodologies in three areas. First, drawing on advances in artificial intelligence and software fuzzing, the project will build new techniques for finding network scan probes that safely identify hardware and software manufacturers, products, and fine-grained versions. The project will build agentic approaches to scalably uncover human misconfigurations that could not otherwise be programmatically identified. Second, new methods to uncover relationships between Internet entities and to extract device owners will be developed. Third, building on the context and relationships derived about Internet assets, new predictive methods for finding Internet services as they come online will be developed. This project will provide new techniques for networking and security researchers to better understand the Internet, for U.S. organizations to more quickly protect themselves against attacks. The project will develop new curriculum to prepare computer science students to work in cybersecurity by providing them hands-on opportunities to understand Internet-security in practice as well as provide research opportunities to both undergraduate and graduate students. Publications, open source software, and other research outputs will be made publicly accessible at http://esrg.stanford.edu/. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.