Yale University

NIH Research Projects · FY 2025 · 2025-07

ABSTRACT Loss-of-function mutations in ribosomal protein S6 kinase 2 (RSK2, encoded by the RPSKA3 gene) cause Coffin- Lowry syndrome, an X-linked dominant condition characterized by craniofacial abnormalities, short stature, and intellectual disability. The RSK2 N-terminal catalytic domain (NTD) shares a consensus phosphorylation site motif with numerous related kinases, including isoforms of AKT, SGK, and S6K. As such, it is difficult to rationalize how RSK2 targets a unique substrate pool, and the utility of this target motif to discover substrates of RSK2 and these other kinases is limited. Recently a cryptic allosteric pocket was identified in the RSK2 NTD that binds to a short linear sequence motif found in several of its interacting proteins. This RSK-interacting motif (RIM) consists of a core Val-Phe sequence, but the full motif has not yet been systematically defined. Furthermore, it is not known whether the analogous binding site in other related kinases might target distinct motifs. Here, we propose to comprehensively define RIM sequence space, screen the human proteome for functional RSK2-interacting sequences, and determine whether other kinases utilize the analogous binding pocket for protein-protein interactions. In preliminary studies, we have found the interaction between the RSK2 NTD and the RIM from its substrate protein SPRED2 is detectable by yeast two-hybrid (Y2H) analysis in a manner sensitive to mutation of a core residue. To comprehensively define the RIM, we will construct a Y2H double positional scanning library consisting of all pairwise combinations of residues at two positions in the context of the SPRED2 sequence. We will screen the library against RSK2 and a panel of related kinases. We will verify the importance of specific residues within the motif through quantitative measurements of kinase binding to synthetic peptides harboring targeted substitutions. Guided by these results, we will construct and screen a human proteomic Y2H library of candidate RIMs consisting of intrinsically disordered fragments of human proteins. RSK2-interacting sequences identified from the screen will be validated as direct binders of RSK2 in vitro. We will then investigate whether the full-length proteins harboring these sequences can act as RSK2 substrates or interactors in a manner dependent on the docking motif. These results will form the basis for subsequent in depth study of novel RSK2 substrates and their importance in normal physiology and in Coffin- Lowry syndrome. Our studies may furthermore reveal that a large number of related protein kinases utilize the corresponding interaction pocket in substrate docking, laying the groundwork for future studies to define their interaction networks and better understand how they function.

NIH Research Projects · FY 2026 · 2025-07

Summary Normal tissue structure and function rely on the regenerative capacities of tissue resident stem cells. These stem cells engage in essential activities like proliferation, migration, and differentiation, which must be carefully coordinated through dynamic signaling interactions to ensure effective tissue regeneration. Disruption of this process can give rise to diseases, including cancer. Thus, comprehending how dynamic signaling activations orchestrate stem cell behaviors to regenerate tissue and sustain tissue function is of paramount importance for advancements in regenerative medicine and cancer therapy. The inability to simultaneously track the dynamic signaling changes and stem cell activities in most mammalian tissues has been limiting our knowledge of the signaling mechanism of stem cell coordination. By generating novel signal sensors and employing our intravital imaging approach on a tissue regeneration model, the hair follicle, our previous studies have overcome this challenge and enable us to interrogate the functional significance of signaling dynamics at the single-cell level in orchestrating regenerative cell behaviors. Previous studies in culture systems and simple epithelial models have discovered regulatory roles of the distinct signal dynamics for specific cell behaviors. It remains largely unclear how dynamic signaling activations orchestrate diverse behaviors during complex tissue regeneration. Here, we study this question by using mouse hair follicle regeneration as a model, which requires both epithelial and mesenchymal signals. We hypothesize that stem cell behaviors are directly coordinated by epithelial ERK signal dynamics and indirectly influenced by TGFβ-mediated mesenchymal organization. We will test this hypothesis by defining the dynamics of the epithelial ERK (Aim 1) and mesenchymal TGFβ (Aim 2) signals and assessing the functional implications of those dynamics on stem cell behaviors and functions. We will apply intravital imaging, genetic, optogenetic, and pharmacological manipulations, as well as novel molecular analyses to gain new insights into the mechanisms by which dynamic signals orchestrate different cell behaviors to support effective tissue regeneration and uphold normal tissue architecture. The outcomes of the research will significantly advance our knowledge of the signaling mechanisms that both promote and disrupt tissue regeneration, which will pave the way for the development of innovative strategies to treat various diseases in the future.

NIH Research Projects · FY 2025 · 2025-07

Project Summary The objective of the Yale Predoctoral Pharmacology Training Program (PPTP) is to train an outstanding and diverse group of graduate students in the pharmacological sciences for research careers in academia, the pharmaceutical and biotechnology industries, or government. The PPTP will accomplish this objective by drawing on the experiences and strengths learned from our previous 45 year-long training program. Our innovative and interdisciplinary curriculum will integrate fundamental pharmacological principles with related areas of modern biological science, including signal transduction, metabolism, neuroscience, and structural biology. Classroom activities in core topics will be complemented with a newly developed course in drug discovery and development that offers hands-on experience in high-throughput screening, proteomics, and functional genomics, as well as interactive workshops conducted by leaders in the pharmaceutical industry. A group of 47 world-class faculty have been assembled from multiple departments, all of whom are dedicated to mentoring and providing a safe and inclusive environment for their trainees. Our mentors will train students in laboratory-based research meeting the highest standards of rigor, reproducibility and ethics. Outside of the laboratory, students will participate in activities to acquire essential skills, including two semesters of teaching and monthly research-in-progress talks. We will promote a cohesive and inclusive PPTP community through regular activities including an annual symposium and retreat, a career development series featuring program alumni, town hall meetings, student organized “research chats”, and a seminar series drawing from leading scientists in pharmacology and related disciplines. Student progress will be assessed through a qualifying examination administered during their second year, and through regular thesis committee meetings occurring through the remainder of their doctoral training. Student feedback will be fully integrated into the delivery of the PPTP training experience in order to optimize program outcomes and to guide future improvements. Additionally, we have developed special assessment tools and surveys that monitor trainee development, which will be augmented by tracking trainee outcomes. The PPTP is unique among T32-supported programs at Yale as it is the only one that provides training in core principles of pharmacology. Experiences provided by our program will equip students for successful careers in biomedicine, meeting a national need for well-trained individuals with expertise in the pharmacological sciences. We request 8 training positions to appoint 4 second year and 4 third year students per year.

NIH Research Projects · FY 2026 · 2025-07

PROJECT SUMMARY Novel long axial PET scanners with axial field of view that can be greater than 1m offer increased sensitivity and allow fast dosimetry and biodistribution for pharmacokinetic studies to enable personalized Targeted Radionuclide Therapy (TRT). However, they pose significant challenges both financially and logistically for siting. In this context, recent and ongoing advances in Time-of-flight (TOF) PET technology afford a rare opportunity to improve signal-to-noise-ratio (SNR) without increasing cost associated with axial coverage as is the case with total body PET technology. In fact, very high timing resolution performance would enable just the opposite: very sparse angular coverage of the patient, long axial field coverage (>1m) for a significantly lower cost total body PET scanner and/or affordable brain PET. The aim of this proposal is to develop and optimize the building block of a modular TOF PET prototype scanner with flexible geometry based on pixelated L(Y)SO scintillators and novel integrated photosensor consisting of SiPMs and electronics with an unprecedented Coincidence Timing Resolution (CTR) mostly limited by the light travel time spread in the scintillator pixels and ~3x better than TOF resolution of leading state-of-the-art scanners.

NIH Research Projects · FY 2025 · 2025-07

There is a compelling national need to train a new generation of scientists who are well prepared to advance translational biology. At the same time, the growing desire among a robust cohort of our most talented students to obtain a sophisticated understanding of human disease and to utilize their training to create novel solutions and expedite application to humans presents tremendous opportunity. To capitalize on this aspiration, we are excited to propose this Medical Research Scholars Program (MRSP), a Certificate-based interdisciplinary training program designed for highly motivated predoctoral students to apply their curiosity, creativity and drive to problems arising from the pathogenesis of human diseases. This new interdisciplinary program is career enriching and will facilitate the long-term engagement of Scholars with mentors and peers during and after the predoctoral period to foster collaboration and expand the network of translational science. The MRSP is designed to be complementary and concurrent with the programs within the aegis of Yale’s Combined Program in Biological and Biomedical Sciences (BBS), allowing scholars to pursue in depth their individualized research and dissertation goals while acquiring the breadth of skills, background, and emotional intelligence needed to contribute productively to future careers that leverage translational science. The major elements of the formal curriculum are: 1. Core Courses providing a foundation in biomedical sciences using contemporary, active learning pedagogical techniques in interactive classrooms; 2. the Mentored Clinical Experience (MCE), the lynchpin of the MRSP, which exposes students to four dimensions of a human disease: 1) clinical manifestations and pathogenesis; 2) contemporary diagnostics & therapeutics; 3) direct interaction with patients in a mentored setting; and 4) cutting edge research & unaddressed scientific challenges. As important as the formal curriculum, 3. Career Development Activities to promote expression of the unique interests of each Scholar through an extended relationship with a clinician mentor, critical thinking exercises with peer teams, sharing of respective research, and community building through supplemental components, such as “beyond the bench” and “lunches with leaders”, engagement with other predoctoral students pursuing translational research at Yale, and an Annual Symposium. The primary objective of the MRSP is to attract and prepare our Scholars to be leaders with sustained careers in academia, industry, and other research-related paths. Outcomes will be measured using self-assessment and reflection, regular surveys of current and alumni scholars, and long-term tracking of career choices and endeavors. This translational research training program enjoys the commitment from our institution manifest by stipend support for the first year of pre-doctoral training, the administrative infrastructure for the MRSP, and the Career Development components. Furthermore, the MRSP will take advantage of the deep biomedical sciences applicant pool at Yale as well as the extraordinary research strengths of the faculty in the basic and translational sciences.

NIH Research Projects · FY 2025 · 2025-07

Project Summary/Abstract In this proposal we hope to build on our efforts showing innate and adaptive immune cell infiltration of labyrinthine tissue in Meniere’s disease. These findings have led us to propose an autoimmune basis for the disease. Our efforts will be directed at confirming our preliminary data that specific pathways including druggable TNF alpha and mTOR pathways are upregulated in Meniere’s disease and establishing that labyrinth infiltrating B cells are autoreactive. We hope that our work will lead to better understanding and rational therapeutics of this poorly understood disease.

NSF Awards · FY 2025 · 2025-07

Although current physics theories successfully explain nearly all laboratory experiments carried out to date, they cannot account for key properties of the universe as determined from astrophysics. For instance, the observed structure of the universe can only be explained through the existence of dark matter—a form of matter that is fundamentally different from atoms, and which has never been detected on Earth because it interacts only extremely weakly with regular matter. In addition, although gravity has been studied for hundreds of years, gravitational forces between microscopic particles that obey the laws of quantum mechanics have never been measured, and theories of gravity may need to be modified in the quantum realm. In this work, the research team will develop new types of force sensors that may allow detection of the tiny forces imparted by dark matter, or from gravity between microscopic particles. Students and postdocs participating in this work will be trained in advanced quantum sensing techniques and will work with the PI to teach a hands-on summer program for local high school students about physics in their everyday lives and connections to state-of-the-art research. This research program will employ arrays of micro- and nano-particles trapped in ultra-high vacuum as sensitive force sensors for searches for physics beyond the Standard Model of Particle Physics. The research team will use these particle arrays to search for dark matter that primarily interacts coherently with nano- or micro-sized particles (rather than single nuclei or electrons), with several orders-of-magnitude improved sensitivity over previous searches for such dark matter models. In addition, the research team will develop techniques to trap solid noble gas particles, such as solid xenon, in a cryogenic optical trap. Solid noble gas particles may provide significant advantages over existing techniques (that primarily employ silica particles) due to their extremely high purity and ability to construct large particle arrays. 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 2025 · 2025-07

Harnessing the power of data collected from a vast amount of geographically distributed and heterogeneous devices, in a manner without moving data around and violating privacy, has great potential in advancing science and technology and improving quality of life. Federated optimization lies at the heart of the practice realizing this vision, encompassing problems such as training large-scale machine learning or artificial intelligence models, delivering insightful data analytics, as well as facilitating decision making under uncertainty, all in distributed manners. There is a significant gap in the algorithmic foundation of federated optimization when interfacing with bandwidth-limited heterogeneous networks, such as internet-of-things, smart healthcare, and edge computing, to meet the unique challenges of taming heterogeneity, privacy, and uncertainty without sacrificing efficiency. This research project will also be tightly integrated with education and workforce developments, through offering new courses, mentoring students at all levels in research projects, and disseminating the research outcomes at suitable conferences and workshops. The goal of the research program is to develop a federated optimization framework to learning and decision making by designing communication-efficient, computation-scalable, and privacy-preserving algorithms that converge provably over highly heterogeneous data and computing environments. Leveraging insights from machine learning, optimization theory, signal processing, and differential privacy, the research program offers an entirely new suite of theoretical and algorithmic tools to enable heterogeneity-embracing and privacy-preserving learning and decision making in federated environments under bandwidth constraints, unveiling fundamental trade-offs among computation, communication, privacy, and utility. The research program will gravitate around a semi-decentralized federated setting suitable to meet the diverse needs of bandwidth-limited heterogeneous networks, and focus on developing bandwidth-limited federated optimization algorithms that are efficient, resilient, and private with rigorous performance guarantees for a wide range of problems arising from machine learning, data analysis, and sequential decision making. 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 2025 · 2025-07

ABSTRACT The Clinical High-Risk Syndrome for Psychosis (CHR-P) is a DSM-5-recognized condition affecting youth and young adults who experience attenuated delusions, hallucinations, and thought disturbances that, while subthreshold for full psychosis, are distressing and impair daily functioning. Attenuated Psychosis Syndrome (APS), listed under "Conditions for Further Study" in DSM-5, affects approximately 1.7% of the general youth population and nearly 20% of youth presenting to psychiatric services, yet remains under-recognized and poorly served by qualified assessment tools. Current clinical outcome assessments (COAs) for APS include the Structured Interview for Psychosis-Risk Syndromes (SIPS) and the Comprehensive Assessment of At-Risk Mental States (CAARMS). Differences between these instruments have historically hindered harmonization across research and clinical trials. In response, the National Institute of Mental Health led a harmonization effort that resulted in the Positive SYmptoms and Diagnostic Criteria for the CAARMS Harmonized with the SIPS (PSYCHS), a semi-structured ClinRO instrument designed to assess 15 distinct attenuated positive symptoms organized into three general concepts: attenuated delusions, hallucinations, and thought disorder. There are currently no FDA-qualified COAs for measuring APS severity in CHR-P clinical trials, presenting a major obstacle to drug development. The PSYCHS is now implemented in the FNIH-funded AMP SCZ observational study and will be further evaluated in the upcoming ProCAN randomized controlled trial, which includes PSYCHS assessments, blinded raters, and ecological momentary assessment. These data provide a unique opportunity to establish the PSYCHS as a reliable and valid ClinRO for use in regulatory trials. The objective of this project is to develop a comprehensive Full Qualification Package (FQP) for the PSYCHS. We aim to: (a) engage the FDA to refine our Qualification Plan through ongoing consultation; (b) systematically evaluate content validity using input from clinicians, researchers, patients, and trainers; and (c) conduct preliminary qualitative and quantitative studies to assess test–retest reliability, recall periods, clinically meaningful change, and within-patient thresholds. The FQP will integrate data from previous studies, ongoing observational research, and a forthcoming clinical trial to support regulatory qualification of the PSYCHS instrument.

NSF Awards · FY 2025 · 2025-07

Minimal hypersurfaces are hypersurfaces that locally minimize area, while mean curvature flow evolves a hypersurface in the direction that decreases its area as fast as possible. This flow often converges to a minimal hypersurface over time. Both concepts—minimal hypersurfaces and mean curvature flow—play an important role in a wide range of scientific disciplines, including physics, materials science, and computer vision. The proposed project aims to deepen our understanding of the singularities that arise in these settings, which are central to current research in the field. In addition to advancing scientific knowledge, the project incorporates an appropriate educational component through teaching, mentoring, and the organization of seminars and conferences. This project has two main parts. The first part focuses on leveraging the theory of self-expanders for mean curvature flow to establish sharp lower bounds for the densities of minimal cones that are topologically nontrivial in specific senses. This work builds on and extends previous results by the PI and Bernstein to higher dimensions. The second part addresses a fundamental question in the study of mean curvature flow: the uniqueness of blow-up limits at singularities. It also explores the behavior of the flow as it emerges from a conical singularity. 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 2025 · 2025-07

This award supports the development of advanced computational methods for tracking and analyzing evolving patterns in large‑scale networks. Patterns of connections among entities, known as subgraphs, underpin insights in domains such as social interactions, biological processes, financial transactions, and communication systems. Real‑time analysis of how these patterns form and dissolve can enable early detection of disease outbreaks, improved understanding of social dynamics, and enhanced network security. By creating scalable and accessible tools for dynamic network analysis, this project will advance the national interest in data‑driven discovery across science, technology, and public welfare. The project will pursue three integrated research thrusts. First, it will develop novel algorithms with provable efficiency guarantees for counting and enumerating subgraphs in the batch‑dynamic model on parallel and distributed systems. Second, it will design and implement high‑level programming frameworks and data structures tailored to dynamic graph workloads, including graphics processing unit (GPU) and distributed implementations, to facilitate practical adoption. Third, it will integrate the new algorithms and frameworks into an open‑source analysis platform and conduct comprehensive evaluations on high‑performance computing clusters and cloud resources. These efforts will yield the first provably‑optimal dynamic subgraph counting algorithms for higher‑order patterns, query‑based enumeration techniques, and user‑friendly software enabling researchers to perform real‑time analysis on evolving networks. 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 2025 · 2025-06

PROJECT SUMMARY Understanding the neural basis of social behavior across development is critical, especially given its implications for a plethora of mental health conditions that result from severe disruptions in social relationships from infancy. Our preliminary studies suggest POMC neurons in the hypothalamus—the brain's source of the potent endogenous opioid, beta-endorphin—might be key players in social behavior during development. For instance, social interactions rapidly activate these neurons in juvenile mice (P24-30) and optogenetic activation of POMC neurons causes strong place preference, suggesting these neurons at this age have rewarding properties. These results—combined with the known molecular heterogeneity and developmental changes in POMC neurons— leads us to hypothesize that POMC neurons are differentially attuned to social behaviors during development to increase social reward; when social behaviors are heightened, as for example during the juvenile period, POMC neurons are recruited, modulating the response to social interactions. On this R21, our overarching goal is to elucidate how POMC neurons respond to social stimuli and regulate social behaviors during development with cellular and circuit resolution. In Aim 1, we will map the in vivo responses of POMC neurons to social cues with unparalleled single cell precision, leveraging novel procedures that our group has developed to perform in vivo two-photon microendoscopy in awake infant (P16-18), juvenile (P24-30) and adult (P65-80) mice. In Aim 2, first, we will record the activity of the POMC neuron projections in the nucleus accumbens to determine the dynamics of circuit-specific POMC neuron activity in relation to social interactions in infant, juvenile, and adult mice. Second, we will explore the effects of inhibiting the same projections on social learning. Overall, by elucidating how POMC neurons respond to social interactions and and influence social behaviors during a formative period of brain development, this research will provide foundational knowledge on how the developing brain regulates age-specific social behaviors. The findings from this research could pave the way for novel approaches to mitigate mental health conditions that stem from disruptions in social behaviors in early life. Moreover, the methodologies developed through this work promise to offer valuable tools for the broader neuroscientific community, enabling investigations into neurodevelopmental processes that play a role in physiological and behavioral processes with cellular and circuit resolution.

NSF Awards · FY 2025 · 2025-06

Combinatorial optimization (CO) problems are pervasive under the hood of modern life. CO problems underlie artificial intelligence, autonomous driving, logistics in healthcare/power grids/transportation, robotic maneuvering, wireless communications, error tolerant data storage, and many other societally important technologies. In recent years, new ways to solve these problems (using "analog oscillator" mechanisms) have emerged that promise far greater solution effectiveness than current techniques can achieve---if appropriate semiconductor "chip" implementations can be devised. The main goal of this project is to design, fabricate and demonstrate such chip implementations, along with systems that utilize them. Achieving this goal will lead to improved efficiencies solving a variety of societally important combinatorial optimization problems. Dissemination and training are also important components of this project. The specific scheme being investigated is called oscillator Ising machines (OIMs). OIM simulations have predicted high success rates solving various combinatorial optimization problems. However, integrated circuit (IC) implementations have had difficulty delivering such predicted levels of performance. In this project, the investigators will identify technological reasons for this discrepancy, and devise measures to address them. A key feature is an IC fabric that supports programmable interconnectivity between analog units. The impact of noise and device variability will be explored, as will specialized IC designs for different types of combinatorial optimization problems. Potential technology and performance benefits offered by novel nanodevices will also be explored. Evaluation metrics will include quality of solution, success rates, and power/energy required. 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 2025 · 2025-06

This REU Site award to Yale University, located in New Haven, CT, will support the training of 10 students for 10 weeks during the summers of 2025-2027. It is anticipated that a total of 30 students, primarily from schools with limited research opportunities, will be trained in the program. The objective of this program is to provide undergraduate students with hands-on interdisciplinary research experiences in systems and quantitative biology at a large research institution to prepare them to pursue careers in STEM. REU students will learn how to apply principles and techniques from the physical sciences and engineering to address important open questions in biology, building the critical thinking skills necessary to perform innovative research. Students will also develop their written and oral communication skills for both scientific and broad audiences, and many will present their results at scientific conferences. Assessment of this program will be done through an online Qualtrics survey and participants will be tracked after the program to monitor their career outcomes. Students should apply to the REU site using NSF ETAP (Education and Training Application: https://etap.nsf.gov) for the summers of 2026 and 2027. Characterization of complex biological systems requires an interdisciplinary approach that integrates methods and principles from the physical sciences and engineering with those from biology. Training in quantitative and systems biology allows researchers to investigate the systems-level impact of each biological perturbation at the molecular and cellular scales. This REU Site will introduce participants to three research areas: quantitative biological imaging, computational modeling of biological systems, and systems biology. Students will be mentored by faculty across 15 departments at Yale, including Physics, Biomedical Engineering, Molecular Biophysics and Biochemistry, Chemistry, Computational Biology and Biomedical Informatics, and Molecular, Cellular, and Developmental Biology. Students will conduct research in a laboratory that fits their interests and will receive supportive one-on-one mentoring by REU faculty, graduate students, and postdoctoral laboratory mentors. Through their research and other program activities, participants will deepen their understanding of interdisciplinary research in systems and quantitative biology and how computational and experimental approaches synergize to drive innovation in these fields. This site is supported by the Department of Defense in partnership with the NSF REU program. 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 2025 · 2025-06

Imagine being able to read the body's most complex signals and patterns inside cells, tissues, organs, and the brain—just as easily as reading text in a book. That is the goal of this research, which merges advanced artificial intelligence methods with cutting-edge mathematics to transform biological and medical studies. By adapting large language models—the same type of technology behind today's most powerful artificial intelligence tools—to biomedical data, this project seeks to unlock critical insights into how cells function, how diseases progress, how the brain operates, and how best to treat a range of health conditions. The research will develop new artificial intelligence methods that track continuous changes in biological systems over time and space, build foundation models that can handle multiple types of data such as brain scans, heart tests, and clinical notes, and create specialized tools for gene expression data to better understand how cells and tissues interact. These approaches could lead to more accurate diagnoses, more efficient drug development, personalized treatments, and a deeper knowledge of both cellular and brain processes—bringing healthcare closer to being more precise, more effective, and more attuned to each individual's unique biology. This project develops a new class of computational frameworks that unify large language models with operator learning techniques to address key challenges in modeling spatiotemporal phenomena in biomedical research. Neural operator learning generalizes deep neural networks from functions to operators, enabling flexible modeling of high-dimensional and continuous dynamical systems governed by integral equations or partial differential equations. By combining these operator-based approaches with the ability of large language models to interpret and generate symbolic or tokenized representations, the proposed methods capture both local and non-local interactions, manage complex boundary conditions, and accommodate the wide variety of scales and data types inherent in biology and medicine. The research extends operator learning to handle continuous dynamical systems characterized by memory effects and long-range dependencies, develops strategies to unify varied biomedical modalities—such as magnetic resonance imaging, electrocardiogram, and clinical metadata—in a text-based format, and builds new frameworks that transform genomic data into context-rich representations, enabling discovery of patterns in single-cell transcriptomics and multicellular interactions. These advances promise to significantly improve the predictive power and mechanistic interpretability of machine learning models in biomedical contexts, supporting breakthroughs in personalized medicine, disease modeling, and drug discovery. 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 · 2025-06

Project Summary/Abstract Biopolymers (proteins, DNA, and RNA) that are stable in solution often change conformations at the aqueous interfaces. Understanding this phenomenon will help elucidate many fundamental biological functions (e.g., membrane protein folding) and help design biomaterials (e.g., drug delivery systems). Aqueous interfaces are the boundaries between water and another medium—such as a cell membrane, biomolecular condensate, or mineral—where the water hydrogen-bonding network is terminated, resulting in asymmetric chemical environments. The key unanswered question is how the asymmetric chemical environments of interfaces modulate hydration and thus structures of the biopolymers. Addressing this question requires a physical method with surface selectivity to suppress signals from the bulk solution, as well as selectivity to distinguish biopolymer folding and isolate water signals from the hydration shells. Current methods are limited in providing such selectivity, hindering the fundamental understanding of biological functions at aqueous interfaces. Our recent progress showed that chiral-selective vibrational sum frequency generation spectroscopy (chiral SFG) can provide the necessary selectivity. It can detect protein and DNA secondary structures and probe their first hydration shell at aqueous interfaces. In this MIRA project, we will first investigate three fundamental questions about protein stability at interfaces: (1) Does the first hydration shell of proteins at interfaces undergo a phase transition melting process during protein melting? (2) How does molecular crowding impact water structures in the first hydration shell of proteins at interfaces? (3) How do denaturants and stabilizers perturb the first hydration shell of proteins at interfaces? Also, we will develop chiral SFG for probing higher- order structures of biopolymers. We will obtain homogeneous isoforms of the amyloid fibrils from Dr. Robert Tycko (NIH) and correlate their distinct molecular symmetry with chiral SFG responses of various vibrational modes. Finally, we will develop chiral SFG for characterizing small-molecule drug binding to DNA. We will detect displacement of water from the first hydration shell using minor groove binders, major groove binders, and intercalators, thus establishing chiral SFG signals of water as reporters for site-specific binding to DNA. In carrying out this MIRA project, we will collaborate with theorists, Profs. Sharon Hammes-Schiffer (Princeton) and Victor Batista (Yale), to simulate the chiral SFG spectra of molecular systems that mimic our experiments in order to build a theoretical basis for interpreting experimental data and advancing chiral SFG as a quantitative approach for elucidating biological function at interfaces at the fundamental level. Hence, this MIRA project will develop and apply chiral SFG for detecting the interplay of interfaces, chirality, and water in modulating structures of biopolymers to understand the principles, mechanisms, and processes taking place at aqueous interfaces in living organisms. Our findings will lay the foundation to develop new technologies and advance fundamental knowledge for solving problems in biomedical sciences.

NIH Research Projects · FY 2025 · 2025-06

Abstract Lysosomes degrade and recycle macromolecules, clearing harmful materials and providing nutrients, while also communicating nutrient availability to the cell. Although these functions are critical in all tissues, the aging central nervous system is particularly sensitive to lysosome disruptions. This sensitivity is evident in the wide range of neurodegenerative diseases caused by mutations in genes encoding lysosomal proteins. A significant challenge in developing new therapies from these genetic insights is the limited understanding of the fundamental cellular processes affected by disease genes and the consequences of perturbing these processes in specialized types of brain cells. Identifying pathways where disease-causing genes converge is thus crucial for developing therapies that target the cellular vulnerabilities underlying disease susceptibility and progression. Our research focuses on TANK-binding kinase 1 (TBK1) and C9orf72, two genes whose mutations cause amyotrophic lateral sclerosis and frontotemporal dementia (ALS-FTD). We have identified a novel pathway where TBK1 signals at the lysosome surface. Additionally, we have found that C9orf72 regulates this lysosomal pool of TBK1 and identified other activators of TBK1 at lysosomes, including nutrients, lysosome damage, and innate immunity signaling. We aim to elucidate the molecular mechanisms that integrate these signals to activate TBK1 at lysosomes. We also seek to understand the contributions of this signaling to lysosome functions in neurons and glia and to identify the targets of TBK1 kinase activity that mediate its effects on lysosomes. To address these issues, we will use genome-edited cell lines to establish the molecular mechanisms by which TBK1 coordinate lysosome function with cellular demand. These mechanistic studies will be complemented with experiments in genome-edited human induced pluripotent stem cells (iPSCs) and neurons and glia derived from them. This research will define a new pathway in which TBK1 helps match lysosome function to cellular demand, potentially revealing novel mechanisms for controlling lysosome function in ALS-FTD. Given the broader role of endolysosomal pathway dysfunction across multiple neurodegenerative diseases, our findings are likely to be relevant beyond diseases directly caused by mutations in C9orf72 and TBK1.

NIH Research Projects · FY 2025 · 2025-06

PROJECT SUMMARY: OVERALL YALE ADRC The Yale Alzheimer’s Disease Research Center (ADRC) seeks to advance our understanding of Alzheimer’s disease (AD) at a cell biological level with the eventual goal of translating laboratory discoveries into novel effective clinical therapies. Seven Cores (Administrative, Clinical, Data, Biomarker, Neuropathology, iPSC and Recruitment) and a Research Education Component will work together to achieve this goal. Our unifying theme is a focus on the disrupted cell biology of specific neural cells in AD, whether measured by proteome-wide methods in biofluids, imaged diagnostically, manifested in behavioral attributes, detected pathologically in brain tissue at autopsy, or observed in cultures of induced pluripotent stem cells. This focus facilitates an assessment of mechanistic variation across diverse populations. The participation of the seven Cores will accelerate and optimize the ability of individual Research Projects both within and beyond the Yale ADRC to ask and answer specific pathophysiological questions and to translate knowledge to therapies. The breadth of Core support for particular projects will allow assessments across the heterogeneity of AD. The Yale ADRC expects to extend its track record of facilitating Research Projects focused on specific organelles and specific neural cell subtypes perturbed in disease while making use of human tissue analysis and human subject imaging to evaluate mechanistic hypotheses. The Biomarker Core will engage a full range of updated imaging and fluid assays while developing novel, sensitive and high-throughput mass spectrometry methods, and new PET tracers coupled with functional MRI connectivity maps to monitor disease mechanisms. The iPSC Core will support human cellular studies defining the molecular mechanisms of specific endophenotypes of AD and their variation across populations. A key emphasis will be the translational development of research findings into therapeutic benefit. To support the future strength of Alzheimer’s research, the Yale ADRC will strive to advance the careers of a diverse cohort of Young Investigators through mentorship from a distinguished Internal Advisory Board, and through Development Project awards coupled with an extensive educational program developed by the Research Education Component. In addition to collecting clinical data and biospecimens of brain, CSF, blood and iPSCs for analysis by members of the Yale ADRC research team, the Yale ADRC will support other NIH-funded research studies on related topics and contribute materials to national NIA-sponsored research networks including NACC, NCRAD, SCAN and CLARITI. The Outreach Recruitment and Engagement Core will connect with the community to provide greater knowledge regarding AD and related dementia and facilitate recruitment of a diverse spectrum of participants for clinical studies.

NIH Research Projects · FY 2026 · 2025-06

Summary Alcohol related liver disease (ALD) and in particular alcohol associated hepatitis (AAH) is a leading cause of liver related deaths worldwide1. AAH involves metabolic alterations in hepatocytes and complex hepatocytes-stromal cell interactions. However, effective treatment options for ALD are very limited due to the lack of suitable in vivo models that recapitulate the full spectrum of ALD. Possible reason for that could be that the humans and rodents liver cells are significantly different in terms of hepatic lipoprotein2, bile acid 3 and in alcohol metabolism rates4. In human AAH, there is severe steatosis, hepatocyte apoptosis, Mallory-Denk hyaline inclusions, cholestasis and peripheral lymphopenia. On the other hand, mouse models can't develop cholestasis with pure alcohol diets and for high degree of steatosis long term of alcohol feeding is required 5. We have already developed a humanized murine system in which mice have been humanized at key loci by knock-in of five human genes (MISTRG6 mouse) and have been further humanized at a cellular level by engraftment of human adult hepatocytes and human CD34+ cord blood (CB) cells. These mice can support human hepatocytes, as well as human immune, endothelial and stellate cell populations (so called human Non-Parenchymal cells, hNPCs) derived from human CD34+ CB cells. Using these mice, we developed an alcoholic model after 10days of alcohol feeding plus one binge of ethanol capturing key features of human disease pathology (severe steatosis, ballooning, Mallory-Denk inclusions, hepatocyte apoptosis, inflammation, mild fibrosis, cholestasis, peripheral lymphopenia). Using this model, we found that when human hepatocytes are engrafted alone and thereby are surrounded only by mouse NPCs disease features in human hepatocytes (steatosis, Mallory denk-bodies, cholestasis) are significantly blocked upon alcohol. These results indicate a species-specific paracrine regulation of hepatocyte alcohol/lipid metabolism and possibly bile acid synthesis/transport that is related to cholestasis. By employing bulk and single-cell RNA sequencing we found NPC ligands-hepatocyte receptors that may be responsible for this phenotype. Our aims are to examine the role of these ligand-receptors interaction in vitro and in vivo by gain and loss of function approaches. In the Aim1 we will examine NPC derived WNT4, WNT10B with hepatocyte FZD6 responsible for the defects in antioxidant defense and bile acid synthesis/transport in hepatocytes. In the Aim2 we will examine NPC derived WNT2, WNT5A with hepatocyte FZD5 responsible for alcohol induced steatosis and mitochondrial dysfunction through alteration in liver cholesterol metabolic fate. In the Aim3 we will examine the role of monocyte/macrophage subsets and their effect on hepatocyte steatosis, cholestasis as well as the role of the monocyte derived Sema3c on hepatocyte Nrp1 for alcohol related steatosis. Our study may reveal important regulators of steatosis and cholestasis in AAH and thereby is in accordance the mission and the scope of NIAAA.

NSF Awards · FY 2025 · 2025-06

With the support of the Chemical Synthesis (SYN) program in the Division of Chemistry Professor Timothy Newhouse of Yale University is studying the development of synthetic routes to diterpenoids. The synthesis of small molecules is one of the rate-limiting steps across disciplines from materials science to medicinal chemistry. To overcome this bottleneck in the discovery process, we need methodological advancements and improvements to the synthesis design process. The long-term goal of this proposal is to apply computational strategies to synthetic planning to access structurally complex natural products, and in this proposal these efforts are focused on synthesis of diterpenoids. The approach to model development described in this proposal can be applied to any synthetic transformation and would be enabling and thus broadly impactful whenever that transformation’s short-term experimental evaluation is not possible. Moreover, strategic partnerships within and around the Yale community will bring science to K-12 audiences. The design of a synthetic pathway to a desired molecule is generally conducted by human analysis although computational approaches are beginning to emerge. This proposal outlines the development of several artificial intelligence-based tools to predict the yield of common carbon-carbon bond forming cyclization reactions. Additionally, generative modelling is proposed to design ligands, substrates, and routes to natural products. These computational methods will enable the planning and synthesis of natural products and analogs. High-risk yet high-reward plans are de-risked through the use of machine learning models, thereby allowing efficient and expedient laboratory access to synthetically challenging molecular scaffolds. 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 2025 · 2025-06

SUMMARY We have characterized Cysteine String Protein α (CSPα), a resident synaptic vesicle (SV) co-chaperone whose activity is essential for presynaptic proteostasis. CSPα chaperones key presynaptic proteins and deletion of CSPα triggers the loss of SVs, synapse loss, neurodegeneration, and early lethality. Significantly, mutations in CSPα causes cause a neurodegenerative disease with presenile dementia (CLN4) and implicated in ADRDs. We discovered that CSPα knockout (KO) brains exhibit increased number of presynaptic autophagosomes that contain SVs as cargo, suggesting that CSPα regulates SV turnover through autophagy. Our single-nucleus RNA- seq on wildtype (WT) and CSPα KO cortex also show elevated autophagy gene expression in neurons. Therefore, we aim to investigate the links between SV degradation, autophagy, and neurodegeneration. Based on these compelling data, we will test the hypothesis that CSPα is important for maintaining the composition of SVs and deletion or CLN4 mutations of CSPα result in SV turnover through autophagy. The objective of this application is to elucidate the mechanisms through which CSPα regulates synaptic autophagy and neurodegeneration. We will use electron microscopy (EM), quantitative mass spectrometry, live imaging Keima assays, western blotting, and immunocytochemistry to test our hypothesis using WT, CSPα KO, CLN4 models. In Aim 1, we will characterize SVs, SVs targeted for autophagy, and synaptic autophagosomes purified from WT and CSPα KO brains by mass spectrometry and EM. In Aim 2, we will investigate the relationship between the SV cycle and autophagosome biogenesis by tagging SV proteins with the Keima autophagy reporter in WT, CSPα KO, and CLN4 cultures. We will also perform EM to monitor distribution of autophagosomes in CLN4 cultures. We expect that realization of this proposal will have major impacts on our knowledge of presynaptic autophagy, SV turnover, and CLN4. Achievement of our aims is likely to advance therapeutic interventions for CLN4 and ADRDs, given the genetic and functional links of CSPα to these neurodegenerative diseases.

NSF Awards · FY 2025 · 2025-06

Algorithms are the building blocks of modern computation that underlies much of the technology on which our society depends. In particular, the successes of modern machine learning systems rely on our ability to solve large-scale computational tasks via efficient algorithms, such as for solving optimization problems for training neural networks, and drawing random samples from probability distributions. Algorithms in practice necessarily operate via discrete iterations, by executing one operation after another. Many algorithms are inspired by the mathematics of dynamical systems, where time flows continuously, but all such derivations have been done manually, often with sophisticated and difficult specialized analysis. A general theory of how to derive algorithms from dynamics that can preserve the desired properties is missing. This project aims to close this gap by developing a rigorous way to systematically translate continuous-time dynamics into discrete-time algorithms that can automatically preserve convergence guarantees, focusing on algorithms for fundamental problems in optimization and sampling motivated by machine learning applications. The outcome of this project will benefit many fields in which computational tasks such as optimization and sampling are used, by providing researchers and practitioners a higher-level language involving dynamical systems to design and reason about algorithms. This project also provides training opportunities for students, outreach events to popularize the dynamical perspective of algorithm design, and a collaborative effort to compile a catalog of dynamics and algorithms. This project will strengthen the matching parallel structures between dynamics and algorithms for optimization and sampling, by studying two complementary directions. In the first direction, this project studies sampling problems as optimization problems on the space of probability distributions. Many recent results in sampling have been derived from this perspective, in particular for the basic greedy methods analogous to gradient descent. This project aims to complete the translation from optimization to sampling by deriving algorithmic results for various sampling problems as inspired by the optimization results, including for constrained sampling, accelerated sampling, and sampling with stronger mixing time guarantees. This project will combine ideas and tools from dynamical systems, geometry, probability, and information theory to extract concrete algorithmic principles to design new algorithms in practice. In the second direction, this project aims to formalize the robust parallel structures between dynamics and algorithms with matching convergence guarantees. This project will study the mechanics of translating dynamics to algorithms, and investigate why certain ideas are successful in the translation, including the idea of speeding up time to achieve faster convergence rates, the optimality of accelerated methods for optimization and sampling, and the geometric structure of the space of algorithms and dynamics that allows such robust parallel structures to be exhibited. 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 · 2025-06

Myocardial infarction (MI) remains a leading cause of morbidity and death in the Western world and often leads to left ventricular remodeling and progressive heart failure with reduced ejection fraction (HFrEF). The local intramyocardial delivery of biomaterials to the MI region and/or the peri-infarct border zone has been shown to reduce post-MI remodeling in both small and large animal models. These injectable biomaterials can be used to release incorporated payload of small molecule therapeutics. However, while these approaches are promising, the delivery strategies and basic understand of these injectable biomaterials in terms of biomechanical properties in the absence of a therapeutic payload remains poorly understood. In our prior work we have demonstrated in porcine models of reperfused and non-reperfused MI that local delivery of hydrogels by minimally invasive surgical approaches to the MI region with and without the local release of inhibitors of matrix metalloproteinases (MMPs) stabilizes hemodynamics, reduces wall stress and results in sustained improvement in regional and global function reducing post-MI remodeling in association with increases in integrin activation a marker of angiogenesis, and decreased activation of MMPs. We have employed multi- isotope hybrid SPECT/CT imaging to non-invasively track the changes in these molecular process in relation to changes in regional mechanics and their therapeutic benefit. Other studies have demonstrated that the degradation kinetics and stiffness of these biomaterials may influence the therapeutic outcomes after MI. We will build upon this past work by development of a percutaneous transthoracic approach for multi-modality image guided intramyocardial delivery of hydrogel of variable stiffness that alter regional mechanics by modulation of fibroblast transformation both locally and remotely reducing post-MI remodeling. We hypothesize that early hydrogel-induced changes in mechanical stress and strain will result in fibroblast transdifferentiation and structural remodeling that can be tracked with targeted molecular imaging of fibroblast activation protein (FAP) along with serial echocardiography and computed tomography. We will employ hybrid SPECT/CT imaging to noninvasively track markers of inflammation, and fibroblast activation with novel SPECT probes that target FAP, interrogating critical pathways involved in mechano-transduction and post-MI remodeling. To address this mechanistic hypothesis and evaluate therapeutic effects of image-guided delivery of therapeutic hydrogel we propose two specific aims. Aim 1 will develop a multimodality imaging approach to guide optimal transthoracic epicardial intramyocardial delivery of therapeutic hydrogels based on early hyperintensity of the MI region on non-contrast CT images for prevention of remodeling early post-MI, using a porcine model of reperfused MI, taking advantage of our integrated translational ultrasound fluoroscopy interventional suite. Aim 2 will apply multi-isotope hybrid SPECT/CT imaging to evaluate early mechanical and molecular mechanisms underlying therapeutic efficacy of soft and stiffening hydrogels and associated longer term therapeutic benefit.

NIH Research Projects · FY 2025 · 2025-06

Abstract Substance use disorders (SUDs) are leading causes of death and disability worldwide. A better understanding of the genetic mechanisms underlying SUD risk could improve their diagnosis, prevention, and treatment, ultimately reducing the costly and disabling addiction-related problems. Recently, substantial progress has been made in identifying genetic variants associated with SUDs through gene discovery efforts, particularly genome-wide association studies (GWAS) in large cohorts. Although these findings considerably advance our knowledge of SUD genetics, key questions remain unanswered. Current GWAS studies predominantly rely on SNP arrays and post hoc imputation to identify common variants, leaving certain genomic regions unexamined due to technical limitations. Whole-genome sequencing (WGS), which can detect both common and rare variants in large cohorts, offers the potential to recover much of the missing heritability. However, to our knowledge, no large WGS study of SUDs has been conducted to date. Another significant gap is the estimated SNP-based heritability of SUDs is low, and as a result, polygenic risk scores (PRS) based on common variants have limited predictive power in independent cohorts. While statistically significant, these PRS are numerically weak and not yet clinically useful. To address these gaps and respond to the Cutting-Edge Basic Research Awards, we propose an innovative study using publicly available WGS data on SUDs from biobanks such as the All of Us and UK Biobank. This project aims to leverage large-scale WGS datasets to identify novel rare and common variants associated with SUDs and recover missing heritability (Aim 1). Additionally, the study will enhance disease prediction through the development of a novel whole-genome multi-ancestry PRS framework (Aim 2). The proposed research is both timely and aligned with the National Institute on Drug Abuse's mission. It will substantially enhance our understanding of the genetic architecture of SUDs and lay the groundwork for the development of precision interventions to prevent and treat alcohol-related diseases.

NIH Research Projects · FY 2026 · 2025-06

ABSTRACT Type 2 diabetes (T2D) is associated with insulin insufficiency. Increased glucagon secretion from pancreatic α-cells exacerbates hyperglycemia by increasing endogenous glucose production. However, a model in which hyperglucagonemia is completely maladaptive does not align well with recent observations indicating that α-cells boost the function of neighboring β-cells. These studies, which identify an insulinotropic role of glucagon under physiological conditions, raise the possibility that increased glucagon secretion mediates an adaptive response to the loss of functional β-cell mass in T2D. Consistent with this model, recent human studies suggest that α-cell signaling via β-cell G-protein coupled receptors (GPCRs) is accentuated in T2D—however these correlative studies do not test causality. To investigate how the loss of functional β-cell mass impacts islet cell crosstalk and glycemic control, we generated a novel strain of β-cell replication-deficient mice (hereafter, abbreviated ‘βRD’). Studies performed in vivo, and in isolated islets, suggest that increased glucagon release from α-cells is essential for the preservation of insulin secretion and organismal glycemia in βRD mice. Single-cell RNA sequencing (scRNAseq) of βRD islet cells indicates α-cell intrinsic reprogramming of metabolism and hormone secretion. Based on these findings, our central hypothesis is that α-cell hypersecretion is essential to preserve insulin secretion in response to β-cell mass deficiency. Testable predictions of this model include: increased glucagon secretion in βRD mice is insulinotropic; signaling through β-cell GLP1R and/or GCGR is essential for the observed β-cell adaptations in βRD mice; and that similar mechanisms occur in human islets with reduced β-cell mass. We will: (1) Determine the signaling mechanisms underlying β-cell compensation in a novel mouse model of β-cell replication deficiency, (2) Determine mechanisms of α-cell compensation in β-cell mass-deficient mouse islets, and finally, by transplanting human pseudoislets with controlled α:β ratios, we will test whether increased α-cell signaling can compensate for β-cell mass deficiency in vivo. Completion of these aims will test the concept of adaptive hyperglucagonemia in T2D. Whether or not this hypothesis proves correct, these studies will identify islet mechanisms of functional compensation for β-cell mass deficiency, elucidate α-cell signals that functionally rescue insulin secretion, and identify GPCR pathways for therapeutic targeting in diabetes.