University of California, Merced

Canonical inflammasome assembly is driven by interactions between sensors and effector procaspase-1, primarily mediated by the adaptor ASC. Homotypic interactions between Death Domains pyrin (PYD) and caspase recruitment and activation domain (CARD) lead to sensor and ASC oligomerization, along with the recruitment of procaspase-1. ASC self-association is essential for inflammasome activation, which initiates inflammatory responses. Therefore, uncontrolled inflammasome activation contributes to chronic inflammatory diseases. ASC-c, an isoform of ASC, acts as a negative regulator of the inflammasome. To better understand ASC-c's regulatory role, we examined its structural properties and interactions with ASC. Our nuclear magnetic resonance data show that ASC-c's CARD is properly folded, whereas the PYD consists of two α-helices instead of the six-helix bundle typical of Death Domains. In addition, a chemical shift perturbation analysis indicates that ASC-c interacts with ASC. We obtained transmission electron micrographs revealing that ASC-c polymerizes into filaments and filament bundles, which display greater heterogeneity than those of ASC based on dynamic light scattering. Overall, our results suggest that ASC-c binding to ASC can have an impact on ASC self-association, thereby affecting inflammasome assembly. Based on these findings, we designed a peptide encompassing the two helices of the ASC-c PYD to target ASC for therapeutic purposes. We demonstrate the interaction between the peptide and ASC by fluorescence anisotropy and show the stability of the complex by molecular dynamics simulations. Finally, cell-based assays measuring inflammasome activation indicate an inhibitory effect of the ASC-c peptide, pointing to its potential use in drug design.

Mexican American adolescents in the U.S. face disproportionately high rates of obesity. Familism, a core Mexican cultural value emphasizing family closeness, and parenting processes such as communication and monitoring are associated with adolescents’ weight-related behaviors such as diet, physical activity, sleep, screen time and weight outcomes. Participants were 439 mother-adolescent dyads (47.2% boys) from the CHAMACOS cohort, a longitudinal study of Mexican-origin families in California’s Salinas Valley. Data were collected at the 14-year study visit. Adolescents reported on familism, maternal communication and monitoring, and their own weight-related behaviors. Anthropometric data were collected to calculate BMI percentiles. Structural equation modeling was used to test hypothesized relationships among variables and examine gender differences. Familism was positively associated with maternal communication and monitoring, as well as with adolescent physical activity and sleep, and negatively with screen time. Mother-youth communication was related to longer sleep duration, while maternal monitoring was not significantly related to any weight-related behaviors. Fruit and vegetable intake and sleep were negatively associated with weight status. Gender differences were observed: for girls, mother-daughter communication was linked to longer sleep duration, and for boys, familism was related to higher physical activity. The results highlight significant associations between cultural values, maternal parenting, and healthy weight-related behaviors among Latino adolescents. These findings emphasize that incorporating gender-sensitive cultural and family dynamics may enhance the design of obesity prevention strategies for this population.

The rapid advancement of large language models (LLMs) has intensified the need for effective mechanisms to transform continuous multimodal data into discrete representations suitable for language-based processing. Discrete tokenization, with vector quantization (VQ) as a central approach, offers both computational efficiency and compatibility with LLM architectures. Despite its growing importance, there is a lack of a comprehensive survey that systematically examines VQ techniques in the context of LLM-based systems. This work fills this gap by presenting the first structured taxonomy and analysis of discrete tokenization methods designed for LLMs. We categorize 8 representative VQ variants that span classical and modern paradigms and analyze their algorithmic principles, training dynamics, and integration challenges with LLM pipelines. Beyond algorithm-level investigation, we discuss existing research in terms of classical applications without LLMs, LLM-based single-modality systems, and LLM-based multimodal systems, highlighting how quantization strategies influence alignment, reasoning, and generation performance. In addition, we identify key challenges including codebook collapse, unstable gradient estimation, and modality-specific encoding constraints. Finally, we discuss emerging research directions such as dynamic and task-adaptive quantization, unified tokenization frameworks, and biologically inspired codebook learning. This survey bridges the gap between traditional vector quantization and modern LLM applications, serving as a foundational reference for the development of efficient and generalizable multimodal systems. A continuously updated version is available at: https://github.com/jindongli-Ai/LLM-Discrete-Tokenization-Survey.

BackgroundDespite the availability of effective vaccines, flu and COVID-19 uptake remains suboptimal, including among Indigenous communities in California who face unique barriers to accessing public health information. While previous research has evaluated health communication message content and design, fewer studies have systematically compared different dissemination strategies for the same intervention, leaving gaps in understanding optimal approaches for reaching marginalized populations.ObjectiveThis protocol describes a three-arm dissemination study designed to compare the effectiveness of different strategies for distributing COVID-19 and flu vaccine promotion videos among Indigenous Peoples residing in California, including American Indian and Alaska Natives, and migrant Indigenous Peoples from Latin America.MethodsFollowing extensive formative work, including cross-sectional surveys, social network analyses, a discrete choice experiment, and qualitative focus groups conducted with guidance from an Indigenous Community Advisory Board, we developed four 30-second vaccine promotion videos available in English and Spanish, two targeting American Indian and Alaska Native communities and 2 targeting Latin American communities. Videos will be translated to four Indigenous languages (Purépecha, Mam, Zapoteco, and Mixteco). The study compared three dissemination strategies implemented over a two-month period: (1) targeted social media advertisements on Facebook and Instagram, (2) distribution through community-based organizations using their established communication channels, and (3) peer-to-peer sharing through Indigenous community members (“seeds”) recruited from previous research. Primary outcomes focus on dissemination effectiveness, defined by measures of reach (eg, link shares, clicks and video views), engagement (eg, watch time and survey completion), and message impact (eg, trust in the message and vaccination intent), assessed using Bitly link tracking, YouTube analytics, and follow-up surveys. Descriptive and comparative analyses will be conducted to assess differences across dissemination strategies.ResultsThe study was funded in January 2023. Intervention development was completed in July 2025. Dissemination activities and data collection across all three study arms were conducted between October 6, 2025, to November 28, 2025, aligned with the COVID-19 and influenza vaccination season in the United States. Data consolidation, cleaning, and analyses began January 2026 and primary results are expected to be submitted for publication in Spring 2026.ConclusionsThis study addresses a critical gap in health communication research by providing a systematic methodology for comparing digital dissemination strategies within Indigenous communities. The combination of community-informed recruitment, multilingual accessibility, and comprehensive digital tracking tools offers a replicable model for evaluating how public health messages spread through different channels, particularly for populations historically excluded from traditional outreach efforts.

Extreme fire weather (hot, dry, and windy conditions) has intensified globally, yet formally attributing this trend to anthropogenic climate change remains challenging. Here, we analyze global trends in extreme fire weather days (FWI95d, annual count of days with Fire Weather Index above the 95th percentile) over 1980–2023, using climate model ensembles, observational data, and fingerprint detection techniques. We find that the observed increase in extreme fire weather bears a clear externally forced signal, detectable at 99% confidence above natural variability and attributable to human-induced climate change. This emerging human-induced fingerprint on extreme fire weather highlights a benchmark for climate science and underscores the urgency of integrating these insights into wildfire risk management and adaptation strategies.

Pervasive evidence supports consistent associations between trait self-control and health behavior, yet few studies have tested the underlying mechanisms. We leveraged meta-analytically synthesized data from the extant research to test a mediational model specifying indirect effects of trait self-control on health behavior mediated by social cognition theory constructs and intention as a candidate mechanism. We also tested the effects of moderators on model effects: behavior control process (impulsive vs. reflective), behavior type, trait self-control and behavior measure type, and measurement lag. Studies reporting relations among trait self-control, social cognition constructs (attitude, subjective norm, and perceived behavioral control), intention, and health behavior were identified in a systematic database search. Our proposed model was fit to the pooled correlation matrix among these variables using multilevel meta-analytic structural equation modeling. Findings supported model predictions including indirect effects of trait self-control on health behavior directed through attitude, perceived behavioral control, and intentions, and a residual direct effect indicating partial mediation. Moderator analysis revealed larger direct and indirect trait self-control-health behavior effects in studies targeting behaviors under impulsive process control relative to those under reflective control. Analyses of behavior type, behavior and self-control measure type, and measurement lag moderators did not reveal systematic variation in model effects. Findings provide preliminary support for the model signaling a candidate mechanism for trait self-control-health behavior relations and providing prior effect size and variability estimates that could be expected in future studies. Impulse-related behaviors represent a key moderating condition that magnifies trait self-control-behavior effects. (PsycInfo Database Record (c) 2026 APA, all rights reserved).