The rapid expansion of biomedical literature demands automated summarization tools that can reliably condense research articles into concise, accurate summaries. We benchmarked 62 text summarization methods, ranging from frequency-based and TextRank extractors to encoder-decoder models (EDMs) and large language models (LLMs), on 1,000 biomedical abstracts with author-generated highlights as reference summaries. Models were evaluated using a composite suite of lexical, semantic, and factual metrics, including ROUGE, BLEU, METEOR, embedding-based similarity, and factuality scores. Our results indicate that general-purpose language models (LMs) achieve the highest overall performance across lexical and semantic dimensions, outperforming both reasoning-oriented and domain-specific models. Notably, medium-sized models often outperform frontier-scale counterparts, suggesting an optimal balance between model capacity and computational efficiency. Statistical extractive methods consistently lag behind neural approaches. These findings provide a systematic reference for selecting biomedical summarization tools and highlight that broad pretraining remains more effective than narrow domain adaptation for generating high-quality scientific summaries.
Modern biomedical imaging workflows generate large volumes of derived images and short videos that must be reviewed, compared, curated, and reused following primary acquisition and analysis. In practice, these assets are often dispersed across nested filesystem hierarchies on local drives, external media, or network storage, limiting efficient retrieval, deduplication, and figure assembly. We present PixelDeck, an open-source, local-first browser application for organizing and interactively browsing large biomedical image and video libraries on commodity workstations. PixelDeck integrates recursive folder import, SHA-256-based duplicate detection, metadata extraction, thumbnail and preview generation, full-text search, and asynchronous export within a responsive interface, supported by a modular ingestion pipeline, managed storage layer, and interactive browsing environment optimized for high-volume media collections. The system is implemented using a Next.js and React frontend, a SQLite metadata store accessed via Prisma, managed local media storage, and a background worker that executes import and export tasks asynchronously, enabling scalable processing on standard hardware. To evaluate performance, we conducted structured benchmark imports using public histopathology images curated from PanopTILs, SICAPv2, and PanNuke datasets, where dataset-specific import behavior, duplicate detection, and ingestion metrics were recorded as reproducible outputs. Embedding-based analysis further demonstrates dataset-level separation consistent with underlying image characteristics. These results show that PixelDeck provides an efficient, scalable local curation layer for heterogeneous biomedical imaging collections, enabling streamlined dataset exploration and preparation for downstream analysis.
Fluid mixing technologies are indispensable in the fields of biomedicine and biotechnology, where precise control over mixing processes is essential for optimal performance. Traditional mixing systems, such as stirred tanks, jet-flow mixers, and mechanical drives, often suffer from limitations in terms of precision, scalability, and energy efficiency, especially when adapted to complex biomedical applications. Recent developments in micro/nanostructured fluid mixing technologies offer promising solutions to these challenges by enabling efficient, controllable, and energy-saving mixing at the microscale. These technologies leverage intricate micro/nanostructures that can enhance diffusion, induce vortices, and generate chaotic convection, thus overcoming the constraints faced by conventional methods. In this paper, we systematically classify these micro/nanostructures based on the fluid mixing phenomena they exploit and review their applications across a range of biomedical fields, including biomaterials fabrication, drug development, cell culture, organs-on-chips, biosensing, and diagnostics. We also examine how these technologies contribute to advancements in precision medicine, personalized treatment, and high-throughput testing. Finally, we discuss the future challenges, opportunities, and interdisciplinary collaboration needed to further advance the clinical and industrial adoption of micro/nanostructured fluid mixing technologies.
High-quality metadata is essential for scientific discovery, yet sparse annotations in rapidly growing repositories leave many biologically relevant details uncaptured. We evaluated whether large language models (LLMs) can accurately infer ion channel and receptor subtype metadata from source code in a neuroscience repository. We extracted 5,133 model files from ModelDB. A subset of 1,100 was manually annotated; 253 were held out for testing, and the remainder split into training (80%) and validation (20%) sets. LLM-based approaches (GPT-5.2 and GPT-mini) were evaluated under zero-shot and heuristic-augmented prompting. Performance was assessed at type and subtype levels using accuracy, precision, recall, and F1 score. A feature-engineered XGBoost model using text- and simulation-derived features served as a baseline. LLMs outperformed the XGBoost baseline. At the type level, GPT-mini with heuristic augmentation achieved the highest performance (accuracy 96.0%, F1 0.962). At the subtype level, both GPT-5.2+heuristics and GPT-mini+heuristics achieved identical accuracy (88.1%), with GPT-5.2+heuristics achieving the highest F1(0.878). Model outputs were consistent across runs and errors confined to related mechanistic families. LLMs demonstrate strong potential for metadata annotation directly from source code, outperforming feature-engineering approaches with minimal tuning. However, performance varied across subtypes, and errors often reflected ambiguity or bias toward more common labels. These findings suggest LLMs may serve as practical tools for scalable metadata generation in biomedical repositories, although careful evaluation and domain-specific validation remain important. While demonstrated in computational neuroscience, this approach may generalize to repository-agnostic metadata annotation in other scientific code repositories.
MXenes are among the most extensively studied materials nowadays due to their functional versatility stemming from their tunable chemical and physical properties. MXenes have been predominantly synthesized by selective wet-chemical etching of parent MAX phases, followed by Li+ intercalation and subsequent delamination. This study demonstrates the substitution of Li+ with Na+ in the preparation of Ti-based MXenes for biomedical and biocatalytic applications, where biologically active Li+ is undesirable. Here, a MILD (Minimally Intensive Layer Delamination) synthesis method of Ti3C2T x and Ti3CNT x is modified by replacing LiF with nontoxic and cost-effective NaF. The produced samples had flake sizes and surface chemistries comparable to those of LiF-MILD samples. The electrical conductivity of Ti3C2T x films made from those flakes exceeded 5500 S/cm. Multiple acid mixtures were investigated, with 12 M HCl producing stable MXene colloids after 48 h of etching without sonication, yielding flakes significantly larger than those obtained using 9 M HCl. The Ti3C2T x flakes exhibited a conventional 2D morphology, while Ti3CNT x scrolled, forming cylindrical nanostructures. With the proper adjustments to the etching conditions, the proposed approach may apply to the synthesis of other Ti-based MXenes.
Tissue engineering requires biomedical devices that stabilize wounds, then degrade as tissue regenerates. However, published material degradation rates are often conflicting. Incorporating monitoring functionality into implanted devices allows real-time assessment of degradation and failure, but requires contrast agents, as most biomedical devices are composed of polymeric materials invisible to medical imaging modalities. Computed tomography (CT)-visible radiopaque composites were created from 5-20wt% tantalum oxide (TaOx) nanoparticles in polymers with distinct degradation profiles: polycaprolactone (PCL), poly(lactide-co-glycolide) (PLGA) 85:15 and PLGA 50:50, representing slow, medium and fast degrading materials respectively. Radiopaque phantoms, mimicking biomedical devices, were implanted into mice intramuscularly or intraperitoneally, and monitored via CT over 20 weeks. Changes in phantom volume, including collapse and swelling, were visualized. Phantom degradation profile was dictated by polymer matrix, regardless of nanoparticle addition. Foreign body response was dependent on implant site and degradation kinetics were significantly affected in mid-degrading materials, transitioning from linear degradation intramuscularly to exponential degradation intraperitoneally, due to differences in inflammatory responses and fluid flow. Nanoparticle excretion via liver and spleen lagged polymer degradation, requiring modulation of nanoparticle clearance. Tracking real time device behavior advances the new era of personalized medicine, allowing individualized treatment plans to biomedical devices.
Categorical variants, or sets of genomic alterations constrained by shared properties, are pervasive across clinical, regulatory, and research domains in the biomedical ecosystem, yet their inconsistent and non-computable representation hinders data interoperability and clinical interpretation. We surveyed genomic knowledgebases spanning regulatory approvals and the biomedical literature and found that categorical variants underpin a substantial proportion of clinical genomics knowledge, but are largely described using incompatible bespoke models. To address this, we developed the GA4GH Categorical Variation Representation Specification (Cat-VRS), a constraint-based framework that provides a unified computable representation for both precise and intentionally broad categories across molecular and systemic variant domains. Cat-VRS enables harmonization of genomic knowledgebases, computable category-based search, and automated matching between assayed variants and categorical entities in clinical and research contexts. By providing a principled, extensible model for categorical variation, Cat-VRS enables computable reasoning over genomic variant categories and establishes a foundation for the standardized representation and exchange of genomic knowledge.
The ability to induce tissue regeneration on demand using biomaterials remains a major goal in biomedical research, yet significant challenges persist. Among the most advanced biomaterial models, the nanofiber-hydrogel composite has demonstrated a striking ability to induce soft adipose tissue remodeling at the injection site without incorporating exogenous biological cues. 1,2 However, the underlying mechanisms that drive such a tissue response remain unclear. Here, we show that biomaterial-induced tissue remodeling is driven by sustained and controlled inflammation mediated by macrophages in strong communication with fibroblasts. Notably, both pro-inflammatory and anti-inflammatory signals remained elevated during this process in the long-term, challenging the prevailing notion that inflammation opposes remodeling. Using macrophage depletion in mice, we demonstrate that macrophages are essential for this process. Single-cell RNA sequencing further revealed robust fibroblast-to-macrophage signaling, contrasting with the conventional macrophage-to-fibroblast paradigm, and identified unique Spp1 ⁺ macrophages and Ctla2a ⁺ fibroblasts within the remodeling niche. These findings provide a comprehensive view of the immune landscape in biomaterial-induced tissue remodeling, highlighting key cellular interactions, prolonged kinetics, and unexpected signaling pathways. By defining key targets and fundamental principles, this work has broad implications for advancing biomaterial-induced tissue regeneration.
Drug-induced interstitial lung disease (DI-ILD) encompasses a heterogeneous spectrum of potentially severe pulmonary toxicities associated with various pharmacological agents. Diagnosis is often delayed due to nonspecific symptoms and imaging findings that can mimic other ILDs or infections, particularly in patients receiving oncological or immunomodulatory therapies. Novel approaches are needed to improve early recognition and management. We conducted a non-systematic, narrative literature review across major biomedical databases up to June 2025, aimed at describing current and emerging applications of artificial intelligence (AI) and radiomics in the early diagnosis, risk stratification, and personalised management of DI-ILD. Radiomics applied to computed tomography and positron emission tomography/computed tomography enables extraction of high-dimensional quantitative features capturing subclinical alterations undetectable by visual assessment. In DI-ILD (particularly immune checkpoint inhibitor-related pneumonitis), radiomic models show potential diagnostic utility in distinguishing overlapping imaging patterns and predicting fibrotic progression. Integration with clinical, radiological and pharmacogenomic data has improved model performance in several studies. Additionally, AI approaches, including convolutional neural networks and ensemble learning methods, demonstrate promise in enhancing pattern recognition and risk stratification. Radiomics and AI are emerging complementary tools in multidisciplinary management of DI-ILD, offering objective imaging biomarkers and facilitating multimodal data integration to improve diagnostic precision and personalised therapeutic decisions. Nonetheless, reproducibility remains limited by variability in imaging protocols and lack of large-scale prospective multicentre validation. Clinical implementation requires standardised protocols to ensure consistency and reliability. The development of transparent, interpretable models seamlessly integrated into healthcare workflows is essential to fully leverage their real-world potential in proactive patient care.
The Chinese translation of "carbohydrate" has long been a topic of considerable debate in chemistry, biomedicine, and nutrition-related disciplines. This issue is not merely linguistic. In Chinese-language contexts, inconsistency among carbohydrate-related expressions may create ambiguity in nutrition education and public understanding, and may introduce practical challenges for literature retrieval and interdisciplinary collaboration, especially in fields such as type 2 diabetes mellitus (T2DM), where distinctions among dietary carbohydrates, sugars, and glucose could be crucial. This article traces the historical evolution of the Chinese translation of "carbohydrate" to clarify its historical trajectory and scientific implications. Historical evidence demonstrates that the term "carbohydrate" did not appear in dictionaries or chemistry books published prior to 1900. However, at the turn of the 20th century, multiple translations emerged, most of which were influenced by the Japanese term "tansuikabutsu/." The earliest recorded Chinese translation appeared in Huaxue Yuanliu Lun. During the early Republic of China, "tanshui huawu/" became the most commonly used term, which was later revised around 1920 with the addition of a semantic radical to the character "tan." In 1932, the National Institute for Compilation and Translation introduced the term "tang/," which gained popularity alongside "tanshui hua(he) wu." However, "tang" was officially abolished in the mid-to-late 1950s and gradually phased out in subsequent decades. By 1980, "tanshui huahe wu/)" and "tang lei/" were officially established as equivalent translations. Currently, "tang lei" is preferred in some disciplinary standards, although "tanshui huahe wu" remains widely used by convention. By reviewing this history, the present work highlights three key principles for addressing terminological ambiguity in nutrition communication. While this historical narrative is anchored in the Chinese context, the communication risks and mitigation strategies discussed might be relevant to other cross-lingual or cross-disciplinary setting, where everyday dietary language interfaces with technical biomedical terminology.
Understanding the molecular impact of ionizing radiation exposure is essential for both biomedical research and public health. Among the possible approaches to study this phenomenon, gene expression profiling via transcriptomics assays has been a valuable approach over the last decades to unravel the mechanisms of cellular responses to radiation. To our knowledge, there is no data package gathering well-curated radiation transcriptomic datasets covering microarrays and, more recently, RNA sequencing. Therefore, we present DoReMiTra, an R/Bioconductor data package that represents the first unified radiation transcriptomics dataset collection integrated with Bioconductor's ExperimentHub for efficient distribution. DoReMiTra standardizes and harmonizes sample-level metadata and provides pre-processed SummarizedExperiment objects to facilitate comparative analyses. Additionally, we introduce a lightweight Shiny app interface for interactive visualization and preliminary exploration. DoReMiTra serves as a valuable resource and tool in radiation research for benchmarking, integrative analyses, and biomarker discovery. DoReMiTra is available under the MIT license at https://bioconductor.org/packages/DoReMiTra.
Electronic Health Record (EHR) data are increasingly used in cancer research, yet the fidelity of this data when exchanged between systems remains poorly quantified. This study investigated the agreement in essential biomarker data after they are passed from the EHR into the cancer registry and Fast Healthcare Interoperability Resources (FHIR) extracts. This single-institution retrospective study compared demographics and 6 biomarkers from 30 lung cancer patients seen between July 2020 and July 2022. Manual review from the EHR served as the gold standard, with concordance tested between the source EHR, Institutional Cancer Registry, and FHIR exports. Demographics showed high concordance across databases. In contrast, biomarker data present in the source EHR were missing in 80%-100% of FHIR extracts. The demographic registry variables were highly concordant. This study reports a significant loss in biomarker data availability across real-world data (RWD) sources. Results underscore critical gaps in RWD extraction or exchange methods and highlight risks of relying on RWD without validation.
The life cycle of Plasmodium falciparum is complex, involving asexual and sexual reproduction in humans and mosquitoes, respectively. The parasite transmission from humans to mosquito vectors requires the formation of female and male gametes through gametogenesis. The MKT1 domain proteins are key molecules involved in posttranscriptional gene regulation and cellular differentiation in protozoans, plants, and animals, and the presence of an MKT1 ortholog suggested a possible role in stage differentiation in Plasmodium. Here, we report that PfMKT1 was expressed in both asexual and sexual stages. Parasites with a MKT1 gene deletion (P. falciparum Pfmkt1¯) proliferated asexually similar to wildtype NF54 parasites and differentiated into gametocytes forming mature male and female gametocytes. Further analysis showed that P. falciparum Pfmkt1¯ gametocytes underwent gametogenesis to form male and female gametes and showed no apparent defect in flagellar gamete formation. This study identifies that MKT1-like protein is dispensable during asexual and sexual stage development.
Radiomics and liquid biopsy represent minimally invasive approaches to assess disease characteristics in solid tumors. We integrated computed tomography (CT) radiomics and circulating tumor DNA (ctDNA) analysis to enhance prognostic stratification and longitudinal monitoring in patients with advanced non-small cell lung cancer (NSCLC). This study prospectively enrolled 91 patients with advanced NSCLC. Baseline molecular profiling was performed on both tumor tissue and plasma ctDNA using targeted next-generation sequencing. Radiomic features were extracted from baseline CT lung lesions using PyRadiomics, and radiomic scores (RS) were developed using LASSO-regularized Cox models. A subgroup of 21 patients with actionable molecular alterations underwent longitudinal CT scans and liquid biopsies during targeted therapy. Clinical, radiomic, and molecular associations with overall survival (OS) and disease-free survival (DFS) were evaluated using log-rank tests and included in multivariable models. Overall concordance between tissue and ctDNA (n = 67 patients) was 85%. In the combined clinical-radiomic-genetic model (C-index: 0.73), the RS (p < 0.001) and the presence of actionable alterations (p = 0.041) were independent OS predictors. For DFS, the integrated model achieved a cross-validated C-index of 0.77, outperforming the clinical-only model (0.59). In patients with EGFR-mutant NSCLC, detectable baseline ctDNA was significantly associated with a higher risk of disease progression (p = 0.018). In this subgroup, the combined clinical-radiogenomic model achieved a cross-validated C-index of 0.80 for DFS. Longitudinal analysis showed that 17 of 21 patients achieved molecular clearance of ctDNA at the first follow-up, correlating with treatment response. Integrating radiomics with liquid biopsy provides a more robust prognostic assessment of advanced NSCLC than clinical or molecular data alone. This multi-modal approach may offer a minimally invasive strategy for personalized risk stratification and monitoring of treatment response in patients with NSCLC.Clinicaltrials.gov identifier: NCT06331975.
Proteolysis-targeting chimeras (PROTACs) represent a powerful therapeutic modality for selective protein degradation but often suffer from poor pharmacokinetics and limited tumor-targeting. To overcome these constraints, we developed albumin-binding BRD4-degrading PROTACs (Alb-TACs) with esterase-cleavable maleimide linkers that hitchhike endogenous albumin and enable esterase-responsive BRD4 degradation in tumors. Alb-TACs were synthesized by conjugating two esterase-cleavable maleimide linkers, bicyclononyne-polyethylene glycol-maleimide (BCN-PEG2-Mal) or N-(2-aminoethyl)maleimide (AE-Mal), to BRD4-degrading PROTAC (ARV-771), resulting in Alb-TAC#1 and Alb-TAC#2, with distinct albumin- and esterase-binding properties. To select effective Alb-TAC, the binding ability to albumin and esterase-specific cleavage of Alb-TACs were carefully assayed using MALDI-TOF, PAGE, and time-course HPLC. Furthermore, the tumor-targeting efficacy of Alb-TACs was assessed by fluorescence imaging in CT26 tumor-bearing BALB/c mice. Next, we investigated the BRD4 degrading efficiency of Alb-TAC in a cell culture system and in CT26 tumor-bearing mice. Finally, the immunogenic cell death (ICD) and reprogrammed immune cells of Alb-TAC-treated tumors were carefully characterized. Alb-TAC#2 containing the AE-Mal linker exhibited rapid albumin binding, accelerated esterase-responsive activation, and enhanced tumor accumulation compared to ARV-771 and Alb-TAC#1 due to its flexible chemical structure. In the CT26 cell culture system, Alb-TAC#2 efficiently degraded BRD4, resulting in BRD4-deficient cell death. Furthermore, in CT26 tumor-bearing mice, Alb-TAC#2 achieved extensive apoptosis through robust BRD4 degradation, leading to marked downregulation of c-Myc, Bcl-2, and PD-L1. Moreover, Alb-TAC#2 induced hallmarks of ICD (elevated surface CRT, extracellular ATP, and HMGB1) and reprogrammed the tumor microenvironment by enhancing CD8⁺ T cell infiltration, promoting dendritic cell maturation, and reducing regulatory T cell function. This esterase-responsive albumin-binding PROTAC design could overcome pharmacokinetic barriers of conventional BRD4-targeting PROTACs by enhancing tumor-specific delivery and esterase-responsive BRD4 degradation in solid tumors. In summary, esterase-responsive albumin-binding PROTAC is proven as a promising strategy that effectively modulates the pharmacokinetics and therapeutic performance of PROTACs for cancer immunotherapy.
Olfactory neuroblastoma (ONB) is a rare head and neck cancer arising from the upper nasal cavity, with limited systemic therapeutic options due to a poor understanding of its genomic landscape. This study aims to utilize a patient-level genomic repository to identify potential therapeutic targets and improve disease modeling in ONB. Retrospective genomic analysis. Data analysis was performed using the American Association for Cancer Research (AACR) Project Genomics Evidence Neoplasia Information Exchange (GENIE) database. Patients with confirmed ONB who have undergone targeted sequencing within GENIE. Data were analyzed for recurrent somatic mutations, along with their clinical and demographic correlations, with significance set at p  < 0.05. A high prevalence of mutations in TP53 (tumor protein p53) and FRK (fibroblast growth factor receptor kinase) genes was identified. A moderate prevalence of mutations in NOTCH3 (notch receptor 3), SMARCA4 (SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily A, member 4), RET (rearranged during transfection), and CTCF (CCCTC-binding factor) was also identified. The mutation patterns differed between pediatric and adult ONB cases. Specific mutations were enriched in metastatic tumors compared with primary tumors. This study provides a genomic profile for ONB, identifying key mutations and potential therapeutic targets. The identification of frequently mutated genes like TP53 and FRK suggests potential targets for novel therapies. The observation that certain genes are mutated in pediatric ONB but not adult ONB (and vice versa), and the presence of specific mutations in metastatic tumors that are absent in primary tumors, offers valuable insights for future precision medicine and the design of targeted therapeutic interventions for these distinct clinical presentations.
Traditional studies of rat communication have focused on a few individuals in simplified laboratory settings, leaving colony-level vocal-behavioral dynamics largely unexplored. Here, we examined the nocturnal lives of a rat colony housed in an enhanced naturalistic habitat (ENH). Using infrared cameras and ultrasound microphones, we recorded 18,412 ultrasonic vocalizations (USVs) alongside social behaviors. Analyses revealed structurally distinct USVs not previously characterized statistically or behaviorally. Notably, 22 kHz calls-typically associated with aversion-occurred across multiple social behaviors lacking obvious aversive context. Brief exposure to cat hair elicited uniquely modified 22 kHz calls as rats fled to their burrows, remaining hidden for two days. The introduction of an unfamiliar caged rat, and the disruption of the colony's social structure then showed how social stressors alter vocal-behavioral dynamics. These findings reveal the structured and context-sensitive nature of rat vocal communication, establishing the ENH as a powerful model for studying social-vocal behavior.
Malignant lung tumors are the leading cause of cancer-related mortality worldwide, and therefore remodeling of the tumor microenvironment (TME) has become an important strategy to overcome anti-tumor therapy resistance in lung cancer. Flavonoid components isolated from Citri Reticulatae Pericarpium (CRP), such as nobiletin, hesperidin, and tangeretin, have been shown to modulate the lung cancer TME in a highly relevant manner. The present narrative review collected literature from the PubMed, Web of Science, Embase, CNKI, and Wanfang databases between 2016 and 2026, and hence discussed the molecular mechanisms by which CRP flavonoids reshape the lung cancer TME, namely their regulation of oxidative stress-inflammation homeostasis, correction of lipid metabolic reprogramming, induction of pyroptosis, and inhibition of epithelial-mesenchymal transition. Discussion of the role of EMT and tumor angiogenesis suppression were also presented. Then evidence regarding the modulation of emerging targets was introduced, namely ferroptosis and the cyclic GMP-AMP synthase-stimulator of interferon genes (cGAS-STING) pathway, which are both promising targets in lung cancer models. Translational prospects of CRP flavonoids were led to enhancing immune checkpoint inhibitor (ICI) efficacy and developing nano-delivery systems. The article first outlined the fundamental barriers, then gave a very systematic and critical review of the contradictory findings, context-dependent effects, and methodological limitations in the existing literature; and then pointed out the gaps in frontier research. Therefore, it provides an excellent theoretical foundation for the research and development of anti-lung cancer drugs from CRP flavonoids, while also objectively identifying the current knowledge gaps and clinical translation bottlenecks.
In recent years, the prevalence of myopia has sharply increased in East Asia, emerging as a major public health issue. This study aimed to identify genetic risk factors for myopia progression and to develop polygenic risk score (PRS) models to predict myopia progression risk. Genotyping was performed using the Asian Screening Array chip among 294 Chinese adolescents who completed a 2-year follow-up. A two-stage (discovery cohort: N = 176; replication cohort: N = 118) genome-wide association study (GWAS) was subsequently conducted. Functional annotation and MAGMA analysis were performed to confirm biological relevance of the identified loci in the progression of myopia. Based on GWAS results from the discovery cohort and a European population from the United Kingdom Biobank (N = 460,536), we constructed single-ancestry and cross-ancestry PRS models with PRSice-2 and PRS-CSx. We evaluated the predictive performance of these models using the replication cohort. Our meta-analysis identified seven novel suggestive loci associated with myopia progression, including FSTL5 on 4q32.2, SMARCA2 on 9p24.3, CCDC3 on 10p13, GALNT6/ACVR1B on 12q13.13, CRY1 on 12q23.3, ULK2 on 17p11.2, and MYL4/EFCAB13-DT on 17q21.32. For myopia progression risk prediction in East Asians, PRS analysis showed that the East Asian training dataset (R 2: 5.69%; OR: 1.61, 95% CI: 1.06-2.44; AUC: 0.66) outperformed both the European and cross-ancestry datasets. This study identified seven promising loci associated with myopia progression and demonstrated that PRS exhibited enhanced predictive performance in genetically and phenotypically matched populations. Our findings expand the genetic understanding of myopia progression in East Asian adolescents and provide new insights for myopia prevention and control.
Gold-centered self-assemblies exhibit unique structures, physicochemical and pharmacokinetic properties, which are widely explored for the applications of catalysis, analytic devices, optical devices, and biomedicines. Most of these materials are based on gold nanoparticles (AuNPs) and gold nanoclusters (AuNCs). However, the use of Au(III) complexes as building blocks for functionally tailored self-assembled architectures remains underexplored. Herein, we develop a facile self-assembly strategy using gold-nicotinamide complex as a building block to construct uniform nanospheres. Remarkably, these nanoparticles can undergo reversible disassembly and reassembly to give cubic nanoparticles upon pH stimulation via alkaline or acidic additions. Through systematic experimental and theoretical calculations, we identified the driving force and rules for these behaviors: the primary building blocks were formed via metal coordination between Au and nicotinamide, which further self-organized into uniform spherical nanoparticles through noncovalent interactions (π-π stacking and hydrogen-bonding N-H···O═C between amide groups). The mechanistic insights reveal that pH-mediated changes alter the balance of these interactions, enabling morphological switching. This work establishes valuable design principles for stimuli-responsive organometallic nanomaterials and highlights the critical role of ligand architecture in controlling self-assembly behaviors.