Eye-movement pattern during text reading has been recognized as a functional predictor of reading comprehension proficiency. However, it remains unclear whether and how the coupling between eye-movement pattern and reading comprehension is influenced by text genre and level of analysis. This study addresses this question by investigating genre-specific eye-movement patterns during narrative and expository reading, as well as evaluating their predictive power for comprehension proficiency at both the keyword and whole-text levels. A total of 76 participants (Mage = 21.02, SD = 2.04; 72.3% female) were recruited for this study. Eye-movement features were extracted at both the whole-text and keyword levels while participants read narrative and expository texts. Support Vector Machine Recursive Feature Elimination was then utilized to build predictive models for reading comprehension and to select optimal feature subsets. Narrative reading and expository reading were associated with two distinctive eye-movement profiles at both levels. While eye-movement pattern during narrative reading exhibited a 'Dual-Stage Balanced Mode' that might reflect synchronized early lexical access and late-stage integration, eye-movement pattern during expository text reading exhibited a 'Late-Stage Compensatory Mode' that reflected backward integration. Moreover, the keyword level analysis outperformed the whole-text level analysis in terms of comprehension proficiency prediction, suggesting that core semantic nodes were more informative of readers' comprehension proficiency. Our findings offer a transformative perspective on how eye-movement patterns can be leveraged to assess reading comprehension and provide important heuristics for developing reading assessment and instructional tools.
To systematically evaluate and compare the quality, readability, and query-model consistency of adenomyosis-related content generated by two large language models, ChatGPT (GPT-5) and DeepSeek (R1). In total, 25 high-frequency patient queries were obtained based on Google Trends. Each query was processed using two interaction modes, namely, three consecutive repetitions and three independent cycles, on both large language models (ChatGPT GPT-5.0-web, released December 2025; DeepSeek R1-web, released November 2025). The generated texts (n = 300) were subsequently assessed for their readability [evaluated by Automated Readability Index (ARI), Flesch Reading Ease Score (FRES), and Gunning Fog Index (GFI)] and quality [assessed by DISCERN score, and Ensuring Quality Information for Patients (EQIP) tool]. Statistical comparisons were performed using non-parametric tests and t-tests. In the cyclic mode, both ChatGPT and DeepSeek maintained stable output text readability and quality. DeepSeek-generated text demonstrated significantly superior readability across both interaction modes (lower ARI: 11.32 vs. 14.56, p < 0.001; higher FRES: 46 vs. 27, p < 0.001; lower GFI: 12.47 vs. 14.16, p < 0.001) and higher information quality (higher DISCERN: 62 vs. 43, p < 0.001; higher EQIP: 75 vs. 70, p < 0.001). Under the repetition mode, DeepSeek's output exhibited significant fluctuations across multiple metrics (ARI: p = 0.021; FRES: p = 0.015; GFI: p = 0.004; DISCERN: p = 0.013; EQIP: p < 0.001), while ChatGPT's output remained stable (all p > 0.05). Notably, the readability scores for both models indicated reading levels equivalent to undergraduate education, which is above the recommended level for general public health information. The findings of this study demonstrate that when generating information on adenomyosis, DeepSeek outperforms ChatGPT in terms of readability and several information quality metrics, whereas ChatGPT exhibits greater consistency in its outputs. However, the reading difficulty of texts generated by both models exceeds the level suitable for the general public, representing a key practical constraint limiting direct public use. Based on these results, AI chatbots may serve as complementary tools in patient education; however, their outputs should undergo expert review and be optimized for comprehensibility before broader clinical application. For clinicians and patients, these findings emphasize the importance of critically appraising AI-generated information and using it as a supplement to, rather than a substitute for, professional medical consultation.
Objective: Online patient education materials (OPEMs) are important resources for patients seeking health information. While the National Institutes of Health (NIH) and American Medical Association (AMA) recommend a sixth-grade readability level for OPEMs, commonly available material often exceeds such criteria. Large language models (LLMs), such as ChatGPT and Gemini, have emerged as tools for health education with potential applications in simplification of health material. This study assesses the utility of ChatGPT and Gemini in enhancing the readability of OPEMs for peripheral nerve surgeries. Methods: Eleven common peripheral nerve surgeries were used as online search terms. The first 20 unique search results were assessed; results were excluded if they did not include patient-facing material. ChatGPT and Gemini were instructed to rewrite the text of the OPEM at or below a sixth-grade reading level. Readability metrics were calculated for original OPEMs, alongside ChatGPT and Gemini rewrites. LLM responses were reviewed for accuracy/quality (five-point scale) and comprehensiveness (three-point scale) using predefined criteria. Results: A total of 220 websites were assessed. In total, 155 OPEMs met the inclusion criteria; 65 websites were excluded because they were academic journal articles or other provider-facing materials. The average Flesch-Kincaid grade level (FKGL) of OPEMs was 11.3, significantly greater than the NIH/AMA-sixth grade recommendations (p < 0.001). The average FKGL of ChatGPT rewrites was significantly lower than that of OPEMs (11.3 vs. 7.5, p < 0.001), as was the average FKGL of Gemini rewrites (11.3 vs. 5.6, p < 0.001). ChatGPT rewrites were of higher accuracy/quality (4.5/5.0 vs. 4.0/5.0, p < 0.001) and comprehensiveness (2.0/3.0 vs. 1.0/3.0, p < 0.001) relative to Gemini rewrites. Conclusions: The readability of online patient education materials for peripheral nerve surgery significantly exceeded NIH/AMA recommendations. ChatGPT and Gemini were able to significantly simplify the reading level of these OPEMs. LLMs may serve as tools to improve the readability of peripheral nerve surgery OPEMs.
Patients with cardiac myxoma require long-term follow-up, and the quality of patient education as well as the ability to recognize early symptoms may influence integrated postoperative cardiovascular and tumor-related management. Although large language models (LLMs) have been increasingly applied in medical education, cross-platform empirical evidence in this intersecting context remains limited. This study evaluated the performance and limitations of widely used LLMs in generating patient education texts for cardiac myxoma using a standardized, expert-curated educational question set. We constructed a standardized dataset of 60 expert-curated educational questions spanning the disease course of cardiac myxoma and used nine widely used LLMs to generate 540 patient education texts. Text quality, readability, and actionability were assessed using the Patient Education Materials Assessment Tool for Printable Materials (PEMAT-P), Ensuring Quality Information for Patients (EQIP-36), the Global Quality Score (GQS), and seven readability formulas. In addition, a regularized Gaussian graphical model-based partial correlation network analysis and latent profile analysis were performed to identify relationships among evaluation metrics and cross-platform text phenotypes. Texts generated across platforms showed significant heterogeneity in information quality, understandability, and objective readability, whereas clinical actionability was generally low. Word count showed the strongest positive correlation with the total EQIP-36 score and occupied a central position in the network. Reading-difficulty indices were consistently negatively correlated with PEMAT-P actionability. Latent profile analysis identified three text phenotypes: moderate-quality/low-readability, high-quality/high-actionability, and low-quality/easy-to-read. Ideally suited patient education texts accounted for only a very small proportion of all outputs. Within this expert-curated educational question set, the current application of LLMs in patient education for cardiac myxoma is primarily limited by reading burden and insufficient behavioral guidance. Although longer outputs were generally associated with higher informational quality scores, greater syntactic complexity was associated with lower actionability of the materials. In addition, latent profile analysis suggested that only a small subset of outputs approached a more favorable quality-actionability profile.
Background/Objectives: Structural variations are important forms of genomic variation and are closely related to genomic diversity and many human diseases. Long-read sequencing has improved the ability to detect structural variations in complex genomic regions, but existing methods still mainly rely on manually designed heuristic rules and often have difficulty jointly modeling local SV signatures and cross-subsegment contextual modeling. To address this problem, we propose CMSV, a structural variation detection and genotyping method for long-read sequencing data. Methods: CMSV extracts multi-channel position-level features from alignment results and combines a multi-scale convolutional encoder with stacked Mamba modules for window-level candidate region detection. Candidate variants are then integrated and optimized through DBSCAN-based density clustering and length-based clustering. Genotypes are inferred based on variant-supporting reads and reference-genome-supporting reads. CMSV is designed to support several major structural variation types, including DEL, INS, DUP, INV, and TRA/BND. In our real-data benchmarks, the strongest validation is provided for DEL and INS, while DUP, INV, and TRA/BND are further evaluated using simulated multi-type datasets. Results: Experiments on real HG002 DEL/INS benchmarks, simulated multi-type datasets, and family-based datasets show that CMSV is competitive across PacBio CCS, PacBio CLR, and ONT platforms within the corresponding evaluation settings. Additional held-out chromosome evaluation on GRCh38 chr13-chr22 was further conducted to assess chromosome-level generalization beyond the full HG002 benchmark. CMSV shows stable performance in DEL/INS detection and genotyping on real-data benchmarks, while simulated multi-type evaluations further support its ability to detect DUP, INV, and TRA/BND. The results also show that CMSV can effectively model complex variant signals and maintain good family-level consistency in trio-based evaluation. Conclusions: CMSV provides an effective deep learning framework for long-read structural variation detection and genotyping across sequencing platforms and coverage levels.
Scar is an inevitable pathological product of tissue injury repair, and pathological scars often occur in exposed areas, bringing severe psychological burden and economic losses to patients. With the popularization of digital healthcare, patients increasingly rely on artificial intelligence (AI) for self-consultation, but the core capabilities of free generative AI in scar management have not been systematically evaluated. This study compared and evaluated the comprehensive performance of ChatGPT-5.4 mini and Gemini 3 Flash in answering clinical and psychological questions of scar patients, investigated multi-dimensional differences, and provided support for the application of AI in patient education. Fifteen core questions from scar patients were extracted and input into ChatGPT-5.4 mini and Gemini 3 Flash, respectively. The DISCERN-AI scale and Global Quality Scale (GQS) were used for evaluation, while multiple standardized tools were applied to quantify text readability and complexity. All data were subjected to a normality test and difference analysis using SPSS software. Both models demonstrated high clinical reliability, with no significant difference in target topic clarity (P=0.806). ChatGPT had better overall quality, with a GQS score of 4.8 (4.5, 4.9), which was significantly higher than Gemini's 4.6 (4.4, 4.7) (P=0.033). ChatGPT was also more rigorous in stating medical limitations and uncertain treatment options (5.0 versus 4.5, P<0.05). In contrast, Gemini performed better in patient demand relevance and empathy (4.5 versus 4.0, P=0.026). Both models achieved moderate scores in shared decision-making support. Readability analysis showed that the reading thresholds of both models were excessively high, far exceeding the internationally recommended 6th- to 8th-grade standard for patient education materials. ChatGPT-5.4 mini and Gemini 3 Flash have complementary advantages and potential as auxiliary tools for digital health education in scar patients, but both have a serious readability gap. For future large-scale applications, readability prompt intervention should be introduced, and it should be clearly stated that AI cannot replace professional diagnosis and treatment to ensure the inclusiveness and safety of digital medical information.
Liquid biopsy, which involves the study of tumor-derived genetic material shed into circulating body fluids, is a rapidly emerging minimally invasive approach for cancer diagnosis and monitoring. Most current cancer liquid biopsy workflows depend on short-read sequencing (SRS). However, SRS methods remain limited in their ability to detect and resolve structural variants (SVs), haplotype phasing, fusion transcripts, and epigenetic modifications. Long-read sequencing (LRS) technologies, including single-molecule real-time (SMRT) and nanopore sequencing, offer opportunities to overcome these limitations by preserving long-range molecular information and enabling multimodal characterization of tumor-derived material in biofluids. In this mini-review, we discuss the emerging role of LRS in cancer liquid biopsy, with primary emphasis on cell-free DNA (cfDNA) and circulating tumor DNA (ctDNA). We summarize recent studies using LRS-based liquid biopsy across multiple cancer types. Particular focus is placed on cancer types most actively investigated to date, such as lung, brain, and pediatric cancers, in which LRS-based liquid biopsy has shown promise in detecting SVs, methylation patterns, and tumor-of-origin (TOF) signals that may not be fully captured by SRS approaches. We also examine current technical and translational barriers of LRS in cancer liquid biopsy, such as pre-analytical variability, cost, and high computational demands. As sequencing technologies and analytical pipelines continue to advance, LRS is likely to serve as a complementary component of multimodal liquid biopsy strategies in precision oncology.
Fragile X syndrome (FXS) is the most common monogenic cause of inherited intellectual disability and is primarily caused by CGG repeat expansion in the FMR1 gene. Conventional diagnostic methods have limited precision for sizing long repeat sequences and cannot resolve AGG interruptions, which are critical for comprehensive risk assessment. Existing national FXS reference materials are based on conventional methods and provide limited molecular information. We developed a targeted long-read sequencing assay for comprehensive FMR1 characterization, termed tLRS-FMR1, and applied it to a panel of 22 national FXS reference materials. The tLRS-FMR1 assay demonstrated 100% concordance with standard methods while overcoming key limitations of conventional approaches. It enabled precise quantification of CGG repeat numbers, including full mutations (>200 repeats) that were only qualitatively reported by traditional techniques and provided comprehensive mapping of AGG interruption patterns. The assay showed high reproducibility, with 100% genotyping concordance across intra- and inter-assay replicates and achieved a detection limit of 3 ng/μL. This study successfully developed tLRS-FMR1 and established a new-generation national FXS reference material system with expanded molecular information and improved precision, providing a foundation for advancing the standardization and accuracy of FXS molecular diagnosis.
Background/Objectives: Cardiopulmonary resuscitation (CPR) skill assessments are susceptible to evaluator subjectivity, cognitive fatigue, and observational limitations. Although recent advances in multimodal artificial intelligence have increased the possibility of automated video-based assessment, its validity for clinical skill evaluation remains insufficiently examined. Methods: In this cross-sectional study, we enrolled 130 laypersons who underwent Basic Life Support training and skill testing. Twenty recordings were used for prompt development and 110 recordings were analyzed. Expert evaluators and GPT-4o independently assessed participants' skills using a 12-item checklist. The manikin sensor data were the reference standard for the four chest compression metrics. Agreement was evaluated using Gwet's agreement coefficient 1 (AC1) and intraclass correlation coefficient (2,1). Diagnostic accuracy, sensitivity, and specificity were compared using McNemar's test. Results: Procedural items such as confirming cardiac arrest, calling 119, and requesting an automated external defibrillator showed a near-perfect agreement between experts and GPT-4o (AC1 > 0.8). However, the agreement was poor for the compression depth (AC1 = 0.374) and full chest recoil (AC1 = 0.355). Experts demonstrated high sensitivity (77.8-84.3%) but low specificity (24.6-47.8%), whereas GPT-4o showed low sensitivity (35.6-40.6%) but high specificity (69.2-76.1%). Conclusions: GPT-4o cannot serve as a standalone evaluator because of its inherent limitations in inferring three-dimensional spatial information from two-dimensional videos. However, its high agreement on procedural items and complementary error patterns with that of human evaluators on compression metrics suggests its potential as a decision support tool to mitigate expert leniency bias in CPR education.
Genomic testing has transformed rare-disease diagnostics, yet a substantial proportion of individuals remain without a molecular diagnosis even after short-read exome sequencing (SR-ES) or short-read genome sequencing (SR-GS) and repeated conventional analysis. To address this persistent gap, we evaluated a coordinated multimodal reanalysis framework for deeply investigated families with suspected monogenic disease. Six families (20 individuals; 8 affected individuals) that had remained unsolved after prior comprehensive testing were reviewed prospectively in weekly interdisciplinary case conferences over one year. Available data included SR-ES, SR-GS, long-read genome sequencing (LR-GS), RNA-seq, optical genome mapping, mobile-element analysis, and mitochondrial genome analysis. The goal was not to test a single modality in isolation, but to assess whether systematic escalation across complementary assays plus continued reinterpretation could improve case resolution. Three families (50%) achieved a reportable molecular diagnosis, two (33%) yielded strong candidate findings requiring additional evidence, and one (17%) remained without a definitive new molecular diagnosis, although reinterpretation of a previously identified NOTCH3 variant provided a possible partial explanation. Resolved cases included compound-heterozygous variants in KLHL40, a 119 kb multi-exon deletion in TTN, and a recurrent insertion in RNU4-2. Candidate findings included biallelic NARS2 variants and a 1.3 kb intragenic deletion involving ZEB2. Functional transcriptomic analyses supported the KLHL40 and TTN diagnoses but did not demonstrate a splicing consequence for the candidate NARS2 intronic variant in cardiac tissue. This small pilot cohort is not intended to estimate general diagnostic yield, but it demonstrates that a coordinated multimodal framework can reveal different sources of added value, including structural variant discovery, orthogonal functional support, and reinterpretation of existing short-read data as knowledge evolves. These findings underscore that archived short-read exome and genome data can retain substantial diagnostic value years after initial testing, particularly when reanalyzed with updated pipelines, expanded disease gene knowledge, and orthogonal multimodal evidence. Adoption of iterative, team-based multimodal strategies may help resolve the most complex unsolved rare-disease cases.
Background/Objectives: Probiotic feed additives are increasingly used in livestock production as antimicrobial-sparing tools, yet viable microbial products should not introduce clinically relevant antimicrobial resistance genes (ARGs) into the intestinal resistome. This study evaluated farm-animal probiotic products using an integrated phenotypic, metagenomic and mobilome-aware safety framework. Methods: Seven commercially available products intended for poultry, pigs, cattle or horses were assessed using product metadata, culture-based recovery, broth microdilution minimum inhibitory concentration (MIC) profiling and Illumina short-read sequencing as a screening-level resistome approach. Reads were quality controlled, assembled, screened using the Comprehensive Antibiotic Research Database (CARD)/Resistance Gene Identifier (RGI) workflow and interrogated for plasmid-, phage- and insertion sequence/mobile genetic element-associated genomic context. Results: MIC profiles were generated for viable bacterial isolates representing Enterococcus faecium, Pediococcus acidilactici, Pediococcus pentosaceus and Bacillus subtilis. One labelled Lactobacillus plantarum component was not recovered as viable culture, and one labelled P. acidilactici component was recorded as P. pentosaceus. Sequencing-based resistome screening identified 30 antimicrobial resistance (AMR)-associated CARD antibiotic-resistant organism (ARO) hits belonging to 13 determinants across six ARG-positive coded products, while one coded product had no retained CARD/RGI hit. Profiles were dominated by recurrent Enterococcus-associated background determinants, including aac(6')-Ii, msrC and eatAv. Plasmid prediction was positive for five hits, whereas no iMGE- or phage-associated ARG context was detected. No vanA/vanB, mcr, optrA, poxtA, cfr, extended-spectrum β-lactamase (ESBL) or carbapenemase gene was detected. Conclusions: The investigated products did not show evidence of high-priority mobile ARG carriage. Nevertheless, AMR-associated determinants and occasional predicted mobile contexts support routine integrated MIC-sequencing-based resistome-mobilome assessment of veterinary probiotic products. Because short-read assemblies do not fully resolve plasmid architecture or transferability, mobile-context predictions should be considered screening-level indicators requiring confirmatory long-read or functional testing for higher-priority findings.
Abortion in cattle entails substantial economic loss, and rapid identification of abortigenic pathogens is critical for timely on-farm response and reduction in human exposure risk. In 2024, two Holstein cows from a small farm in Inner Mongolia aborted in close succession without an obvious cause. Vulvar swabs from both cows, one afterbirth sample, and whole blood from one aborted fetus were collected. Shotgun metagenomic sequencing was performed, followed by host-read removal, taxonomic profiling with Kraken2, de novo assembly of Brucella-aligned reads, and whole-genome comparison. Serological tests, Gram-stained smears, and Brucella genus- and species-specific qPCR assays were used as orthogonal verification. Putative resistance and virulence determinants were screened against CARD and VFDB. Brucella reads were detected in all samples, with the highest relative abundance in the 138-afterbirth (96%). qPCR assays detected Brucella DNA and B. abortus-specific signals in all four samples. A draft Brucella genome was assembled from the 138-afterbirth sample and was phylogenetically placed within B. abortus, showing relatedness to previously circulating Chinese lineages. Cows 138 and 198 were RBT-positive with SAT titres of 1:100 (++). No acquired Brucella resistance genes were identified in CARD. Within 72 h of sample receipt, B. abortus was reported to the farm and local authorities and emergency biosecurity measures were implemented. This field investigation shows that metagenomic sequencing, when combined with conventional serology, microscopy, and targeted qPCR, can support rapid etiological investigation when culture is delayed, hazardous, or biosafety level 3 facilities are unavailable.
Mitochondrial genomes are widely used in insect taxonomy and phylogenetics, but their signals may conflict with morphology and nuclear genomic evidence because the mitochondrial genome represents a single maternally inherited locus. Here, we assembled complete mitochondrial genomes of four pygmy grasshoppers, Zhengitettix transpicula, Formosatettix sp., Gibbotettix parvipulvillus, and Bolivaritettix sp., using PacBio HiFi reads. The four mitogenomes ranged from 15,152 to 17,976 bp and contained the typical 37 mitochondrial genes. Mitochondrial phylogenies inferred by maximum likelihood and Bayesian methods were topologically identical and recovered several well-supported tetrigid relationships, including a close relationship between Formosatettix sp. and Bolivaritettix sp. However, Z. transpicula was unexpectedly placed near Macromotettixoides rather than close to other Zhengitettix representatives. In contrast, a morphology-based tree recovered Z. transpicula with Z. triangularis, and comparison with a published nuclear single-copy ortholog tree based on 1962 loci supported a non-mitochondrial placement of Zhengitettix inconsistent with the anomalous mitochondrial position of Z. transpicula. Independent assembly from the original HiFi reads, read-depth inspection, protein-coding gene checks, and nuclear-genome screening for NUMT-like sequences supported the authenticity of the assembled Z. transpicula mitogenome. These results document mito-nuclear and cyto-morphological discordance in Tetrigidae and highlight the need for integrative interpretation of mitochondrial phylogenies in taxonomically complex insect groups.
Precision medicine is increasingly applied in the cancer clinic, adapting treatment to genomic alterations of the tumor. However, whether alterations disrupt the function of a protein can depend on if both alleles of a gene are altered. While massively parallel sequencing technologies can identify sequence aberrations, they are limited in resolving the corresponding haplotype information. In this proof-of-concept case study, we applied the linked-read droplet barcode sequencing (DBS) technology to resolve the haplotype complexity of colorectal cancer genomes on paired tumor and normal samples. Several cancer-related genes carried multiple mutations in either one or both haplotypes. Additionally, a number of haplotype-resolved large structural variants and copy number alterations were detected and phased with short somatic variants. Nearly all characterized oncogenic pathways harbored some of the identified short somatic variants. The study demonstrates that linked-read DBS technology can characterize complex genetic variations in a haplotype context and may provide essential information for personalized approaches.
Upland cotton (Gossypium hirsutum) is a major natural fiber crop and an important model for studying genome evolution and gene function in polyploid plants. However, its large and highly redundant genome presents substantial challenges for efficient and coordinated multiplex genome editing. Here, we developed a high-efficiency CRISPR/Cas12a-based multiplex genome editing system in cotton by integrating a tRNA-crRNA polycistronic expression strategy with a Bean yellow dwarf virus (BeYDV)-derived replicon. This platform enabled coordinated expression of multiple crRNAs and simultaneous targeting of 16 loci within a centromere-proximal region of chromosome D03 (18.65-24.47 Mb). In individual transgenic lines, up to 10 target sites were edited concurrently, with nine targets exhibiting editing efficiencies above 56% and the highest efficiency reaching 96.46%. High-density multiplex editing predominantly induced small insertions and deletions at target loci. Notably, edited plants exhibited reduced growth and pronounced cytological abnormalities, including chromosome bridges, lagging chromosomes, and abnormal meiotic products. Transcriptome analysis revealed widespread dysregulation of genes involved in chromosome segregation and cell cycle regulation. Despite these functional perturbations, HiFi long-read sequencing detected no large-scale chromosomal rearrangements, indicating that genome instability arises from cumulative local perturbations rather than global structural alterations. Together, our results establish an efficient multiplex genome editing platform in cotton and highlight potential constraints of high-density editing on genome stability in complex plant genomes.
Background/Objectives: Melophagus ovinus is an economically important ectoparasite of small ruminants with a broad global distribution. Although mitochondrial genomes are widely used in population genetic studies, the D-loop region of M. ovinus remains poorly characterized because its high AT content and repetitive structure complicate amplification, assembly, and sequencing. Methods: We sequenced the mitochondrial genome of M. ovinus collected from Qinghai using an integrative approach combining Illumina paired-end sequencing, targeted PCR amplification, and Nanopore long-read sequencing. Comparative genomic analysis was performed against published mitogenomes from Gansu (MH024396) and Xinjiang (NC_037368). Results: The Qinghai mitochondrial genome contained the typical 37 mitochondrial genes within a 14,728 bp conserved region. Comparative analysis revealed exceptionally high conservation (>99.6% sequence identity) among Qinghai, Gansu, and Xinjiang isolates outside the D-loop region. Notably, the D-loop exhibited length polymorphism, with different assembly strategies or samples yielding lengths ranging from 317 bp to 2385 bp. Targeted long-read sequencing of ten individuals identified a predominant D-loop variant of approximately 844 bp in nine samples and a markedly shorter variant of approximately 164 bp in one sample. The short variant was characterized by extensive deletions and a novel 45 bp insertion. Support for this variant was obtained from independent Illumina DNA-seq, RNA-seq, Nanopore sequencing, and de novo assembly analyses. Conclusions: This study provides preliminary evidence for D-loop structural heterogeneity in M. ovinus, suggesting remarkable length polymorphism and complex indel patterns that require further validation. These findings significantly expand the genomic resources available for this important veterinary parasite and establish a foundation for future population genetic and evolutionary studies.
The genus Sesamum (Pedaliaceae) comprises a wide range of cultivated and wild species. Sesame (Sesamum indicum) is recognized as one of the oldest oilseed crops cultivated worldwide, whereas S. schinzianum is a wild relative closely associated with the evolutionary history of cultivated sesame. Although the nuclear and chloroplast genomes of Sesamum species have been investigated in previous studies, mitochondrial genome evolution within the genus has received relatively limited attention. In this study, we assembled and comparatively analyzed the mitochondrial genomes of S. indicum and S. schinzianum by integrating BGI short-read sequencing and Oxford Nanopore long-read sequencing data. Genome assembly was performed using Flye and Unicycler, and annotation was conducted using PMGA together with manual curation. Comparative analyses were then carried out to examine genome organization, gene content, repetitive sequences, codon usage, RNA editing sites, chloroplast-derived sequences, phylogenetic relationships, and collinearity patterns. The mitochondrial genomes of S. indicum and S. schinzianum were assembled into one circular molecule and two major circular contigs, respectively, and both contained 36 conserved protein-coding genes. Abundant simple sequence repeats dominated by tetranucleotide motifs and notable repeat variation were detected. Codon usage showed moderate bias, and 478 and 455 RNA editing sites were predicted in S. indicum and S. schinzianum, respectively. Chloroplast-derived sequences accounted for 8.50% and 6.96% of the mitochondrial genomes, respectively. Phylogenetic and collinearity analyses supported a close relationship between the two Sesamum species and identified synteny-based structural differences between their mitogenomes. Comparative analysis of mitochondrial- and chloroplast-based phylogenies showed that the two datasets were largely congruent at the family level and consistently supported the close relationship between the two Sesamum species, although they differed in the placement of several deeper lineages. These results suggest that Sesamum mitogenomes retain conserved gene content while showing assembly- and synteny-supported structural differences. This study provides useful genomic resources for comparative and evolutionary studies of mitochondrial genomes in Pedaliaceae.
Research continues to inadequately integrate gender and diversity, reflecting the longstanding model in which the male body has been considered the universal standard. To address these gaps, initiatives like the SAGER (Sex And Gender Equity in Research) and SAGER-swissethics guidelines have emerged, to encourage the integration of sex and gender into research, from the design to the dissemination of results. Clinical guidelines rarely include sex- or gender-specific recommendations. It is unclear whether this lack reflects the absence of differences or, conversely, a lack of data to establish specific differences. While awaiting an update in the data, clinicians are invited to critically read the literature, adjust dosages to body compositions, and value patients' experiences to ensure equitable care. La recherche continue à manquer de données intégrant le genre et la diversité, du fait d’un modèle masculin érigé en norme universelle. Pour combler ces lacunes, plusieurs initiatives existent. Les recommandations SAGER (Sex And Gender Equity in Research) et SAGER-swissethics sont un guide pour l’intégration du sexe et du genre dans la recherche, de la conception à la diffusion des résultats. Les recommandations cliniques incluent rarement des indications selon le sexe ou genre. On ignore donc si ces lacunes reflètent une réelle absence de différences ou un manque de données pour le prouver. En attendant de nouvelles données, les clinicien-ne-s sont invité-es à adopter une lecture critique de la littérature, adapter les posologies selon la composition corporelle et valoriser l’expérience des patient-e-s pour garantir des soins équitables.
Background/Objectives: Accurate estimation of nutritional content from food images has important applications in dietary assessment and public health surveillance. While large language models (LLMs) have shown promise for this task, the effects of prompt design and model selection on estimation accuracy remain poorly characterized. Methods: We evaluated three Claude models (Haiku 4.5, Sonnet 4.6, Opus 4.6) for visual estimation of five mandatory nutritional components (energy, protein, fat, carbohydrate, and salt equivalent) across three datasets: NutriImage (691 Japanese meal photographs with dietitian-validated ground truth, after OCR-mask quality filtering), SNAPMe (1463 US meal photographs from a publicly available benchmark), and the Japan Branded Food Database (JBFD; 989-1000 packaged food product images). We systematically compared a default prompt and a visual estimation prompt explicitly instructing the model not to read any text or numbers visible in the image. Results: The visual estimation prompt substantially improved accuracy when paired with a sufficiently capable model (energy R2: 0.23 for Haiku to 0.60 for Sonnet, JBFD). Sonnet and Opus substantially outperformed Haiku across all datasets, while differences between Sonnet and Opus were small (MedAPE difference 1-3 percentage points). Packaged food images (JBFD) yielded higher R2 than meal photographs. Salt equivalent showed consistently poor accuracy (MedAPE 34-64%). On SNAPMe, Sonnet achieved lower energy MAE (116.9 vs. 123.0 kcal, -4.9%) and lower MAE for protein (5.9 vs. 7.9 g, -25.7%) and fat (6.6 vs. 8.7 g, -24.5%) compared with a recent ChatGPT-5 study. Conclusions: Claude Sonnet offers the best cost-performance balance for LLM-based nutritional estimation. Prompt design substantially affects accuracy, but only when paired with a sufficiently capable model; model visual recognition capability appears to be a key determinant of performance. These findings highlight the inherent difficulty of this task and provide practical guidance for dietary assessment system development.
Background: Pulmonary tuberculosis (TB) is a significant trigger of acute exacerbations of chronic obstructive pulmonary disease (AECOPD), so its timely and accurate diagnosis is essential. Also, the risk factors for TB occurrence in this population remain unclear. This study aimed to evaluate the performance of metagenomic next-generation sequencing (mNGS) for TB diagnosis in AECOPD patients, as well as to identify the associated risk factors. Methods: A retrospective observational cohort of 659 AECOPD patients with suspected pulmonary infection was enrolled. The microbial cell-free nucleic acids in bronchoalveolar lavage fluid samples were extracted and subjected to mNGS detection. The clinical data for each patient were collected from the hospital information system. The statistical analyses were performed with SPSS version 25.0. Results: A total of 170 cases, included for final analyses, were categorized into TB (n = 41), bacterial infection (n = 73), and non-infective control (n = 56) groups. Among these groups, the TB group had the highest intensive care unit (ICU) admission rate (46.34%) and longest median hospital stay (19.50 days) (p < 0.01). For TB diagnosis, mNGS demonstrated a greater sensitivity (86.00%), a lower specificity (93.30%), and a higher area under the curve (AUC, 0.877) than TB-DNA detection (70.21%, 100%, 0.848, respectively) and Xpert Mycobacterium tuberculosis/rifampicin (MTB/RIF) assay (63.83%, 100.00%, 0.870, respectively). Notably, mNGS identified the bacterial or viral co-infections in 18.00% of TB cases. Furthermore, the stringently mapped read number determined by mNGS showed a positive correlation with ICU admission rate (r = 0.76) and in-hospital mortality (r = 0.77). The lower body mass index (BMI) and reduced natural killer (NK) cell count were identified as the independent risk factors in the TB group (both p < 0.05). Conclusions: For the diagnosis of pulmonary TB in AECOPD patients, mNGS demonstrated comparable performance to TB-DNA detection and Xpert MTB/RIF assay, and also mNGS identified co-infections. In addition, a lower BMI and reduced NK cell count were identified as the independent risk factors for TB occurrence in this cohort.