Reducing sedentary behavior through personalized digital interventions holds particular promise for individuals with chronic pelvic pain disorders (CPPDs), who face unique barriers to physical activity. We present a self-contained, missing data-resilient framework for real-time forecasting of a physical activity score (PAS) using wearable Fitbit data from 134 females with CPPDs. Comparing online and offline learning approaches, we demonstrate that models leveraging recent activity and daily recurring patterns perform the best. When applied to 15-min sedentary bouts (SBs), the PAS forecasts support timely alerts, yielding approximately one true alert per day vs 0.6 false alerts at a conservative operating point. By integrating real-time imputation, the system supports forecasting of SBs for potential use in just-in-time adaptive interventions. Our work demonstrates a privacy-preserving, scalable pathway for integrating precision forecasting into just-in-time adaptive interventions, laying the groundwork for more equitable and effective digital health solutions for women with CPPDs.

Metabolic and bariatric surgery (MBS) is the most effective treatment for severe obesity, yet it remains profoundly underutilized. Patients seeking MBS face pretreatment requirements, including mandatory psychological evaluation, documented failed diet attempts, and months of lifestyle interventions that are substantially more extensive than those required for most common gastrointestinal procedures. These requirements were initially developed with patient safety in mind, but evidence has accumulated that many lack outcome-specific support and have become barriers to care, particularly as third-party payers have increasingly used them to restrict access. Meanwhile, potent obesity management medications (OMMs) are prescribed with relatively limited oversight, often to individuals with preclinical disease, at costs that raise questions about economic sustainability. The Lancet Diabetes & Endocrinology Commission's distinction between preclinical obesity (excess adiposity with preserved organ function) and clinical obesity (a systemic illness characterized by objective organ dysfunction or substantial limitations of daily activities) provides a framework for evaluating current treatment access patterns. A critical implication of this framework, which warrants further exploration, is that it may narrow the population for whom medical treatment is considered necessary: for individuals with preclinical obesity, excess adiposity represents a risk state requiring monitoring and evidence-based counseling, not generally an ongoing illness requiring pharmacotherapy or surgery. This paper examines whether the current pattern, relatively low barriers to expensive, chronic pharmacotherapy for preclinical or early-stage disease; relatively high barriers to durable, cost-effective surgery for established clinical obesity, represents a departure from severity-matched care that warrants systematic reconsideration.

Bacteria have evolved diverse anti-phage defence systems to counter the constant threat of viral infection. Like all immune systems, these defences must accurately detect infection while avoiding inappropriate activation that could harm the uninfected host. Recent years have witnessed a rapid expansion in the discovery of bacterial immune mechanisms, yet fundamental questions remain about how these systems recognize phage invasion. What molecules or events are reliable indicators of infection? How do the sensory mechanisms underlying immunity ensure rapid and robust activation? And what evolutionary trade-offs do phages face when mutating to evade detection? In this Review, we examine the experimental approaches used to identify phage-based triggers and organize known triggers into three classes: phage nucleic acids, phage proteins and perturbations to host processes. For each class, we summarize current evidence and highlight emerging principles that connect diverse defence mechanisms. By focusing on sensing rather than downstream responses, we aim to provide a cohesive framework for understanding how bacteria discriminate self from non-self, a challenge universal to immune systems across all domains of life.

We present a classical algorithm for approximating the expectation values of observables in linear optical circuits with arbitrary product input states, achieving additive-error accuracy. This result indicates that current applications of photonic systems aimed at demonstrating practical quantum advantage through expectation value estimation, such as photonic variational algorithms, may face challenges in attaining the computational advantage. It also implies the output probabilities of boson sampling with arbitrary product input states can be efficiently approximated by our method, so that boson sampling becomes efficiently simulable when its output probability distribution is polynomially sparse. We also develop an efficient classical algorithm for estimating transition amplitudes of arbitrary product states in linear optical circuits. This provides additive-error approximation algorithms for matrix functions associated with linear optical circuits, such as the (loop-)Hafnian, which are of independent interest. As an application, it solves the generalized molecular vibronic spectra problem [C. Oh et al., Nat. Phys. 20, 225 (2024)10.1038/s41567-023-02308-9], previously suggested as a candidate for practical quantum advantage. Finally, we extend our framework to near-Clifford circuits, enabling classical approximation of their expectation values.

Bone tissue engineering is an emerging medical technology that utilizes scaffold materials to promote bone tissue regeneration and repair. Current scaffolds include biodegradable porous polymers, inorganic ceramics, and growth factors, but they face limitations such as difficulty in maintaining long-term biological activity in vivo and restricted applications. DNA hydrogels, composed of natural polymer chains and small molecules, exhibit excellent biocompatibility, biodegradability, bioactivity, high mechanical strength, and low toxicity. They have been applied in tissue repair, particularly in bone regeneration, serving as cell scaffolds, guiding tissue regeneration, and enabling gene therapy. This review summarizes recent advances in DNA hydrogels for bone repair, outlines their fabrication methods and performance characteristics, discusses applications involving growth factors, bone defect repair, bone regeneration, drug delivery, and organoid development, and explores their integration with 3D/4D/5D/6D printing, big data, cloud computing, artificial intelligence, machine learning, precision manufacturing, and automation. Despite existing challenges, DNA hydrogels offer new insights into bone tissue engineering due to their programmability and superior biological properties, and hold great promise as a key tool in bone repair.

Post-traumatic trigeminal neuropathic pain (PTNP) is challenging to treat, as conventional therapies have limited efficacy and side-effects, requiring individualized management. N-of-1 trials focus on individualized/personalized medicine. Theta burst stimulation (TBS), a modified form of high-frequency repetitive transcranial magnetic stimulation (rTMS), has shown promising effects on chronic pain. This study explored the potential of N-of-1 trials to characterize individual responses across active (continuous TBS) and sham rTMS sessions in patients with PTNP. Eight patients (48±16 years) with PTNP received active or sham rTMS sessions for four weeks targeting primary somatosensory cortex (face area) in a single-blinded, randomized cross-over N-of-1 design (active-sham-active-sham or sham-active-sham-active). Pain intensity (primary outcome) and secondary outcomes including perceptual distortion (PD), somatic symptoms, anxiety, jaw function, quality of life, sleep, and catastrophizing were assessed at baseline, end of treatment (EOT), and 1- and 3-month follow-up. Individual pain and PD trajectories were analyzed using segmented regression models including all repeated measurements and accounting for serial correlation. Pain intensity decreased in five patients at the EOT and 1-month follow-up and in four patients at 3 months, although changes varied across participants and sessions. PD magnitude decreased in three patients at the EOT and 1 month, while increasing in one. Psychosocial and clinical outcomes also showed heterogeneous individual changes. Improvements were observed after both active and sham sessions, without a consistent treatment-specific pattern. Thus, N-of-1 trials allow identification of individual treatment responses to rTMS in PTNP, while highlighting marked interindividual variability that should be considered when planning future randomized clinical trials. TRIAL REGISTRATION: This study is registered at ClinicalTrials.gov (NCT06577649) PERSPECTIVE: This study explored the potential of N-of-1 trials to evaluate neuromodulation outcomes in post-traumatic trigeminal neuropathic pain (PTNP). N-of-1 trials enable identification of patient-specific treatment responses to theta-burst stimulation in PTNP, supporting personalized treatment strategies, improving clinical decision-making and informing the design of future randomized clinical trials.

Jiekarzi semi-fine-wool sheep are indigenous to the Qinghai-Tibet Plateau, and may face substantial demands on digestive efficiency under low temperatures, hypoxia and seasonal forage limitation. However, the molecular organization of the gastrointestinal tract in this breed remains insufficiently characterized. Here, we profiled mucosal transcriptomes and digestive-fluid metabolomes from 10 gastrointestinal segments collected from six 12-month-old male Jiekarzi. The segments were grouped into four digestive regions: forestomach, true stomach, small intestine and large intestine. Transcriptomes analysis separated clearly by regions, with PC1 and PC2 explaining 40.82% and 15.00% of the variation, respectively. Pairwise differential expression analysis identified 4,941-9,598 differentially expressed genes, with the largest difference between the small intestine and forestomach. Region-specific enrichment indicated differences in immune processes, epithelial organization, transmembrane transport, digestion and absorption, lipid metabolism and xenobiotic associated pathways. Untargeted LC-MS detected 1,757 metabolites in positive-ion mode and 1547 metabolites in negative-ion mode. Integrated transcriptome and metabolomic patterns suggested region associated immune function, nutrient transport and metabolic specialization. These results provide a baseline molecular map of gastrointestinal regionalization for a wider area of the tract, and a reference resource for future research on digestive physiology, nutrient utilization and environmental resilience in plateau-adapted ruminants.

Adolescents with hearing impairments may face elevated emotional risks related to personal traits and family environments. Guided by the differential susceptibility model, this study examined how environmental sensitivity was associated with depressive symptoms and whether parental overprotection showed indirect associations in hearing and hearing-impaired adolescents. A cross-sectional questionnaire survey was conducted with 2008 adolescents (1386 hearing and 622 hearing-impaired; 1006 males and 1002 females; mean age = 15.25 years, SD = 3.97) who completed measures of environmental sensitivity, perceived parental overprotection, and depressive symptoms. Mediation and network analyses were performed separately for the two groups. Results indicated that hearing adolescents reported higher environmental sensitivity, whereas hearing-impaired adolescents reported higher perceived parental overprotection and depressive symptoms. Maternal overprotection showed a significant indirect association between environmental sensitivity and depressive symptoms among hearing adolescents (β = 0.07, p < .001), whereas both paternal and maternal overprotection showed significant indirect associations among hearing-impaired adolescents (βs = 0.08, p < .001). Network analysis further showed distinct item-level network patterns: bridge nodes were distributed across three communities in hearing adolescents, but were primarily concentrated within parental overprotection and depressive symptom items in hearing-impaired adolescents. These findings suggest that parenting processes may be differently related to depressive symptoms in adolescents with and without hearing impairments. They also highlight the potential value of family-centered support tailored to adolescents with hearing impairments.

Protein sequence similarity search underpins function annotation, evolutionary analysis, and metagenomic mining, yet conventional alignment-based methods face two persistent difficulties. First, their computational cost grows linearly with database size, making them impractical for billion-scale repositories. Second, their sensitivity declines below 30% sequence identity (the "twilight zone"), limiting remote homolog detection. Protein language models (PLMs) produce evolutionarily informative embeddings through self-supervised learning; however, the systematic integration of PLM embeddings with vector search into online retrieval services, together with experimental characterisation across model variants and index types, has received limited attention. We present ProtFaiss, combining ESM2 embeddings with FAISS approximate nearest neighbour search. On SCOPe40, we evaluate classification accuracy, ranking quality, remote homology sensitivity, retrieval speed, and storage footprint. Ablations spanning index type, model capacity, database scale, pooling strategy, and layer selection characterise the design space. ProtFaiss vector search completes in ∼0.02ms per query - roughly 3×105-fold faster than a BLASTp database scan - and, including ESM2 encoding, the end-to-end latency is ∼30ms, approximately 200-fold faster than BLASTp. ProtFaiss achieves remote homology sensitivity of 7.7% (BLASTp: 6.7%), concentrated in the 15%-25% identity range. A notable finding is that intermediate Transformer layers substantially outperform the final layer: ESM2-650M layer 22 achieves Recall@10 of 0.780, compared with 0.431 for the final layer, suggesting later layers trade global semantic signal for local structural detail. A learned attention-pooling analysis and a preliminary Flat-index evaluation on multi-domain Swiss-Prot proteins further qualify this interpretation: learned pooling partially recovers final-layer performance, whereas intermediate layers remain advantageous in both evaluations. These results support the feasibility of PLM-based retrieval for large-scale protein search; however, generalisability is constrained by the single-domain benchmark.

Pulmonary embolism (PE) remains a major cause of cardiovascular morbidity and mortality, but the central management question clinicians increasingly face is not whether to anticoagulate, but how to anticoagulate well. Rapid expansion of the evidence base has transformed anticoagulation management from a largely protocol-driven approach into a more individualised and dynamic model of care. Contemporary treatment decisions extend beyond initial drug selection and increasingly require consideration of dose intensity, treatment duration, bleeding phenotype, cancer status, renal function, drug interactions, and evolving thrombotic risk. Direct oral anticoagulants (DOACs) are now established as first-line therapy for most haemodynamically stable patients, while recent studies have refined selection within the DOAC class and challenged traditional assumptions regarding extended therapy. Emerging evidence suggests that some patients with apparently strongly provoked venous thromboembolism may nevertheless benefit from prolonged anticoagulation when enduring prothrombotic risk factors persist. Other studies support selective dose reduction during long-term therapy, while highlighting important limitations of this strategy in obesity. In cancer-associated thrombosis, contemporary trials increasingly support dynamic and phenotype-driven approaches to prolonged anticoagulation rather than fixed-duration treatment paradigms. This review synthesises evolving evidence on anticoagulation management in PE into a practical framework for clinicians involved in longitudinal venous thromboembolism care. Throughout this review, we distinguish established practice supported by contemporary international guidelines from emerging evidence that may refine future management but has not yet been incorporated into guideline recommendations.