搜索结果：Starts

BackgroundVirologic failure in children is a significant public health concern in sub-Saharan Africa. This study aims to assess the incidence rate of virologic failure and its predictors among children undergoing first-line antiretroviral therapy (ART) in Ethiopia.MethodsA multicenter retrospective follow-up study was conducted in HIV-infected children on first-line ART from January 1, 2013, to December 31, 2022, in Ethiopia. A simple random sampling method was employed to select the sample. Data entry was performed using EpiData, and analysis was conducted using STATA version 14. Kaplan-Meier curves and log-rank tests were utilized for survival analysis.ResultAmong 537 HIV-infected children followed over the study period, 12.29% developed virologic failure, yielding an incidence rate of 17 per 10 000 person-month observations (95% confidence interval [CI]: 13.4, 21.7). Factors independently associated with an increased risk of virologic failure included poor ART adherence (adjusted hazard ratio [AHR] = 2.63; 95% CI: 1.38, 4.97), advanced World Health Organization (WHO) Treatment (T) stages III and IV (AHR = 2.71; 95% CI: 1.15, 6.37), no history of regimen change (AHR = 5.88; 95% CI: 3.23, 10.71), and age at ART initiation above 10 years (AHR = 2.97; 95% CI: 1.30, 6.78). In contrast, having a caregiver younger than 40 years was associated with a significantly lower risk of virologic failure (AHR = 0.42; 95% CI: 0.25, 0.72). These findings emphasize the importance of monitoring adherence, providing closer follow-up for children in advanced treatment stages, and considering caregiver-related factors to reduce virologic failure.ConclusionsVirologic failure among HIV-infected children on first-line ART in Northwest Ethiopia was a relatively low incidence. Interventions targeting poor adherence, children with advanced WHO Treatment (T) stage, older age at ART initiation, and those with no history of regimen change are essential. Additionally, caregiver characteristics, such as age below 40 years, play a protective role and should be considered when designing adherence support and monitoring strategies to further reduce the risk of virologic failure. How Often HIV Treatment Fails and Why Among Children Receiving First-Line Antiretroviral Therapy in Northwest EthiopiaPlain Language SummaryChildren living with HIV need lifelong treatment to keep the virus under control. Antiretroviral therapy (ART) helps reduce the amount of HIV in the body, allowing children to grow and live healthier lives. However, in some children, the treatment stops working well. This is known as virologic failure, meaning the virus is no longer adequately suppressed.This study looked at how often virologic failure occurred and what factors were linked to it among children receiving first-line ART in several health facilities in Northwest Ethiopia. We reviewed the medical records of children who had been on ART for a period of time and followed their treatment outcomes.We found that a notable number of children experienced virologic failure while on first-line treatment. Several factors were linked to a higher risk of treatment failure. These included poor adherence to medication, advanced stage of HIV disease at the start of treatment, low immune status, and treatment-related factors such as drug substitutions and missed clinic visits.Understanding why HIV treatment fails in children is important because early identification of these risk factors can help healthcare providers intervene sooner. Improving medication adherence, providing close follow-up for high-risk children, and strengthening routine viral load monitoring may help prevent treatment failure.The findings of this study can support healthcare workers, program managers, and policymakers in improving pediatric HIV care and strengthening treatment programs, especially in resource-limited settings like Ethiopia.

Financial news recommender systems are crucial for helping investors and financial analysts access essential market information. Recently, the incorporation of external knowledge into recommendation algorithms serves as supplementary data, aiming to mitigate the challenges associated with cold-start and data scarcity. However, the current research faces challenges in assimilating knowledge from diverse sources into the prompt-based framework. Additionally, multi-task learning methods for recommendation tasks and related tasks enhance recommendation effectiveness. However, the comprehensive consideration of the attributes of the news itself, such as sentiment, topic, and popularity, is neglected. To tackle these challenges, we present a novel multi-task prompt large language model (LLM) approach for financial news recommendation that effectively unifies external knowledge. In this study, we first develop a dual knowledge enhancement strategy to integrate structured and unstructured financial knowledge into the recommendation process to enrich the semantic understanding of news. Second, we design hierarchical knowledge prompt templates to enable LLM to learn diverse knowledge for specific tasks. Finally, we implement a multi-task prompt integration mechanism that jointly optimizes recommendation, sentiment analysis, topic classification, and popularity prediction tasks, leveraging inter-task dependencies. Experimental results show that our approach improves performance significantly on real financial news datasets, particularly in few-shot scenarios.

Accurate drug-target affinity (DTA) prediction is pivotal for virtual screening, yet practical reliability is often limited by the static treatment of molecular interactions and the "black-box" nature of deep learning models. In biological reality, binding involves dynamic structural adaptations known as induced-fit effects, which are frequently overlooked by conventional static representations. To address this, we propose a geometry-aware and interpretable framework (DCR-DTA) that explicitly models bidirectional induced-fit interactions, prioritizing stable structural anchors rather than computing raw 3D displacements. By integrating Dynamic Contextual Regularization, our model not only mitigates feature anisotropy in pretrained representations but also captures the intrinsic geometry of the interaction manifold. Extensive experiments on the Davis and KIBA benchmarks demonstrate that our method consistently outperforms state-of-the-art baselines, particularly in challenging cold-start scenarios. Notably, beyond achieving competitive Mean Squared Error (MSE), our model achieves superior external validation metrics, including a higher rm2 score (0.787 on KIBA) and a Concordance Index (CI) of 0.902. Crucially, visualization analyses reveal that the model learns discriminative interaction patterns that align with biological intuition, providing interpretable insights into the binding mechanism. These results suggest that explicitly modeling interaction dynamics and representation geometry is essential for robust and explainable DTA prediction. The source code and datasets are available at https://github.com/ycqmyxh/DCR-DTA.

Preeclampsia (PE) is a pregnancy-specific hypertensive disorder that could lead to serious maternal and fetal complications, yet early identification of women at risk remains challenging because reliable biomarkers are limited. Here we show that generating relatively stable cell-free DNA (cfDNA) fragmentomic metrics, including transcription start site (TSS) coverage, TSS score, and Gini coefficient, required 600 million whole-genome sequencing reads of plasma cfDNA. These metrics exhibited observable differences among genes with varying expression levels in blood cells and placental tissues. In a cohort of 1,058 pregnant women, cfDNA fragmentomics could distinguish pregnancies that subsequently developed PE. When integrated with maternal risk factors, predictive models in two independent test sets achieved mean area under the curves of 0.903 and 0.850 for early-onset and late-onset PE, respectively, with sensitivities of 0.731 and 0.607 at a 10% false positive rate. Importantly, these models also performed well in samples collected before or at 16 weeks of gestation, supporting the potential of cfDNA fragmentomics in early PE risk assessment.

Prolonged sedentary behavior and slump sitting posture may impair neuromuscular function; however, acute effects on postural sway and potential differences between athletes and non-athletes remain unclear. Therefore, this study aims to examine the acute effects of prolonged slump sitting on postural sway in female athletes and non-athletes. In this study, 24 females (12 athletes, 12 non-athletes; age 18-30 years) completed balance assessments using the Biodex Balance System, followed by a 30-minute standardized slump sitting protocol. Outcomes included static postural stability test (PST), Modified Clinical Test of Sensory Interaction on Balance (mCTSIB), and limits of stability (LOS). Group differences were analyzed using analysis of covariance (ANCOVA). Prolonged slump sitting significantly influenced selected postural sway outcomes. Athletes demonstrated greater post-intervention sway compared with non-athletes in the Overall Stability Index (p = 0.01) and the eyes-closed firm-surface mCTSIB condition (p = 0.02). A significant group difference was also observed in static left LOS performance (p = 0.01). In a nutshell, the present study demonstrated that prolonged slump sitting may influence postural sway in female athletes and non-athletes, with specific differences emerging between groups under certain sensory and stability conditions. By highlighting the potential impact of slump sitting on postural stability, this research contributes to the fields of sports science, ergonomics, and rehabilitation, emphasizing the need for strategies that mitigate the negative effects of prolonged sedentary behavior.

The atherogenic index of plasma (AIP) and estimated glucose disposal rate (eGDR) are two composite indices derived from routine metabolic measurements and are associated with cardiocerebrovascular disease risk. In individuals with Cardiovascular-Kidney-Metabolic (CKM) syndrome stages 0-3, however, it remains unclear whether joint stratification by these markers helps summarize gradients of cardiovascular disease, heart disease, and stroke risk beyond single-marker assessment. Using data from the China Health and Retirement Longitudinal Study (CHARLS), 5,925 participants without CVD at the start and in CKM stages 0-3 were analyzed. Participants were grouped by median AIP and/or eGDR values. Kaplan-Meier curves and Cox models assessed the link between these indicators and new CVD, heart disease, and stroke cases. Furthermore, both multiplicative and additive interactions between AIP and eGDR were assessed. The predictive value was assessed using the time-dependent Harrell's C index, integrated discrimination improvement (IDI), and net reclassification improvement (NRI). A cohort of 5,925 participants aged 45 years and older (mean age: 57.92 ± 8.52 years) was analyzed, with 54.65% of the cohort being female. During the nine-year follow-up period, 1,467 (24.76%) participants developed incident CVD, including 1,106 (18.67%) with heart disease and 525 (8.86%) with stroke. The high AIP and low eGDR group had the highest risk, with CVD hazard ratios (HRs) of 1.35 (95% CI 1.14-1.59), heart disease HRs of 1.32 (95% CI 1.08-1.62), and stroke HRs of 1.59 (95% CI 1.19-2.12), using the low AIP and high eGDR group as the reference. Neither multiplicative nor additive interaction was statistically significant. The combined application of AIP and eGDR provided a modest improvement in predictive capability for cardiovascular disease, heart disease, and stroke. In individuals with CKM stages 0-3, combined AIP and eGDR stratification captured gradients of cardiovascular risk. The combined application of these indicators may provide modest incremental value for risk stratification within CKM stages, thereby aiding in the identification of high-risk individuals during the early stages of CKM.

To understand circumstances surrounding transitions between initial drug use, non-prescribed opioid use, and injection drug use (IDU) among people who use non-prescribed opioids in the rural U.S. We interviewed adults who use non-prescribed opioids in 10 states. We coded transcript data regarding age, drug, modality, circumstances for each "first use" event and transitions to illicit opioid use/IDU. Transition-related themes were categorized using the Social Ecological Model (SEM) through iterative qualitative memoing. We calculated frequencies and measures of central tendency. Participants (n = 304, mean age [MA] = 36 years, 56% male) reported marijuana or opioid pills as first drug used (60%, reported MA = 13, and 16%, MA = 19, respectively). First opioid use (MA = 20) was typically pills (74%); 43% were initially prescribed opioids for pain. Of participants reporting IDU (92%, MA = 24), first drugs injected were methamphetamine (26%), opioid pills (22%), and heroin (22%). Reasons for first injection of these three substances were recreational (65%, 84%, 66%, respectively) followed by mental health coping (35%, 16%, 33%). Interviews revealed factors influencing transition to illicit opioid use/IDU at environmental, interpersonal, and intrapersonal levels of the SEM, including early household exposure to drugs, and family/peer encouragement to use opioids for coping. Participants described initial use of prescribed opioids following surgery or workplace injury, followed by illicit use of opioids and later IDU for recreational as well as practical goals. IDU was facilitated by intimate partners and friends in the context of diminished local access to illicit opioid pills. Among people who use opioids and inject drugs in the rural U.S., drug use typically started in pre-adolescence, first opioid use in early adulthood, and IDU a few years later, highlighting the need for early intervention. Methamphetamine comprised a substantial proportion of initial IDU raising concerns for polysubstance use. Drug use for coping highlights the need for increased mental health resources in rural areas.

Kinetic resolution has been a cornerstone for accessing enantioenriched molecules, but its application in radical chemistry has remained elusive due to the high reactivity of radical intermediates. Here, we present a new approach enabling precise Kinetic resolution in radical addition processes, yielding enantioenriched products and recovered starting materials with high efficiency. Two examples are provided: the Kinetic resolution of Minisci reactions between N-heterobiaryls or biaryls and glycine-derived redox-active esters under visible light irradiation with a chiral Brønsted acid catalyst, achieving high yields and enantioselectivities. The second example involves the reductive coupling of aldehydes with N-heterobiaryl-based olefins, enabling efficient synthesis of axially chiral heterobiaryls featuring both axial and remote central chirality. This work represents a conceptual breakthrough in asymmetric radical reactions, inspiring future developments in radical transformations using accessible racemic feedstocks.

In order to understand barriers related to GDMT treatment of hyperkalemia, the RPA convened a nephrology expert panel to review and discuss hyperkalemia across a continuum of care providers and settings. The panel focused on the real-world solutions and identifying opportunities for more education. The panel discussion started with a review of goals, a presentation on the epidemiology of hyperkalemia, and a review of the latest KDIGO Guidelines for Diabetes Management in Chronic Kidney Disease. Following presentations by leading experts in the field, the panelists engaged in a facilitated discussion of incidence and patient management pathways, K+ binders, current KDIGO Guidelines and the ISN Optimization of RAASi Therapy Toolkit, barriers to optimal care, and possible solutions to overcome barriers. Consensus emerged that prescriptive protocols, potassium-restrictive diets and RAASi discontinuation are not appropriate treatments.

The aims of this study were to 1) quantitatively compare various k-space interpolation-based simultaneous multislice (SMS) reconstruction algorithms, including linear GRAPPA and nonlinear, data-driven RAKI methods, and 2) provide an open-source toolbox for GPU-accelerated SMS reconstructions. For a single phantom, fully sampled k-space reference data and SMS-accelerated, RSMS, data were collected for RSMS factors of 2 through 5, with in-plane acceleration, Rip, added retrospectively. Slice-GRAPPA, split-slice-GRAPPA, readout-SENSE-GRAPPA, and their hyperparameter-tuned analogous deep learning-based RAKI reconstructions were performed at different combinations of RSMS and Rip factors. Performance of the reconstruction methods were compared quantitatively by testing for significant differences in whole-image SSIM and regions of interest-based measurements of coefficient of variation (CV). The number of epochs, hidden layer size, and the interaction between the two were statistically significant factors for a Type II ANOVA test with SSIM performance (p=0.00005, p=0.00687, and p=0.000281, respectively) for RAKI reconstructions with post hoc tests suggesting that 500 epochs, 3 hidden layers, and 128 neurons per layer strike an ideal balance of speed and performance for this dataset. CV testing through Kruskal-Wallis tests did not yield any significant differences between reconstruction algorithms at any given RSMS×Rip. Limited testing indicated that a RAKI method generally resulted in larger SSIM values at any net acceleration factor, (Rnet=RSMS×Rip), and all RAKI methods had a larger SSIM value than their respective GRAPPA counterparts at Rnet>4, with the exception of ROSR at RSMS=5 and Rnet=10, and SPSR at RSMS=2,Rip=5. The difference between GRAPPA and RAKI was typically large enough (ΔSSIM of >0.02) to be obvious at Rnet≥8. Slice-GRAPPA, slice-RAKI, or readout-SENSE-GRAPPA are performed well for net accelerations Rnet≤6, and readout SENSE-RAKI or split-slice-RAKI performed well for Rnet≥8. Caution should be taken when using these generalizations as observations were dataset-specific and collected on a single vendor, coil array, pulse sequence, and phantom. They may, however, serve as a useful starting point for hyperparameter tuning. More experiments need to be conducted before these results can be translated into in vivo observations.

ObjectiveTo synthesise and analyse qualitative evidence relevant to the question: What are the experiences and perspectives of healthcare professionals on goal setting in stroke rehabilitation?Data sourcesPubMed, PsycINFO, MEDLINE and CINAHL were systematically searched in May 2025, supplemented by backward and forward citation searching.Review methodsThis systematic review was pre-registered on PROSPERO (CRD420251038210). Eligibility criteria included peer-reviewed qualitative or mixed methods studies with qualitative data from healthcare professionals outlining experiences of goal setting in stroke rehabilitation. Non-English publications were excluded. The Critical Appraisal Skills Programme (CASP) checklist was used to appraise quality. Data were analysed using thematic synthesis.ResultsEight studies, published between 1999 and 2020, were included. These comprised 108 clinicians of various rehabilitation disciplines, from multiple countries, working across acute, inpatient and community settings. Most data were collected via semi-structured interviews. Methodological rigour of identified studies was generally high. Nine descriptive themes emerged from the thematic synthesis. From these descriptive themes, three analytical themes were derived: (1) Who leads, who follows?, (2) Between hope and reality, (3) Starting with the person, not the problem. Eight of the descriptive themes were directly related to analytical themes, whereas one theme was a stand-alone theme. Confidence in the thematic synthesis findings was assessed as moderate.ConclusionThis synthesis of qualitative studies from various rehabilitation settings in stroke found that experiences of goal setting from the perspective of healthcare professionals describe directive and collaborative approaches, emotional aspects of goal setting in time-limited contexts and a commitment to person-centred care.

Despite extensive efforts by scientists, academic institutions, and pharmaceutical companies, a safe and effective HIV/AIDS vaccine remains elusive. Most HIV-1 envelope peptide vaccine strategies have concentrated on Gp120, gp140, or gp160. HIV-1 Env trimer binding to the CD4-receptor initiates structural changes promoting the envelope's transition from a closed to an open state via an intermediate step. Broadly neutralizing antibodies target the state-1 Env conformation, while less effective antibodies typically recognize open states. However, due to virus variability, an optimal vaccine has not yet been successfully developed. In this study, focusing on the pivotal role of Gp41 in various vaccine strategies, a very large sequence dataset was utilized. These sequences were obtained from drug-naïve individuals or those undergoing antiretroviral therapy (ART). Gp41 amino acid variability was characterized genetically using a starting pool dataset of 24,505 full-length Env sequences from HIV-1 Subtype-B infected individuals. The dataset underwent hydropathy analysis, genetic distance evaluation, non-synonymous/synonymous substitution rate estimation, Shannon-Entropy calculation, and N-linked glycosylation (NLG) analysis. Similar variability between viral sequences retrieved from drug-naïve and antiretroviral-treated individuals was observed. In our dataset, ART selection pressures observed at gp41 level are minimal: 7 positions with dN/dS > 1, significant increases in entropy values, and a comparable value of glycosylation sites were highlighted. This study reinforces the importance of identifying specific single sensitizing mutations in HIV control. Gp41 remains an important vaccine target for understanding virus-host immunological interactions. Further analyses may reveal specific mechanisms related to host antiviral responses and viral regions with strong masking activity.

The antibody-drug conjugate Temab-A comprises the c-Met-targeting antibody telisotuzumab conjugated to a novel topoisomerase 1 inhibitor payload, adizutecan. A first-in-human phase I study (ClinicalTrials.gov identifier: NCT05029882) of Temab-A in patients with advanced solid tumors whose disease has progressed is currently ongoing. We report results from all patients in the dose escalation and the monotherapy metastatic colorectal cancer (mCRC) dose expansion. Temab-A was administered intravenously once every 3 weeks as a monotherapy starting at 1.6 mg/kg in dose escalation. In mCRC dose expansion, patients with confirmed BRAF wild-type, microsatellite stable/mismatch repair-proficient mCRC were randomly assigned to 1.6 mg/kg, 2.4 mg/kg, or 3.0 mg/kg Temab-A once every 3 weeks. Primary end points included safety, pharmacokinetics, recommended phase II dose of Temab-A monotherapy, and Temab-A efficacy in patients with mCRC. In total, 57 patients received ≥1 dose of Temab-A in dose escalation; 3.0 mg/kg once every 3 weeks was established as the maximum tolerated dose. Collectively, in dose escalation and dose expansion, 122 patients with mCRC received Temab-A (dose escalation, N = 29; randomized dose optimization expansion, N = 93). All patients experienced ≥1 treatment-emergent adverse event; the most frequent were gastrointestinal (78%) and hematologic (71%) toxicities. Treatment-related discontinuations and deaths were infrequent (10% and 3%, respectively). Across all doses in patients with mCRC, overall response rate was 15.6% (95% CI, 9.6 to 23.2), disease control rate was 74.6% (95% CI, 65.9 to 82.0), and duration of response was 5.9 months (95% CI, 4.1 to 10.5); responses were more frequent at doses of 2.4 mg/kg and 3.0 mg/kg once every 3 weeks. Median progression-free survival was 4.6 months (95% CI: 4.0, 5.4), and median overall survival was 10.4 months (95% CI, 8.9 to 13.1). Temab-A at 2.4 mg/kg once every 3 weeks has a tolerable and manageable safety profile, with promising antitumor activity.

Adolescent pregnancy is increasingly understood as closely associated with pre-existing disadvantage, yet critical gaps remain regarding which adverse childhood experiences (ACEs) are associated with early childbearing, partner characteristics, and parenting trajectories. To examine associations among ACEs, partner characteristics, and parenting practices in adolescent versus adult-onset mothers. We recruited 1019 mothers of children aged 24-48&#xa0;months from poverty-focused programs in Santo Domingo, Dominican Republic (November 2024-January 2025), classified as current adolescent mothers (n&#xa0;=&#xa0;91), former adolescent mothers (n&#xa0;=&#xa0;316), or adult-onset mothers (n&#xa0;=&#xa0;598). Data collected via Audio Computer-Assisted Self-Interview using validated instruments: ACE questionnaire, HITS violence screen, Edinburgh Postnatal Depression Scale, and standardized parenting assessments. Analyses included correlations, chi-square tests, t-tests, ANOVA, and regression models. Emotional neglect and physical neglect showed the most consistent associations with younger maternal age at first birth, surviving FDR correction across the 10 ACE indicators (r&#xa0;=&#xa0;-0.11, pFDR&#xa0;=&#xa0;0.007 and r&#xa0;=&#xa0;-0.09, pFDR&#xa0;=&#xa0;0.021, respectively). In a simultaneous regression model including all 10 ACE indicators, emotional neglect remained independently associated with earlier childbearing (&#x3b2;&#xa0;=&#xa0;-0.96, p&#xa0;=&#xa0;.014); no other ACE indicators reached significance after accounting for co-occurring adversities. Mothers who started childbearing before 18 had partners who averaged 9.1&#xa0;years older, compared with 2.5&#xa0;years for mothers &#x2265;26 (p&#xa0;<&#xa0;.001, &#x3b7;p2&#xa0;=&#xa0;0.089); partners also had lower educational attainment (40.4% primary-only vs 13.0%). ACE scores were independently associated with intimate partner violence and depression after covariate adjustment; maternal age was not. Adult-onset mothers maintained 2.85 times the odds of university education after age adjustment. Former adolescent mothers used less violent discipline than adult-onset mothers (d&#xa0;=&#xa0;-0.22, p&#xa0;<&#xa0;.001), robust to adjustment. The findings support the view that preexisting adversity rather than early childbearing is associated with higher psychosocial risk. Educational inequalities persisted after age adjustment. Former adolescent mothers' lower use of violent discipline is consistent with possible adaptation, though alternative explanations cannot be excluded. Evidence before this study. We searched PubMed, EBSCOhost, EMBASE, and SciSpace online databases for "adolescent pregnancy" and "early childhood development." We included publications in any language between 2015 and 2024 and screened 283 studies, plus 23 earlier studies cited in these publications; we excluded all other publications. Most examined cumulative ACE scores rather than specific adversities. Only 11 studies included the child's biological father data, focusing solely on age differences without educational characteristics. No studies examined whether parenting deficits persist as adolescent mothers mature or linked childhood emotional neglect to partner characteristics. The selection effects hypothesis suggests that preexisting disadvantage may explain poor outcomes, but evidence remained limited regarding specific mechanisms and long-term trajectories. Added value of this study. We identify specific ACEs-emotional neglect and physical neglect-that show stronger associations with early childbearing than cumulative ACE scores. We document that adolescent mothers' partners average 9.1&#xa0;years older at the time of the first childbirth and have lower education (40.4% primary-only versus 13.0% for older mothers' partners). In exploratory analyses, different childhood adversities show opposing associations with partner characteristics: feeling unloved was associated with younger partners, while sexual abuse and witnessing maternal intimate partner violence were associated with older partners, independent of reproductive timing-associations that did not survive correction for multiple comparisons and require replication. Comparing former adolescent mothers with adult-onset mothers, we find that former adolescent mothers use less violent discipline despite persistent adversity-a pattern consistent with a possible adaptation that was robust to covariate adjustment-while educational inequalities persist after age adjustment, with adult-onset mothers having 2.85 times the odds of university education compared to former adolescent mothers. Implications of all the available evidence. Adolescent pregnancy may be better understood as closely associated with pre-existing disadvantage rather than as an independent source of disadvantage. Prevention efforts may benefit from addressing childhood trauma, particularly emotional and physical neglect, rather than focusing solely on pregnancy prevention. Programs should consider how to address power dynamics in age-asymmetric relationships while expanding opportunities for young women. The lower use of violent discipline among former adolescent mothers, if replicated, would challenge deficit perspectives and support strengths-based approaches. Healthcare systems may consider screening for childhood adversity and relationship dynamics during prenatal care as one strategy to identify women who could benefit from trauma-informed services. Future longitudinal research should examine adaptation trajectories and protective factors that enable positive outcomes despite adversity.

Optical Chemical Structure Recognition (OCSR) aims to convert two-dimensional molecular images into machine-readable formats such as SMILES strings. Deep learning has substantially improved OCSR performance, yet most methods rely on synthetic training data and struggle to generalize to real-world inputs, especially hand-drawn diagrams, where stroke width, geometry, and drawing conventions vary widely across individuals. In this work, we propose an image-to-graph model AdaptMol that enables effective transfer from synthetic to real-world data without requiring manual graph annotations in the target domains. AdaptMol is an integrated pipeline that starts with training a base model on synthetic data, and then refines model representations through unsupervised domain adaptation and self-training. Our key insight is that bond features are domain-invariant in nature; they encode structural relationships between atoms that are independent of visual variations across domains. Thus, during domain adaptation, we align bond-level feature distributions via class-conditional Maximum Mean Discrepancy (MMD) to enforce cross-domain consistency. We also design a comprehensive data augmentation strategy to enhance the robustness of the base model, facilitating stable self-training on unlabeled target samples. On hand-drawn molecular images, our model achieves 82.6% accuracy and outperforms the best prior method by 10.7 points, while maintaining competitive performance across four benchmarks comprising molecular images from scientific literature and patent documents.Scientific contributionWe propose AdaptMol, an image-to-graph model that predicts molecular structures as graphs of atoms and bonds, achieving effective transfer from synthetic to hand-drawn molecular images without requiring target domain graph annotations. We combine class-conditional Maximum Mean Discrepancy to align bond features across domains with comprehensive data augmentation to increase training data variation, jointly improving base model accuracy sufficiently for self-training and addressing the critical failure mode of prior approaches that begin with insufficient accuracy. We further introduce a dual position representation that supervises atom positions through both discrete coordinate tokens and continuous spatial heatmaps to reduce false positives in atom localization.

Living tissues strengthen under repeated mechanical loading, yet replicating such adaptive growth in synthetic materials remains a formidable challenge. Here, we report a protein-based hydrogel that undergoes mechanochemically induced self-growth, autonomously reinforcing its baseline mechanical properties under applied stress. This strategy harnesses the copper-storage protein Csp1, whose force-regulated unfolding releases Cu(I) that catalyzes in situ azide-alkyne cycloaddition, generating secondary crosslinks under mechanical load. Upon unloading, Csp1 refolds and re-sequesters Cu(I), halting catalysis and restoring growth capacity. This mechano-catalytic feedback loop enables stress- and time-dependent self-reinforcement within a closed system, without external monomer supply. The hydrogel exhibits programmable mechanical memory via leveraging Cu(I) homeostasis in cyclic growth-pause-growth transitions. By coupling force-dependent protein conformational dynamics with catalytic activity, this strategy establishes a generalizable mechanochemical framework for designing self-adapting biomaterials whose structure and function evolve under mechanical stimulation.

The classification of functional brain networks plays an important role in the diagnosis of neurodegenerative diseases, brain decoding and other fields. Functional brain networks can effectively reflect the functional connection relationships between brain regions or neurons and accurately represent brain activities. Therefore, a large number of problems related to the classification of functional brain networks have been studied. However, the traditional functional brain network merely measures the static correlation between brain regions or neurons in a simple way, and does not reflect the causal transmission effect between brain regions. This directionality is crucial for the regulatory relationship between brain regions. Furthermore, since the brain is constantly in a state of dynamic change, the dynamics of functional connectivity also plays a very important role in the classification of functional brain networks. Therefore, we propose a classification framework named Dynamic Directed Propagation Networks (DDPN) for functional brain networks considering the dynamic directed propagation mechanism. This method effectively captures the dynamics and directionality of the dynamic directed brain network and further improves the classification accuracy of the functional brain network. To verify the effectiveness of the proposed method, we conduct experiments on real datasets. The experiments show that the proposed method improved by 3.1-4.1% compared with state-of-the art methods in two datasets.

Chemoresistance remains a major cause of treatment failure in colorectal cancer (CRC), yet the metabolic mechanisms sustaining efflux-mediated drug resistance are not fully defined. Here, we identify ATP-citrate lyase (ACLY) as a metabolic regulator linking citrate-dependent acetyl-CoA production to epigenetic control of MDR1/ABCB1 expression. Using genetic and pharmacologic approaches, we show that ACLY catalytic activity contributes to the maintenance of histone acetylation at H3K9 and H4K16 and supports MDR1 transcription in CRC cells. Consistently, ACLY overexpression enhances, whereas its inhibition reduces, MDR1 expression and associated resistance-related transcriptional programs. In human CRC specimens, ACLY and MDR1 levels positively correlate, with a stronger association observed in advanced-stage tumors, supporting clinical relevance of this metabolic-epigenetic axis. Metabolic tracing with 13C-glucose suggests that perturbation of citrate flux influences ACLY-associated pathways and acetyl-CoA availability. In this context, vitamin C treatment reduces citrate-derived acetyl-CoA and ACLY phosphorylation and is associated with global histone deacetylation and decreased MDR1 expression in vitro and in KRAS-mutant patient-derived xenografts. Together, these findings highlight ACLY-dependent acetyl-CoA production as a potential metabolic vulnerability linked to epigenetic regulation of drug efflux programs in CRC. Targeting this metabolic-chromatin axis may represent a strategy to modulate MDR1-associated chemoresistance.

The SARS-CoV-2 main protease (Mpro, also known as 3CLpro) is an attractive antiviral drug target due to its essential role in viral replication and absence of human homologues. Development of new coronavirus-specific Mpro inhibitors will be important as SARS-CoV-2 continues to evolve. Leveraging the rapidly expanding pool of diverse, experimental Mpro-inhibitor data, we developed a target-specific deep learning workflow to accelerate the discovery of new Mpro inhibitor compounds and fragment-like starting points. This workflow combined a fine-tuned inhibitor prediction model with solubility (logS) and lipophilicity (logP) models, molecular similarity analysis, and literature mining to prioritize novel, drug-like candidates. Applied to a purchasable library of over 500,000 compounds, the approach rapidly identified 24 candidates for experimental testing. Biochemical assays revealed a novel, small covalent inhibitor fragment (A02) with an apparent IC50 of 1.5 &#x3bc;M, prior to any synthetic optimization or derivatization. A 1.76 &#xc5; crystal structure of Mpro bound to A02 confirmed covalent modification of the catalytic Mpro cysteine (C145), unique engagement of the underutilized Mpro S3' pocket, and the potential for derivatives of this scaffold to interact with additional Mpro pockets in future optimization efforts. Together, these results demonstrate the potential for target-specific deep learning approaches to guide the rapid screening and discovery of new inhibitor leads or drug scaffolds.