Explore the vast range of titles available.

Background Preeclampsia (PE), a multisystem hypertensive disorder complicating ~ 5% of pregnancies, remains a major global cause of maternal and perinatal morbidity and mortality. First-trimester screening methods are limited in accessibility and sensitivity, underscoring the need for affordable biomarkers. The pan-immune-inflammation value (PIV), derived from routine blood counts, shows potential as a predictor, but its independent predictive utility and dose–response association with PE are not yet fully defined. Methods In this retrospective cohort study, the association between first-trimester PIV and subsequent development of PE was evaluated. Multivariable logistic regression was used to assess PIV as an independent predictor, adjusting for confounders including maternal age, body mass index (BMI), parity, multifetal pregnancy and mode of conception (natural vs. in vitro fertilization [IVF, a method of assisted reproductive technology]). Dose–response patterns were explored using restricted cubic spline and two-piecewise linear regression models. Subgroup and interaction analyses were conducted to assess consistency across clinical strata. To improve interpretability, all PIV values in the study were divided by 100 during analysis and result presentation. Results A total of 11,894 pregnant women were included (non-PE 11409, PE 485), among whom elevated first-trimester PIV/100 was significantly associated with an increased risk of PE (adjusted OR, 1.076; 95% CI, 1.041–1.112; P  < 0.001). A dose-dependent increase in PE prevalence was observed across PIV quartiles, with the highest quartile associated with a 60.2% increased risk (OR, 1.602; 95% CI, 1.210–2.119; P < 0.001) compared to the lowest. Restricted cubic spline analysis suggested a nonlinear association in unadjusted models; however, an adequate linear fit was confirmed after adjustment ( P for non-linearity > 0.05). Two-piecewise linear regression identified an inflection point at PIV/100 = 3.6889, but no significant threshold effect was detected ( P  = 0.274). The association between PIV/100 and PE remained consistent across subgroups defined by age, BMI, parity, IVF status, and plurality (all P for interaction > 0.05). Conclusion Elevated first-trimester PIV is independently and linearly associated with increased risk of PE, with no identifiable threshold effect. These findings highlight the potential utility of PIV as an accessible, inflammation-based biomarker for early PE risk stratification. Future prospective studies are warranted to validate its clinical applicability and explore integration into prenatal screening protocols.

The Caco-2 cell model has been widely used to assess the intestinal permeability of drug candidates in vitro , owing to its morphological and functional similarity to human enterocytes. While Caco-2 cell assay is considered safe and cost-effective, it is also characterized by being time-consuming. Therefore, computational models that achieve high accuracies in predicting Caco-2 permeability are crucial for enhancing the efficiency of oral drug development. In this study, we conducted an in-depth analysis of the characteristics of an augmented Caco-2 permeability dataset, and evaluated a diverse range of machine learning algorithms in combination with different molecular representations. The results indicated that XGBoost generally provided better predictions than comparable models for the test sets. In addition, we investigated the transferability of machine learning models trained on publicly available data to internal pharmaceutical industry datasets. Our findings, based on the Shanghai Qilu’s in-house dataset, showed that the boosting models retained a degree of predictive efficacy when applied to industry data. Furthermore, Y-randomization test and applicability domain analysis were employed to assess the robustness and generalizability of these models. Matched Molecular Pair Analysis (MMPA) was utilized to extract chemical transformation rules. We believe that the model developed in this study could represent a reliable tool for assessing Caco-2 permeability during early-stage drug discovery and the chemical transformation rules derived here could provide insights for optimizing Caco-2 permeability. Scientific contribution A comprehensive validation of various machine learning algorithms combined with diverse molecular representations on a large dataset for predicting Caco-2 permeability was reported. The transferability of machine learning models trained on publicly available data to internal pharmaceutical industry datasets was also investigated. Matched molecular pair analysis was carried out to provide reasonable suggestions for researchers to improve the Caco-2 permeability of compounds. Graphical Abstract

Objective: To assess the association between lipid metabolism and fetal fraction, which is a critical factor in ensuring a highly accurate non-invasive prenatal testing (NIPT), and on the rate of screen failures or “no calls” in NIPT. Methods: A total of 4,514 pregnant women at 12–26 weeks of gestation underwent NIPT sequencing and serum lipid measurements. Univariate analysis and multivariate regression models were used to evaluate the associations of serum lipid concentrations with the fetal fraction and the rate of screen failures. Results: The fetal fraction decreased with increased low-density lipoprotein cholesterol and triglyceride (TG) levels, which were significant factors (standardized coefficient: −0.11). Conversely, high-density lipoprotein cholesterol and the interval between the two tests were positively correlated with the fetal fraction. The median fetal fraction was 10.88% (interquartile range, 8.28–13.89%) and this decreased with TG from 11.56% at ≤1.10 mmol/L to 9.51% at >2.30 mmol/L. Meanwhile, multivariate logistic regression analysis revealed that increased TG levels were independently associated with the risk of screen failures. The rate of screen failures showed an increase with TG levels from 1.20% at ≤1.70 mmol/L to 2.41% at >2.30 mmol/L. Conclusions: The fetal fraction and the rate of screen failures in NIPT are affected by TG levels. Meanwhile, in pregnant women with high TG levels, delaying the time between NIPT blood collections can significantly increase the fetal fraction.

ObjectiveTo determine the effects of alanine transaminase (ALT) levels on the screening failure rates or “no calls” due to low fetal fraction (FF) to obtain a result in non-invasive prenatal screening (NIPS).MethodsNIPS by sequencing and liver enzyme measurements were performed in 7,910 pregnancies at 12–26 weeks of gestation. Univariate and multivariable regression models were used to evaluate the significant predictors of screening failure rates among maternal characteristics and relevant laboratory parameters.ResultsOf the 7,910 pregnancies that met the inclusion criteria, 134 (1.69%) had “no calls.” Multiple logistic regression analysis demonstrated that increased body mass index, ALT, prealbumin, albumin levels, and in vitro fertilization (IVF) conception rates were independently associated with screening failures. The test failure rate was higher (4.34 vs. 1.41%; P < 0.001) in IVF pregnancies relative to those with spontaneous conceptions. Meanwhile, the screening failure rates increased with increasing ALT levels from 1.05% at ≤10 U/L to 3.73% at >40 U/L. In particular, IVF pregnancies with an ALT level of >40 U/L had a higher test failure rate (9.52%). Compared with that for an ALT level of ≤10 U/L, the adjusted odds ratio of “no calls” for ALT levels of 10–20, 21–40, and >40 U/L was 1.204 [95% confidence interval (CI), 0.709–2.045], 1.529 (95% CI, 0.865–2.702), and 2.764 (95% CI, 1.500–5.093) ( P trend < 0.001), respectively.ConclusionsIncreased ALT and IVF conceptions were associated with a higher screening failure rates in NIPS. Therefore, a feasible strategy to adjust these factors to reduce the probability of “no calls” due to low FF would be of great clinical significance.

preeclampsia (PE) remains a leading cause of maternal and perinatal morbidity and mortality worldwide. While metabolic and inflammatory factors are increasingly recognized in its pathogenesis, the clinical utility of composite biomarkers remains underexplored. This study aimed to investigate the association between the C-reactive protein-triglyceride-glucose (CRP-TG-glucose) index (CTGI), a novel marker of metabolic-inflammation stress, and the risk of preeclampsia. This retrospective cohort study included 11,916 pregnant women, of whom 486 developed preeclampsia. Maternal baseline characteristics were compared between the PE and non-PE groups. Logistic regression analyses were conducted to identify factors associated with PE. The relationship between CTGI and PE risk was further explored using quartile stratification, restricted cubic spline regression, and threshold effect analyses. Subgroup analyses were also performed to assess interaction effects across maternal and obstetric variables. Women with PE had significantly higher maternal age, body mass index (BMI), in vitro fertilization (IVF) conception, multifetal pregnancies, and elevated CTGI levels compared to non-PE counterparts (all P < .001). Multivariate logistic regression identified CTGI as an independent risk factor for PE (adjusted OR, 1.78; 95% CI, 1.51-2.09; P < .001), alongside BMI, maternal age, IVF, and multifetal gestation. A dose-response relationship was observed across CTGI quartiles, with the highest quartile showing a markedly increased PE risk (adjusted OR, 2.06; 95% CI, 1.52-2.81). Restricted cubic spline models and threshold analysis revealed a nonlinear association with a significant inflection point at CTGI = 2.244. Above this threshold, the risk of PE rose sharply (OR, 3.93; 95% CI, 2.09-7.39; P < .001). Subgroup analyses demonstrated consistent associations across maternal age, BMI, parity, plurality, and IVF status, without significant interaction. Elevated CTGI in early pregnancy is independently and nonlinearly associated with an increased risk of preeclampsia, particularly above a critical threshold of 2.244. These findings underscore the potential clinical value of CTGI as an early risk stratification biomarker for PE, enabling timely intervention in high-risk pregnancies.

Most molecular generative models based on artificial intelligence for de novo drug design are ligand-centric and do not consider the detailed three-dimensional geometries of protein binding pockets. Pocket-aware three-dimensional molecular generation is challenging due to the need to impose physical equivariance and to evaluate protein–ligand interactions when incrementally growing partially built molecules. Inspired by multiscale modelling in condensed matter and statistical physics, we present a three-dimensional molecular generative model conditioned on protein pockets, termed ResGen, for designing organic molecules inside of a given target. ResGen is built on the principle of parallel multiscale modelling, which can capture higher-level interaction and achieve higher computational efficiency (about eight-times faster than the previous best art). The generation process is formulated as a hierarchical autoregression, that is, a global autoregression for learning protein–ligand interactions and atomic component autoregression for learning each atom’s topology and geometry distributions. We demonstrate that ResGen has a higher success rate than existing state-of-the-art approaches in generating novel molecules that can bind to unseen targets more tightly than the original ligands. Moreover, retrospective computational experiments on de novo drug design in real-world scenarios show that ResGen successfully generates drug-like molecules with lower binding energy and higher diversity than state-of-the-art approaches. Generating novel molecules that bind to specific protein targets is a challenging but important task in computational drug design. Zhang and colleagues present a molecular generation method based on hierarchical auto-regression.

Background Preeclampsia (PE) affects 2–4% of pregnancies, and early detection and intervention can reduce its incidence. Dysregulation of the maternal immune response and red blood cells (RBCs) are key to its development, although early alterations remain unclear. Methods This study analyzed data from 17,955 pregnant women across two centers to explore the relationships among inflammatory markers, RBC indices, and PE via multivariate logistic regression and restricted cubic splines (RCSs). Machine learning integrated inflammatory markers, RBC indices, and maternal risk factors to predict PE risk at 14 weeks, as validated by receiver operating characteristic (ROC) curve analysis. Results After adjusting for confounders, the lymphocyte (LYMPH) count (OR = 1.27, 95% CI: 1.05–1.53, P  = 0.013), monocyte (MONO) count (OR = 2.57, 95% CI: 1.31–5.03, P  = 0.006), systemic inflammatory response index (SIRI) (OR = 1.11, 95% CI: 1.01–1.21, P  = 0.032), and systemic immune inflammatory index (SII) (OR = 1.01, 95% CI: 1.01–1.01, P  = 0.002) were identified as significant risk factors for PE. Nonlinear associations between white blood cell (WBC) count, neutrophil (NEUT) count, platelet (PLT) count, RBC count, and hemoglobin (HGB) and PE were observed via RCS (nonlinear P  < 0.05). Further analysis revealed threshold effects for WBC ( P  = 0.034), with an inflection point at 8.44. Below 8.44, no significant association was found (OR = 0.92, P  = 0.307), but above 8.44, each unit increase was linked to a 0.14-fold rise in PE risk (OR = 1.14, P  < 0.001). Similar threshold effects were found for the PLT, RBC, and HGB ( P  < 0.001). A prediction model based on inflammatory markers, RBC indices, and maternal risk factors achieved high performance (ROC = 0.82). Conclusions LYMPH, MONO, SIRI, and SII were linearly associated with PE, whereas WBC, NEUT, PLT, RBC, and HGB showed nonlinear associations with threshold effects. Early prediction using these indicators is a cost-effective strategy for PE prevention.

Accurate tourist flow forecasting is an important issue in tourist destinations management. Given the influence of various factors on varying degrees, tourist flow with strong nonlinear characteristics is difficult to forecast accurately. In this study, a deep learning method, namely, Gated Recurrent Unit (GRU) is used for the first time for tourist flow forecasting. GRU captures long-term dependencies efficiently. However, GRU’s ability to pay attention to the characteristics of sub-windows within different related factors is insufficient. Therefore, this study proposes an improved attention mechanism with a horizontal weighting method based on related factors importance. This improved attention mechanism is introduced to the encoding–decoding framework and combined with GRU. A competitive random search is also used to generate the optimal parameter combination at the attention layer. In addition, we validate the application of web search index and climate comfort in prediction. This study utilizes the tourist flow of the famous Huangshan Scenic Area in China as the research subject. Experimental results show that compared with other basic models, the proposed Improved Attention-based Gated Recurrent Unit (IA-GRU) model that includes web search index and climate comfort has better prediction abilities that can provide a more reliable basis for tourist destinations management.

While separately leveraging monocular 3D object detection and 2D multi-object tracking can be straightforwardly applied to sequence images in a frame-by-frame fashion, stand-alone tracker cuts off the transmission of the uncertainty from the 3D detector to tracking while cannot pass tracking error differentials back to the 3D detector. In this work, we propose jointly training 3D detection and 3D tracking from only monocular videos in an end-to-end manner. The key component is a novel spatial-temporal information flow module that aggregates geometric and appearance features to predict robust similarity scores across all objects in current and past frames. Specifically, we leverage the attention mechanism of the transformer, in which self-attention aggregates the spatial information in a specific frame, and cross-attention exploits relation and affinities of all objects in the temporal domain of sequence frames. The affinities are then supervised to estimate the trajectory and guide the flow of information between corresponding 3D objects. In addition, we propose a temporal -consistency loss that explicitly involves 3D target motion modeling into the learning, making the 3D trajectory smooth in the world coordinate system. Time3D achieves 21.4\\% AMOTA, 13.6\\% AMOTP on the nuScenes 3D tracking benchmark, surpassing all published competitors, and running at 38 FPS, while Time3D achieves 31.2\\% mAP, 39.4\\% NDS on the nuScenes 3D detection benchmark.

Structure-based molecule generation represents a significant advancement in AI-aided drug design (AIDD). However, progress in this domain is constrained by the scarcity of structural data on protein-ligand complexes, a challenge we term the Paradox of Sparse Chemical Space Generation. To address this limitation, we propose a novel latent variable approach that bridges the data gap between ligand-only and protein-ligand complexes, enabling the target-aware generative models to explore a broader chemical space and enhancing the quality of molecular generation. Drawing inspiration from quantum molecular simulations, we introduce ECloudGen, a generative model that leverages electron clouds as meaningful latent variables—an innovative integration of physical principles into deep learning frameworks. ECloudGen incorporates modern techniques, including latent diffusion models, Llama architectures, and a newly proposed contrastive learning task, which organizes the chemical space into a structured and highly interpretable latent representation. Benchmark studies demonstrate that ECloudGen outperforms state-of-the-art methods by generating more potent binders with superior physiochemical properties and by covering a significantly broader chemical space. The incorporation of electron clouds as latent variables not only improves generative performance but also introduces model-level interpretability, as illustrated in a case study designing V2R inhibitors. Furthermore, ECloudGen’s structurally ordered modeling of chemical space enables the development of a model-agnostic optimizer, extending its utility to molecular optimization tasks. This capability has been validated through a single-objective oracle benchmark and a complex multi-objective optimization scenario involving the redesign of endogenous BRD4 ligands. In conclusion, ECloudGen effectively addresses the Paradox of Sparse Chemical Space Generation through its integration of theoretical insights, advanced generative techniques, and real-world validation. The newly proposed technique of leveraging physical entities (such as electron clouds) as latent variables within a deep learning framework may prove useful for computational biology fields beyond AIDD.