Additive Model Research Articles

Multivariate and hybrid data-driven models to predict thermoelectric power plants fuel consumption

Ensuring the reliable operation of diesel/Heavy fuel oil (HFO) engines in Thermoelectric Power Plants (TPPs) requires constant monitoring and control of operational parameters. Time series forecasting methods can predict physical system characteristics based on historical data, which can help regulate engine operational parameters. Despite the widespread use of these techniques in coal and natural gas powered TPPs, further examination of their application in large diesel/HFO engines in Brazilian TPPs operating under dispatch conditions is needed. This work investigated the fuel consumption generalization capacity of linear, nonlinear, and hybrid prediction models, both considering univariate and multivariate approaches, related to a TPP engine-driven generator in Pernambuco, Brazil. For multivariate model development, exogenous variables were selected based on change of causality overtime analysis, which sought not only to infer the relationship between signals but also to identify the influence regime of each causality during operations. The feature selection step initially identified a performance improvement when eight additional features were used. However, each feature causality throughout the operations indicated that only four signals helped significantly predict consumption in different ways. The exogenous variables introduction to ARIMA and NAR models resulted in significant improvements only to the nonlinear approach, NARX, which could recognize operational disturbances more quickly when compared to other analyzed models. In addition, the application of additive and multivariate hybrid models provided the ability to detect more complex variations and, at the same time, maintain model stability during full-load operation, benefiting from linear and nonlinear characteristics capturing ability related to the combined model.

Machine learning and deep learning to identifying subarachnoid haemorrhage macrophage-associated biomarkers by bulk and single-cell sequencing.

We investigated subarachnoid haemorrhage (SAH) macrophage subpopulations and identified relevant key genes for improving diagnostic and therapeutic strategies. SAH rat models were established, and brain tissue samples underwent single-cell transcriptome sequencing and bulk RNA-seq. Using single-cell data, distinct macrophage subpopulations, including a unique SAH subset, were identified. The hdWGCNA method revealed 160 key macrophage-related genes. Univariate analysis and lasso regression selected 10 genes for constructing a diagnostic model. Machine learning algorithms facilitated model development. Cellular infiltration was assessed using the MCPcounter algorithm, and a heatmap integrated cell abundance and gene expression. A 3 × 3 convolutional neural network created an additional diagnostic model, while molecular docking identified potential drugs. The diagnostic model based on the 10 selected genes achieved excellent performance, with an AUC of 1 in both training and validation datasets. The heatmap, combining cell abundance and gene expression, provided insights into SAH cellular composition. The convolutional neural network model exhibited a sensitivity and specificity of 1 in both datasets. Additionally, CD14, GPNMB, SPP1 and PRDX5 were specifically expressed in SAH-associated macrophages, highlighting its potential as a therapeutic target. Network pharmacology analysis identified some targeting drugs for SAH treatment. Our study characterised SAH macrophage subpopulations and identified key associated genes. We developed a robust diagnostic model and recognised CD14, GPNMB, SPP1 and PRDX5 as potential therapeutic targets. Further experiments and clinical investigations are needed to validate these findings and explore the clinical implications of targets in SAH treatment.

An acoustic study on age-related changes in vowel production of Chinese

This paper investigates the relationship between vowel production and age using speech data from 109 Chinese L1 speakers (61 females and 48 males) covering an age range of 20 to 80 years. Acoustical estimation of vocal tract length (VTL) as well as multiple acoustic metrics are analyzed with generalized additive mixed models (GAMM). The results indicate that: (1) After controlling for VTL, F0 and duration, vowels show a centralization trend with increasing age, with a more significant effect observed in female speakers; (2) VTL does not significantly change with age; (3) The patterns observed in vowel distinctiveness and duration may present evidence contradicting the notion of vowel lengthening as a compensatory mechanism; (4) The patterns of age-related changes in different measurements and different genders are diverse. The U-shaped change patterns are found in the male speakers and the age around 50 may serve as a turning point. Based on these findings, we have explored some possible reasons for inconsistent conclusions in previous studies. The physiological aging phenomena of vowel production and potential compensatory mechanisms on motor control abilities, as well as other possible influencing factors, are also discussed.

Forecasting short-term methane based on corrected numerical weather prediction outputs

Methane (CH4) represents a significant greenhouse gas, and the control of its emissions is crucial in impacting global climate change. Accurate forecasting of CH4 emissions is important for identifying future sources of high CH4 emissions, rational design of experiments, and information gathering. This study presented a method to improve numerical weather prediction (NWP) outputs using the Extreme Gradient Boosting (XGB) algorithm. In contrast to previous conventional correction methods, the XGB model greatly enhances the accuracy and stability of the NWP outputs. Short-term (1-3 d) CH4 forecasting models were built for 53 stations worldwide using three machine learning algorithms: Convolutional Neural Network (CNN), Random Forest (RF), and XGB. The models were validated with flux data from the stations. The results showed that the CNN model performed best in forecasting CH4 for all lead times (mean NRMSE=0.34), followed by the RF model (mean NRMSE=0.68) and the XGB model performed worst (mean NRMSE=0.74). For land cover types (e.g., savanna (WSA) and forest (EBF)), the accuracy of the model may be greater for 2 d lead times than for 1 d lead times, which is closely related to the regional climate and terrain conditions. For lead times, the CNN model achieves high CH4 forecast accuracy at 1-2 d lead times. In addition, the RF and XGB models with 1 d lead times exhibit significant overestimation at most stations in North America (PBIAS>0.5). This work quantified the performance of machine learning models for correcting NWP outputs, providing a technical approach to the correction of NWP outputs; and assessed the performance of deep learning models for forecasting short-term CH4, providing a reference for model selection for CH4 forecasting.

A data and physical model dual-driven based trajectory estimator for long-term navigation

Long-term navigation ability based on consumer-level wearable inertial sensors plays an essential role towards various emerging fields for instance smart healthcare, emergency rescue, soldier positioning et al. The performance of existing long-term navigation algorithm is limited by the cumulative error of inertial sensors, disturbed local magnetic field, and complex motion modes of the pedestrian. This paper develops a robust data and physical model dual-driven based trajectory estimation (DPDD-TE) framework, which can be applied for long-term navigation tasks. A Bi-directional Long Short-Term Memory (Bi-LSTM) based quasi-static magnetic field (QSMF) detection algorithm is developed for extracting useful magnetic observation for heading calibration, and another Bi-LSTM is adopted for walking speed estimation by considering hybrid human motion information under a specific time period. In addition, a data and physical model dual-driven based multi-source fusion model is proposed to integrate basic INS mechanization and multi-level constraint and observations for maintaining accuracy under long-term navigation tasks, and enhanced by the magnetic and trajectory features assisted loop detection algorithm. Real-world experiments indicate that the proposed DPDD-TE outperforms than existing algorithms, and final estimated heading and positioning accuracy indexes reaches 5° and less than 2 m under the time period of 30 minutes, respectively.

YBX1, Targeted By Microrna-382-5p, Promotes Laryngeal Squamous Cell Carcinoma Progression via Modulating RAS/MAPK Signaling.

Laryngeal squamous cell carcinoma (LSCC) is the most common cancer of head and neck cancer. Y-box binding protein-1 (YBX1) has tumor-promoting effects in some types of cancers. However, its role in LSCC remains unknown. This study set out to identify the role of YBX1 in LSCC. Bioinformatics analysis of the Gene Expression Omnibus (GEO) database and our cohort data were used to explore the association of YBX1 expression with clinicopathological factors in LSCC. Then, cells with stably or transiently transfected with plasmid or siRNA were constructed to assess the effect of loss and gain of YBX1 on the biological phenotypes of LSCC cells in vitro. In addition, subcutaneous xenograft and orthotopic liver tumor mouse models were constructed for validation. The interrogated miRNA databases and subsequent luciferase reporter assays were used to confirm the miR-382-5p target of YBX1. At last, KEGG enrichment annotation from TGCA data was used for downstream analyses of miR-382-5p/YBX1 and verified by PCR and Western immunoblotting. The results showed that significant upregulation of YBX1 in LSCC tumors was correlated with advanced TNM stage and poor prognosis. Knockdown of YBX1 markedly impaired the proliferative, invasive, and migratory activity of Tu212 cells. We confirmed that miR-382-5p targets YBX1 to mediate LSCC progression both in vitro and in vivo. We further confirmed that miR-382-5p/YBX1 modulated the Ras/MAPK signaling axis to regulate the progression of LSCC. Together, our results indicated that YBX1 is an important promoter of LSCC progression. And miR-382-5p/YBX1/RAS/MAPK signaling pathway can be perceived as a promising target in the treatment of LSCC.

Doppler ultrasound surveillance of recently formed haemodialysis arteriovenous fistula: the SONAR observational cohort study.

Arteriovenous fistulas are considered the best option for haemodialysis provision, but as many as 30% fail to mature or suffer early failure. To assess the feasibility of performing a randomised controlled trial that examines whether, by informing early and effective salvage intervention of fistulas that would otherwise fail, Doppler ultrasound surveillance of developing arteriovenous fistulas improves longer-term arteriovenous fistula patency. A prospective multicentre observational cohort study (the 'SONAR' study). Seventeen haemodialysis centres in the UK. Consenting adults with end-stage renal disease who were scheduled to have an arteriovenous fistula created. Participants underwent Doppler ultrasound surveillance of their arteriovenous fistulas at 2, 4, 6 and 10 weeks after creation, with clinical teams blinded to the ultrasound surveillance findings. Fistula maturation at week 10 defined according to ultrasound surveillance parameters of representative venous diameter and blood flow (wrist arteriovenous fistulas: ≥ 4 mm and > 400 ml/minute; elbow arteriovenous fistulas: ≥ 5 mm and > 500 ml/minute). Mixed multivariable logistic regression modelling of the early ultrasound scan data was used to predict arteriovenous fistula non-maturation by 10 weeks and fistula failure at 6 months. A total of 333 arteriovenous fistulas were created during the study window (47.7% wrist, 52.3% elbow). By 2 weeks, 37 (11.1%) arteriovenous fistulas had failed (thrombosed), but by 10 weeks, 219 of 333 (65.8%) of created arteriovenous fistulas had reached maturity (60.4% wrist, 67.2% elbow). Persistently lower flow rates and venous diameters were observed in those fistulas that did not mature. Models for arteriovenous fistulas' non-maturation could be optimally constructed using the week 4 scan data, with fistula venous diameter and flow rate the most significant variables in explaining wrist fistula maturity failure (positive predictive value 60.6%, 95% confidence interval 43.9% to 77.3%), whereas resistance index and flow rate were most significant for elbow arteriovenous fistulas (positive predictive value 66.7%, 95% confidence interval 48.9% to 84.4%). In contrast to non-maturation, both models predicted fistula maturation much more reliably [negative predictive values of 95.4% (95% confidence interval 91.0% to 99.8%) and 95.6% (95% confidence interval 91.8% to 99.4%) for wrist and elbow, respectively]. Additional follow-up and modelling on a subset (n = 192) of the original SONAR cohort (the SONAR-12M study) revealed the rates of primary, assisted primary and secondary patency arteriovenous fistulas at 6 months were 76.5, 80.7 and 83.3, respectively. Fistula vein size, flow rate and resistance index could identify primary patency failure at 6 months, with similar predictive power as for 10-week arteriovenous fistula maturity failure, but with wide confidence intervals for wrist (positive predictive value 72.7%, 95% confidence interval 46.4% to 99.0%) and elbow (positive predictive value 57.1%, 95% confidence interval 20.5% to 93.8%). These models, moreover, performed poorly at identifying assisted primary and secondary patency failure, likely because a subset of those arteriovenous fistulas identified on ultrasound surveillance as at risk underwent subsequent successful salvage intervention without recourse to early ultrasound data. Although early ultrasound can predict fistula maturation and longer-term patency very effectively, it was only moderately good at identifying those fistulas likely to remain immature or to fail within 6 months. Allied to the better- than-expected fistula patency rates achieved (that are further improved by successful salvage), we estimate that a randomised controlled trial comparing early ultrasound-guided intervention against standard care would require at least 1300 fistulas and would achieve only minimal patient benefit. This trial is registered as ISRCTN36033877 and ISRCTN17399438. This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: NIHR135572) and is published in full in Health Technology Assessment; Vol. 28, No. 24. See the NIHR Funding and Awards website for further award information.

Exploring the role of HLA variants in neuroblastoma susceptibility through whole exome sequencing.

Although a number of susceptibility loci for neuroblastoma (NB) have been identified by genome-wide association studies, it is still unclear whether variants in the HLA region contribute to NB susceptibility. In this study, we conducted a comprehensive genetic analysis of variants in the HLA region among 724 NB patients and 2863 matched controls from different cohorts. We exploited whole-exome sequencing data to accurately type HLA alleles with an ensemble approach on the results from three different typing tools, and carried out rigorous sample quality control to ensure a fine-scale ancestry matching. The frequencies of common HLA alleles were compared between cases and controls by logistic regression under additive and non-additive models. Population stratification was taken into account adjusting for ancestry-informative principal components. We detected significant HLA associations with NB. In particular, HLA-DQB1*05:02 (OR = 1.61; padj = 5.4 × 10-3) and HLA-DRB1*16:01 (OR = 1.60; padj = 2.3 × 10-2) alleles were associated to higher risk of developing NB. Conditional analysis highlighted the HLA-DQB1*05:02 allele and its residue Ser57 as key to this association. DQB1*05:02 allele was not associated to clinical features worse outcomes in the NB cohort. Nevertheless, a risk score derived from the allelic combinations of five HLA variants showed a substantial predictive value for patient survival (HR = 1.53; p = 0.032) that was independent from established NB prognostic factors. Our study leveraged powerful computational methods to explore WES data and HLA variants and to reveal complex genetic associations. Further studies are needed to validate the mechanisms of these interactions that contribute to the multifaceted pattern of factors underlying the disease initiation and progression.

Influence of topography and synoptic weather patterns on air quality in a valley basin city of Northwest China

To clarify the mechanism underlying the effects of weather patterns and topography on air pollution, this study conducted the obliquely rotated principal component analysis in the T-mode to analyze ERA5 reanalysis data and categorize typical weather patterns at a 700-hPa geopotential height from 2015 to 2022. The probability of worsened air pollution attributable to weather patterns was quantitatively assessed using a generalized additive model. The results indicated that due to the influence of topography, Lanzhou was affected by an extended period of downdraft (with weak convective intensity) and the delayed formation of a convective boundary layer during the daytime by 1–2 h relative to other areas. Under the combined effect of low trough patterns (south low pressure type [SL] and south low weak pressure type [SL−]) and topography, the formation of a stable layer above the planetary boundary layer (PBL) would weaken the vertical exchange of the local airflow and inhibit the development of the PBL. The type of SL led to the most severe pollution, causing a 61.9 % (95 % confidence interval [CI]: 46.3 %–79.3 %) increase in PM2.5 concentration. For southwest high pressure patterns (south high [SH], southwest weak high [SWH−], southwest high [SWH], and southwest strong high [SWH+] pressure types), the prevailing northwest wind was the main transport path for pollutants. For the high pressure patterns (north high [NH] and northwest high [NWH] pressure types) and south wind patterns (southeast weak high [SEH−], southeast high [SEH], and northeast high [NEH] pressure types), the enhancement of vertical convection, deepening of the PBL, and reduction of pollution transport led to improved air quality. The NH, NWH, and NEH pressure types caused PM2.5 concentration to decrease by 18.4 % (95 % CI: 8.8 %–27.1 %), 14.9 % (95 % CI: 4.7 %–24.0 %), and 35.9 % (95 % CI: 9.7 %–54.6 %), respectively.

Nonlinear time effects of vegetation response to climate change: Evidence from Qilian Mountain National Park in China

Vegetation responses to climate change are typically nonlinear with varied time effects, yet current research lacks comprehensiveness and precise definitions, hindering a deeper understanding of the underlying mechanisms. This study focuses on the mountain-type Qilian Mountains National Park (QMNP), investigating the characteristics and patterns of these nonlinear time effects using a generalized additive model (GAM) based on MODIS-NDVI, growing season temperature, and precipitation data. The results show that 1) The time effects of climate change on vegetation exhibit significant spatial variations, differing across vegetation types and topographic conditions. Accounting for optimal time effects can increase the explanatory power of climate on vegetation change by 6.8 %. Precipitation responses are mainly characterized by time-lag and time-accumulation effects, notably in meadows and steppes, while temperature responses are largely cumulative, especially in steppes. The altitude and slope significantly influence the pattern of vegetation response to climate, particularly in areas with high altitudes and steep slopes. 2) There is a significant nonlinear relationship between vegetation growth and both precipitation and temperature, with the nonlinear relationship between precipitation and vegetation being stronger than that with temperature, particularly in the western and central regions of the park. Different vegetation types exhibit significant variations in their response to climate change, with deserts and steppes being more sensitive to precipitation. 3) Precipitation is the primary driver of vegetation change in the QMNP, particularly for high-elevation vegetation and herbaceous vegetation. The complex temporal patterns of vegetation response to climate change in the QMNP not only deepen the understanding of the intricate relationship between regional vegetation and climate variability but also provide a methodological reference for global studies on vegetation responses to climate change.

Cracking the code: Spatial heterogeneity as the missing piece for modeling granular fluidized bed drying

Pharmaceutical twin-screw wet granulation is a multifaceted and intricate process pivotal to drug product development. Accurate modeling of this process is indispensable for optimizing manufacturing parameters and ensuring product quality. The fluid bed dryer, an integral component of this granulation process, significantly influences the granular critical quality attributes. This study builds upon prior research by integrating experimental findings on granule segregation during fluid bed drying into an existing compartmental model, enhancing its predictive capabilities. An additional model layer on granule segregation behavior is composed and integrated into the existing model structure in this study. The added model compartment describes probability distributions on the vertical position of granules within each granule size class considered. To beware of overfitting, predictions of both the moisture content after drying and the granule bed temperature throughout drying are discussed in this study relative to experimental data from earlier published studies. These independent analyses demonstrated a marked improvement in prediction accuracy compared to earlier published model structures. The refined model accurately predicts the residual moisture content after drying for an untrained formulation. Moreover, it simultaneously makes accurate predictions of the granular bed temperature, which emboldens its structural correctness. This advancement makes it a powerful tool for predicting the behavior of the pharmaceutical fluid bed drying, which holds significant promise to facilitate pharmaceutical product development.

THE VALIDITY OF USING THE FULLTRACK AI APP TO DETERMINE CRICKET BOWLING LINE AND LENGTH COMPARED TO 3D MOTION CAPTURE

INTRODUCTION A bowler’s ability to manipulate line and length of a delivery is a key performance attribute in cricket. 3D-motion capture (3DMocap), the gold standard in quantifying ball trajectories1 lacks ecological validity2. Fulltrack AI app may provide a feasible alternative to quantify and provide immediate ball trajectory feedback. This study explored the validity of Fulltrack AI to measure ball landing position compared with 3DMocap. METHODS 836 deliveries under various conditions (pace, spin; bowled, thrown (SidearmTM); batter, no batter), were recorded using a Qualisys 8-camera 3DMocap system and iPad running Fulltrack AI (version 1.13.1). Line and length were extracted from Quintic Spline filtered 3DMocap data and tabulated with Fulltrack AI data. Statistical analyses were conducted in R Statistical Software following removal of outliers. Bland Altman’s, 95% limits of agreement (LOA)3 were calculated, with line and length interaction further explored using 95% confidence ellipse area (CEA) to assess practical difference between 3DMocap and Fulltrack AI relative to ball landing position. Validity between Fulltrack AI and 3DMocap, and the interaction of conditions, were assessed with generalised additive models (GAMs). RESULTS Whilst LOA (line= -0.15 to 0.10; length= -0.88 to 0.27) demonstrated good agreement, CEA ranged from 0.17m2 to 0.42m2 depending on ball landing location relative to the stumps (2-4m or >8m, respectively). GAMs ‘average model’ established no significant (p>0.05) difference between 3DMocap and Fulltrack AI, or the interaction effects of training conditions. CONCLUSION Although Fulltrack AI appears statistically and ecologically valid in identifying cricket ball landing position, practitioners should be cognisant that practically, error varies relative to where on the wicket a ball lands. Thus, hindering the accuracy of feedback if needing to precisely quantify ball landing position when using Fulltrack AI app.

Low replicability of testing the stress–dominance hypothesis using a trait convergence/divergence pattern

AbstractAimsEcological theories predict that assembly processes are driven by two deterministic forces: environmental filtering and limiting similarity. Their relative importance under different environmental conditions is still not completely obvious. Therefore, in this paper the predictions of the stress–dominance hypothesis (SDH) are tested in several sites.LocationKiskunság in Hungary, and Deliblato Sands in Serbia, Central Europe.MethodsWe studied a productivity gradient from open sand grasslands to meadows. The cover of species was estimated visually in plots with a size of 2 m × 2 m, resulting in 344 vegetation plots. Four trait values (height, seed mass, specific leaf area, and leaf size) were collected from field measurements and databases. The weighted median of interspecies distances in traits (a robust alternative to Rao's quadratic entropy) was used to determine functional diversity. The convergence and divergence of each trait in communities were evaluated by randomization tests, and effect sizes were calculated for each plot. We used hierarchical general additive models (HGAM) to determine whether the trend of effect sizes along the productivity gradient is the same in different sites.ResultsThe HGAM approach indicated that trait variations follow global trends but are influenced by site‐specific effects. The exception is seed mass, whose variation did not have any trend. Both environmental filtering and limiting similarity exist in the communities, and mainly a shift from trait convergence to a divergence pattern along the productivity gradient was observed.ConclusionThe results are mainly congruent with theoretical expectations, but the results from the different sites did not lead to the same conclusion. Although traits follow a global trend, the site effect is not negligible. Critical evaluation of SDH using trait convergence/divergence patterns for exploring rules of community assembly points out the weaknesses of this hypothesis. Therefore, alternative ways of studying trait patterns should be found to better understand community organization.

Spatiotemporal Patterns and Environmental Drivers of Particulate Inorganic Carbon (PIC) Variability in the Southern Waters of Sumatra‐Java Islands

AbstractThe southern waters of the Sumatra‐Java Islands, within the Indo‐Pacific Warm Pool, are highly susceptible to global climate change impacts. Understanding particulate inorganic carbon (PIC) dynamics and environmental responses in this region is critical for assessing climate change effects on marine ecosystems. Utilizing two decades of remote sensing PIC data (2003–2022), we identified five sub‐regions via K‐means clustering: the Sunda Strait, southern coastal/offshore waters of Sumatra Island, and southern coastal/offshore waters of Java Island. Self‐Organizing Map explored interannual PIC variations, while Generalized Additive Models delineated driving factors. In the Sunda Strait, PIC dynamics are influenced by water exchanges between the Indonesian Seas (ISs) and eastern Indian Ocean (EIO). Seasonally, PIC concentrations peak during the east monsoon (boreal summer); decreasing during positive Indian Ocean Dipole (+IOD) and El Niño events. Along the southern Sumatra, PIC variations positively correlate with rainfall and photosynthetically active radiation (PAR). Along the southern Sumatra, upwelling drives PIC variations, with higher concentrations during mature summer monsoon induced upwelling periods. PIC increases during +IOD and decreases during El Niño. Our study highlights the nexus between nutrients, physical factors (e.g., rainfall), and climatic events (e.g., El Niño) on PIC concentrations. Climate‐induced changes in oceanic physical processes modulate nutrient concentrations, thereby governing PIC variations. We propose a mechanistic model to elucidate PIC variations in upwelling areas. Findings suggest that prospective climate variability, encompassing alterations in climate events, rainfall, and temperature, may escalate PIC concentrations within the study area, advocating PIC as a forward‐looking climate change indicator.

Characterization and Determinants of Long-Term Immune Recovery Under Suppressive Antiretroviral Therapy.

We developed a robust characterization of immune recovery trajectories in people living with HIV on antiretroviral treatment (ART) and relate our findings to epidemiological risk factors and bacterial pneumonia. Using data from the Swiss HIV Cohort Study and the Zurich Primary HIV Infection Cohort Study (n = 5907), we analyzed the long-term trajectories of CD4 cell and CD8 cell counts and their ratio in people living with HIV on ART for at least 8 years by fitting nonlinear mixed-effects models. The determinants of long-term immune recovery were investigated using generalized additive models. In addition, prediction accuracy of the modeled trajectories and their impact on the fit of a model for bacterial pneumonia was assessed. Overall, our population showed good immune recovery (median plateau [interquartile range]-CD4: 718 [555-900] cells/μL, CD8: 709 [547-893] cells/μL, CD4/CD8: 1.01 [0.76-1.37]). The following factors were predictive of recovery: age, sex, nadir/zenith value, pre-ART HIV-1 viral load, hepatitis C, ethnicity, acquisition risk, and timing of ART initiation. The fitted models proved to be an accurate and efficient way of predicting future CD4 and CD8 cell recovery dynamics: Compared with carrying forward the last observation, mean squared errors of the fitted values were lower by 1.3%-18.3% across outcomes. When modeling future episodes of bacterial pneumonia, using predictors derived from the recovery dynamics improved most model fits. We described and validated a method to characterize individual immune recovery trajectories of people living with HIV on suppressive ART. These trajectories accurately predict long-term immune recovery and the occurrence of bacterial pneumonia.

The ecological risk and fate of microplastics in the environmental matrices of marine ranching area in coastal water

The debate surrounding "source" and "sink" of microplastics (MPs) in coastal water has persisted for decades. While the transportation of MPs is influenced by surface runoff and currents, the precise transport patterns remain inadequately defined. In this study, the typical coastal habitat – marine ranching in Haizhou Bay (Jiangsu Province, China) were selected as a case study to assess the ecological risk of MPs. An enhanced framework was employed to assess the entire community characteristics of MPs in various environmental compartments, including surface water (SW), middle water (MW), bottom water (BW), sea bottom sediment (SS), and intertidal sediment (IS). The results of the assessment showed a low risk in the water column and a high risk in the sediment. PERMANOVA based on size and polymer of MPs revealed significant differences between IS and other compartments (SW, MW, BW, and SS) (P<0.001). The co–occurrence network analysis for MP size indicated that most sites occupied central positions, while the analysis for MP polymer suggested that sites near the marine ranching area held more central positions, with sites in MW, BW, and SS being somewhat related to IS. Generalized additive model (GAM) demonstrated that MP concentration in the water correlated with Chla and nutrients, whereas MPs in sediment exhibited greater susceptibility to dissolved oxygen (DO) and salinity. We believe that except for the natural sedimentation and re–suspension of MPs in the vertical direction, MPs in bottom water may migrate to the surface water due to upwelling mediated by artificial reefs. Additionally, under the combined influence of surface runoff, currents, and tides, MPs may migrate horizontally, primarily occurring between middle and bottom water and sediments. The study recommends limiting and reducing wastewater and sewage discharge, as well as regulating fishing and aquaculture activities to control the sources and sinks of MPs in coastal water. Moreover, it advocates the implementation and strengthening of marine monitoring activities to gain a better understanding of the factors driving MP pollution in marine ranching area.

Solid as a rock: The main drivers of changes in natural, rocky plant communities

AbstractQuestionChanges caused by climate warming and nitrogen pollution are observed in forest, grassland and alpine ecosystems worldwide. However, still little is known about the impact of these globally influencing factors on natural rocky plant communities. Has species composition of natural rocky communities changed over time? What is the role of large‐scale and fine‐scale environmental factors in shaping the compositional, functional and habitat patterns in studied plant communities over time?LocationSudetes Mountains, southwestern Poland.MethodsWe used 214 pairs of replots (collected between 1989 and 2022) of rocky plant communities, with a mean timespan of 14.2 years. The changes in species composition, environmental conditions and functional traits were analysed using ordination techniques and generalised additive models (GAMs) and with reference to large‐scale factors (mean maximum temperatures, actual evapotranspiration, N deposition) and fine‐scale factors (light availability, bedrock type, initial site conditions) for each locality.ResultsSpecies composition of the studied communities has not changed significantly over time. Only for 11 out of 258 species, statistically significant increases (from 2% to 8%) in their proportion were recorded. The changes in environmental conditions were significantly influenced mainly by fine‐scale factors such as changes in light availability and baseline site conditions. Plots that were initially less thermophilic or nitrophilic showed stronger signals of thermophilisation and eutrophication in the resurvey. The influence of large‐scale factors was considerably less pronounced. Similarly, the key role in explaining changes in plant traits for the data set under study falls to local factors, particularly changes in light availability.ConclusionsOur study confirms the validity of considering both large‐ and fine‐scale factors as well as initial site conditions in research on long‐term changes in plant communities. Rocky plant communities respond to global changes in a different way than other types of phytocoenose, showing high stability of species composition and functional traits over time.

RELIABILITY AND VALIDITY OF FULLTRACK AI APP TO MEASURE CRICKET BALL SPEED UNDER TRAINING CONDITIONS

INTRODUCTION Radar guns are commonly used to accurately and reliably measure ball speed(1), a key cricket bowling performance indicator. App-based approaches, such as Fulltrack AI, are gaining popularity. This study investigated the reliability and validity of Fulltrack AI to measure cricket ball speed compared to a validated radar gun(1). METHODS Ball speed of 1081 deliveries (pace=783; spin=298) from a range of training sessions and conditions (batter, no batter; indoor and outdoor wickets) were recorded simultaneously using a radar gun (Stalker ATS2) and iPad running Fulltrack AI (version 1.13.1). Fulltrack AI data (ball speed (km/hr), line, length (m)) were extracted post-session for tabulation with radar gun data. Statistical analyses were conducted in R Statistical Software independently for bowling type (pace, spin) following exclusion of outliers. Reliability was assessed with standard error of measurement (SEM), coefficient of variation (CV) and intraclass correlation coefficient (ICC). Agreement was assessed using Bland Altman’s, 95% limits of agreement (LOA)(2). Validity was assessed using generalised additive models (GAM), controlling for line, length and interaction of training conditions. RESULTS Whilst reliability coefficients for pace deliveries demonstrated very good agreement (ICC=0.90; SEM=2.61) and lower variability (CV=2.56%) in contrast to spin (ICC=0.76; SEM=2.17; CV=3.08%); LOA demonstrated poor to fair levels of agreement, exceeding maximal allowable differences (&amp;gt;3%). When controlling for line, length and training conditions, GAMs ‘average model’ identified Fulltrack AI significantly (p&amp;lt;0.05) overestimated ball speed (pace: estimate 2.58km/hr, SE=1.24; spin: estimate 3.93km/hr, SE=0.81) when compared to the radar gun. CONCLUSION Fulltrack AI is a reliable method for monitoring ball speed where accuracy is not of paramount importance. Significant overestimation of ball speed in contrast with a radar gun, even after controlling for different training conditions, suggests software refinement is required before such technology is readily adopted for the measurement of speed. 1.Smith &amp; Burke (2021) 2.Bland &amp; Altman (1986)

A deep learning computational fluid dynamics solver for simulating liquid hydrogen jets

Modeling and simulating the sudden depressurization of liquids inside nozzles is a significant challenge because of the plethora of the associated complex phenomena. This pressure drop together with the rapid phase change of the liquid is important characteristics of flash boiling. Computational fluid dynamics (CFD) multiscale simulations of flashing jets usually deploy additional models for modeling heat and mass transfer with long computational times. Intermediate steps such as volumetric meshing in mesh-based methods can also significantly increase the computational cost. This paper aims at providing academia and industry with a modeling tool to simulate and investigate the complex multi-facet phenomenon of flash-boiling atomization deploying a machine-learning method that could save thousand Central Processing Unit hours offering instantaneous CFD predictions. The presented machine-learning CFD method completely replaces the traditional CFD simulations workflow and requires little simulation expertise from the end-user. Notably, this is a novel model that couples for the first time the thermodynamic non-equilibrium with convolutional neural networks to simulate flashing liquid hydrogen jets thousand times faster than the standalone CFD solver. The accuracy of the novel approach is evaluated, demonstrating adequate accuracy compared to different unseen simulations and experiments. This work offers the groundwork for further accelerating CFD predictions in multiphase flow problems and could significantly improve testing flash-boiling scenarios in various industrial settings.

Predicting the site productivity of forest tree species using climate niche models

Species occurrence-based climate niche models (CNMs) serve as valuable tools for predicting the future ranges of species’ suitable habitats, aiding the development of climate change adaptation strategies. However, these models do not address an essential aspect - productivity, which holds economic significance for timber production and ecological importance for carbon sequestration and ecosystem services. In this study, we investigated the potential to extend the CNMs to predict species productivity under various climate conditions. Lodgepole pine (Pinus contorta Dougl. ex Loud.) and Douglas-fir (Pseudotsuga menziesii Franco.) were selected as our model species due to their comprehensive range-wide occurrence data and measurement of site productivity. To achieve this, we compared and optimized the performance of four individual modeling algorithms (Random Forest (RF), Maxent, Generalized Boosted Models (GBM), and Generalized Additive Model (GAM)) in reflecting site productivity by evaluating the effect of spatial filtering, and the ratio of presence to absence (p/a ratio) observations. Additionally, we applied a binning process to capture the overarching trend of climatic effects while minimizing the impact of other factors. We observed consistency in optimal performance across both species when using the unfiltered data and a 1:1.5 p/a ratio, which could potentially be extended to other species. Among the modeling algorithms explored, we selected the ensemble model combining RF and Maxent as the final model to predict the range-wide site productivity for both species. The predicted range-wide site productivity was validated with an independent dataset for each species and yielded promising results (R2 above 0.7), affirming our model’s credibility. Our model introduced an innovative approach for predicting species productivity with high accuracy using only species occurrence data, and significantly advanced the application of CNMs. It provided crucial tools and insights for evaluating climate change's impact on productivity and holds a better potential for informed forest management and conservation decisions.