PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “predictive model”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 109 records · Page 6Linked to original sources
PubMed1986

Cat lung hemodynamics: comparison of experimental results and model predictions.

Commonly, attempts have been made to learn about the structure and function of the pulmonary vascular bed from measurements of arterial and venous pressures and blood flow rate under steady-state conditions (e.g., from pressure vs. flow data) or dynamic conditions (e.g., from vascular occlusion data). Zhuang et al. (J. Appl. Physiol. 55: 1341-1348, 1983) have presented a detailed model of steady-state cat lung hemodynamics based on direct measurements of anatomical and elasticity data. This model provides an opportunity to better understand the information content of the hemodynamic data. Therefore, in the present study we carried out a series of steady-state and dynamic experiments on isolated cat lungs. We then compared the results with those predicted by the model. We found that the model provided a good fit to the steady-state data. However, to fit the dynamic data, some modifications were necessary to account for the viscous behavior of the vessel walls and to move the first moment of the distribution of vascular resistance toward the arterial end of the vascular bed relative to that of the distribution of vascular compliance. Due to the sensitivity of the vascular resistance to small changes in vessel diameters and branching ratio, the modifications in morphometry represent small changes in morphometric data and are probably within the range of uncertainty in such data. The modifications had little effect on the steady-state model simulations but substantially improved the dynamic model simulations, suggesting that the dynamic data are quite sensitive to small changes in the relative distributions of vessel diameters and elasticity.

Animals↗
PubMed1989

A predictive model for combined temperature and water activity on microbial growth during the growth phase.

An empirical and generalized model is presented, based on a modified Arrhenius equation, for predicting the combined effect of temperature and water activity on the growth rate of bacteria. When it was applied to seven separate sets of wide ranging published results, spanning some 50 years and including a spore-former and a silage micro-organism, predictions explained between 92.9 and 99.0% of the variation in the results with an overall mean of 96.6%. Advantages over existing models are that it is relatively easy to fit to data using least squares regression and requires only five coefficients. These, together with its simplicity and demonstrated wide application, will facilitate its practical use.

Bacteria↗
PubMed1986

Grading scores and survivorship functions in liver cirrhosis: a comparative statistical analysis of various predictive models.

In a group of followed-up liver cirrhotics we evaluated the reliability of prognostic estimates predicted on the basis of a previously described multivariate statistical model (MSM). In the same subjects we also compared theoretical survival estimates obtained by fitting some other liver cirrhosis grading scores (Child-Turcotte's, McCormick's and Orrego's) to prognostic purposes. No statistical difference between actual and MSM-estimated survivorship functions was found (employing a life-table method with Logrank test), thus confirming the prognostic reliability of this multivariate classification model. Such a global and prognosis-correlated index may be recommendable both for comparing different groups of patients, and for assessing treatment effectiveness. Or results also substantially confirm the other investigated classificative methods such as reliable liver cirrhosis severity indexes, although their use for prognostic purposes seems to be less suitable.

Actuarial Analysis↗
PubMed1985

Importance of glucose per se to intravenous glucose tolerance. Comparison of the minimal-model prediction with direct measurements.

Glucose disappearance after an oral or intravenous challenge is a function of the effects of both endogenously secreted insulin and of glucose itself. We previously introduced the term "glucose effectiveness," or SG, defined as the ability of glucose per se to enhance its own disappearance independent of an increment in plasma insulin. The present investigation, performed in conscious dogs, was undertaken to quantify this glucose effect by minimal-model-based analysis of insulin and glucose dynamics after a frequently sampled intravenous glucose tolerance test (FSIGT). The values from the standard FSIGT were then compared with direct measurements obtained from experiments in which the dynamic insulin response to glucose was suppressed with somatostatin (SRIF). In addition, we examined SG values from the modified FSIGT protocol, which involves both glucose and tolbutamide injections. Protocol l (N = 9): FSIGTs were performed and the glucose and insulin data were analyzed by computer. KG was 2.65 +/- 0.28 min-1, S1 was 4.09 +/- 0.34 X 10(4) min-1/(microU/ml), and SG was 0.033 +/- 0.004 min-1. Protocol II (N = 6): FSIGTs were performed on animals in which SRIF was infused (0.8 micrograms/min X kg) to obliterate the dynamic insulin response to glucose injection. Before the FSIGt, insulin and glucagon were infused intraportally to reattain basal glycemia. Without dynamic insulin, KG was reduced to 0.96 +/- 0.18 min-1 (P less than 0.0001). However, SG, estimated from the exponential rate of fall of plasma glucose in the absence of dynamic insulin, was similar to the standard FSIGTs: 0.025 +/- 0.004 (P greater than 0.25). Protocol III (N = 6): modified FSIGTs were performed using glucose and tolbutamide injections for a better estimate of model parameters. Model parameters Sl and SG, and the KG were not different from standard FSIGTs (P greater than 0.3). In fact, the value of SG (0.028 +/- 0.003 min-1) was nearly identical to the direct measure from protocol II. Therefore, the effect of glucose per se on glucose decline, estimated by modeling the standard and modified FSIGTs, was confirmed by a direct measurement with the endogenous insulin response suppressed with SRIF. Also, the time course of the insulin effect to enhance net glucose disappearance from plasma [Ieff(t)] was calculated from the data of protocol II, and was the same as the time course predicted by the model. These studies demonstrate the ability of the computer modeling approach to separate insulin-dependent and glucose-dependent glucose disappearance, and represent a direct confirmation of the minimal model.(ABSTRACT TRUNCATED AT 400 WORDS)

Animals↗
PubMed2026

How negative sampling shapes the performance of transcription factor binding site prediction models.

MOTIVATION: Transcription factors (TFs) are key players in gene regulation and development, where they activate and repress gene expression through DNA binding. Predicting transcription factor binding sites (TFBSs) has long been an active area of research, with many deep learning methods developed to tackle this problem. These models are often trained on TF ChIP-seq data, which is generally seen as only providing positive samples. The choice of datasets and negative sampling techniques is a critical yet often overlooked aspect of this work. RESULTS: In this study, we investigate the impact of different negative sampling techniques on TFBS prediction performance. We create high-quality test datasets based on ChIP-seq and ATAC-seq data, where true negatives can be identified as positions that are accessible but not bound by the TF in question. We then train models using various negative sampling techniques, including genomic sampling, shuffling, dinucleotide shuffling, neighborhood sampling, and cell line specific sampling, simulating cases where matching ATAC-seq data is not available. Our results show that, generally, metrics calculated on training datasets give inflated performance scores. Of the tested techniques, genomic sampling of negatives based on similarity to the positives performed by far the best, although still not reaching the performance of baseline models trained on high-quality datasets. Models trained on dinucleotide shuffled negatives performed poorly, despite being a common practice in the field. Our findings highlight the importance of carefully selecting negative sampling techniques for TFBS prediction, as they can significantly impact model performance and the interpretation of results. AVAILABILITY AND IMPLEMENTATION: The code used in this study is available at https://github.com/NatanTourne/TFBS-negatives (DOI: 10.5281/zenodo.18007567).

Binding Sites↗
PubMed1989

[Congenital malformations: a model predictive based on risk factors].

Several risk factors were studied in regard to congenital malformations. Malformed newborns (n = 1200) and controls (n = 1200) seen at the Universidad de Chile Hospital between 1969 and 1979 were examined. Their mothers were asked about possible risk factors. Parenteral age and birth order was significantly higher for malformed newborns than for controls. A family history of congenital malformations was more frequent in malformed newborns. Infertility, metrorrhagia and maternal diseases during pregnancy were more frequent in malformed newborns than in controls. A function that discriminates between controls mothers and mothers of malformed newborns was obtained by a logistic regression model. This function correctly predicted 65% of cases.

Congenital Abnormalities↗
PubMed1987

Predictive models for deposition of inhaled diesel exhaust particles in humans and laboratory species.

Mathematical and computer models of the respiratory tracts of human beings and of laboratory animals (rats, hamsters, guinea pigs) were used to estimate the deposition patterns of inhaled diesel exhaust particles from automobile emissions. The accuracy of these models was tested by comparing the calculated depositions in laboratory animals with actual laboratory data. Our goal was to be able to predict the relation between exposure to diesel exhaust particles and the deposition of these particles in the lungs of humans of various ages. Diesel exhaust particles are aggregates with a mass median aerodynamic diameter of approximately 0.2 micron. Their actual size depends on the conditions under which they are generated. Using an appropriate particle model, we derived mathematical expressions that describe the effects of diffusion, sedimentation, impaction, and interception on the deposition of these particles. Because of their small size, we found that most diesel exhaust particles deposited through diffusion, and that the role of the other mechanisms was minor. Anatomical models of the human lung from birth to adulthood, as well as models of the lungs of laboratory species were formulated mathematically using available morphometric data. We used these lung models, together with the corresponding ventilation conditions of each species, to calculate deposition of diesel exhaust particles in the lungs. Under normal breathing conditions, we calculated that 7 to 13 percent (depending on particle size) of inhaled diesel exhaust particles deposit in the alveolar region of the adult human lung. Although the breathing mode (nose or mouth breathing) did not appear to affect alveolar deposition, increasing the minute ventilation (the number of breaths per minute multiplied by the tidal volume) increased alveolar deposition significantly. The calculated deposition patterns for diesel exhaust particles in younger humans (under age 25) were similar. However, with the exception of alveolar deposition in very young children (under age two), predicted deposition was greater in the lungs of younger humans than in the lungs of humans age 25 or older. For an equal exposure, the surface minute dose (particle mass deposited per minute per unit surface area) of unciliated airways appeared to change profoundly with age. Predicted dose was maximal in the lung models of two-year-old children. At this age, the calculated dose was approximately twice as high as in the mature adult lung. Deposition predictions for laboratory species compared favorably with existing data. Distribution of deposition was found to be similar among all species studied, although surface minute dose decreased with body weight.

Adolescent↗
PubMed1992

Distinguishing between persistent and transient impaired glucose tolerance using a prediction model.

Screening for impaired glucose tolerance (IGT) and Type 2 (non-insulin dependent) diabetes was carried out in 777 people and those with high blood glucose levels completed three 2-h oral glucose tolerance tests (OGTT). Blood lipid levels, fasting and 2-h insulin levels, body mass index, and blood pressure were also measured and family history of Type 2 diabetes recorded. Fifty people were identified with IGT and of these 21 were found to have persistent IGT and 29 transient IGT. A model including the variables body mass index, fasting and 2-h insulin levels, fasting triglycerides and family history of Type 2 diabetes was developed using the Speigelhalter-Knill-Jones weighting method to predict subjects with persistent IGT. This model could be useful in identifying people with persistent IGT and therefore eliminate the need for repeat OGTTs which are time consuming and expensive.

Blood Glucose↗
PubMed2026

Machine learning-based clinical prediction model and multi-omics integration for assessing pancreatic cancer risk in new-onset diabetes.

BACKGROUND: Given that pancreatic cancer (PC) is typically diagnosed at an advanced stage but is often preceded by new-onset diabetes mellitus (NODM), providing a window for early detection, we sought to develop and validate an interpretable machine-learning model integrated with multi-omics profiling to identify early biomarkers of NODM-associated PC. METHODS: In a population-based cohort, individuals with NODM-associated PC and NODM without PC were identified and randomly divided (70:30) into training and validation sets after feature selection. Eight machine learning (ML) classifiers were compared using fivefold cross-validation, and model performance was evaluated in terms of discrimination, calibration, and decision curve–based clinical utility. We evaluated interpretability using the Shapley additive explanations (SHAP) analyses. Mechanistically, Olink proteomic profiling and metabolomics were analyzed through clinical classifications and model-defined risk strata. RESULTS: Categorical boosting achieved the best performance in the independent validation set (AUROC = 0.844). The NODM cohort was stratified into high- (n = 2,362) and low-risk (n = 5,030) groups, and internal validation together with SHAP analyses demonstrated consistent model performance and identified clinically interpretable predictors. Proteomic and metabolomic analyses under clinical and risk-based grouping identified 39 overlapping differentially expressed proteins and 145 overlapping metabolites with enriched across 11 shared KEGG pathways. Cross-platform validation highlighted PLTP, CRTAC1, and ITGAV as serum biomarkers with a strong potential for early NODM-PC detection. CONCLUSIONS: We developed an interpretable ML framework centered on NODM enables practical risk stratification for early PC detection by multi-omics and provides a pathway of ML-based triage followed by biomarker confirmation for earlier detection and diagnosis.

Humans↗
PubMed2026

CLASPP: A unified model for predicting post-translational modifications.

Post-Translational Modifications (PTMs) are a fundamental mechanism for regulating cellular pathways and increasing the functional diversity of the proteome. Accurately predicting the PTM types that are likely to occur at a given site in the primary sequence is a key challenge in functional proteomics. Existing PTM prediction models predominantly focus on either single PTM types or employ ensemble methods that combine multiple models to predict different PTM types. This fragmentation is largely driven by the vast imbalance in data availability across PTM types, making it difficult to predict multiple PTM types with a single model. To address this limitation, we present the Contrastively Learned Attention-based Stratified PTM Predictor (CLASPP), a unified PTM prediction model. CLASPP addresses imbalance challenges by leveraging unsupervised clustering-based undersampling and a novel contrastive learning framework tailored to PTM data. Additionally, our hierarchical data organization and curation are shown to improve CLASPP's performance by balancing the representation of individual PTM types and provides a standardized dataset to train and validate future model designs. Drawing inspiration from advancements in image and natural language processing, the CLASPP model employs a multi-stage training strategy and a high-quality, curated training dataset to improve PTM prediction performance. To uncover what is learned during the contrastive learning stage, the CLASPP model is shown to distinguish known protein kinase substrate specificity profiles as a form of explainability. Finally, we evaluate the application of CLASPP in predicting PTMs in different model organisms and experimentally validated ubiquitination sites in the understudied DCLK3 kinase. Overall, CLASPP represents a unified model for PTM prediction that addresses key bottlenecks in data imbalance and offers new strategies for biological data curation, thereby improving PTM-type prediction performance across diverse organisms.

Protein Processing, Post-Translational↗
PubMedDate not supplied

Errors between two- and three-dimensional thermal model predictions of hyperthermia treatments.

A simulation program to study the three-dimensional temperature distributions produced by hyperthermia in anatomically realistic inhomogenous tissue models has been developed using the bioheat transfer equation. The anatomical data for the inhomogeneous tissues of the human body are entered on a digitizing tablet from serial computed tomography (CT) scans. Power deposition patterns from various heating modalities must be calculated independently. The program has been used to comparatively evaluate two- and three-dimensional simulations in a series of parametric calculations based on a simple inhomogeneous tissue model for uniform power deposition. The conclusions are that two-dimensional simulations always lead to significant errors at the ends of tumors (up to tens of degrees). However, they can give valid results for the central region of large tumors, but only with tumor blood perfusions greater than approximately 1 kg/m3/s. These conclusions from the geometrically simple model are substantiated by the results obtained using the full three-dimensional model for actual patient anatomical simulations. In summary, three-dimensional simulations will be necessary for accurate patient treatment planning. The effect of the thermal conductivity, used in the models, on the temperature field has also been studied. The results show that using any thermal conductivity value in the range of 0.4 to 0.6 W/m/degrees C sufficiently characterizes most soft tissues, especially in the presence of high blood perfusion. However, bone (thermal conductivity of 1.16 W/m/degrees C) and fat (thermal conductivity of 0.2 W/m/degrees C) do not fit this generalization and significant errors result if soft tissue values are used.

Blood Circulation↗
PubMed1992

Dynamic model prediction of the value of reduced solubility of alfalfa silage protein for lactating dairy cows.

A net carbohydrate and protein system was used to develop model diets for lactating dairy cattle with various protein solubilities in the alfalfa silage component of the diet. The objective was to determine the level to which alfalfa silage could be used to replace supplemental protein sources as the silage protein solubility decreased and to estimate the value of silage treatments needed to reduce protein solubility. Four cow groups were considered: early lactation multiparous cows, primiparous cows, midlactation cows, and late lactation cows. Diets were balanced for metabolizable protein, metabolizable energy, and ammonia and peptides for rumen bacteria; limits on DMI and effective NDF were enforced. Lower protein solubility was predicted to increase the yield of bacteria per unit of alfalfa silage DM and the yield of metabolizable protein per unit of alfalfa silage CP. Because of reduced protein supplements, diet costs were decreased. The savings per unit of silage in these rations increased as alfalfa silage protein solubility decreased. For example, with a reduction in solubility from 61 to 51% of CP, the savings ranged from $2.96 to $3.26/tonne of silage across the four cow groups. The value of acid treatment of silage needed to effect these reductions exhibited diminishing returns as application rate increased and appeared to be most cost effective when used on high quality alfalfa fed to high producing cows with application rates less than 2 kg/tonne. Management practices that reduce silage temperatures were predicted to save $.50 to $1.50/tonne of silage when the diets were balanced to account for protein degradability.

Animals↗
PubMedDate not supplied

Single-dose model for predicting steady-state valproic acid serum concentrations in seizure patients.

Valproic acid serum concentrations predicted by a single-dose prediction model were compared with steady-state serum concentrations measured after the start of therapy in seizure patients. Ten patients receiving valproic acid for the first time or who had not been taking the drug for two or more weeks were entered into the study. The patients' therapies were initiated with the prescribed doses of valproic acid and then maintained on fixed doses and dosing intervals until steady-state trough serum samples were obtained. Initial 5-ml blood samples were collected six to 14 hours after the ingestion of the first dose; a 5-ml steady-state trough sample was drawn in the same manner three to seven days later. Both free and total drug concentrations were determined within 48 hours of sample collection using gas-liquid chromatography. The elimination rate constant was estimated from age-specific population half-life values found in the literature. Six patients (five children, aged four to 16 years) and one adult (aged 87 years) completed the study. There was a statistically significant correlation between predicted and measured steady-state valproic acid serum concentrations for both free and total concentrations. The single-dose prediction model accurately predicted steady-state valproic acid serum concentrations in these seizure patients.

Adolescent↗
PubMed2026

Massively parallel characterization and predictive modelling of neuronal regulatory variation.

Disease-associated variants reside frequently in noncoding cis-regulatory elements (CREs), yet their functional consequences remain poorly understood. We performed a large-scale lentiMPRA in human excitatory neurons, quantifying the impact of >46,000 naturally occurring variants across >27,000 candidate CREs near 524 disease-associated genes. These data improved regulatory variant effect predictions beyond state-of-the-art models. Significant allelic effects occurred at comparable rates across common, rare, and singleton variants, demonstrating that, within MPRA-measurable effects, population frequency carries limited information about per-variant regulatory impact. Variant effect detectability and magnitude were governed primarily by baseline activity of the enclosing regulatory element and local sequence context. Regulatory effects were distributed across numerous transcription factors rather than concentrated in master regulators, consistent with a combinatorial enhancer architecture. We establish a large-scale functional variant catalog and provide a complementary benchmark and resource for developing and evaluating models of noncoding regulatory variation.

Journal Article↗
PubMed1988

A prediction model of suicide among youth.

Epidemiological relationships were studied between adolescent (ages 15 to 24 years) suicide rates and population shifts among adolescents. Suicide rates among adolescents tripled from 1956 to 1977 and have subsequently leveled off. Increases (and decreases) in adolescent suicide rates corresponded to increases (and decreases, respectively) in the proportion of adolescents in the United States. Opposite trends have been found among older age groups. Projected population fluctuations were used to predict trends in adolescent suicide rates to the year 2000: the current decrease in rates is predicted to continue until the mid-1990s. Recent data indicating a decrease in adolescent suicide rates tend to support the population model hypothesis. The data suggest that demographic variables may be of explanatory and predictive use in understanding the epidemiological trends of suicide. Early intervention and prevention strategies emerge from this model, and various social, public health, and research implications exist. However, the results must be viewed with caution because of the methodological problems inherent in using national mortality data, the possibility that other variables may account for the observed relationships, and the length of time required to test such prospective epidemiological propositions.

Adolescent↗