Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “model selection”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

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

At least 1,081 records · Page 60Linked to original sources
PubMed2004

Cross-validation for linear model with unequal variances in genomic analysis.

In recent years, genomic studies are usually conducted to identify genes that may have an impact on clinical outcomes. The identified genes are then used to establish a predictive model for identifying subjects who are most likely to respond to the test treatment in clinical trials. This information is useful in early and later phases of clinical development. The United States Food and Drug Administration (FDA) requires that such a predictive model be validated before it can be used in clinical development. Shao [Shao, J. (1993). Linear model selection by cross-validation. J. Amer. Statist. Assoc. 88(422):486-494] proposed a cross-validation method for linear model with equal variances, which is found useful in genomic studies. In practice, however, genomic data may be obtained from different sources with unequal variances. As a result, Shao's method may not be applied directly. In this paper, we extend Shao's method for cross-validation of a linear model with unequal variances. Along this line, two re-sampling methods were proposed to account for the heterogeneity in variance. Several simulations were performed to evaluate the finite samples performances of the proposed methods. An example concerning a breast cancer research is present to illustrate the use of the proposed methods.

Algorithms↗
PubMed2004

Fully Bayesian spatio-temporal modeling of FMRI data.

We present a fully Bayesian approach to modeling in functional magnetic resonance imaging (FMRI), incorporating spatio-temporal noise modeling and haemodynamic response function (HRF) modeling. A fully Bayesian approach allows for the uncertainties in the noise and signal modeling to be incorporated together to provide full posterior distributions of the HRF parameters. The noise modeling is achieved via a nonseparable space-time vector autoregressive process. Previous FMRI noise models have either been purely temporal, separable or modeling deterministic trends. The specific form of the noise process is determined using model selection techniques. Notably, this results in the need for a spatially nonstationary and temporally stationary spatial component. Within the same full model, we also investigate the variation of the HRF in different areas of the activation, and for different experimental stimuli. We propose a novel HRF model made up of half-cosines, which allows distinct combinations of parameters to represent characteristics of interest. In addition, to adaptively avoid over-fitting we propose the use of automatic relevance determination priors to force certain parameters in the model to zero with high precision if there is no evidence to support them in the data. We apply the model to three datasets and observe matter-type dependence of the spatial and temporal noise, and a negative correlation between activation height and HRF time to main peak (although we suggest that this apparent correlation may be due to a number of different effects).

Bayes Theorem↗
PubMed2000

Bounds on error expectation for support vector machines.

We introduce the concept of span of support vectors (SV) and show that the generalization ability of support vector machines (SVM) depends on this new geometrical concept. We prove that the value of the span is always smaller (and can be much smaller) than the diameter of the smallest sphere containing the support vectors, used in previous bounds (Vapnik, 1998). We also demonstrate experimentally that the prediction of the test error given by the span is very accurate and has direct application in model selection (choice of the optimal parameters of the SVM).

Learning↗
PubMed2006

An oscillatory neural model of multiple object tracking.

An oscillatory neural network model of multiple object tracking is described. The model works with a set of identical visual objects moving around the screen. At the initial stage, the model selects into the focus of attention a subset of objects initially marked as targets. Other objects are used as distractors. The model aims to preserve the initial separation between targets and distractors while objects are moving. This is achieved by a proper interplay of synchronizing and desynchronizing interactions in a multilayer network, where each layer is responsible for tracking a single target. The results of the model simulation are presented and compared with experimental data. In agreement with experimental evidence, simulations with a larger number of targets have shown higher error rates. Also, the functioning of the model in the case of temporarily overlapping objects is presented.

Animals↗
PubMed2026

A multi-ancestry polygenic risk score for body mass index predicts longitudinal weight change.

BACKGROUND: Identifying individuals at risk for future weight gain is challenging, partly because associations with traditional clinical risk factors may be biased by confounding and reverse causation. Polygenic risk scores (PRS) provide a stable, lifelong measure of genetic predisposition to obesity. However, existing PRS have not been evaluated for their association with longitudinal weight change in adulthood and often lack generalizability across diverse genetic ancestry groups. METHODS: We conducted ancestry-specific genome-wide association study meta-analyses of body mass index (BMI) in populations of European, African or African American, Admixed American, East Asian, and South Asian ancestries and developed ancestry-specific PRS. A multi-ancestry polygenic risk score (MAPRS) was trained using ancestry-specific PRS in a model selection dataset (N = 39,685) from the All of Us Research Program (AoU). We evaluated the MAPRS in an independent AoU model evaluation dataset (N = 158,743) for BMI prediction and in a separate AoU test dataset (N = 78,219) with repeated measurements over 1.5-2.5 years for weight change prediction. The outcomes included change in BMI and ≥ 10% or ≥ 5% total body weight (TBW) gain. We further examined the relationship between MAPRS and 12 clinical risk factors commonly comorbid with obesity in relation to weight change. RESULTS: The MAPRS captured 7.05% of the variance in measured BMI in the AoU model evaluation dataset and demonstrated improved generalizability across all non-European genetic ancestry groups. In the AoU test dataset, conditioned on baseline BMI at the second-to-last measurement, a one SD increase in MAPRS was associated with a 0.16 kg/m2 increase in future BMI (standard error = 0.012 kg/m2; p-value = 2.2 × 10-39), 1.27-fold increased odds of experiencing ≥ 10% TBW gain (95% CI: 1.24-1.31; p-value = 1.4 × 10-55), and 1.15-fold increased odds of experiencing ≥ 5% TBW gain (95% CI: 1.13-1.18; p-value = 2.8 × 10-39). These associations were observed across all genetic ancestry groups and remained highly consistent after adjustment for any clinical risk factor. In contrast, most clinical risk factors demonstrated inconsistent or weaker associations with weight change outcomes. CONCLUSIONS: We developed an MAPRS for BMI that represents a robust and generalizable risk factor for longitudinal weight gain in adulthood, providing a foundation for genetically informed risk stratification and earlier, more targeted obesity prevention strategies.

Humans↗
PubMed2003

Modeling the human lumbar spine for assessing spinal loads, stability, and risk of injury.

This article provides a critical review of biomechanical modeling approaches used to estimate spinal loads, stability, and risk of injury. The complete biomechanical analysis of the spine requires a two-stage approach: (1) equilibrium analysis for estimating tissue loads, and (2) stability analysis for estimating structural tolerance of the spine. For each level of analysis, basic principles are reviewed and relevant published models and studies are referenced. Throughout the article, implications for lumbar function and dysfunction derived from the various modeling methods are presented, and their applications are discussed. At the end, future directions for research are identified. Detailed descriptions of selected modeling approaches are provided in the appendices.

Compressive Strength↗
PubMed2003

Supervised detection of regulatory motifs in DNA sequences.

Identification of transcription factor binding sites (regulatory motifs) is a major interest in contemporary biology. We propose a new likelihood based method, COMODE, for identifying structural motifs in DNA sequences. Commonly used methods (e.g. MEME, Gibbs motif sampler) model binding sites as families of sequences described by a position weight matrix (PWM) and identify PWMs that maximize the likelihood of observed sequence data under a simple multinomial mixture model. This model assumes that the positions of the PWM correspond to independent multinomial distributions with four cell probabilities. We address supervising the search for DNA binding sites using the information derived from structural characteristics of protein-DNA interactions. We extend the simple multinomial mixture model to a constrained multinomial mixture model by incorporating constraints on the information content profiles or on specific parameters of the motif PWMs. The parameters of this extended model are estimated by maximum likelihood using a nonlinear constraint optimization method. Likelihood-based cross-validation is used to select model parameters such as motif width and constraint type. The performance of COMODE is compared with existing motif detection methods on simulated data that incorporate real motif examples from Saccharomyces cerevisiae. The proposed method is especially effective when the motif of interest appears as a weak signal in the data. Some of the transcription factor binding data of Lee et al. (2002) were also analyzed using COMODE and biologically verified sites were identified.

Journal Article↗
PubMed2000

From guideline modeling to guideline execution: defining guideline-based decision-support services.

We describe our task-based approach to defining the guideline-based decision-support services that the EON system provides. We categorize uses of guidelines in patient-specific decision support into a set of generic tasks--making of decisions, specification of work to be performed, interpretation of data, setting of goals, and issuance of alert and reminders--that can be solved using various techniques. Our model includes constructs required for representing the knowledge used by these techniques. These constructs form a toolkit from which developers can select modeling solutions for guideline task. Based on the tasks and the guideline model, we define a guideline-execution architecture and a model of interactions between a decision-support server and clients that invoke services provided by the server. These services use generic interfaces derived from guideline tasks and their associated modeling constructs. We describe two implementations of these decision-support services and discuss how this work can be generalized. We argue that a well-defined specification of guideline-based decision-support services will facilitate sharing of tools that implement computable clinical guidelines.

Asthma↗
PubMed2002

Evolutionary experiments on mate recognition in the Drosophila serrata species complex.

It is becoming increasingly apparent that at least some aspects of the evolution of mate recognition may be amenable to manipulation in evolutionary experiments. Quantitative genetic analyses that focus on the genetic consequences of evolutionary processes that result in mate recognition evolution may eventually provide an understanding of the genetic basis of the process of speciation. We review a series of experiments that have attempted to determine the genetic basis of the response to natural and sexual selection on mate recognition in the Drosophila serrata species complex. The genetic basis of mate recognition has been investigated at three levels: (1) between the species of D. serrata and D. birchii using interspecific hybrids, (2) between populations of D. serrata that are sympatric and allopatric with respect to D. birchii, and (3) within populations of D. serrata. These experiments suggest that it may be possible to use evolutionary experiments to observe important events such as the reinforcement of mate recognition, or the generation of the genetic associations that are central to many sexual selection models.

Animals↗
PubMed1986

Immunological variables as predictors of prognosis in patients with Kaposi's sarcoma and the acquired immunodeficiency syndrome.

Multivariate analysis was used to identify which of a large number of pretreatment immunological parameters correlated with therapeutic response, subsequent development of opportunistic infection, and survival from the time of diagnosis in a group of 70 patients with Kaposi's sarcoma and acquired immunodeficiency syndrome treated with recombinant leukocyte A interferon. In a logistic regression model, delayed type hypersensitivity response to one or more recall antigens and high proliferative response to Escherichia coli were significant predictors for response to recombinant leukocyte A interferon (for the model, P = 0.01). For prediction of the development of opportunistic infection, the model selected low proliferative responses to phytohemagglutinin and E. coli (P less than 0.001). Favorable factors predicting survival in the Cox regression model were the absence of endogenous serum interferon activity and a high proliferative response to E. coli (P less than 0.001). The estimated median survival for the group with endogenous serum interferon activity and low E. coli response was 12 months; the median has not yet been reached for the group with no serum interferon and a high E. coli response. We conclude that immunological parameters may be useful in predicting prognosis in patients with Kaposi's sarcoma and acquired immunodeficiency syndrome.

Acquired Immunodeficiency Syndrome↗
PubMed2002

Mutation-selection balance: ancestry, load, and maximum principle.

We analyze the equilibrium behavior of deterministic haploid mutation-selection models. To this end, both the forward and the time-reversed evolution processes are considered. The stationary state of the latter is called the ancestral distribution, which turns out as a key for the study of mutation-selection balance. We find that the ancestral genotype frequencies determine the sensitivity of the equilibrium mean fitness to changes in the corresponding fitness values and discuss implications for the evolution of mutational robustness. We further show that the difference between the ancestral and the population mean fitness, termed mutational loss, provides a measure for the sensitivity of the equilibrium mean fitness to changes in the mutation rate. The interrelation of the loss and the mutation load is discussed. For a class of models in which the number of mutations in an individual is taken as the trait value, and fitness is a function of the trait, we use the ancestor formulation to derive a simple maximum principle, from which the mean and variance of fitness and the trait may be derived; the results are exact for a number of limiting cases, and otherwise yield approximations which are accurate for a wide range of parameters. These results are applied to threshold phenomena caused by the interplay of selection and mutation (known as error thresholds). They lead to a clarification of concepts, as well as criteria for the existence of error thresholds.

Animals↗
PubMed2004

Comparison of toxicological impacts of integrated and chemical pest management in Mediterranean greenhouses.

The goal of this paper is to assess the relative impacts of pest-control methods in greenhouses, based on current LCA tools. As a case study, the relative impacts of two tomato production methods, chemical pest management (CPM) and integrated pest management (IPM), are assessed. The amount of the active ingredients applied, the fate of the ingredients in the various greenhouse and environmental compartments, the human exposure routes via the various compartments and the inherent toxicity of the ingredients were taken into account in the relative impact calculations. To assess the importance of model selection in the assessment, pesticide-specific fate and exposure factors for humans and aquatic and terrestrial ecosystems, used to aggregate pesticide emissions, were calculated with two different models: (1) the USES-LCA model, adapted in order to calculate the pesticide transfer from greenhouse air and soil to fruits, and (2) the empirical model critical surface time (CST). Impact scores have in general shown a higher level of potential contamination in greenhouses treated with CPM compared to IPM (a factor of 1.4 to 2.3). Relative impacts have been shown highly dependent on the selection of specific pesticides and crop stage development at the moment of pesticide application. This means that both CPM and IPM could be improved by a careful selection of pesticides. In order to improve the relative impact calculations, future research in pesticide transfer to food will be necessary.

Agriculture↗
PubMed2000

Body weight and tail length divergence in mice selected for rate of development.

A series of mouse lines has been produced by 19 generations of restricted index selection for rate of development during early and late ontogeny. The selection program was based on an index with the following four replicated selection treatments: E(+) and E(-) were selected to alter birth to 10-day body weight gain while holding late gain for both selection lines constant; correspondingly, L(+) and L(-) were selected to alter 28- to 56-day body weight gain holding early gain for both lines constant. Herein, we characterize response to selection for growth rate by analyzing age-specific mouse body weight and tail lengths and for growth curves using a logistics model. Selection on developmental rate has resulted in divergence in both age-specific and growth curve traits. E(+) and L(+) lines reached identical weights during the late selection interval, then diverged to unique mature weights. E(-) and L(-) lines similarly achieved identical weights during late selection and diverged to unique mature weights. However, the shapes of early and late growth curves were significantly divergent, and at least two distinct growth patterns are shown to result from selection. Response in body weight gain was accompanied by similar, though less pronounced, change in tail length traits. Significant response during intervals of restricted growth was also found, especially in lines selected for late gain. The evolution of the growth trajectory under restricted index selection is discussed in terms of drift and available additive genetic variation and covariation.

Aging↗
PubMed1998

[Epidemiologic and statistical interaction models and consequences for regression analysis].

Different ways to define interaction between exposition factors in epidemiological studies as well as the choice between additive and multiplicative no interaction leads frequently to confusion during data analysis. In their standard form methods of event data analysis such as Poisson or logistic regression assume a multiplicative parameterization of no interaction. However, evidence from empirical investigations as well as causal models of disease etiology, e.g. the simple independent action model of Finney or the sufficient-component-causes model of Rothman, suggest additive or other kinds of non-multiplicative concepts of no interaction. For additive structured data we illustrate the asymptotic bias ("interaction-bias") of main effect estimates which are based on inappropriate data analysis using multiplicative models and omitting significant or non-significant interaction terms. We show that both the epidemiological study design as well as the underlying causal model are determinants of the interaction structure of the data and should be considered in the model selection process. The definition of interaction should distinguish between risk, rate and odds if risks are not very small. Using generalized linear models with parametrical link functions we are able to analyze non-multiplicative interaction structures.

Adult↗
PubMed2006

Three dimensional model of severe acute respiratory syndrome coronavirus helicase ATPase catalytic domain and molecular design of severe acute respiratory syndrome coronavirus helicase inhibitors.

The modeling of the severe acute respiratory syndrome coronavirus helicase ATPase catalytic domain was performed using the protein structure prediction Meta Server and the 3D Jury method for model selection, which resulted in the identification of 1JPR, 1UAA and 1W36 PDB structures as suitable templates for creating a full atom 3D model. This model was further utilized to design small molecules that are expected to block an ATPase catalytic pocket thus inhibit the enzymatic activity. Binding sites for various functional groups were identified in a series of molecular dynamics calculation. Their positions in the catalytic pocket were used as constraints in the Cambridge structural database search for molecules having the pharmacophores that interacted most strongly with the enzyme in a desired position. The subsequent MD simulations followed by calculations of binding energies of the designed molecules were compared to ATP identifying the most successful candidates, for likely inhibitors - molecules possessing two phosphonic acid moieties at distal ends of the molecule.

Amino Acid Sequence↗
PubMed2005

Mortality after cardiac bypass surgery: prediction from administrative versus clinical data.

BACKGROUND: Risk-adjusted outcome rates frequently are used to make inferences about hospital quality of care. We calculated risk-adjusted mortality rates in veterans undergoing isolated coronary artery bypass surgery (CABS) from administrative data and from chart-based clinical data and compared the assessment of hospital high and low outlier status for mortality that results from these 2 data sources. STUDY POPULATION: We studied veterans who underwent CABS in 43 VA hospitals between October 1, 1993, and March 30, 1996 (n=15,288). METHODS: To evaluate administrative data, we entered 6 groups of International Classification of Diseases (ICD)-9-CM codes for comorbid diagnoses from the VA Patient Treatment File (PTF) into a logistic regression model predicting postoperative mortality. We also evaluated counts of comorbid ICD-9-CM codes within each group, along with 3 common principal diagnoses, weekend admission or surgery, major procedures associated with CABS, and demographic variables. Data from the VA Continuous Improvement in Cardiac Surgery Program (CICSP) were used to create a separate clinical model predicting postoperative mortality. For each hospital, an observed-to-expected (O/E) ratio of mortality was calculated from (1) the PTF model and (2) the CICSP model. We defined outlier status as an O/E ratio outside of 1.0 (based on the hospital's 90% confidence interval). To improve the statistical and predictive power of the PTF model, selected clinical variables from CICSP were added to it and outlier status reassessed. RESULTS: Significant predictors of postoperative mortality in the PTF model included 1 group of comorbid ICD-9-CM codes, intraortic balloon pump insertion before CABS, angioplasty on the day of or before CABS, weekend surgery, and a principal diagnosis of other forms of ischemic heart disease. The model's c-index was 0.698. As expected, the CICSP model's predictive power was significantly greater than that of the administrative model (c=0.761). The addition of just 2 CICSP variables to the PTF model improved its predictive power (c=0.741). This model identified 5 of 6 high mortality outliers identified by the CICSP model. Additional CICSP variables were statistically significant predictors but did not improve the assessment of high outlier status. CONCLUSIONS: Models using administrative data to predict postoperative mortality can be improved with the addition of a very small number of clinical variables. Limited clinical improvements of administrative data may make it suitable for use in quality improvement efforts.

Cardiopulmonary Bypass↗
PubMed1990

The perception of psychiatric-hospital ward sisters/charge nurses towards nursing models.

A search of the literature revealed a dearth of research on hospital ward-based practitioners' opinions/perceptions towards nursing models. This is surprising considering that such frameworks are being introduced into patient care settings on an international basis. As part of a larger investigation pertaining to model selection, the researcher interviewed a population of 95 ward managers, giving each of them the opportunity to put forward their views concerning models of nursing. The research area was psychiatric continuing-care/rehabilitation facilities in Northern Ireland. The majority of opinions (61.1%) were positive in nature while 23.2% of respondents voiced negative perceptions. Implications for nursing practice, education and administration are raised from the study's findings.

Adult↗
PubMedDate not supplied

[Receptor selectivity profile of short enkephalin analogs containing D-ornithine at the second position of the molecule].

The radioreceptor binding assay using a membrane fraction from the rat brain was applied to study binding activity and opiate receptor selectivity of truncated enkephalin analogs (with a free or modified C-terminal carboxyl group) bearing D-ornithine at the second position. D-ornithine introduction does not lead to the increase of mu-receptor selectivity as is proposed by the Schwyzer membrane selection model. Some additional modifications (C-terminal amidation or ornithine side chain acetylation) of tetrapeptide molecule were required to reach mu-selectivity. Simultaneous replacement of glycine by D-ornithine and C-terminal carboxyl amidation resulted in the potent analog whose selectivity toward the mu-receptors was 113 times as high as that of leucine-enkephalin. D-ornithine side chain prolongation by means of the attachment of some amino acid residues (methionine, leucine, proline, asparagine) led to the original branched enkephalin analogs. Studies of their binding activity showed that the best branched analog was only 27 times more selective for mu-vs delta-opiate receptors.

Animals↗