SEARCH · Search PubMed

Results for “statistical inference”

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.

SourceResearch connectionFrom yearThrough year

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

At least 1,495 records · Page 83Linked to original sources

PubMed2005

Efficient kernel density estimation of shape and intensity priors for level set segmentation.

We propose a nonlinear statistical shape model for level set segmentation which can be efficiently implemented. Given a set of training shapes, we perform a kernel density estimation in the low dimensional subspace spanned by the training shapes. In this way, we are able to combine an accurate model of the statistical shape distribution with efficient optimization in a finite-dimensional subspace. In a Bayesian inference framework, we integrate the nonlinear shape model with a nonparametric intensity model and a set of pose parameters which are estimated in a more direct data-driven manner than in previously proposed level set methods. Quantitative results show superior performance (regarding runtime and segmentation accuracy) of the proposed nonparametric shape prior over existing approaches.

Algorithms↗

PubMed2004

Pitfalls of and controversies in cluster randomization trials.

It is now well known that standard statistical procedures become invalidated when applied to cluster randomized trials in which the unit of inference is the individual. A resulting consequence is that researchers conducting such trials are faced with a multitude of design choices, including selection of the primary unit of inference, the degree to which clusters should be matched or stratified by prognostic factors at baseline, and decisions related to cluster subsampling. Moreover, application of ethical principles developed for individually randomized trials may also require modification. We discuss several topics related to these issues, with emphasis on the choices that must be made in the planning stages of a trial and on some potential pitfalls to be avoided.

Bias↗

PubMed2003

More "mapping" in brain mapping: statistical comparison of effects.

The term "mapping" in the context of brain imaging conveys to most the concept of localization; that is, a brain map is meant to reveal a relationship between some condition or parameter and specific sites within the brain. However, in reality, conventional voxel-based maps of brain function, or for that matter of brain structure, are generally constructed using analyses that yield no basis for inferences regarding the spatial nonuniformity of the effects. In the normal analysis path for functional images, for example, there is nowhere a statistical comparison of the observed effect in any voxel relative to that in any other voxel. Under these circumstances, strictly speaking, the presence of significant activation serves as a legitimate basis only for inferences about the brain as a unit. In their discussion of results, investigators rarely are content to confirm the brain's role, and instead generally prefer to interpret the spatial patterns they have observed. Since "pattern" implies nonuniform effects over the map, this is equivalent to interpreting results without bothering to test their significance, a practice most of the experimentally-trained would eschew in other contexts. In this review, we appeal to investigators to adopt a new standard of data presentation that facilitates comparison of effects across the map. Evidence for sufficient effect size difference between the effects in structures of interest should be a prerequisite to the interpretation of spatial patterns of activation.

Brain Mapping↗

PubMed1980

MANOVA of statistical moments in biopharmaceutical studies: a numerical example with three equally spaced doses of amoxicillin.

To assess the applicability of statistical moments in biopharmacuetical studies, an urinary excretion trial was conducted with three doses--400, 500, and 600 mg of amoxicillin trihidrate capsules--in 15 healthy volunteers, using a latin square design. The statistical analyses were a two-way MANOVA and a growth curve technique. It was inferred, in the dose range studied, that the amount excreted during the 12 hr following administration, was proportional to dose and that the dose had no effect on either the mean or the variance of the residence time. The calculated minimum significant difference in the amount of amoxicillin excreted was 15% with respect to the mean value.

Absorption↗

PubMed2006

Is there a star tree paradox?

Concerns have been raised that posterior probabilities on phylogenetic trees can be unreliable when the true tree is unresolved or has very short internal branches, because existing methods for Bayesian phylogenetic analysis do not explicitly evaluate unresolved trees. Two recent papers have proposed that evaluating only resolved trees results in a "star tree paradox": when the true tree is unresolved or close to it, posterior probabilities were predicted to become increasingly unpredictable as sequence length grows, resulting in inflated confidence in one resolved tree or another and an increasing risk of false-positive inferences. Here we show that this is not the case; existing Bayesian methods do not lead to an inflation of statistical confidence, provided the evolutionary model is correct and uninformative priors are assumed. Posterior probabilities do not become increasingly unpredictable with increasing sequence length, and they exhibit conservative type I error rates, leading to a low rate of false-positive inferences. With infinite data, posterior probabilities give equal support for all resolved trees, and the rate of false inferences falls to zero. We conclude that there is no star tree paradox caused by not sampling unresolved trees.

Bayes Theorem↗

PubMedDate not supplied

Parametric bootstrap and penalized quasi-likelihood inference in conditional autoregressive models.

This paper discusses a variety of conditional autoregressive (CAR) models for mapping disease rates, beyond the usual first-order intrinsic CAR model. We illustrate the utility and scope of such models for handling different types of data structures. To encourage their routine use for map production at statistical and health agencies, a simple algorithm for fitting such models is presented. This is derived from penalized quasi-likelihood (PQL) inference which uses an analogue of best-linear unbiased estimation for the regional risk ratios and restricted maximum likelihood for the variance components. We offer the practitioner here the use of the parametric bootstrap for inference. It is more reliable than standard maximum likelihood asymptotics for inference purposes since relevant hypotheses for the mapping of rates lie on the boundary of the parameter space. We illustrate the parametric bootstrap test of the practically relevant and important simplifying hypothesis that there is no spatial autocorrelation. Although the parametric bootstrap requires computational effort, it is straightforward to implement and offers a wealth of information relating to the estimators and their properties. The proposed methodology is illustrated by analysing infant mortality in the province of British Columbia in Canada.

Algorithms↗

PubMed2005

Problems of allometric scaling analysis: examples from mammalian reproductive biology.

Biological scaling analyses employing the widely used bivariate allometric model are beset by at least four interacting problems: (1) choice of an appropriate best-fit line with due attention to the influence of outliers; (2) objective recognition of divergent subsets in the data (allometric grades); (3) potential restrictions on statistical independence resulting from phylogenetic inertia; and (4) the need for extreme caution in inferring causation from correlation. A new non-parametric line-fitting technique has been developed that eliminates requirements for normality of distribution, greatly reduces the influence of outliers and permits objective recognition of grade shifts in substantial datasets. This technique is applied in scaling analyses of mammalian gestation periods and of neonatal body mass in primates. These analyses feed into a re-examination, conducted with partial correlation analysis, of the maternal energy hypothesis relating to mammalian brain evolution, which suggests links between body size and brain size in neonates and adults, gestation period and basal metabolic rate. Much has been made of the potential problem of phylogenetic inertia as a confounding factor in scaling analyses. However, this problem may be less severe than suspected earlier because nested analyses of variance conducted on residual variation (rather than on raw values) reveals that there is considerable variance at low taxonomic levels. In fact, limited divergence in body size between closely related species is one of the prime examples of phylogenetic inertia. One common approach to eliminating perceived problems of phylogenetic inertia in allometric analyses has been calculation of 'independent contrast values'. It is demonstrated that the reasoning behind this approach is flawed in several ways. Calculation of contrast values for closely related species of similar body size is, in fact, highly questionable, particularly when there are major deviations from the best-fit line for the scaling relationship under scrutiny.

Analysis of Variance↗

PubMed1992

Poisson sampling-based inference for single ion channel data with time interval omission.

Patch-clamp recording allows investigations of the gating kinetics of single ion channels. Statistical analysis of kinetic data can enhance our understanding of channel gating at a molecular level. Experimental channel records suffer from time interval omission, i.e. failure to detect brief channel openings and closings. It is important to incorporate this phenomenon into statistical analyses of ion channel data. When time interval omission is ignored, the method of maximum likelihood can usually be used to estimate gating parameters from a single channel record. However, it is far more difficult to apply this method when time interval omission is incorporated. We present an alternative approach to parameter estimation based on Poisson sampling. A simulated homogeneous Poisson process is superimposed onto the channel record and inference is based on the numbers of points in successive open and closed sojourns, rather than on the sojourn times themselves. We describe the method for the two-state Markov model C<-->O, although it is applicable to more general models. Computer-simulated data are used to demonstrate the efficacy of the method. Modifications of the method are discussed briefly.

Animals↗

PubMed2006

Mapping quantitative trait loci in noninbred mosquito crosses.

The identification of genes that affect quantitative traits has been of great interest to geneticists for many decades, and many statistical methods have been developed to map quantitative trait loci (QTL). Most QTL mapping studies in experimental organisms use purely inbred lines, where the two homologous chromosomes in each individual are identical. As a result, many existing QTL mapping methods developed for experimental organisms are applicable only to genetic crosses between inbred lines. However, it may be difficult to obtain inbred lines for certain organisms, e.g., mosquitoes. Although statistical methods for QTL mapping in outbred populations, e.g., humans, can be applied for such crosses, these methods may not fully take advantage of the uniqueness of these crosses. For example, we can generally assume that the two grandparental lines are homozygous at the QTL of interest, but such information is not be utilized through methods developed for outbred populations. In addition, mating types and phases can be relatively easy to establish through the analysis of adjacent markers due to the large number of offspring that can be collected, substantially simplifying the computational need. In this article, motivated by a mosquito intercross experiment involving two selected lines that are not genetically homozygous across the genome, we develop statistical methods for QTL mapping for genetic crosses involving noninbred lines. In our procedure, we first infer parental mating types and use likelihood-based methods to infer phases in each parent on the basis of genotypes of offspring and one parent. A hidden Markov model is then employed to estimate the number of high-risk alleles at marker positions and putative QTL positions between markers in each offspring, and QTL mapping is finally conducted through the inferred QTL configuration across all offspring in all crosses. The performance of the proposed methods is assessed through simulation studies, and the usefulness of this method is demonstrated through its application to a mosquito data set.

Alleles↗

PubMed2001

Alternative solutions to the problem of selection bias in an analysis of federal residential drug treatment programs.

In an evaluation of prison-based residential drug treatment programs, the authors use three different regression-based approaches to estimating treatment effects. Two of the approaches, the instrumental variable and the Heckman approach, attempt to minimize selection bias as an explanation for treatment outcomes. Estimates from these approaches are compared with estimates from a regression in which treatment is represented by a dummy variable. The article discusses the advantage of using more than one method to increase confidence in findings when possible selection bias is a concern. Three-year outcome data for 2,315 federal inmates are used in analyses where the authors separately examine criminal recidivism and relapse to drug use for men and women. Statistical tests lead the authors to conclude that treatment reduces criminal recidivism and relapse to drug use. The treatment effect was largest when the inference was based on the Heckman approach, somewhat smaller when based on the instrumental variable approach, and smallest when based on the traditional dummy variable approach. Treatment effects for females were not statistically significant.

Crime↗

PubMed2006

Evolutionary model selection with a genetic algorithm: a case study using stem RNA.

The choice of a probabilistic model to describe sequence evolution can and should be justified. Underfitting the data through the use of overly simplistic models may miss out on interesting phenomena and lead to incorrect inferences. Overfitting the data with models that are too complex may ascribe biological meaning to statistical artifacts and result in falsely significant findings. We describe a likelihood-based approach for evolutionary model selection. The procedure employs a genetic algorithm (GA) to quickly explore a combinatorially large set of all possible time-reversible Markov models with a fixed number of substitution rates. When applied to stem RNA data subject to well-understood evolutionary forces, the models found by the GA 1) capture the expected overall rate patterns a priori; 2) fit the data better than the best available models based on a priori assumptions, suggesting subtle substitution patterns not previously recognized; 3) cannot be rejected in favor of the general reversible model, implying that the evolution of stem RNA sequences can be explained well with only a few substitution rate parameters; and 4) perform well on simulated data, both in terms of goodness of fit and the ability to estimate evolutionary rates. We also investigate the utility of several distance measures for comparing and contrasting inferred evolutionary models. Using widely available small computer clusters, our approach allows, for the first time, to evaluate the performance of existing RNA evolutionary models by comparing them with a large pool of candidate models and to validate common modeling assumptions. In addition, the new method provides the foundation for rigorous selection and comparison of substitution models for other types of sequence data.

Algorithms↗

PubMed2026

Is a Win-Win possible? Achieving pareto-optimal privacy-utility balance in fine-tuned genome language model embeddings against embedding reconstruction attacks.

MOTIVATION: Genomic data is among the most sensitive categories of personal information, and the growing adoption of language models for sequence analysis raises significant privacy concerns. Prior work demonstrated that embeddings from general-purpose language models adapted for genomic sequences leak substantial single-nucleotide information under reconstruction attacks, and that fine-tuning embeddings can reduce this vulnerability at certain positions. However, three critical questions remain unaddressed: (i) whether privacy-utility tradeoffs are inherent constraints or configuration-dependent phenomena; (ii) whether genomic-specialized models such as DNABERT-base and Nucleotide Transformer exhibit different vulnerabilities than adapted general-purpose models; and (iii) how to statistically validate whether observed privacy improvements represent meaningful gains. Addressing these gaps is essential for guiding model selection in privacy-sensitive genomic applications. RESULTS: We systematically evaluated 13 transformer architectures, 9 general-purpose and 4 genomic-specialized, under position-specific embedding reconstruction attacks. We assessed the vulnerabilities of both pre-trained and fine-tuned models to the single-nucleotide inference-reconstruction attack using our new metrics, including error-based privacy gain and Pareto dominance scores, and statistically validated the results via paired t-tests. XLNet-Large achieved the best observed privacy protection among all evaluated models (+19.5% mean privacy gain) while maintaining competitive prediction performance. General-purpose models outperformed genomic-specialized models in 56% of pairwise comparisons. Tokenization strategy, rather than domain specialization, emerged as the primary determinant of the privacy-utility balance. These findings provide evidence-based guidance for selecting models in privacy-sensitive short-window genomic applications. All privacy claims in this work are specific to position-wise embedding reconstruction attacks and do not extend to other privacy risks, such as membership inference or training data extraction, which may respond differently to fine-tuning. AVAILABILITY AND IMPLEMENTATION: The code is publicly available at https://github.com/AnonymousISCBConf/Win-Win-Privacy-Utility-Analysis.

Genomics↗

PubMed2006

Optimization of analytical parameters for inferring relationships among Escherichia coli isolates from repetitive-element PCR by maximizing correspondence with multilocus sequence typing data.

Repetitive-element PCR (rep-PCR) is a method for genotyping bacteria based on the selective amplification of repetitive genetic elements dispersed throughout bacterial chromosomes. The method has great potential for large-scale epidemiological studies because of its speed and simplicity; however, objective guidelines for inferring relationships among bacterial isolates from rep-PCR data are lacking. We used multilocus sequence typing (MLST) as a "gold standard" to optimize the analytical parameters for inferring relationships among Escherichia coli isolates from rep-PCR data. We chose 12 isolates from a large database to represent a wide range of pairwise genetic distances, based on the initial evaluation of their rep-PCR fingerprints. We conducted MLST with these same isolates and systematically varied the analytical parameters to maximize the correspondence between the relationships inferred from rep-PCR and those inferred from MLST. Methods that compared the shapes of densitometric profiles ("curve-based" methods) yielded consistently higher correspondence values between data types than did methods that calculated indices of similarity based on shared and different bands (maximum correspondences of 84.5% and 80.3%, respectively). Curve-based methods were also markedly more robust in accommodating variations in user-specified analytical parameter values than were "band-sharing coefficient" methods, and they enhanced the reproducibility of rep-PCR. Phylogenetic analyses of rep-PCR data yielded trees with high topological correspondence to trees based on MLST and high statistical support for major clades. These results indicate that rep-PCR yields accurate information for inferring relationships among E. coli isolates and that accuracy can be enhanced with the use of analytical methods that consider the shapes of densitometric profiles.

Alleles↗

PubMed2006

Singularities affect dynamics of learning in neuromanifolds.

The parameter spaces of hierarchical systems such as multilayer perceptrons include singularities due to the symmetry and degeneration of hidden units. A parameter space forms a geometrical manifold, called the neuromanifold in the case of neural networks. Such a model is identified with a statistical model, and a Riemannian metric is given by the Fisher information matrix. However, the matrix degenerates at singularities. Such a singular structure is ubiquitous not only in multilayer perceptrons but also in the gaussian mixture probability densities, ARMA time-series model, and many other cases. The standard statistical paradigm of the Cramér-Rao theorem does not hold, and the singularity gives rise to strange behaviors in parameter estimation, hypothesis testing, Bayesian inference, model selection, and in particular, the dynamics of learning from examples. Prevailing theories so far have not paid much attention to the problem caused by singularity, relying only on ordinary statistical theories developed for regular (nonsingular) models. Only recently have researchers remarked on the effects of singularity, and theories are now being developed. This article gives an overview of the phenomena caused by the singularities of statistical manifolds related to multilayer perceptrons and gaussian mixtures. We demonstrate our recent results on these problems. Simple toy models are also used to show explicit solutions. We explain that the maximum likelihood estimator is no longer subject to the gaussian distribution even asymptotically, because the Fisher information matrix degenerates, that the model selection criteria such as AIC, BIC, and MDL fail to hold in these models, that a smooth Bayesian prior becomes singular in such models, and that the trajectories of dynamics of learning are strongly affected by the singularity, causing plateaus or slow manifolds in the parameter space. The natural gradient method is shown to perform well because it takes the singular geometrical structure into account. The generalization error and the training error are studied in some examples.

Journal Article↗

PubMed2001

Genetic and environmental adaptation in high altitude natives. Conceptual, methodological, and statistical concerns.

A great number of physiological and anthropological studies have investigated Andean and Himalayan populations native to high altitude (HA). A non-scientific survey of the extant literature reveals a relatively liberal tradition of inferring genetic (evolutionary) adaptation to HA in these groups, often based on limited evidence and/or based on study designs insufficient to fully address the issue. Rather than review the evidence for or against genetic adaptation, and in order to provide some perspective, this paper will review relevant conceptual, methodological, and statistical issues that are germane to the study of HA native human groups. In particular, focus will be on the limitations of the most common research approach which bases evolutionary inference on the comparison of phenotypic mean differences between highland and lowland native populations. The migrant study approach is discussed, as is a relatively new approach based on genetic admixture in hybrid populations.

Acclimatization↗

PubMed1989

A statistical test of phylogenies estimated from sequence data.

A simple approach to testing the significance of the branching order, estimated from protein or DNA sequence data, of three taxa is proposed. The branching order is inferred by the transformed-distance method, under the assumption that one or two outgroups are available, and the branch lengths are estimated by the least-squares method. The inferred branching order is considered significant if the estimated internodal distance is significantly greater than zero. To test this, a formula for the variance of the internodal distance has been developed. The statistical test proposed has been checked by computer simulation. The same test also applies to the case of four taxa with no outgroup, if one considers an unrooted tree. Formulas for the variances of internodal distances have also been developed for the case of five taxa. Conditions are given under which it is more efficient to add the sequence of a fifth taxon than to do 25% more nucleotide sequencing in each of the original four. A method is presented for combining analyses of disparate data to get a single P value. Finally, the test, applied to the human-chimpanzee-gorilla problem, shows that the issue is not yet resolved.

Animals↗

PubMed2000

Nature of altered pulsatile hormone release and neuroendocrine network signalling in human ageing: clinical studies of the somatotropic, gonadotropic, corticotropic and insulin axes.

Recent clinical investigations have implemented an array of new analytical tools to evaluate the neuroregulation of endocrine axes. These studies demonstrate multifold disruption within the growth hormone (GH), luteinizing hormone (LH)-testosterone, adrenocorticotropin (ACTH)-cortisol and the insulin axes in healthy ageing men and women. Novel research strategies in ageing include such developments as the indirect in vivo assessment of neuroendocrine network integration, via the approximate entropy (ApEn) statistic to monitor the unihormonal orderliness and bihormonal synchronicity of hormone release, and thus infer stability of network-integrative processes. For example, ApEn calculations show that the individual orderliness of GH, insulin or LH release falls progressively in older men and women, and the conditional synchrony between LH and testosterone (or LH and follicle-stimulating hormone/prolactin) release, and LH secretion and the neurogenically organized signal, nocturnal penile tumescence (NPT), all decline markedly in older men. Evaluation of the ACTH-cortisol axis points additionally to disrupted bihormonal synchrony within this stress-responsive system in healthy ageing. A complementary investigative tool, viz. a stochastic differential equation random-effects feedback construct of the interactive male gonadotropin-releasing hormone-LH-testosterone axis, predicts that only certain extant postulates of ageing in the male reproductive axis will give rise to the observed erosion of LH-testosterone synchrony. Collectively, available clinical data suggest a general model of early neuroendocrine ageing in the human, in both the male and female, wherein ageing is marked by variable disruption in the time-delayed feedback and feedforward interconnections among neuroendocrine glands, which constitute an integrated axis and which control the joint synchrony of hormone release.

Adrenocorticotropic Hormone↗

PubMed2002

Spike-timing dependent synaptic plasticity: a phenomenological framework.

In this paper a phenomenological model of spike-timing dependent synaptic plasticity (STDP) is developed that is based on a Volterra series-like expansion. Synaptic weight changes as a function of the relative timing of pre- and postsynaptic spikes are described by integral kernels that can easily be inferred from experimental data. The resulting weight dynamics can be stated in terms of statistical properties of pre- and postsynaptic spike trains. Generalizations to neurons that fire two different types of action potentials, such as cerebellar Purkinje cells where synaptic plasticity depends on correlations in two distinct presynaptic fibers, are discussed. We show that synaptic plasticity, together with strictly local bounds for the weights, can result in synaptic competition that is required for any form of pattern formation. This is illustrated by a concrete example where a single neuron equipped with STDP can selectively strengthen those synapses with presynaptic neurons that reliably deliver precisely timed spikes at the expense of other synapses which transmit spikes with a broad temporal distribution. Such a mechanism may be of vital importance for any neuronal system where information is coded in the timing of individual action potentials.

Action Potentials↗