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 973 records · Page 54Linked to original sources

PubMed2000

Approximate maximum entropy joint feature inference consistent with arbitrary lower-order probability constraints: application to statistical classification

We propose a new learning method for discrete space statistical classifiers. Similar to Chow and Liu (1968) and Cheeseman (1983), we cast classification/inference within the more general framework of estimating the joint probability mass function (p.m.f.) for the (feature vector, class label) pair. Cheeseman's proposal to build the maximum entropy (ME) joint p.m.f. consistent with general lower-order probability constraints is in principle powerful, allowing general dependencies between features. However, enormous learning complexity has severely limited the use of this approach. Alternative models such as Bayesian networks (BNs) require explicit determination of conditional independencies. These may be difficult to assess given limited data. Here we propose an approximate ME method, which, like previous methods, incorporates general constraints while retaining quite tractable learning. The new method restricts joint p.m.f. support during learning to a small subset of the full feature space. Classification gains are realized over dependence trees, tree-augmented naive Bayes networks, BNs trained by the Kutato algorithm, and multilayer perceptrons. Extensions to more general inference problems are indicated. We also propose a novel exact inference method when there are several missing features.

Journal Article↗

PubMed2005

Phylogenetic analyses of fat body endosymbionts reveal differences in invasion times of blaberid wood-feeding cockroaches (Blaberidae: Panesthiinae) into the Japanese archipelago.

Cockroaches have endosymbiotic bacteria in their fat bodies. Recent molecular phylogenetic analyses on both hosts and endosymbionts have revealed that co-evolution has occurred throughout the history of cockroaches and termites. Co-cladogenesis was also shown among closely related taxa (woodroach genus Cryptocercus; Cryptocercidae), and thus endosymbiont data are likely to be useful for biogeographical analyses. To test the possibility of co-cladogenesis among inter-and intraspecific taxa, as well as the utility of endosymbiont data for inferring biogeographical scenarios, we analyzed rRNA genes of endosymbionts of Japanese and Taiwanese Panesthiinae (Salganea and Panesthia; Blaberidae), on which phylogenetic analyses previously had been performed based on the mitochondrial genes. Statistical analyses on the topologies inferred from both endosymbiont and host mitochondria genes showed that co-cladogenesis has occurred. The endosymbiont sequences examined appear to have evolved in a clock-like manner, and their rate of evolution based on the host fossil data showed a major difference in the time of invasion of the two Japanese genera, that is congruent with the recent analyses of their mitochondrial genes.

Animals↗

PubMed1981

Biological models and statistical interactions: an example from multistage carcinogenesis.

From the assessment of statistical interaction between risk factors it is tempting to infer the nature of the biologic interaction between the factors. However, the use of statistical analyses of epidemiologic data to infer biologic processes can be misleading. as an example, we consider the multistage model of carcinogenesis. Under this biologic model, it is shown, by means of simple hypothetical examples, that even if carcinogenic factors act independently, some pairs may fit an additive statistical model, some a multiplicative statistical model, and some neither. The elucidation of biological interactions by means of statistical models requires the imaginative and prudent use of inductive and deductive reasoning; it cannot be done mechanically.

Cell Transformation, Neoplastic↗

PubMed2005

Using weighted permutation scores to detect differential gene expression with microarray data.

A class of nonparametric statistical methods, including a nonparametric empirical Bayes (EB) method, the Significance Analysis of Microarrays (SAM) and the mixture model method (MMM) have been proposed to detect differential gene expression for replicated microarray experiments. They all depend on constructing a test statistic, for example, a t-statistic, and then using permutation to draw inferences. However, due to special features of microarray data, using standard permutation scores may not estimate the null distribution of the test statistic well, leading to possibly too conservative inferences. We propose a new method of constructing weighted permutation scores to overcome the problem: posterior probabilities of having no differential expression from the EB method are used as weights for genes to better estimate the null distribution of the test statistic. We also propose a weighted method to estimate the false discovery rate (FDR) using the posterior probabilities. Using simulated data and real data for time-course microarray experiments, we show the improved performance of the proposed methods when implemented in MMM, EB and SAM.

Algorithms↗

PubMed2004

Prevalence of antimitochondrial antibody in Japanese corporate workers in Kanagawa prefecture.

BACKGROUND: The prevalence of antimitochondrial antibody (AMA) in humans and its relationship to the development of primary biliary cirrhosis (PBC) are not well known. We have estimated the frequency of AMA in the general population, and studied its association with PBC. METHODS: We studies 1714 corporate workers (median age, 48 years; range, 30 to 59 years) who had an annual health check from 1998 to 1999 at Kawasaki Social Insurance Hospital in Japan. We used an indirect immunofluorescence method for screening serum AMA. We applied the prevalence of AMA-positive people in the study group to the general population in Japan. Then the inferred AMA-positive population was compared to the actual number of patients with PBC in statistics published by the Japanese Government. RESULTS: AMA was detected in 11 of 1714 people (0.64%; 95% confidence interval, 0.26% to 1.02%). All these 11 sera reacted with 2-oxoacid-dehydrogenase complex by immunoblotting. Of these 11 individuals, none had subjective symptoms, all had normal serum bilirubin levels, and 6 had abnormal liver function test results. Using published statistics for the Japanese population, we inferred that there were approximately 336,472 AMA-positive people in Japan from age 30 to 59 years. The number of patients with symptomatic PBC recorded by the nationwide epidemiological survey of the Japanese Government was 2459. Thus, we inferred the rate of symptomatic PBC among AMA-positive persons to be about 0.73% (2459/336,472). CONCLUSIONS: AMA is not a rare antibody in the general population, but few people develop recognizable PBC even if they have AMA.

Adult↗

PubMed1995

Inference for the association between coefficients in a multivariate growth curve model.

This paper generalizes the work of Blomqvist (1977, Journal of the American Statistical Association 72, 746-749) on inference for the relationship between the individual-specific slope and the individual-specific intercept in a linear growth curve model. The paper deals with longitudinal data involving one or more response variables and irregular follow-up times, with each response variable postulated to follow a linear growth curve model. The problem considered is inference concerning the association between one growth curve coefficient and another--for example, the slope and intercept for a selected response variable, or the two slopes for two different response variables--after adjusting for all remaining coefficients among all of the response variables. An inferential approach based on the method of moments and an inferential approach based on maximum likelihood are described, and the asymptotic properties of these procedures are presented. Extensions of the methodology to allow polynomial growth curves and baseline covariates are outlined. The methodology is illustrated with a practical example arising from a clinical trial in lung disease.

Body Height↗

PubMed2002

Bayesian model adequacy and choice in phylogenetics.

Bayesian inference is becoming a common statistical approach to phylogenetic estimation because, among other reasons, it allows for rapid analysis of large data sets with complex evolutionary models. Conveniently, Bayesian phylogenetic methods use currently available stochastic models of sequence evolution. However, as with other model-based approaches, the results of Bayesian inference are conditional on the assumed model of evolution: inadequate models (models that poorly fit the data) may result in erroneous inferences. In this article, I present a Bayesian phylogenetic method that evaluates the adequacy of evolutionary models using posterior predictive distributions. By evaluating a model's posterior predictive performance, an adequate model can be selected for a Bayesian phylogenetic study. Although I present a single test statistic that assesses the overall (global) performance of a phylogenetic model, a variety of test statistics can be tailored to evaluate specific features (local performance) of evolutionary models to identify sources failure. The method presented here, unlike the likelihood-ratio test and parametric bootstrap, accounts for uncertainty in the phylogeny and model parameters.

Animals↗

PubMed2006

Estimating the prevalence and regulatory potential of the telomere looping effect in yeast transcription regulation.

Telomeres have long been implicated in the regulation of gene expression. Some studies have reported that telomere looping effect (TLE) can juxtapose genes and regulatory sequences that are far apart and facilitate long-distance control of gene expression. In this work, we report a detailed investigation on the prevalence and regulatory potential of TLE on a genomic scale by assembling data on protein-DNA interactions from several large-scale ChIp-chip experiments in Saccharomyces cerevisiae. Analysis of the assembled data revealed that a statistically significant number of DNA segments that were inferred to be bound by ten or more transcription factors in these experiments physically mapped to the ends of several chromosomes (19 of 32 chromosome ends). For the 83 transcription factors that were inferred to interact with these DNA segments, we found a statistically significant skew in the distribution of their internal binding sites over the length of the entire chromosome, such that more than expected binding events occurred proximal to chromosomal ends than elsewhere. Taken together these observations suggest that the telomere looping effect is their most likely explanation and imply that a notable fraction of the internally bound yeast transcription factors potentially interact with looped back telomeres. Further, we also identified several components of the basal transcriptional machinery that are also frequently linked to these chromosome end segments, strengthening the proposal for a direct interaction between the chromosome ends and internally located transcriptional complexes. We observed that certain chromatin factors might participate in the TLE and potentially modulate gene expression by chromatin modifications such as histone deacetylation. Our findings provide the first computational evidence for a significant role of long-range regulatory interactions due to telomere looping. Based on these observations, we also propose that genome-wide chromatin immunoprecipitation data might be useful to systematically uncover long-range chromatin looping effects in gene expression.

Binding Sites↗

PubMed2003

A classifier for the SNP-based inference of ancestry.

Ancestral inference from DNA could serve as an important adjunct for both standard and future human identity testing procedures. However, current STR methods for the inference of ancestral affiliation have inherent statistical and technical limitations. In an effort to identify bi-allelic markers that can be used to infer ancestral affiliation from DNA, we screened 211 SNPs in the human pigmentation and xenobiotic metabolism genes. Allele frequencies of 56 SNPs (most from pigmentation genes) were dramatically different between groups of unrelated individuals of Asian, African, and European descent, and both observed and simulated log likelihood ratios revealed that the markers were of exceptional value for ancestral inference. Log likelihood ratios of the multilocus estimates of biological ancestry (EAE/EBA) ranged from 7 to 10, which are on par with the best of the STR batteries yet described. A linear classification method was developed for incorporating these SNPs into a classifier model that was 99, 98, and 100% accurate for identifying individuals of European, African, and Asian descent, respectively. The methods and markers we describe are therefore an important first step for the development of a practical multiplex test for the inference of ancestry in a forensics setting.

Algorithms↗

PubMed2005

Statistical issues in the analysis of neuronal data.

Analysis of data from neurophysiological investigations can be challenging. Particularly when experiments involve dynamics of neuronal response, scientific inference can become subtle and some statistical methods may make much more efficient use of the data than others. This article reviews well-established statistical principles, which provide useful guidance, and argues that good statistical practice can substantially enhance results. Recent work on estimation of firing rate, population coding, and time-varying correlation provides improvements in experimental sensitivity equivalent to large increases in the number of neurons examined. Modern nonparametric methods are applicable to data from repeated trials. Many within-trial analyses based on a Poisson assumption can be extended to non-Poisson data. New methods have made it possible to track changes in receptive fields, and to study trial-to-trial variation, with modest amounts of data.

Action Potentials↗

PubMed1997

Use of the Wei-Lin-Weissfeld method for the analysis of a recurring and a terminating event.

We consider application of the Wei-Lin-Weissfeld (WLW) method for multiple failure time data when analysing a disease process consisting of a recurring outcome, such as clinical progression, and a terminating outcome, such as death. In order to adapt WLW for this situation, 'events' must be specified that define multiple failure times and whether these are censored. Various choices of events are possible, and each corresponds to inferences about a different aspect of the underlying disease process. Definitions which regard the terminating outcome as a censor of the recurring outcome focus on specific cause-specific hazard functions, while event definitions which make no distinction between a recurring and terminating outcome focus on hazard functions of the induced failure times. Some event definitions require strong statistical assumptions to yield valid inferences and are not recommended. The application of WLW for recurring/terminating processes is illustrated with the results of two recently conducted clinical trials in persons with HIV.

Acquired Immunodeficiency Syndrome↗

PubMed1998

Strategies for analyzing ecological momentary assessment data.

Studies incorporating repeated observations of momentary phenomena are becoming more common in behavioral and medical science. Analysis of such data requires the use of statistical techniques that are unfamiliar to many investigators. Some common ways of analyzing momentary data are reviewed--aggregation strategies, repeated measures analysis of variance, pooled within-person regression, and two-stage estimation procedures for multilevel models--and are found to be usually suboptimal, possibly leading to incorrect inferences. A broad class of statistical models for multilevel data that can address many research questions typically asked of momentary data are then described. Analytic issues that merit careful consideration include the scaling of momentary variables, allowance for serial autocorrelation of residuals, and the treatment of coefficients that vary across individuals as fixed versus random effects.

Analysis of Variance↗

PubMedDate not supplied

The role of body image in patient compliance and education.

This study investigated the relation between one aspect of body image (body cathexis) and the achievement of knowledge about life after a heart attack, as well as subsequent compliance with posthospitalization prescriptions. There was no significant correlation between body cathexis and achieved knowledge. The hypothesis which proposed that a relationship exists between body cathexis and compliance with posthospitalization prescriptions was statistically significant, body-cathexis effects are inferred to be present in compliance behaviors, and the hypothesis was supported. A statistically significant correlation was not found to exist between achieved knowledge and subsequent compliance behavior.

Adult↗

PubMed2005

Pros and cons in the use of SNPs in forensic kinship investigation: a comparative analysis with STRs.

Recent advances in single nucleotide polymorphisms (SNPs) research have raised the possibility that these markers could replace the forensically established short tandem repeats (STRs). In this work, we compare STRs and SNPs applicability for kinship investigation in terms of expected informative content and probability of occurrence of "difficult cases" (when isolated Mendelian incompatibilities between alleged father and child are found). Since SNPs have a much lower mutation rate than STRs, these difficulties were expected to occur less frequently if SNPs were used instead of STRs. The purpose of this paper is to make some simulations allowing the estimation of how often such difficult cases are expected to occur using both types of markers and how serious can be their impact in routine work. Our results demonstrate that a battery based exclusively on SNPs matching the informative power of current STR kits would be prone, if applied to routine paternity investigation, to the occurrence of cases where the statistical evidence would be inconclusive. We infer that the introduction of a SNP based strategy, as a substitute to the now classical STR approach poses statistical problems that must be carefully evaluated.

Adult↗

PubMed2005

The utility of prior information and stratification for parameter estimation with two screening tests but no gold standard.

When a gold standard screening or diagnostic test is not routinely available, it is common to apply two different imperfect tests to subjects from a study population. There is a considerable literature on estimating relevant parameters from the resultant data. In the situation that test sensitivities and specificities are unknown, several inferential strategies have been proposed. One suggestion is to use rough knowledge about the unknown test characteristics as prior information in a Bayesian analysis. Another suggestion is to obtain the statistical advantage of an identified model by splitting the population into two strata with differing disease prevalences. There is some division of opinion in the epidemiological literature on the relative merits of these two approaches. This article aims to shed light on the issue, by applying some recently developed theory on the performance of Bayesian inference in non-identified statistical models.

Bayes Theorem↗

PubMed1992

Research methods for measurement of heart rate and respiration.

Strategies for heart rate and respiration quantification are dependent upon a knowledge of physiology and statistics. To develop appropriate and sensitive psychophysiological measures, it is necessary to understand the neural control of the autonomic nervous system and the statistical characteristics of physiological data. Moreover, it is proposed that inferences derived from physiological measures are dependent upon a statistical interaction between the characteristics of the physiological data and the measurement techniques. The steps required to analyze heart rate and respiration data are presented, ranging from data acquisition and editing to data analysis and physiological inference. A variety of techniques are described and contrasted with recommendations for future research.

Arousal↗

PubMed2006

Adaptive web sampling.

A flexible class of adaptive sampling designs is introduced for sampling in network and spatial settings. In the designs, selections are made sequentially with a mixture distribution based on an active set that changes as the sampling progresses, using network or spatial relationships as well as sample values. The new designs have certain advantages compared with previously existing adaptive and link-tracing designs, including control over sample sizes and of the proportion of effort allocated to adaptive selections. Efficient inference involves averaging over sample paths consistent with the minimal sufficient statistic. A Markov chain resampling method makes the inference computationally feasible. The designs are evaluated in network and spatial settings using two empirical populations: a hidden human population at high risk for HIV/AIDS and an unevenly distributed bird population.

Animals↗

PubMed2025

Detecting Interspecific Positive Selection Using Convolutional Neural Networks.

Traditional statistical methods using maximum likelihood and Bayesian inference can detect positive selection from an interspecific phylogeny and a codon sequence alignment based on model assumptions, but they are prone to false positives due to alignment errors and can lack power. These problems are particularly pronounced when faced with high levels of indels and divergence. To address these issues, we trained and tested convolutional neural network models on simulated data and achieved higher accuracy in detecting selection across a specific range of phylogenetic scenarios and evolutionary modes. This advantage is particularly evident when performing inference on noisy data prone to misalignments. Our method shows some ability to account for these errors, where most statistical frameworks fail to do so in a tractable manner. We explore the generalizability of our convolutional neural network models to unseen evolutionary scenarios and identify future avenues to achieve broader utility. Once trained, our convolutional neural network model is faster at test time, making it a scalable alternative to traditional statistical methods for large-scale, multigene analyses. In addition to binary classification (inference of the presence or absence of positive selection during the evolution of the sequences), we use saliency maps to understand what the model learns and observe how this could be leveraged for sitewise inference of positive selection.

Neural Networks, Computer↗