Model-based Cluster Analysis Research Articles

Two subtypes of psychopathic violent offenders that parallel primary and secondary variants.

Although psychopathy usually is treated as a unitary construct, a seminal theory posits that there are 2 variants: Primary psychopathy is underpinned by an inherited affective deficit, whereas secondary psychopathy reflects an acquired affective disturbance. The authors investigated whether psychopathy phenotypically may be disaggregated into such types in a sample of 367 prison inmates convicted of violent crimes. Model-based cluster analysis of the Revised Psychopathy Checklist (PCL-R; R. D. Hare, 2003) and trait anxiety scores in the psychopathic subgroup (n = 123; PCL-R > or = 29) revealed 2 clusters. Relative to primary psychopaths, secondary psychopaths had greater trait anxiety, fewer psychopathic traits, and comparable levels of antisocial behavior. Across validation variables, secondary psychopaths manifested more borderline personality features, poorer interpersonal functioning (e.g., irritability, withdrawal, poor assertiveness), and more symptoms of major mental disorder than primary psychopaths. When compared with the nonpsychopathic subgroup (n = 243), the 2 psychopathic variants manifested a theoretically coherent pattern of differences. Implications for etiological research and violence prevention are discussed.

Penalized Model-Based Clustering with Application to Variable Selection

Variable selection in clustering analysis is both challenging and important. In the context of model-based clustering analysis with a common diagonal covariance matrix, which is especially suitable...

Development and evaluation of a new microsatellite multiplex system for parental allocation and management of rainbow trout ( Oncorhynchus mykiss) broodstocks

Examination of microsatellite variation at several polymorphic loci is a widely accepted method for determining parentage and examining genetic diversity within rainbow trout ( Oncorhynchus mykiss) breeding programs. Genotyping costs are considerable; therefore, we developed a single-step method of co-amplifying twelve microsatellites in two hexaplex reactions to reduce costs and improve efficiency. The protocol is explicitly described to ensure reproducible results. We applied the protocol to samples analyzed at the National Center for Cool and Coldwater Aquaculture (NCCCWA) with previously reported marker sets for a comparison of results. Each marker within the multiplex system was evaluated for duplication, null alleles, physical linkage, and probability of genotyping errors. Data from four of the 12 markers were excluded from parental analysis based on these criteria. Parental assignments were compared to those of a previous study that used five independently amplified microsatellites. Percentages of progeny assigned to parents were higher using the subset of eight markers from the multiplex system than with five markers used in the previous study (98% vs. 92%). Through multiplexing, additional markers were used to improve parental allocation while also improving efficiency by reducing the number of required PCR reactions and genotyping runs. We evaluated our methods further through estimation of F-statistics, pairwise genetic distances, and cluster analysis among brood-years at the NCCCWA facility. These measures were compared to estimates from nine independently amplified microsatellites used in a previous study. The F st results calculated between brood-years show similar values of genetic differentiation using both marker sets. Estimates of individual pairwise genetic distances were used for constructing neighbor-joining trees. Both marker sets yielded trees that show similar subpopulation structuring and agreed with results from a model-based cluster analysis and available pedigree information. These approaches for detecting population substructure and admixture portions within individuals are particularly useful for new breeding programs where the founders' relatedness is unknown.

Bayesian Model Based Clustering Analysis: Application to a Molecular Dynamics Trajectory of the HIC‐1 Integrase Catalytic Core.

AbstractChemInform is a weekly Abstracting Service, delivering concise information at a glance that was extracted from about 200 leading journals. To access a ChemInform Abstract, please click on HTML or PDF.

ALLOZYME AND MITOCHONDRIAL DNA VARIATION IN THE TAILED FROG (ANURA: ASCAPHUS): THE INFLUENCE OF GEOGRAPHY AND GENE FLOW

In this paper we support the division of the previously monotypic genus Ascaphus into two species based on an analysis of 23 allozyme loci from 34 populations. We use maximum likelihood to estimate the Ascaphus phylogeny from the allozyme data and find strongly supported monophyletic Rocky Mountain and Pacific clades. In a nonhierarchal, model-based cluster analysis of the data, each of the 1085 individual genotypes is correctly assigned to either Ascaphus montanus or A. truei with a high probability. We also find a virtually fixed difference between the species at the Pgm-2 locus. Within A. truei, we find a lack of significant pairwise FST values among populations from the Coast and central Cascades Mountains, suggesting relatively recent range expansion or contemporary gene flow among these populations. Olympic Mountains populations form a discrete clade in the allozyme topology and are fixed for a unique allele at the Lap locus. These populations remain isolated from the remainder of the species' range...

Speed-up for the expectation-maximization algorithm for clustering categorical data

In model-based cluster analysis, the expectation-maximization (EM) algorithm has a number of desirable properties, but in some situations, this algorithm can be slow to converge. Some variants are proposed to speed-up EM in reducing the time spent in the E-step, in the case of Gaussian mixture. The main aims of such methods is first to speed-up convergence of EM, and second to yield same results (or not so far) than EM itself. In this paper, we compare these methods from categorical data, with the latent class model, and we propose a new variant that sustains better results on synthetic and real data sets, in terms of convergence speed-up and number of misclassified objects.

Bayesian Model Based Clustering Analysis: Application to a Molecular Dynamics Trajectory of the HIV-1 Integrase Catalytic Core

This work describes the application of a Bayesian method for clustering protein conformations sampled during a molecular dynamics simulation of the HIV-1 integrase catalytic core. A clustering analysis is carried out under the assumption of normal distribution without fixing the number of clusters in advance. Some performance measures, such as posterior probability and class cross entropy, are used to determine the most probable set of clusters. The Bayesian clustering method results in meaningful groups identifying transitions between conformational ensembles. The dihedral angles involved in such transitions are also examined in detail. The conformations in high dimensional space are projected into 3D space employing a multidimensional scaling technique to provide a visual inspection.

Cluster analysis using multivariate normal mixture models to detect differential gene expression with microarray data

DNA microarrays make it possible to study simultaneously the expression of thousands of genes in a biological sample. Univariate clustering techniques have been used to discover target genes with differential expression between two experimental conditions. Because of possible loss of information due to use of univariate summary statistics, it may be more effective to use multivariate statistics. We present multivariate normal mixture model based clustering analyses to detect differential gene expression between two conditions. Deviating from the general mixture model and model-based clustering, we propose mixture models with specific mean and covariance structures that account for special features of two-condition microarray experiments. Explicit updating formulas in the EM algorithm for three such models are derived. The methods are applied to a real dataset to compare the expression levels of 1176 genes of rats with and without pneumococcal middle-ear infection to illustrate the performance and usefulness of this approach. About 10 genes and 20 genes are found to be differentially expressed in a six-dimensional modeling and a bivariate modeling, respectively. Two simulation studies are conducted to compare the performance of univariate and multivariate methods. Depending on data, neither method can always dominate the other. The results suggest that multivariate normal mixture models can be useful alternatives to univariate methods to detect differential gene expression in exploratory data analysis.

Hybridization and genetic variation in Danish populations of European crab apple (Malus sylvestris)

The aim of the present study was to investigate the genetic variation in Danish populations of the endangered European crab apple (Malus sylvestris). Special emphasis was given to hybridization between the wild species and its cultivated relative Malus ×domestica. A total of 178 wild individuals from four Danish populations were studied along with a reference sample of 29 old cultivars. The genetic variation within and among samples was studied at ten microsatellite marker loci. Additionally, a morphological analysis was carried out to identify hybrids and test for correspondence between phenotypic and genotypic indices of hybridization. From application of ordination and a model-based cluster analysis to the molecular data, two clusters were identified consisting of wild and cultivated individuals, respectively. This indicates that pronounced admixture between the two species is not present. At the population level, a high correspondence was found between geographic isolation from M. ×domestica and genotypic and morphological indices of hybridization. As expected, isolated populations appeared less affected by hybridization than poorly isolated populations. Isolated “pure” M. sylvestris populations could thus be identified. However, morphological and molecular evidences of hybridization were found to be divergent at the individual level. This is suggestive of some historical introgression into the M. sylvestris gene pool and indicates that relying exclusively on either morphological or molecular characters as diagnostic markers in studies of hybridization between M. ×domestica and M. sylvestris might lead to fallible results. Combined application of genetic and morphological markers is therefore recommended.

Model-based cluster and discriminant analysis with the MIXMOD software

The Mixture Modeling ( MIXMOD) program fits mixture models to a given data set for the purposes of density estimation, clustering or discriminant analysis. A large variety of algorithms to estimate the mixture parameters are proposed (EM, Classification EM, Stochastic EM), and it is possible to combine these to yield different strategies for obtaining a sensible maximum for the likelihood (or complete-data likelihood) function. MIXMOD is currently intended to be used for multivariate Gaussian mixtures, and fourteen different Gaussian models can be distinguished according to different assumptions regarding the component variance matrix eigenvalue decomposition. Moreover, different information criteria for choosing a parsimonious model (the number of mixture components, for instance) are included, their suitability depending on the particular perspective (cluster analysis or discriminant analysis). Written in C++, MIXMOD is interfaced with S CILAB and M ATLAB. The program, the statistical documentation and the user guide are available on the internet at the following address: http://www-math.univ-fcomte.fr/mixmod/index.php.

Determination of the genetic structure of remnant Morus boninensis Koidz. trees to establish a conservation program on the Bonin Islands, Japan

BackgroundMorus boninensis, is an endemic plant of the Bonin (Ogasawara) Islands of Japan and is categorized as "critically endangered" in the Japanese red data book. However, little information is available about its ecological, evolutionary and genetic status, despite the urgent need for guidelines for the conservation of the species. Therefore, we adopted Moritz's MU concept, based on the species' current genetic structure, to define management units and to select mother tree candidates for seed orchards.ResultsNearly all individuals of the species were genotyped on the basis of seven microsatellite markers. Genetic diversity levels in putative natural populations were higher than in putative man-made populations with the exception of those on Otouto-jima Island. This is because a limited number of maternal trees are likely to have been used for seed collection to establish the man-made populations. A model-based clustering analysis clearly distinguished individuals into nine clusters, with a large difference in genetic composition between the population on Otouto-jima Island, the putative natural populations and the putative man-made populations. The Otouto-jima population appeared to be genetically differentiated from the others; a finding that was also supported by pairwise FST and RST analysis. Although multiple clusters were detected in the putative man-made populations, the pattern of genetic diversity was monotonous in comparison to the natural populations.ConclusionThe genotyping by microsatellite markers revealed strong genetic structures. Typically, artificial propagation of this species has ignored the genetic structure, relying only on seeds from Otouto-jima for replanting on other islands, because of a problem with inter-specific hybridization on Chichi-jima and Haha-jima Islands. However, this study demonstrates that we should be taking into consideration the genetic structure of the species when designing a propagation program for the conservation of this species.

A Semiparametric Approachto Modeling Nonlinear Relations Among Latent Variables

To date, finite mixtures of structural equation models (SEMMs) have been developed and applied almost exclusively for the purpose of providing model-based cluster analyses. This type of analysis constitutes a direct application of the model wherein the estimated component distributions of the latent classes are thought to represent the characteristics of distinct unobserved subgroups of the population. This article instead considers an indirect application of the SEMM in which the latent classes are estimated only in the service of more flexibly modeling the characteristics of the aggregate population as a whole. More specifically, the SEMM is used to semiparametrically model nonlinear latent variable regression functions. This approach is first developed analytically and then demonstrated empirically through analyses of simulated and real data.

Mixture of linear mixed models for clustering gene expression profiles from repeated microarray experiments

Data variability can be important in microarray data analysis. Thus, when clustering gene expression profiles, it could be judicious to make use of repeated data. In this paper, the problem of analysing repeated data in the model-based cluster analysis context is considered. Linear mixed models are chosen to take into account data variability and mixture of these models are considered. This leads to a large range of possible models depending on the assumptions made on both the covariance structure of the observations and the mixture model. The maximum likelihood estimation of this family of models through the EM algorithm is presented. The problem of selecting a particular mixture of linear mixed models is considered using penalized likelihood criteria. Illustrative Monte Carlo experiments are presented and an application to the clustering of gene expression profiles is detailed. All those experiments highlight the interest of linear mixed model mixtures to take into account data variability in a cluster analysis context.

Fuzzy clustering to estimate the parameters of block mixture models

Finite mixture models are being increasingly used to provide model-based cluster analysis. To tackle the problem of block clustering which aims to organize the data into homogeneous blocks, recently we have proposed a block mixture model; we have considered this model under the classification maximum likelihood approach and we have developed a new algorithm for simultaneous partitioning based on the classification EM algorithm. From the estimation point of view, classification maximum likelihood approach yields inconsistent estimates of the parameters and in this paper we consider the block clustering problem under the maximum likelihood approach; unfortunately, the application of the classical EM algorithm for the block mixture model is not direct: difficulties arise due to the dependence structure in the model and approximations are required. Considering the block clustering problem under a fuzzy approach, we propose a fuzzy block clustering algorithm to approximate the EM algorithm. To illustrate our approach, we study the case of binary data by using a Bernoulli block mixture.

Population structure and long-range linkage disequilibrium in a durum wheat elite collection

A collection of 134 durum wheat accessions, mainly including cultivars (cvs.) representative of the major gene pools, was assembled and characterized with 70 SSRs for genetic diversity and level of long-range linkage disequilibrium (LD). Results of both a distance-based and a model-based (Bayesian) cluster analysis evidenced the presence of a structured diversity. In the model-based analysis, six to eight main distinct subpopulations were identified based on the molecular data. Only a relatively small portion (20%) of the molecular variation was accounted for by the geographical origin of the accessions. Major differences were detected between the North American and the Mediterranean cvs., while a considerable overlap characterized the cvs. from CIMMYT-ICARDA and Italy. The North American cvs. showed the highest within group mean genetic similarity (GSm = 0.68). French cvs. revealed sizeable similarities with both the North American as well as the Italian and CIMMYT-ICARDA pools. Considering the germplasm as a whole, high levels of LD were found both at locus pairs with an intrachromosomal distance <50 cM as well as at those with distances more than 50 cM and independent (86, 52 and 54% of SSR pairs at p < 0.01, respectively). After re-evaluating LD within each of the three main subgroups identified through the analysis of the germplasm structure, the LD level remained high for tightly to moderately linked locus pairs (<20 cM apart), but was greatly reduced in the loosely linked (more than 50 cM apart) and independent locus pairs. The implications of these findings as to the possibility of using association mapping for gene/QTL discovery in durum wheat are discussed.

Scalable model-based cluster analysis using clustering features

We present two scalable model-based clustering systems based on a Gaussian mixture model with independent attributes within clusters. They first summarize data into sub-clusters, and then generate Gaussian mixtures from their clustering features using a new algorithm—EMACF. EMACF approximates the aggregate behavior of each sub-cluster of data items in the Gaussian mixture model. It provably converges. The experiments show that our clustering systems run one or two orders of magnitude faster than the traditional EM algorithm with few losses of accuracy.

Identifying Psychopathy Subtypes on the Basis of Personality Structure.

The authors used model-based cluster analysis to identify subtypes of criminal psychopaths on the basis of differences in personality structure. Participants included 96 male prisoners diagnosed as psychopathic, using the Psychopathy Checklist Revised (PCL-R; R. D. Hare, 1991). Personality was assessed using the brief form of the Multidimensional Personality Questionnaire (MPQ-BF; C. J. Patrick, J. J. Curtin, and A. Tellegen, 2002). The best-fitting model yielded two clusters. Emotionally stable psychopaths were characterized by low Stress Reaction and high Agency. Aggressive psychopaths were characterized by high Negative Emotionality, low Constraint, and low Communion. These results suggest that psychopaths as defined by the PCL-R includes distinct subtypes, distinguishable in terms of personality structure, that may reflect different etiologies.

Mixture density separation as a tool for high-quality interpretation of multi-source remote sensing data and related issues

The end user often needs to define extremely complex interpretation tasks and to require the analysis results to be quantitatively proven for each pixel without the test data. To this end, this paper extends the ideas underlying the model-based unsupervised classification method previously proposed by us (Koltunov and Ben-Dor 2001). Consistently with that method, the quality of assigning a pixel to a cluster is defined as the lower confidence bound (l.c.b.) of the corresponding posterior probability estimate. We propose to compute the l.c.b.s in an approximate way using the Fisher information matrix instead of the bootstrap scheme suggested previously, leading to an l.c.b.-estimation procedure that is faster by a factor of hundreds to thousands, while being reasonably accurate. The issue of selecting the number of clusters is considered in accordance with the quantitative requirements for the level of detail and the reliability of the thematic interpretation. Specifically, the l.c.b.s form a novel selection criterion that allows the most detailed landscape descriptions to be provided with at least a pre-specified value of confidence. This implies detection of highly overlapping clusters, leading to very detailed segmentations. Consequently, numerous thematic classification and object detection problems can be solved, based on single clustering and assuming that thematic classes are unions of components. Then the thematic classification accuracy can be computed in a well-founded manner for each separate pixel, using the obtained covariance matrices of the posterior probability estimators of component membership. The procedures for thematic mapping and object detection are described. The accuracy of the l.c.b. estimation and stability of the new criterion in choosing the number of clusters are illustrated on simulated datasets. The hyperspectral data analysis experiment performed demonstrates part of the developments described in this paper. Several issues that are relevant for remote sensing data interpretation are addressed constructively. In particular, we draft the novel algorithms that use model-based cluster analysis for detection and recognition of remotely sensed objects based on prior information on their size and shape. In addition, we introduce a generalized approach to unsupervised feature extraction from data acquired by a plurality of sensors of different physical nature. A new data model generalizing the traditional Gaussian mixture model is also presented.

Pottery production during the Late Jomon period: insights from the chemical analyses of Kasori B pottery

The minor and trace element chemistry of 175 Late Jomon pottery sherds was determined using energy dispersive X-ray fluorescence (EDXRF). The sherds came from six contemporary sites in the Kanto region and belong to the Kasori B style of pottery. Principal component analysis (PCA) was used to reduce the number of variables and model-based cluster analysis was used to find clusters in the PCA scores. Model-based clustering found two clusters in the data set; these two clusters correspond to the local sedimentary raw materials near the sites.

Vascular plant compositional gradients within and between Iowa fens

Abstract Question: What is the nature and relative importance of compositional gradients within- and between fens? Location: Iowa, USA. Methods: 506 0.5 m × 0.5 m quadrats were sampled from 31 fens across a 550 km extent. Presence/absence of all vascular plant taxa, plus the non-vascular genera Sphagnum and Chara, and values for 24 environmental variables were noted. Global Non-Metric Multidimensional Scaling and Monte Carlo tests were used to describe compositional variation and identify significant environmental co-variables. Model-based cluster analysis was used to identify the optimal number of groups supported by the data, while k-means clustering was used to assign each quadrat to a group. The number of occurrences (and frequency) of each species within each group was calculated. Two-dimensional 95% Gaussian confidence intervals, ANOVA, correlation coefficient homogeneity tests, log-linear modelling, and Fisher's exact tests were used to document patterns of compositional change. Results: Two stable...