On generalized latent factor modeling and inference for high-dimensional binomial data

We explore a hierarchical generalized latent factor model for discrete and bounded response variables and in particular, binomial responses. Specifically, we develop a novel two-step estimation procedure and the corresponding statistical inference that is computationally efficient and scalable for the high dimension in terms of both the number of subjects and the number of features per subject. We also establish the validity of the estimation procedure, particularly the asymptotic properties of the estimated effect size and the latent structure, as well as the estimated number of latent factors. The results are corroborated by a simulation study and for illustration, the proposed methodology is applied to analyze a dataset in a gene–environment association study.     

Matching consumer feeding behaviours and resource traits: a fourth‐corner problem in food‐web theory

For decades, food web theory has proposed phenomenological models for the underlying structure of ecological networks. Generally, these models rely on latent niche variables that match the feeding behaviour of consumers with their resource traits. In this paper, we used a comprehensive database to evaluate different hypotheses on the best dependency structure of trait‐matching patterns between consumers and resource traits. We found that consumer feeding behaviours had complex interactions with resource traits; however, few dimensions (i.e. latent variables) could reproduce the trait‐matching patterns. We discuss our findings in the light of three food web models designed to reproduce the multidimensionality of food web data; additionally, we discuss how using species traits clarify food webs beyond species pairwise interactions and enable studies to infer ecological generality at larger scales, despite potential taxonomic differences, variations in ecological conditions and differences in species abundance between communities.     

The latent structure of soccer in the phases of attack and defense

With the aim of establishing the latent structure of tactical elements in the attack and defense phases of soccer 117 tactical elements of soccer were defined and their importance assessed by means of 30 variables that determine the basic segments of the game of soccer. 93 attack and 24 defense tactical elements were chosen as the entity sample and described by the 15 variables of the attack phase and 15 variables of the defense phase. Ten competent soccer experts determined the characteristics of the aforementioned entities by means of 30 variables. The experts graded from 0 to 5 the impact of every entity (tactical technique) on the individual variables that describe soccer in its phases of either attack and defense. A high level of inter-expert agreement was reached in regard to the properties of attack and defense techniques, as demonstrated by the objectivity coefficients. According to principal component factor analysis and the Kaiser and Guttman rule a total of five significant latent dimensions were obtained: finishing efficiency, ball possession performance, counter-attack efficiency, combined defense performance, and obstruction and redirection of the opposing team's attack build-up. The research partly resolved the issue of the hypothetical structure of tactical techniques in soccer by dividing the game into phases and sub-phases, attack and defense players'positions, and types (styles) of play in the attack and defense. If it is clear which movement structures have the most significant influence on the efficiency on a particular playing position and performance in the sub-phases and styles of play, it would be possible to create such training operators that will facilitate the formation of the most important motor skills in soccer.     

Depression scores associate with chronotype and social jetlag in a rural population

In public health, mood disorders are among the most important mental impairments. Patients with depressive episodes exhibit daily mood variations, abnormal patterns in sleep-wake behavior, and in the daily rhythms of several endocrine-metabolic parameters. Although the relationship between the sleep/circadian processes and mood disorders is poorly understood, clock-related therapies, such as light therapy, sleep deprivation, and rigid sleep schedules, have been shown to be effective treatments. Several studies investigated the relationship between circadian phenotype (chronotype) and depression. These focused mainly on urban populations and assessed diurnal preferences (Morningness-Eveningness score) rather than the actual timing of sleep and activity. Here, we used the Beck Depression Inventory (BDI) in an essentially rural population (N?=?4051), and investigated its relation to circadian phenotype (chronotype and social jetlag), assessed with the Munich Chronotype Questionnaire (MCTQ). In our study design, we (i) normalized both chronotype and BDI scores for age and sex (MSF(sas) and BDI(as), respectively); (ii) calculated individual social jetlag (misalignment of the biological and social time); and (iii) investigated the relationship between circadian phenotypes and BDI scores in a population homogeneous in respect to culture, socioeconomic factors, and daily light exposure. A 15.65% (N?=?634) of the participants showed mild to severe depressive BDI scores. Late chronotypes had a higher BDI(as) than intermediate and early types, which was independent of whether or not the participants were smokers. Both chronotype and BDI(as) correlated positively with social jetlag. BDI(as) was significantly higher in subjects with >2?h of social jetlag than in the rest of the population?again independent of smoking status. We also compared chronotype and social jetlag distributions between BDI categories (no symptoms, minimal symptoms, and mild to severe symptoms of depression) separately for men and women and for four age groups; specifically in the age group 31?40 yrs, subjects with mild to severe BDI scores were significantly later chronotypes and suffered from higher social jetlag. Our results indicate that misalignment of circadian and social time may be a risk factor for developing depression, especially in 31- to 40-yr-olds. These relationships should be further investigated in longitudinal studies to reveal if reduction of social jetlag should be part of prevention strategies. (Author correspondence: karla.allebrandt@med.uni-muenchen.de ).     

Exploring the Factor Structure of Neurocognitive Measures in Older Individuals

Here we focus on factor analysis from a best practices point of view, by investigating the factor structure of neuropsychological tests and using the results obtained to illustrate on choosing a reasonable solution. The sample (n=1051 individuals) was randomly divided into two groups: one for exploratory factor analysis (EFA) and principal component analysis (PCA), to investigate the number of factors underlying the neurocognitive variables; the second to test the “best fit” model via confirmatory factor analysis (CFA). For the exploratory step, three extraction (maximum likelihood, principal axis factoring and principal components) and two rotation (orthogonal and oblique) methods were used. The analysis methodology allowed exploring how different cognitive/psychological tests correlated/discriminated between dimensions, indicating that to capture latent structures in similar sample sizes and measures, with approximately normal data distribution, reflective models with oblimin rotation might prove the most adequate.     

Measuring hiking specialization and identification of latent profiles of hikers

Previous specialization studies have only measured single items or the sum of responses across dimensions, making it impossible to classify recreationists by their degree of multidimensional specialization. The purpose of the present study is to test a three-dimensional model and its traditional components (i.e., items) of specialization in hikers. Applying a latent profile analysis, the present study examined hikers who shared similar profiles based on multiple dimensions of specialization and classified them based on latent class characteristics and structures. Data from 587 hikers on Namhansanseong Trail in South Korea were analyzed. Four subgroups were identified: novice (32 %), behavior-oriented (18 %), veteran (30 %), and potential veteran (20 %). These groups showed differential patterns in the behavioral and cognitive dimensions of specialization. In particular, the component items of hiking experience and setting experience in the behavioral dimension differed by group.     

Factor analytic models of clustered multivariate data with informative censoring

This article describes a general class of factor analytic models for the analysis of clustered multivariate data in the presence of informative missingness. We assume that there are distinct sets of cluster-level latent variables related to the primary outcomes and to the censoring process, and we account for dependency between these latent variables through a hierarchical model. A linear model is used to relate covariates and latent variables to the primary outcomes for each subunit. A generalized linear model accounts for covariate and latent variable effects on the probability of censoring for subunits within each cluster. The model accounts for correlation within clusters and within subunits through a flexible factor analytic framework that allows multiple latent variables and covariate effects on the latent variables. The structure of the model facilitates implementation of Markov chain Monte Carlo methods for posterior estimation. Data from a spermatotoxicity study are analyzed to illustrate the proposed approach.     

Depression Dimensions: Integrating Clinical Signs and Symptoms from the Perspectives of Clinicians and Patients

Background

Several studies have recognized that depression is a multidimensional construct, although the scales that are currently available have been shown to be limited in terms of the ability to investigate the multidimensionality of depression. The objective of this study is to integrate information from instruments that measure depression from different perspectives–a self-report symptomatic scale, a clinician-rated scale, and a clinician-rated scale of depressive signs–in order to investigate the multiple dimensions underlying the depressive construct.

Methods

A sample of 399 patients from a mood disorders outpatient unit was investigated with the Beck Depression Inventory (BDI), the Hamilton Depression Rating Scale (HDRS), and the Core Assessment of Psychomotor Change (CORE). Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) were used to investigate underlying dimensions of depression, including item level analysis with factor loadings and item thresholds.

Results

A solution of six depression dimensions has shown good-fit to the data, with no cross-loading items, and good interpretability. Item-level analysis revealed that the multidimensional depressive construct might be organized into a continuum of severity in the following ascending order: sexual, cognitive, insomnia, appetite, non-interactiveness/motor retardation, and agitation.

Conclusion

An integration of both signs and symptoms, as well as the perspectives of clinicians and patients, might be a good clinical and research alternative for the investigation of multidimensional issues within the depressive syndrome. As predicted by theoretical models of depression, the melancholic aspects of depression (non-interactiveness/motor retardation and agitation) lie at the severe end of the depressive continuum.     

Dynamic modeling of multivariate dimensions and their temporal relationships using latent processes: Application to Alzheimer's disease

Alzheimer's disease gradually affects several components including the cerebral dimension with brain atrophies, the cognitive dimension with a decline in various functions, and the functional dimension with impairment in the daily living activities. Understanding how such dimensions interconnect is crucial for Alzheimer's disease research. However, it requires to simultaneously capture the dynamic and multidimensional aspects and to explore temporal relationships between dimensions. We propose an original dynamic structural model that accounts for all these features. The model defines dimensions as latent processes and combines a multivariate linear mixed model and a system of difference equations to model trajectories and temporal relationships between latent processes in finely discrete time. Dimensions are simultaneously related to their observed (possibly multivariate) markers through nonlinear equations of observation. Parameters are estimated in the maximum likelihood framework enjoying a closed form for the likelihood. We demonstrate in a simulation study that this dynamic model in discrete time benefits the same causal interpretation of temporal relationships as models defined in continuous time as long as the discretization step remains small. The model is then applied to the data of the Alzheimer's Disease Neuroimaging Initiative. Three longitudinal dimensions (cerebral anatomy, cognitive ability, and functional autonomy) measured by six markers are analyzed, and their temporal structure is contrasted between different clinical stages of Alzheimer's disease.     

Merging the Idiographic Filter With Dynamic Factor Analysis to Model Process

To model processes we propose merging idiographic filter measurement with dynamic factor analysis. This involves testing whether or not the same latent dynamics (concurrent and lagged factor interrelations) can describe different individuals' observed multivariate time series. The methodology allows fitting, across different individuals, dynamic factor models that are invariant with respect to the latent dynamics, but not necessarily the factor loadings (measurement model). This methodology allows the same latent process to manifest differently from one individual to another, thus recognizing that the process is general but its realization in a given person is to some degree idiosyncratic. The approach is illustrated with empirical data.     

Behavior Based Social Dimensions Extraction for Multi-Label Classification

Classification based on social dimensions is commonly used to handle the multi-label classification task in heterogeneous networks. However, traditional methods, which mostly rely on the community detection algorithms to extract the latent social dimensions, produce unsatisfactory performance when community detection algorithms fail. In this paper, we propose a novel behavior based social dimensions extraction method to improve the classification performance in multi-label heterogeneous networks. In our method, nodes’ behavior features, instead of community memberships, are used to extract social dimensions. By introducing Latent Dirichlet Allocation (LDA) to model the network generation process, nodes’ connection behaviors with different communities can be extracted accurately, which are applied as latent social dimensions for classification. Experiments on various public datasets reveal that the proposed method can obtain satisfactory classification results in comparison to other state-of-the-art methods on smaller social dimensions.     

Structure of the human myostatin precursor and determinants of growth factor latency

Myostatin, a key regulator of muscle mass in vertebrates, is biosynthesised as a latent precursor in muscle and is activated by sequential proteolysis of the pro‐domain. To investigate the molecular mechanism by which pro‐myostatin remains latent, we have determined the structure of unprocessed pro‐myostatin and analysed the properties of the protein in its different forms. Crystal structures and SAXS analyses show that pro‐myostatin adopts an open, V‐shaped structure with a domain‐swapped arrangement. The pro‐mature complex, after cleavage of the furin site, has significantly reduced activity compared with the mature growth factor and persists as a stable complex that is resistant to the natural antagonist follistatin. The latency appears to be conferred by a number of distinct features that collectively stabilise the interaction of the pro‐domains with the mature growth factor, enabling a regulated stepwise activation process, distinct from the prototypical pro‐TGF‐β1. These results provide a basis for understanding the effect of missense mutations in pro‐myostatin and pave the way for the design of novel myostatin inhibitors.     

The Use of Bayesian Latent Class Cluster Models to Classify Patterns of Cognitive Performance in Healthy Ageing

The main focus of this study is to illustrate the applicability of latent class analysis in the assessment of cognitive performance profiles during ageing. Principal component analysis (PCA) was used to detect main cognitive dimensions (based on the neurocognitive test variables) and Bayesian latent class analysis (LCA) models (without constraints) were used to explore patterns of cognitive performance among community-dwelling older individuals. Gender, age and number of school years were explored as variables. Three cognitive dimensions were identified: general cognition (MMSE), memory (MEM) and executive (EXEC) function. Based on these, three latent classes of cognitive performance profiles (LC1 to LC3) were identified among the older adults. These classes corresponded to stronger to weaker performance patterns (LC1>LC2>LC3) across all dimensions; each latent class denoted the same hierarchy in the proportion of males, age and number of school years. Bayesian LCA provided a powerful tool to explore cognitive typologies among healthy cognitive agers.     

Characterization of human factor VIII and interaction with von Willebrand factor. An electron microscopic study

Blood coagulation factor VIII is a large glycoprotein that circulates in plasma at relative low concentration (0.1 microgram/ml). It consists of a heterogeneous mixture of a series heavy-chain peptides (90-200 kDa), each associated with a light chain of 80 kDa. To gain insight into the physical properties of the protein, we have characterized purified human factor VIII by electron microscopy and rotary shadowing. Electron microscopy of rotary shadowed factor VIII molecules showed predominantly a single globular domain structure, with a somewhat asymmetric shape, while two-domain structures were also encountered. The overall dimensions of the globular domains ranged from 4 x 6 nm to 8 x 12 nm. EDTA treatment of factor VIII reduced the overall dimensions (2.5 x 5 nm to 6 x 10 nm) while treatment with thrombin reduced the dimensions to a small extent. In complexes with von Willebrand factor, factor VIII appeared localized at the globular domains of von Willebrand factor multimers. In addition, incubation of factor VIII with Staphylococcus aureus V8 protease fragments SpII and SpIII revealed only binding to the globular domains of SpIII. In this study, the first morphological characterization of human factor VIII is presented, together with its direct localization on von Willebrand factor multimers.     

Modeling adverse birth outcomes via confirmatory factor quantile regression

We describe a Bayesian quantile regression model that uses a confirmatory factor structure for part of the design matrix. This model is appropriate when the covariates are indicators of scientifically determined latent factors, and it is these latent factors that analysts seek to include as predictors in the quantile regression. We apply the model to a study of birth weights in which the effects of latent variables representing psychosocial health and actual tobacco usage on the lower quantiles of the response distribution are of interest. The models can be fit using an R package called factorQR.     

Generation of recombinant human large latent transforming growth factor-beta 1 and monoclonal antibodies to it

Transforming growth factor beta 1 (TGF-beta 1) is a regulator of cell growth and differentiation. It is produced in various of cells and tissues as a biologically latent complex, whose significance is still unknown. We established a Chinese hamster ovary cells that produced recombinant human large latent TGF-beta 1. The growth factor was purified from serum-free conditioned medium of the cell line was purified to apparent homogeneity by four steps of column chromatography. The purified protein gave a single band with the apparent molecular weight of 210,000 on SDS-PAGE, and had four subunits, of 12.5, 40, 53, and 150-190 kDa. These components were identical to TGF-beta 1, the N-terminal remnant of pro-TGF-beta 1, pro-TGF-beta 1, and latent TGF-beta 1 binding protein, respectively. The purified growth factor had biological activity similar to that of the growth factor purified from human platelets. We prepared four monoclonal antibodies by immunization of mice with the recombinant protein. In western blotting, two of the antibodies bound to latent TGF-beta 1 binding protein. The two other antibodies reacted with the N-terminal remnant of pro-TGF-beta 1. Recombinant large latent TGF-beta 1 and its monoclonal antibodies could be used for detailed structural and functional studies of the large latent TGF-beta 1 complex.     

Structural characterisation of the bromouracil.guanine base pair mismatch in a Z-DNA fragment.

The deoxyoligonucleotide d(BrU-G-C-G-C-G) was crystallised at pH 8.2 and its structure analysed by X-ray diffraction. The unit cell, of dimensions a = 17.94, b = 30.85, c = 49.94A contains four DNA duplexes in space group P2(1)2(1)2(1). The duplexes are in the Z conformation, with four Watson-Crick G.C base pairs and two BrU.G base pairs. The structure was refined to an R factor of 0.16 at a resolution of 2.2A with 64 solvent molecules located. The BrU.G base pair mismatch is of the wobble type, with both bases in the major tautomer form and hydrogen bonds linking 0-2 of BrU with N-1 of G and N3 of BrU with 0-6 of G. There is no indication of the presence of ionised base pairs, in spite of the high pH of crystallisation. The results are discussed in terms of the mutagenic properties of 5- bromouracil.     

Thermal effects on an enzymatically latent conformation of coagulation factor VIIa.

Activation of the zymogen factor VII yields an enzyme form, factor VIIa, with only modest activity. The thermal effect on this low activity of factor VIIa and its enhancement by the cofactor tissue factor was investigated. Factor VIIa activity measured with a chromogenic peptide substrate is characterized by an unusual temperature dependency which indicates that the activated protease exists in an equilibrium between a latent (enzymatically inactive) and an active conformation. As shown by calorimetry and activity measurements the thermal effects on factor VIIa are fully reversible below the denaturation temperature of 58.1 degrees C. A model for factor VIIa has been proposed [Higashi, S., Nishimura, H., Aita, K. & Iwanaga, S. (1994) J. Biol. Chem. 269, 18891-18898] in which the protease is supposed to exist primarily as a latent enzyme form because of the poor incorporation into the protease structure of the N-terminal Ile153 released by proteolytic cleavage during activation of factor VII. Binding of tissue factor to factor VIIa is assumed to shift the equilibrium towards an active conformation in which the N-terminal Ile153 forms a salt bridge with Asp343. We corroborate the validity of this model by: (a) chemical modification of factor VIIa; this suggests that the thermal effect on the equilibrium between the active and inactive conformation is reflected in the relative accessibility of the active site and the N-terminal Ile153; (b) measurements of factor VIIa binding to tissue factor indicating that complex formation is favoured by stabilization of the active conformation; and (c) activity measurements of a cross-linked factor VIIa-tissue factor complex; this showed that cross-linking stabilized the active conformation of factor VIIa and essentially prevented its thermally-induced transformation into the inactive state.     

The crystal structure of the mouse apoptosis-inducing factor AIF

Mitochondria play a key role in apoptosis due to their capacity to release potentially lethal proteins. One of these latent death factors is cytochrome c, which can stimulate the proteolytic activation of caspase zymogens. Another important protein is apoptosis-inducing factor (AIF), a flavoprotein that can stimulate a caspase-independent cell-death pathway required for early embryonic morphogenesis. Here, we report the crystal structure of mouse AIF at 2.0 A. Its active site structure and redox properties suggest that AIF functions as an electron transferase with a mechanism similar to that of the bacterial ferredoxin reductases, its closest evolutionary homologs. However, AIF structurally differs from these proteins in some essential features, including a long insertion in a C-terminal beta-hairpin loop, which may be related to its apoptogenic functions.