Supervised Gaussian Process Latent Variable Model for Dimensionality Reduction

The Gaussian process latent variable model (GP-LVM) has been identified to be an effective probabilistic approach for dimensionality reduction because it can obtain a low-dimensional manifold of a data set in an unsupervised fashion. Consequently, the GP-LVM is insufficient for supervised learning tasks (e. g., classification and regression) because it ignores the class label information for dimensionality reduction. In this paper, a supervised GP-LVM is developed for supervised learning tasks, and the maximum a posteriori algorithm is introduced to estimate positions of all samples in the latent variable space. We present experimental evidences suggesting that the supervised GP-LVM is able to use the class label information effectively, and thus, it outperforms the GP-LVM and the discriminative extension of the GP-LVM consistently. The comparison with some supervised classification methods, such as Gaussian process classification and support vector machines, is also given to illustrate the advantage of the proposed method.

Latent tuberculosis in a newborn: diagnostic challenges.

In countries where the incidence of tuberculosis is low, perinatal tuberculosis is seldom diagnosed. With increasing numbers of human immunodeficiency virus-infected people and increasing immigrant population from high tuberculosis incidence countries, one might expect perinatal tuberculosis to become more frequent. Early recognition of newborns at risk for perinatal tuberculosis infection is of utmost importance to prevent disease by chemoprophylaxis. We describe a case of latent perinatal tuberculosis infection in a newborn infected from a mother with extrapulmonary primary tuberculosis. Tuberculin skin test was negative, and latent tuberculosis infection was eventually diagnosed by specific immunological tests. We discuss the difficulties in diagnosis of recent tuberculosis infection in neonates and infants, and the risk factors for vertical transmission of tuberculosis, which need to be taken into account in considering the need for chemoprophylaxis in the newborn. Although perinatal TB infection is a rare condition and diagnosis is difficult due to poor diagnostic testing in pregnancy and newborns, a high index of suspicion is needed to limit the diagnostic delay and to avoid progression to perinatal TB disease.

Equity trading volume and volatility: Latent information arrivals and common long-run dependencies

This article examines the behavior of equity trading volume and volatility for the individual firms composing the Standard & Poor's 100 composite index. Using multivariate spectral methods, we find that fractionally integrated processes best describe the long-run temporal dependencies in both series. Consistent with a stylized mixture-of-distributions hypothesis model in which the aggregate "news"-arrival process possesses long-memory characteristics, the long-run hyperbolic decay rates appear to be common across each volume-volatility pair.

Latent Stick-Breaking Processes.

We develop a model for stochastic processes with random marginal distributions. Our model relies on a stick-breaking construction for the marginal distribution of the process, and introduces dependence across locations by using a latent Gaussian copula model as the mechanism for selecting the atoms. The resulting latent stick-breaking process (LaSBP) induces a random partition of the index space, with points closer in space having a higher probability of being in the same cluster. We develop an efficient and straightforward Markov chain Monte Carlo (MCMC) algorithm for computation and discuss applications in financial econometrics and ecology. This article has supplementary material online.

Latent factor analysis to discover pathway-associated putative segmental aneuploidies in human cancers.

Tumor microenvironmental stresses, such as hypoxia and lactic acidosis, play important roles in tumor progression. Although gene signatures reflecting the influence of these stresses are powerful approaches to link expression with phenotypes, they do not fully reflect the complexity of human cancers. Here, we describe the use of latent factor models to further dissect the stress gene signatures in a breast cancer expression dataset. The genes in these latent factors are coordinately expressed in tumors and depict distinct, interacting components of the biological processes. The genes in several latent factors are highly enriched in chromosomal locations. When these factors are analyzed in independent datasets with gene expression and array CGH data, the expression values of these factors are highly correlated with copy number alterations (CNAs) of the corresponding BAC clones in both the cell lines and tumors. Therefore, variation in the expression of these pathway-associated factors is at least partially caused by variation in gene dosage and CNAs among breast cancers. We have also found the expression of two latent factors without any chromosomal enrichment is highly associated with 12q CNA, likely an instance of "trans"-variations in which CNA leads to the variations in gene expression outside of the CNA region. In addition, we have found that factor 26 (1q CNA) is negatively correlated with HIF-1alpha protein and hypoxia pathways in breast tumors and cell lines. This agrees with, and for the first time links, known good prognosis associated with both a low hypoxia signature and the presence of CNA in this region. Taken together, these results suggest the possibility that tumor segmental aneuploidy makes significant contributions to variation in the lactic acidosis/hypoxia gene signatures in human cancers and demonstrate that latent factor analysis is a powerful means to uncover such a linkage.

Bayesian Analysis of Latent Threshold Dynamic Models

We discuss a general approach to dynamic sparsity modeling in multivariate time series analysis. Time-varying parameters are linked to latent processes that are thresholded to induce zero values adaptively, providing natural mechanisms for dynamic variable inclusion/selection. We discuss Bayesian model specification, analysis and prediction in dynamic regressions, time-varying vector autoregressions, and multivariate volatility models using latent thresholding. Application to a topical macroeconomic time series problem illustrates some of the benefits of the approach in terms of statistical and economic interpretations as well as improved predictions. Supplementary materials for this article are available online. © 2013 Copyright Taylor and Francis Group, LLC.

Cross-Domain Multitask Learning with Latent Probit Models

Learning multiple tasks across heterogeneous domains is a challenging problem since the feature space may not be the same for different tasks. We assume the data in multiple tasks are generated from a latent common domain via sparse domain transforms and propose a latent probit model (LPM) to jointly learn the domain transforms, and the shared probit classifier in the common domain. To learn meaningful task relatedness and avoid over-fitting in classification, we introduce sparsity in the domain transforms matrices, as well as in the common classifier. We derive theoretical bounds for the estimation error of the classifier in terms of the sparsity of domain transforms. An expectation-maximization algorithm is derived for learning the LPM. The effectiveness of the approach is demonstrated on several real datasets.

Tuberculin skin test versus blood immunologic diagnosis of latent mycobacterium tuberclosis infection among old subjects

info:eu-repo/semantics/published

Effector functions of heparin-binding hemagglutinin-specific CD8+ T lymphocytes in latent human tuberculosis.

BACKGROUND: Most individuals infected with Mycobacterium tuberculosis do not develop tuberculosis (TB) and can be regarded as being protected by an appropriate immune response to the infection. The characterization of the immune responses of individuals with latent TB may thus be helpful in the definition of correlates of protection and the development of new vaccine strategies. The highly protective antigen heparin-binding hemagglutinin (HBHA) induces strong interferon (IFN)- gamma responses during latent, but not active, TB. Because of the recently recognized importance of CD8(+) T lymphocytes in anti-TB immunity, we characterized the CD8(+) T lymphocyte responses to HBHA in subjects with latent TB. RESULTS: HBHA-specific CD8(+) T lymphocytes expressed memory cell markers and synthesized HBHA-specific IFN- gamma .They also restricted mycobacterial growth and expressed cytotoxicity by a granule-dependent mechanism. This activity was associated with the intracellular expression of HBHA-induced perforin. Surprisingly, the perforin-producing CD8(+) T lymphocytes were distinct from the IFN- gamma -producing CD8(+) T lymphocytes. CONCLUSION: During latent TB, the HBHA-specific CD8(+) T lymphocyte population expresses all 3 effector functions associated with CD8(+) T lymphocyte-mediated protective immune mechanisms, which supports the notion that HBHA may be protective in humans and suggests that markers of HBHA-specific CD8(+) T lymphocyte responses may be useful in the monitoring of protection.

Heparin-binding-hemagglutinin-induced IFN-gamma release as a diagnostic tool for latent tuberculosis.

BACKGROUND: The detection of latent tuberculosis infection (LTBI) is a major component of tuberculosis (TB) control strategies. In addition to the tuberculosis skin test (TST), novel blood tests, based on in vitro release of IFN-gamma in response to Mycobacterium tuberculosis-specific antigens ESAT-6 and CFP-10 (IGRAs), are used for TB diagnosis. However, neither IGRAs nor the TST can separate acute TB from LTBI, and there is concern that responses in IGRAs may decline with time after infection. We have therefore evaluated the potential of the novel antigen heparin-binding hemagglutinin (HBHA) for in vitro detection of LTBI. METHODOLOGY AND PRINCIPAL FINDINGS: HBHA was compared to purified protein derivative (PPD) and ESAT-6 in IGRAs on lymphocytes drawn from 205 individuals living in Belgium, a country with low TB prevalence, where BCG vaccination is not routinely used. Among these subjects, 89 had active TB, 65 had LTBI, based on well-standardized TST reactions and 51 were negative controls. HBHA was significantly more sensitive than ESAT-6 and more specific than PPD for the detection of LTBI. PPD-based tests yielded 90.00% sensitivity and 70.00% specificity for the detection of LTBI, whereas the sensitivity and specificity for the ESAT-6-based tests were 40.74% and 90.91%, and those for the HBHA-based tests were 92.06% and 93.88%, respectively. The QuantiFERON-TB Gold In-Tube (QFT-IT) test applied on 20 LTBI subjects yielded 50% sensitivity. The HBHA IGRA was not influenced by prior BCG vaccination, and, in contrast to the QFT-IT test, remote (>2 years) infections were detected as well as recent (<2 years) infections by the HBHA-specific test. CONCLUSIONS: The use of ESAT-6- and CFP-10-based IGRAs may underestimate the incidence of LTBI, whereas the use of HBHA may combine the operational advantages of IGRAs with high sensitivity and specificity for latent infection.

Risk stratification of latent tuberculosis defined by combined interferon gamma release assays.

Most individuals infected with Mycobacterium tuberculosis develop latent tuberculosis infection (LTBI). Some may progress to active disease and would benefit from preventive treatment yet no means currently exists to predict who will reactivate. Here, we provide an approach to stratify LTBI based on IFN-γ responses to two antigens, the recombinant Early-Secreted Antigen Target-6 (rESAT-6) and the latency antigen Heparin-Binding Haemagglutinin (HBHA).

Heparin-Binding Haemagglutinin, a New Tool for the Detection of Latent Mycobacterium tuberculosis Infection in Hemodialysis Patients

Background:Patients with end-stage renal disease (ESRD) and latently infected with Mycobacterium tuberculosis (LTBI) are at higher risk to develop tuberculosis (TB) than healthy subjects. Interferon-gamma release assays (IGRAs) were reported to be more sensitive than tuberculin skin tests for the detection of infected individuals in dialysis patients.Methods:On 143 dialysis patients prospectively enrolled, we compared the results from the QuantiFERON®-TB Gold assay (QFT), to those of an IGRA in response to in vitro stimulation of circulating mononuclear cells with the mycobacterial latency antigen Heparin-Binding Haemagglutinin purified from Mycobacterium bovis BCG (native HBHA, nHBHA).Results:Seven patients had a past history of active TB and 1 had an undetermined result with both IGRAs. Among the other 135 patients, 94 had concordant results with the QFT and nHBHA-IGRA, 40.0% being negative and therefore not latently infected, and 29.6% being positive and thus LTBI. Discrepant results between these tests were found for 36 patients positive only with the nHBHA-IGRA and 5 only with the QFT.Conclusions:The nHBHA-IGRA is more sensitive than the QFT for the detection of LTBI dialysis patients, and follow-up of the patients will allow us to define the clinical significance of discrepant results between the nHBHA-IGRA and the QFT. © 2013 Dessein et al.

Key role of effector memory CD4+ T lymphocytes in a short-incubation heparin-binding hemagglutinin gamma interferon release assay for the detection of latent tuberculosis

info:eu-repo/semantics/published