Hierarchical Bayes models for cDNA microarray gene expression

cDNA microarrays are used in many contexts to compare mRNA levels between samples of cells. Microarray experiments typically give us expression measurements on 1000-20 000 genes, but with few replicates for each gene. Traditional methods using means and standard deviations to detect differential expression are not satisfactory in this context. A handful of alternative statistics have been developed, including several empirical Bayes methods. In the present paper we present two full hierarchical Bayes models for detecting gene expression, of which one (D) describes our microarray data very well. We also compare the full Bayes and empirical Bayes approaches with respect to model assumptions, false discovery rates and computer running time. The proposed models are compared to existing empirical Bayes models in a simulation study and for a set of data (Yuen et al., 2002), where 27 genes have been categorized by quantitative real-time PCR. It turns out that the existing empirical Bayes methods have at least as good performance as the full Bayes ones.     

On hierarchical Bayes procedures for predicting simple exponential survival

The situation considered is the prediction of a future observation from a simple exponential survival distribution in a hierarchical Bayes context. It is shown that when the hyperparameters need to be estimated from the data, a sample reuse approach is superior to maximum likelihood and method of moments estimation procedures.     

Methods for optimizing antiviral combination therapies

MOTIVATION: Despite some progress with antiretroviral combination therapies, therapeutic success in the management of HIV-infected patients is limited. The evolution of drug-resistant genetic variants in response to therapy plays a key role in treatment failure and finding a new potent drug combination after therapy failure is considered challenging. RESULTS: To estimate the activity of a drug combination against a particular viral strain, we develop a scoring function whose independent variables describe a set of antiviral agents and viral DNA sequences coding for the molecular targets of the respective drugs. The construction of this activity score involves (1) predicting phenotypic drug resistance from genotypes for each drug individually, (2) probabilistic modeling of predicted resistance values and integration into a score for drug combinations, and (3) searching through the mutational neighborhood of the considered strain in order to estimate activity on nearby mutants. For a clinical data set, we determine the optimal search depth and show that the scoring scheme is predictive of therapeutic outcome. Properties of the activity score and applications are discussed.     

An adaptive multimeme algorithm for designing HIV multidrug therapies

This paper proposes a period representation for modeling the multidrug HIV therapies and an adaptive multimeme algorithm (AMmA) for designing the optimal therapy. The period representation offers benefits in terms of flexibility and reduction in dimensionality compared to the binary representation. The AMmA is a memetic algorithm which employs a list of three local searchers adaptively activated by an evolutionary framework. These local searchers, having different features according to the exploration logic and the pivot rule, have the role of exploring the decision space from different and complementary perspectives and, thus, assisting the standard evolutionary operators in the optimization process. Furthermore, the AMmA makes use of an adaptation which dynamically sets the algorithmic parameters in order to prevent stagnation and premature convergence. The numerical results demonstrate that the application of the proposed algorithm leads to very efficient medication schedules which quickly stimulate a strong immune response to HIV. The earlier termination of the medication schedule leads to lesser unpleasant side effects for the patient due to strong antiretroviral therapy. A numerical comparison shows that the AMmA is more efficient than three popular metaheuristics. Finally, a statistical test based on the calculation of the tolerance interval confirms the superiority of the AMmA compared to the other methods for the problem under study     

Impact of new antiretroviral combination therapies in HIV infected patients in Switzerland: prospective multicentre study. Swiss HIV Cohort Study.

OBJECTIVES: To examine trends in disease progression and survival among patients enrolled in the Swiss HIV cohort study during 1988-96 and to assess the influence of new antiretroviral combination therapies. DESIGN: Prospective multicentre study, with follow up visits planned at six monthly intervals. SETTING: Seven HIV units at university centres and cantonal hospitals in Switzerland. PATIENTS: 3785 men (mean age 35.0 years) and 1391 women (30.3 years) infected with HIV. 2023 participants had a history of intravenous drug misuse; 1764 were men who had sex with men; 1261 were infected heterosexually; and 164 had other or unknown modes of transmission. 601 participants had had an AIDS defining illness. RESULTS: During more than 15,000 years of follow up, there were 1456 first AIDS defining diagnoses and 1903 deaths. Compared with those enrolled during 1988-90, the risk of progression to a first AIDS diagnosis was reduced by 18% (relative risk 0.82 (95% confidence interval 0.73 to 0.93)) among participants enrolled in 1991-2, by 23% (0.77 (0.65 to 0.91)) among those enrolled in 1993-4, and by 73% (0.27 (0.18 to 0.39)) among those enrolled in 1995-6. Mortality was reduced by 19% (0.81 (0.73 to 0.90)), 26% (0.74 (0.63 to 0.87)), and 62% (0.38 (0.25 to 0.97)) respectively. Compared with no antiretroviral treatment, the risk of an initial AIDS diagnosis after CD4 lymphocyte counts fell to < 200 cells x 10(6)/1 was reduced by 16% (0.84 (0.73 to 0.97)) with monotherapy, 24% (0.76 (0.63 to 0.91)) with dual therapy, and 42% (0.58 (0.37 to 0.92)) with triple therapy. Mortality was reduced by 23% (0.77 (0.68 to 0.88)), 31% (0.69 (0.60 to 0.80)), and 65% (0.35 (0.20 to 0.60)) respectively. CONCLUSIONS: The introduction of antiretroviral combination therapies outside the selected patient groups included in clinical trials has led to comparable reductions in disease progression and mortality.     

A critical examination of current HIV therapies

This review critically examines the current methods of eliminating and preventing human immunodeficiency virus (HIV) infection. It illustrates both the experimental and practical limitations that each approach faces, and how they may be overcome. An overview of the HIV, including its structure and life cycle is presented. Subsequently, the two main methods of post-infection treatment, drug and gene therapy are outlined. The development of HIV vaccination is discussed with an analysis of conventional vaccination techniques leading into the novel approaches. The final option examined describes the potential for a combined vaccination regimen. Finally, the question of why these approaches have met with little success is addressed. This includes practical research limitations, as well as an examination of the qualities of HIV that make it so elusive.     

Computational Identification of MoRFs in Protein Sequences Using Hierarchical Application of Bayes Rule

Motivation

Intrinsically disordered regions of proteins play an essential role in the regulation of various biological processes. Key to their regulatory function is often the binding to globular protein domains via sequence elements known as molecular recognition features (MoRFs). Development of computational tools for the identification of candidate MoRF locations in amino acid sequences is an important task and an area of growing interest. Given the relative sparseness of MoRFs in protein sequences, the accuracy of the available MoRF predictors is often inadequate for practical usage, which leaves a significant need and room for improvement. In this work, we introduce MoRF_{CHiBi_Web}, which predicts MoRF locations in protein sequences with higher accuracy compared to current MoRF predictors.

Methods

Three distinct and largely independent property scores are computed with component predictors and then combined to generate the final MoRF propensity scores. The first score reflects the likelihood of sequence windows to harbour MoRFs and is based on amino acid composition and sequence similarity information. It is generated by MoRF_CHiBi using small windows of up to 40 residues in size. The second score identifies long stretches of protein disorder and is generated by ESpritz with the DisProt option. Lastly, the third score reflects residue conservation and is assembled from PSSM files generated by PSI-BLAST. These propensity scores are processed and then hierarchically combined using Bayes rule to generate the final MoRF_{CHiBi_Web} predictions.

Results

MoRF_{CHiBi_Web} was tested on three datasets. Results show that MoRF_{CHiBi_Web} outperforms previously developed predictors by generating less than half the false positive rate for the same true positive rate at practical threshold values. This level of accuracy paired with its relatively high processing speed makes MoRF_{CHiBi_Web} a practical tool for MoRF prediction.

Availability

http://morf.chibi.ubc.ca:8080/morf/.     

Hierarchical Dirichlet process model for gene expression clustering

Clustering is an important data processing tool for interpreting microarray data and genomic network inference. In this article, we propose a clustering algorithm based on the hierarchical Dirichlet processes (HDP). The HDP clustering introduces a hierarchical structure in the statistical model which captures the hierarchical features prevalent in biological data such as the gene express data. We develop a Gibbs sampling algorithm based on the Chinese restaurant metaphor for the HDP clustering. We apply the proposed HDP algorithm to both regulatory network segmentation and gene expression clustering. The HDP algorithm is shown to outperform several popular clustering algorithms by revealing the underlying hierarchical structure of the data. For the yeast cell cycle data, we compare the HDP result to the standard result and show that the HDP algorithm provides more information and reduces the unnecessary clustering fragments.     

Using ensemble of classifiers for predicting HIV protease cleavage sites in proteins

The focus of this work is the use of ensembles of classifiers for predicting HIV protease cleavage sites in proteins. Due to the complex relationships in the biological data, several recent works show that often ensembles of learning algorithms outperform stand-alone methods. We show that the fusion of approaches based on different encoding models can be useful for improving the performance of this classification problem. In particular, in this work four different feature encodings for peptides are described and tested. An extensive evaluation on a large dataset according to a blind testing protocol is reported which demonstrates how different feature extraction methods and classifiers can be combined for obtaining a robust and reliable system. The comparison with other stand-alone approaches allows quantifying the performance improvement obtained by the ensembles proposed in this work.     

A biological marker model for predicting disease transitions

For patients with chronic myelogenous leukemia (CML), the effect of elevated blood levels of adenosine deaminase (ADA) is studied as a marker for transitions from stable disease to blast crisis and then to death. Data in the form of snapshots over time, with day, state of disease, and ADA level, are analyzed for 55 patients. A simple three-state Markov model with one-way transition probabilities dependent on ADA is used to determine if the marker has a significant effect on the prediction of changes from stable disease to blast crisis.     

A novel means of using gene clusters in a two-step empirical Bayes method for predicting classes of samples

MOTIVATION: The classification of samples using gene expression profiles is an important application in areas such as cancer research and environmental health studies. However, the classification is usually based on a small number of samples, and each sample is a long vector of thousands of gene expression levels. An important issue in parametric modeling for so many gene expression levels is the control of the number of nuisance parameters in the model. Large models often lead to intensive or even intractable computation, while small models may be inadequate for complex data.Methodology: We propose a two-step empirical Bayes classification method as a solution to this issue. At the first step, we use the model-based cluster algorithm with a non-traditional purpose of assigning gene expression levels to form abundance groups. At the second step, by assuming the same variance for all the genes in the same group, we substantially reduce the number of nuisance parameters in our statistical model. RESULTS: The proposed model is more parsimonious, which leads to efficient computation under an empirical Bayes estimation procedure. We consider two real examples and simulate data using our method. Desired low classification error rates are obtained even when a large number of genes are pre-selected for class prediction.     

An integrated electromechanical-growth heart model for simulating cardiac therapies

An emerging class of models has been developed in recent years to predict cardiac growth and remodeling (G&R). We recently developed a cardiac G&R constitutive model that predicts remodeling in response to elevated hemodynamics loading, and a subsequent reversal of the remodeling process when the loading is reduced. Here, we describe the integration of this G&R model to an existing strongly coupled electromechanical model of the heart. A separation of timescale between growth deformation and elastic deformation was invoked in this integrated electromechanical-growth heart model. To test our model, we applied the G&R scheme to simulate the effects of myocardial infarction in a realistic left ventricular (LV) geometry using the finite element method. We also simulate the effects of a novel therapy that is based on alteration of the infarct mechanical properties. We show that our proposed model is able to predict key features that are consistent with experiments. Specifically, we show that the presence of a non-contractile infarct leads to a dilation of the left ventricle that results in a rightward shift of the pressure volume loop. Our model also predicts that G&R is attenuated by a reduction in LV dilation when the infarct stiffness is increased.     

A new mouse model to explore therapies for preeclampsia

Background

Pre-eclampsia, a pregnancy-specific multisystemic disorder is a leading cause of maternal and perinatal mortality and morbidity. This syndrome has been known to medical science since ancient times. However, despite considerable research, the cause/s of preeclampsia remain unclear, and there is no effective treatment. Development of an animal model that recapitulates this complex pregnancy-related disorder may help to expand our understanding and may hold great potential for the design and implementation of effective treatment.

Methodology/Principal Findings

Here we show that the CBA/J x DBA/2 mouse model of recurrent miscarriage is also a model of immunologically-mediated preeclampsia (PE). DBA/J mated CBA/J females spontaneously develop many features of human PE (primigravidity, albuminuria, endotheliosis, increased sensitivity to angiotensin II and increased plasma leptin levels) that correlates with bad pregnancy outcomes. We previously reported that antagonism of vascular endothelial growth factor (VEGF) signaling by soluble VEGF receptor 1 (sFlt-1) is involved in placental and fetal injury in CBA/J x DBA/2 mice. Using this animal model that recapitulates many of the features of preeclampsia in women, we found that pravastatin restores angiogenic balance, ameliorates glomerular injury, diminishes hypersensitivity to angiotensin II and protects pregnancies.

Conclusions/Significance

We described a new mouse model of PE, were the relevant key features of human preeclampsia develop spontaneously. The CBA/J x DBA/2 model, that recapitulates this complex disorder, helped us identify pravastatin as a candidate therapy to prevent preeclampsia and its related complications. We recognize that these studies were conducted in mice and that clinical trials are needed to confirm its application to humans.     

Computational methods for the design of effective therapies against drug resistant HIV strains

The development of drug resistance is a major obstacle to successful treatment of HIV infection. The extraordinary replication dynamics of HIV facilitates its escape from selective pressure exerted by the human immune system and by combination drug therapy. We have developed several computational methods whose combined use can support the design of optimal antiretroviral therapies based on viral genomic data.     

A combination of compositional index and genetic algorithm for predicting transmembrane helical segments

Transmembrane helix (TMH) topology prediction is becoming a focal problem in bioinformatics because the structure of TM proteins is difficult to determine using experimental methods. Therefore, methods that can computationally predict the topology of helical membrane proteins are highly desirable. In this paper we introduce TMHindex, a method for detecting TMH segments using only the amino acid sequence information. Each amino acid in a protein sequence is represented by a Compositional Index, which is deduced from a combination of the difference in amino acid occurrences in TMH and non-TMH segments in training protein sequences and the amino acid composition information. Furthermore, a genetic algorithm was employed to find the optimal threshold value for the separation of TMH segments from non-TMH segments. The method successfully predicted 376 out of the 378 TMH segments in a dataset consisting of 70 test protein sequences. The sensitivity and specificity for classifying each amino acid in every protein sequence in the dataset was 0.901 and 0.865, respectively. To assess the generality of TMHindex, we also tested the approach on another standard 73-protein 3D helix dataset. TMHindex correctly predicted 91.8% of proteins based on TM segments. The level of the accuracy achieved using TMHindex in comparison to other recent approaches for predicting the topology of TM proteins is a strong argument in favor of our proposed method. Availability: The datasets, software together with supplementary materials are available at: http://faculty.uaeu.ac.ae/nzaki/TMHindex.htm.     

Cancer treatment: the combination of vaccination with other therapies

Harnessing of the immune system by the development of ‘therapeutic’ vaccines, for the battle against cancer has been the focus of tremendous research efforts over the past two decades. As an illustration of the impressive amounts of data gathered over the past years, numerous antigens expressed on the surface of cancer cells, have been characterized. To this end, recent years research has focussed on characterization of antigens that play an important role for the growth and survival of cancer cells. Anti-apoptotic molecules like survivin that enhance the survival of cancer cells and facilitate their escape from cytotoxic therapies represent prime vaccination candidates. The characterization of a high number of tumor antigens allow the concurrent or serial immunological targeting of different proteins associated with such cancer traits. Moreover, while vaccination in itself is a promising new approach to fight cancer, the combination with additional therapy could create a number of synergistic effects. Herein we discuss the possibilities and prospects of vaccination when combined with other treatments. In this regard, cell death upon drug exposure may be immunogenic or non-immunogenic depending on the specific chemotherapeutics. Also, chemotherapy represents one of several options available for clearance of CD4⁺ Foxp3⁺ regulatory T cells. Moreover, therapies based on monoclonal antibodies may have synergistic potential in combination with vaccination, both when used for targeting of tumor cells and endothelial cells. The efficacy of therapeutic vaccination against cancer will over the next few years be studied in settings taking advantage of strategies in which vaccination is combined with other treatment modalities. These combinations should be based on current knowledge not only regarding the biology of the cancer cell per se, but also considering how treatment may influence the malignant cell population as well as the immune system. This article is a symposium paper from the conference “Progress in Vaccination against Cancer 2007 (PIVAC 7)”, held in Stockholm, Sweden, on 10–11 September 2007.