Hsun-Hsien Chang

CGBayesNets: Conditional Gaussian Bayesian Network Learning and Inference with Mixed Discrete and Continuous Data

Bayesian Networks (BN) have been a popular predictive modeling formalism in bioinformatics, but their application in modern genomics has been slowed by an inability to cleanly handle domains with mixed discrete and continuous variables. Existing free BN software packages either discretize continuous variables, which can lead to information loss, or do not include inference routines, which makes prediction with the BN impossible. We present CGBayesNets, a BN package focused around prediction of a clinical phenotype from mixed discrete and continuous variables, which fills these gaps. CGBayesNets implements Bayesian likelihood and inference algorithms for the conditional Gaussian Bayesian network (CGBNs) formalism, one appropriate for predicting an outcome of interest from, e.g., multimodal genomic data. We provide four different network learning algorithms, each making a different tradeoff between computational cost and network likelihood. CGBayesNets provides a full suite of functions for model exploration and verification, including cross validation, bootstrapping, and AUC manipulation. We highlight several results obtained previously with CGBayesNets, including predictive models of wood properties from tree genomics, leukemia subtype classification from mixed genomic data, and robust prediction of intensive care unit mortality outcomes from metabolomic profiles. We also provide detailed example analysis on public metabolomic and gene expression datasets. CGBayesNets is implemented in MATLAB and available as MATLAB source code, under an Open Source license and anonymous download at http://www.cgbayesnets.com.

Predicting Inhaled Corticosteroid Response in Asthma with Two Associated SNPs

Inhaled corticosteroids (ICS) are the most commonly used controller medications prescribed for asthma. Two single-nucleotide polymorphisms (SNPs), rs1876828 in corticotrophin releasing hormone receptor 1 and rs37973 in GLCCI1, have previously been associated with corticosteroid efficacy. We studied data from four existing clinical trials of asthmatics, who received ICS and had lung function measured by forced expiratory volume in 1 s (FEV1) before and after the period of such treatment. We combined the two SNPs rs37973 and rs1876828 into a predictive test of FEV1 change using a Bayesian model, which identified patients with good or poor steroid response (highest or lowest quartile, respectively) with predictive performance of 65.7% (P=0.039 vs random) area under the receiver–operator characteristic curve in the training population and 65.9% (P=0.025 vs random) in the test population. These findings show that two genetic variants can be combined into a predictive test that achieves similar accuracy and superior replicability compared with single SNP predictors.

Transcriptional network classifiers

BackgroundGene interactions play a central role in transcriptional networks. Many studies have performed genome-wide expression analysis to reconstruct regulatory networks to investigate disease processes. Since biological processes are outcomes of regulatory gene interactions, this paper develops a system biology approach to infer function-dependent transcriptional networks modulating phenotypic traits, which serve as a classifier to identify tissue states. Due to gene interactions taken into account in the analysis, we can achieve higher classification accuracy than existing methods.ResultsOur system biology approach is carried out by the Bayesian networks framework. The algorithm consists of two steps: gene filtering by Bayes factor followed by collinearity elimination via network learning. We validate our approach with two clinical data. In the study of lung cancer subtypes discrimination, we obtain a 25-gene classifier from 111 training samples, and the test on 422 independent samples achieves 95% classification accuracy. In the study of thoracic aortic aneurysm (TAA) diagnosis, 61 samples determine a 34-gene classifier, whose diagnosis accuracy on 33 independent samples achieves 82%. The performance comparisons with three other popular methods, PCA/LDA, PAM, and Weighted Voting, confirm that our approach yields superior classification accuracy and a more compact signature.ConclusionsThe system biology approach presented in this paper is able to infer function-dependent transcriptional networks, which in turn can classify biological samples with high accuracy. The validation of our classifier using clinical data demonstrates the promising value of our proposed approach for disease diagnosis.

Automatic Detection of Regional Heart Rejection in USPIO-Enhanced MRI

Contrast-enhanced magnetic resonance imaging (MRI) is useful to study the infiltration of cells in vivo. This research adopts ultrasmall superparamagnetic iron oxide (USPIO) particles as contrast agents. USPIO particles administered intravenously can be endocytosed by circulating immune cells, in particular, macrophages. Hence, macrophages are labeled with USPIO particles. When a transplanted heart undergoes rejection, immune cells will infiltrate the allograft. Imaged by T2*-weighted MRI, USPIO-labeled macrophages display dark pixel intensities. Detecting these labeled cells in the image facilitates the identification of acute heart rejection. This paper develops a classifier to detect the presence of USPIO-labeled macrophages in the myocardium in the framework of spectral graph theory. First, we describe a USPIO-enhanced heart image with a graph. Classification becomes equivalent to partitioning the graph into two disjoint subgraphs. We use the Cheeger constant of the graph as an objective functional to derive the classifier. We represent the classifier as a linear combination of basis functions given from the spectral analysis of the graph Laplacian. Minimization of the Cheeger constant based functional leads to the optimal classifier. Experimental results and comparisons with other methods suggest the feasibility of our approach to study the rejection of hearts imaged by USPIO-enhanced MRI.

A Transcriptional Network Signature Characterizes Lung Cancer Subtypes

Transcriptional networks play a central role in cancer development. The authors described a systems biology approach to cancer classification based on the reverse engineering of the transcriptional network surrounding the 2 most common types of lung cancer: adenocarcinoma (AC) and squamous cell carcinoma (SCC).

Phenotype prediction by integrative network analysis of SNP and gene expression microarrays

A long-term goal of biomedical research is to decipher how genetic processes influence disease formation. Ubiquitous and advancing microarray technology can measure millions of DNA structural variants (single-nucleotide polymorphisms, or SNPs) and thousands of gene transcripts (RNA expression microarrays) in cells. Both of these information modalities can be brought to bear on disease etiology. This paper develops a Bayesian network-based approach to integrate SNP and expression microarray data. The network models SNP-gene interactions using a phenotype-centric network. Inferring the network consists of two steps: variable selection and network learning. The learned network illustrates how functionally dependent SNPs and genes influence each other, and also serves as a predictor of the phenotype. The application of the proposed method to a pediatric acute lymphoblastic leukemia dataset demonstrates the feasibility of our approach and its impact on biological investigation and clinical practice.

Mapping transcription mechanisms from multimodal genomic data

BackgroundIdentification of expression quantitative trait loci (eQTLs) is an emerging area in genomic study. The task requires an integrated analysis of genome-wide single nucleotide polymorphism (SNP) data and gene expression data, raising a new computational challenge due to the tremendous size of data.ResultsWe develop a method to identify eQTLs. The method represents eQTLs as information flux between genetic variants and transcripts. We use information theory to simultaneously interrogate SNP and gene expression data, resulting in a Transcriptional Information Map (TIM) which captures the network of transcriptional information that links genetic variations, gene expression and regulatory mechanisms. These maps are able to identify both cis- and trans- regulating eQTLs. The application on a dataset of leukemia patients identifies eQTLs in the regions of the GART, PCP4, DSCAM, and RIPK4 genes that regulate ADAMTS1, a known leukemia correlate.ConclusionsThe information theory approach presented in this paper is able to infer the dependence networks between SNPs and transcripts, which in turn can identify cis- and trans-eQTLs. The application of our method to the leukemia study explains how genetic variants and gene expression are linked to leukemia.

Transcriptional network predicts viral set point during acute HIV-1 infection

BACKGROUND HIV-1-infected individuals with higher viral set points progress to AIDS more rapidly than those with lower set points. Predicting viral set point early following infection can contribute to our understanding of early control of HIV-1 replication, to predicting long-term clinical outcomes, and to the choice of optimal therapeutic regimens. METHODS In a longitudinal study of 10 untreated HIV-1-infected patients, we used gene expression profiling of peripheral blood mononuclear cells to identify transcriptional networks for viral set point prediction. At each sampling time, a statistical analysis inferred the optimal transcriptional network that best predicted viral set point. We then assessed the accuracy of this transcriptional model by predicting viral set point in an independent cohort of 10 untreated HIV-1-infected patients from Malawi. RESULTS The gene network inferred at time of enrollment predicted viral set point 24 weeks later in the independent Malawian cohort with an accuracy of 87.5%. As expected, the predictive accuracy of the networks inferred at later time points was even greater, exceeding 90% after week 4. The composition of the inferred networks was largely conserved between time points. The 12 genes comprising this dynamic signature of viral set point implicated the involvement of two major canonical pathways: interferon signaling (p<0.0003) and membrane fraction (p<0.02). A silico knockout study showed that HLA-DRB1 and C4BPA may contribute to restricting HIV-1 replication. CONCLUSIONS Longitudinal gene expression profiling of peripheral blood mononuclear cells from patients with acute HIV-1 infection can be used to create transcriptional network models to early predict viral set point with a high degree of accuracy.

TCR Affinity Associated with Functional Differences between Dominant and Subdominant SIV Epitope-Specific CD8+ T Cells in Mamu-A*01+ Rhesus Monkeys

Many of the factors that contribute to CD8+ T cell immunodominance hierarchies during viral infection are known. However, the functional differences that exist between dominant and subdominant epitope-specific CD8+ T cells remain poorly understood. In this study, we characterized the phenotypic and functional differences between dominant and subdominant simian immunodeficiency virus (SIV) epitope-specific CD8+ T cells restricted by the major histocompatibility complex (MHC) class I allele Mamu-A*01 during acute and chronic SIV infection. Whole genome expression analyses during acute infection revealed that dominant SIV epitope-specific CD8+ T cells had a gene expression profile consistent with greater maturity and higher cytotoxic potential than subdominant epitope-specific CD8+ T cells. Flow-cytometric measurements of protein expression and anti-viral functionality during chronic infection confirmed these phenotypic and functional differences. Expression analyses of exhaustion-associated genes indicated that LAG-3 and CTLA-4 were more highly expressed in the dominant epitope-specific cells during acute SIV infection. Interestingly, only LAG-3 expression remained high during chronic infection in dominant epitope-specific cells. We also explored the binding interaction between peptide:MHC (pMHC) complexes and their cognate TCRs to determine their role in the establishment of immunodominance hierarchies. We found that epitope dominance was associated with higher TCR:pMHC affinity. These studies demonstrate that significant functional differences exist between dominant and subdominant epitope-specific CD8+ T cells within MHC-restricted immunodominance hierarchies and suggest that TCR:pMHC affinity may play an important role in determining the frequency and functionality of these cell populations. These findings advance our understanding of the regulation of T cell immunodominance and will aid HIV vaccine design.

Immune Cells Detection of the In Vivo Rejecting Heart in USPIO-Enhanced Magnetic Resonance Imaging

Contrast-enhanced magnetic resonance imaging (MRI) is useful to study the infiltration of immune cells, in particular macrophages. Contrast agents, for example ultra-small superparamagnetic iron oxide (USPIO) particles, administered intravenously into the blood stream will be engulfed by macrophages circulating in the circulation system. When a transplanted heart rejects, macrophages and other immune cells will infiltrate the rejecting tissue. Imaged by T2* weighted MRI, USPIO-labeled macrophages will display dark pixel intensities. Detecting the presence of USPIO particles in the images facilitates the study of heart rejection. We cast the problem of detecting the presence of USPIO-labeled myocardium in the framework of spectral graph theory, and treat our decision function as a level set function on the image. The pixels with positive level set values correspond to the presence of immune cells, and negative to the absence. When the image is modeled by a graph, the spectral analysis of the graph Laplacian provides a basis to represent the level set function. We develop from the Cheeger constant of the graph an objective functional of the level set function. The minimization of the objective leads to the optimal level set function. Experimental results suggest the feasibility of our approach in the study of rejecting hearts