MOTIVATION: Identifying disease associated taxa and constructing networks for bacteria interactions are two important tasks usually studied separately. In reality, differentiation of disease associated taxa and correlation among taxa may affect each other. One genus can be differentiated because it is highly correlated with another highly differentiated one. In addition, network structures may vary under different clinical conditions. Permutation tests are commonly used to detect differences between networks in distinct phenotypes, and they are time-consuming. RESULTS: In this manuscript, we propose a multilevel regularized regression method to simultaneously identify taxa and construct networks. We also extend the framework to allow construction of a common network and differentiated network together. An efficient algorithm with dual formulation is developed to deal with the large-scale n ≪ m problem with a large number of taxa (m) and a small number of samples (n) efficiently. The proposed method is regularized with a general Lp (p ∈ [0, 2]) penalty and models the effects of taxa abundance differentiation and correlation jointly. We demonstrate that it can identify both true and biologically significant genera and network structures. AVAILABILITY AND IMPLEMENTATION: Software MLRR in MATLAB is available at http://biostatistics.csmc.edu/mlrr/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

With the development of various high throughput technologies and analysis methods, researchers can study different aspects of a biological phenomenon simultaneously or one aspect repeatedly with different experimental techniques and analysis methods. The output from each study is a rank list of components of interest. Aggregation of the rank lists of components, such as proteins, genes and single nucleotide variants (SNV), produced by these experiments has been proven to be helpful in both filtering the noise and bringing forth a more complete understanding of the biological problems. Current available rank aggregation methods do not consider the network information that has been observed to provide vital contributions in many data integration studies. We developed network tuned rank aggregation methods incorporating network information and demonstrated its superior performance over aggregation methods without network information. The methods are tested on predicting the Gene Ontology function of yeast proteins. We validate the methods using combinations of three gene expression data sets and three protein interaction networks as well as an integrated network by combining the three networks. Results show that the aggregated rank lists are more meaningful if protein interaction network is incorporated. Among the methods compared, CGI_RRA and CGI_Endeavour, which integrate rank lists with networks using CGI [ 1 ] followed by rank aggregation using either robust rank aggregation (RRA) [ 2 ] or Endeavour [ 3 ] perform the best. Finally, we use the methods to locate target genes of transcription factors.

MOTIVATION: Direct sequencing of microbes in human ecosystems (the human microbiome) has complemented single genome cultivation and sequencing to understand and explore the impact of commensal microbes on human health. As sequencing technologies improve and costs decline, the sophistication of data has outgrown available computational methods. While several existing machine learning methods have been adapted for analyzing microbiome data recently, there is not yet an efficient and dedicated algorithm available for multiclass classification of human microbiota. RESULTS: By combining instance-based and model-based learning, we propose a novel sparse distance-based learning method for simultaneous class prediction and feature (variable or taxa, which is used interchangeably) selection from multiple treatment populations on the basis of 16S rRNA sequence count data. Our proposed method simultaneously minimizes the intraclass distance and maximizes the interclass distance with many fewer estimated parameters than other methods. It is very efficient for problems with small sample sizes and unbalanced classes, which are common in metagenomic studies. We implemented this method in a MATLAB toolbox called MetaDistance. We also propose several approaches for data normalization and variance stabilization transformation in MetaDistance. We validate this method on several real and simulated 16S rRNA datasets to show that it outperforms existing methods for classifying metagenomic data. This article is the first to address simultaneous multifeature selection and class prediction with metagenomic count data. AVAILABILITY: The MATLAB toolbox is freely available online at http://metadistance.igs.umaryland.edu/. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Partial AUC (pAUC) represents the area with a restricted range of specificity (e.g. low false positive rate). It may identify important regional differentiated genes missed by full-range analysis. Unlike the popular t-test, which is based on the mean difference and the standard deviation between the disease and health groups, pAUC based test statistic relies on the rank of a gene in different samples. It can effectively detect genes that are not significant in a t-test and only differentiated in a subset of the disease groups. Our experiments with real gene expression data show that the proposed pAUC statistic is appealing in terms of both detection power and the biological relevance of the results.

BACKGROUND: Most genomic data have ultra-high dimensions with more than 10,000 genes (probes). Regularization methods with L₁ and L(p) penalty have been extensively studied in survival analysis with high-dimensional genomic data. However, when the sample size n << m (the number of genes), directly identifying a small subset of genes from ultra-high (m > 10, 000) dimensional data is time-consuming and not computationally efficient. In current microarray analysis, what people really do is select a couple of thousands (or hundreds) of genes using univariate analysis or statistical tests, and then apply the LASSO-type penalty to further reduce the number of disease associated genes. This two-step procedure may introduce bias and inaccuracy and lead us to miss biologically important genes. RESULTS: The accelerated failure time (AFT) model is a linear regression model and a useful alternative to the Cox model for survival analysis. In this paper, we propose a nonlinear kernel based AFT model and an efficient variable selection method with adaptive kernel ridge regression. Our proposed variable selection method is based on the kernel matrix and dual problem with a much smaller n x n matrix. It is very efficient when the number of unknown variables (genes) is much larger than the number of samples. Moreover, the primal variables are explicitly updated and the sparsity in the solution is exploited. CONCLUSIONS: Our proposed methods can simultaneously identify survival associated prognostic factors and predict survival outcomes with ultra-high dimensional genomic data. We have demonstrated the performance of our methods with both simulation and real data. The proposed method performs superbly with limited computational studies.

The development of high-throughput technology has generated a massive amount of high-dimensional data, and many of them are of discrete type. Robust and efficient learning algorithms such as LASSO [1] are required for feature selection and overfitting control. However, most feature selection algorithms are only applicable to the continuous data type. In this paper, we propose a novel method for sparse support vector machines (SVMs) with L_(p) (p < 1) regularization. Efficient algorithms (LpSVM) are developed for learning the classifier that is applicable to high-dimensional data sets with both discrete and continuous data types. The regularization parameters are estimated through maximizing the area under the ROC curve (AUC) of the cross-validation data. Experimental results on protein sequence and SNP data attest to the accuracy, sparsity, and efficiency of the proposed algorithm. Biomarkers identified with our methods are compared with those from other methods in the literature. The software package in Matlab is available upon request.

In an association study, empirical evidences support the commonality of gene-gene interactions. Although genetic factors play an important role in many human diseases, multiple genes or genes and environmental factors may ultimately influence individual risk for these disease. However, such interactions are difficult to detect. In this paper, we propose a penalized area under ROC curve (AUC) maximization (LpAUC) to detect gene-gene interactions. The proposed approach is demonstrated by a simulation study and real data analysis. Analyses of both real data and simulated data show the effectiveness of our approach.

BACKGROUND: Identifying genes and pathways associated with diseases such as cancer has been a subject of considerable research in recent years in the area of bioinformatics and computational biology. It has been demonstrated that the magnitude of differential expression does not necessarily indicate biological significance. Even a very small change in the expression of particular gene may have dramatic physiological consequences if the protein encoded by this gene plays a catalytic role in a specific cell function. Moreover, highly correlated genes may function together on the same pathway biologically. Finally, in sparse logistic regression with Lp (p < 1) penalty, the degree of the sparsity obtained is determined by the value of the regularization parameter. Usually this parameter must be carefully tuned through cross-validation, which is time consuming. RESULTS: In this paper, we proposed a simple Bayesian approach to integrate the regularization parameter out analytically using a new prior. Therefore, there is no longer a need for parameter selection, as it is eliminated entirely from the model. The proposed algorithm (BLpLog) is typically two or three orders of magnitude faster than the original algorithm and free from bias in performance estimation. We also define a novel similarity measure and develop an integrated algorithm to hunt the regulatory genes with low expression changes but having high correlation with the selected genes. Pathways of those correlated genes were identified with DAVID http://david.abcc.ncifcrf.gov/. CONCLUSION: Experimental results with gene expression data demonstrate that the proposed methods can be utilized to identify important genes and pathways that are related to cancer and build a parsimonious model for future patient predictions.

Advances in high throughput technology provide massive high dimensional data. It is very important and challenging to study the association of genes with various clinical outcomes. Due to large variability in time to certain clinical event among patients, studying possibly censored survival data can be more informative than classification. We proposed the Cox's proportional hazards model with Lppenalty method for simultaneous feature (gene) selection and survival prediction. Lppenalty shrinks coefficients and produces some coefficients that are exactly zero. It has been shown that Lp(p1in the regression and classification framework (Knight & Fu 2000, Liu et al. 2007). Experimental results with different data demonstrate that the proposed procedures can be used for identifying important genes (features) that are related to time to death due to cancer and for building parsimonious model for predicting the survival of future patients.

OBJECTIVE: The development and application of data-driven decision-support systems for medical triage, diagnostics, and prognostics pose special requirements on physiologic data. In particular, that data are reliable in order to produce meaningful results. The authors describe a method that automatically estimates the reliability of reference heart rates (HRr) derived from electrocardiogram (ECG) waveforms and photoplethysmogram (PPG) waveforms recorded by vital-signs monitors. The reliability is quantitatively expressed through a quality index (QI) for each HRr. DESIGN: The proposed method estimates the reliability of heart rates from vital-signs monitors by (1) assessing the quality of the ECG and PPG waveforms, (2) separately computing heart rates from these waveforms, and (3) concisely combining this information into a QI that considers the physical redundancy of the signal sources and independence of heart rate calculations. The assessment of the waveforms is performed by a Support Vector Machine classifier and the independent computation of heart rate from the waveforms is performed by an adaptive peak identification technique, termed ADAPIT, which is designed to filter out motion-induced noise. RESULTS: The authors evaluated the method against 158 randomly selected data samples of trauma patients collected during helicopter transport, each sample consisting of 7-second ECG and PPG waveform segments and their associated HRr. They compared the results of the algorithm against manual analysis performed by human experts and found that in 92% of the cases, the algorithm either matches or is more conservative than the human's QI qualification. In the remaining 8% of the cases, the algorithm infers a less conservative QI, though in most cases this was because of algorithm/human disagreement over ambiguous waveform quality. If these ambiguous waveforms were relabeled, the misclassification rate would drop from 8% to 3%. CONCLUSION: This method provides a robust approach for automatically assessing the reliability of large quantities of heart rate data and the waveforms from which they are derived.

Gene expression profiles have been successfully applied to class prediction. Due to a large number of genes (features) and a small number of samples in gene expression data, feature selection is essential when performing the prediction task. Many methods have been proposed to select features in microarray data analysis, but there is no unique method which performs uniformly well for all the learning algorithms. It is then practical to find a feature selection method and a learning algorithm that give superior performance. In this paper, we present an integrated scheme to perform the task of class prediction based on gene expression profiles. The scheme incorporates a simple novel feature selection procedure into naive Bayes models. Each selected gene has a high score of discriminatory power determined by the Brown-Forsythe test statistics. Any pair of selected genes have a low correlation. This facilitates the use of the conditional independence among genes assumed by the naive Bayes models. To demonstrate the effectiveness, the proposed scheme was applied to three commonly used expression data sets COLON, OVARIAN, and LEUKEMIA. The results show that the numbers of misclassified samples are 0, and 4, respectively.