Search

The ensemble method random forests has become a popular classification tool in bioinformatics and related fields. The out-of-bag error is an error estimation technique often used to evaluate the accuracy of a random forest and to select appropriate values for tuning parameters, such as the number of candidate predictors that are randomly drawn for a split, referred to as mtry. However, for binary classification problems with metric predictors it has been shown that the out-of-bag error can overestimate the true prediction error depending on the choices of random forests parameters. Based on simulated and real data this paper aims to identify settings for which this overestimation is likely. It is, moreover, questionable whether the out-of-bag error can be used in classification tasks for selecting tuning parameters like mtry, because the overestimation is seen to depend on the parameter mtry. The simulation-based and real-data based studies with metric predictor variables performed in this paper show that the overestimation is largest in balanced settings and in settings with few observations, a large number of predictor variables, small correlations between predictors and weak effects. There was hardly any impact of the overestimation on tuning parameter selection. However, although the prediction performance of random forests was not substantially affected when using the out-of-bag error for tuning parameter selection in the present studies, one cannot be sure that this applies to all future data. For settings with metric predictor variables it is therefore strongly recommended to use stratified subsampling with sampling fractions that are proportional to the class sizes for both tuning parameter selection and error estimation in random forests. This yielded less biased estimates of the true prediction error. In unbalanced settings, in which there is a strong interest in predicting observations from the smaller classes well, sampling the same number of observations from each class is a promising alternative. [ABSTRACT FROM AUTHOR]

The process of Nuclei detection in high-grade breast cancer images is quite challenging in the case of image processing techniques due to certain heterogeneous characteristics of cancer nuclei such as enlarged and irregularly shaped nuclei, highly coarse chromatin marginalized to the nuclei periphery and visible nucleoli. Recent reviews state that existing techniques show appreciable segmentation accuracy on breast histopathology images whose nuclei are dispersed and regular in texture and shape; however, typical cancer nuclei are often clustered and have irregular texture and shape properties. This paper proposes a novel segmentation algorithm for detecting individual nuclei from Hematoxylin and Eosin (H&E) stained breast histopathology images. This detection framework estimates a nuclei saliency map using tensor voting followed by boundary extraction of the nuclei on the saliency map using a Loopy Back Propagation (LBP) algorithm on a Markov Random Field (MRF). The method was tested on both whole-slide images and frames of breast cancer histopathology images. Experimental results demonstrate high segmentation performance with efficient precision, recall and dice-coefficient rates, upon testing high-grade breast cancer images containing several thousand nuclei. In addition to the optimal performance on the highly complex images presented in this paper, this method also gave appreciable results in comparison with two recently published methods—Wienert et al. (2012) and Veta et al. (2013), which were tested using their own datasets. [ABSTRACT FROM AUTHOR]

For studying cancer and genetic diseases, the issue of identifying high correlation genes from high-dimensional data is an important problem. It is a great challenge to select relevant biomarkers from gene expression data that contains some important correlation structures, and some of the genes can be divided into different groups with a common biological function, chromosomal location or regulation. In this paper, we propose a penalized accelerated failure time model CHR-DE using a non-convex regularization (local search) with differential evolution (global search) in a wrapper-embedded memetic framework. The complex harmonic regularization (CHR) can approximate to the combination and ℓq (1 ≤ q < 2) for selecting biomarkers in group. And differential evolution (DE) is utilized to globally optimize the CHR’s hyperparameters, which make CHR-DE achieve strong capability of selecting groups of genes in high-dimensional biological data. We also developed an efficient path seeking algorithm to optimize this penalized model. The proposed method is evaluated on synthetic and three gene expression datasets: breast cancer, hepatocellular carcinoma and colorectal cancer. The experimental results demonstrate that CHR-DE is a more effective tool for feature selection and learning prediction. [ABSTRACT FROM AUTHOR]

Several modern genomic technologies, such as DNA-Methylation arrays, measure spatially registered probes that number in the hundreds of thousands across multiple chromosomes. The measured probes are by themselves less interesting scientifically; instead scientists seek to discover biologically interpretable genomic regions comprised of contiguous groups of probes which may act as biomarkers of disease or serve as a dimension-reducing pre-processing step for downstream analyses. In this paper, we introduce an unsupervised feature learning technique which maps technological units (probes) to biological units (genomic regions) that are common across all subjects. We use ideas from fusion penalties and convex clustering to introduce a method for Spatial Convex Clustering, or SpaCC. Our method is specifically tailored to detecting multi-subject regions of methylation, but we also test our approach on the well-studied problem of detecting segments of copy number variation. We formulate our method as a convex optimization problem, develop a massively parallelizable algorithm to find its solution, and introduce automated approaches for handling missing values and determining tuning parameters. Through simulation studies based on real methylation and copy number variation data, we show that SpaCC exhibits significant performance gains relative to existing methods. Finally, we illustrate SpaCC’s advantages as a pre-processing technique that reduces large-scale genomics data into a smaller number of genomic regions through several cancer epigenetics case studies on subtype discovery, network estimation, and epigenetic-wide association. [ABSTRACT FROM AUTHOR]

Microarray technology is important to simultaneously express multiple genes over a number of time points. Multiple classifier models, such as sparse representation (SR)-based method, have been developed to classify microarray gene expression data. These methods allocate the gene data points to different clusters. In this paper, we propose a novel collaborative representation (CR)-based classification with regularized least square to classify gene data. First, the CR codes a testing sample as a sparse linear combination of all training samples and then classifies the testing sample by evaluating which class leads to the minimum representation error. This CR-based classification approach is remarkably less complex than traditional classification methods but leads to very competitive classification results. In addition, compressive sensing approach is adopted to project the high-dimensional gene expression dataset to a lower-dimensional space which nearly contains the whole information. This compression without loss is beneficial to reduce the computational load. Experiments to detect subtypes of diseases, such as leukemia and autism spectrum disorders, are performed by analyzing the gene expression. The results show that the proposed CR-based algorithm exhibits significantly higher stability and accuracy than the traditional classifiers, such as support vector machine algorithm. [ABSTRACT FROM AUTHOR]

Integrative clustering is used to identify groups of samples by jointly analysing multiple datasets describing the same set of biological samples, such as gene expression, copy number, methylation etc. Most existing algorithms for integrative clustering assume that there is a shared consistent set of clusters across all datasets, and most of the data samples follow this structure. However in practice, the structure across heterogeneous datasets can be more varied, with clusters being joined in some datasets and separated in others. In this paper, we present a probabilistic clustering method to identify groups across datasets that do not share the same cluster structure. The proposed algorithm, Clusternomics, identifies groups of samples that share their global behaviour across heterogeneous datasets. The algorithm models clusters on the level of individual datasets, while also extracting global structure that arises from the local cluster assignments. Clusters on both the local and the global level are modelled using a hierarchical Dirichlet mixture model to identify structure on both levels. We evaluated the model both on simulated and on real-world datasets. The simulated data exemplifies datasets with varying degrees of common structure. In such a setting Clusternomics outperforms existing algorithms for integrative and consensus clustering. In a real-world application, we used the algorithm for cancer subtyping, identifying subtypes of cancer from heterogeneous datasets. We applied the algorithm to TCGA breast cancer dataset, integrating gene expression, miRNA expression, DNA methylation and proteomics. The algorithm extracted clinically meaningful clusters with significantly different survival probabilities. We also evaluated the algorithm on lung and kidney cancer TCGA datasets with high dimensionality, again showing clinically significant results and scalability of the algorithm. [ABSTRACT FROM AUTHOR]

Breast cancer is an all too common disease in women, making how to effectively predict it an active research problem. A number of statistical and machine learning techniques have been employed to develop various breast cancer prediction models. Among them, support vector machines (SVM) have been shown to outperform many related techniques. To construct the SVM classifier, it is first necessary to decide the kernel function, and different kernel functions can result in different prediction performance. However, there have been very few studies focused on examining the prediction performances of SVM based on different kernel functions. Moreover, it is unknown whether SVM classifier ensembles which have been proposed to improve the performance of single classifiers can outperform single SVM classifiers in terms of breast cancer prediction. Therefore, the aim of this paper is to fully assess the prediction performance of SVM and SVM ensembles over small and large scale breast cancer datasets. The classification accuracy, ROC, F-measure, and computational times of training SVM and SVM ensembles are compared. The experimental results show that linear kernel based SVM ensembles based on the bagging method and RBF kernel based SVM ensembles with the boosting method can be the better choices for a small scale dataset, where feature selection should be performed in the data pre-processing stage. For a large scale dataset, RBF kernel based SVM ensembles based on boosting perform better than the other classifiers. [ABSTRACT FROM AUTHOR]

Identification of disease-causing genes is a fundamental challenge for human health studies. The phenotypic similarity among diseases may reflect the interactions at the molecular level, and phenotype comparison can be used to predict disease candidate genes. Online Mendelian Inheritance in Man (OMIM) is a database of human genetic diseases and related genes that has become an authoritative source of disease phenotypes. However, disease phenotypes have been described by free text; thus, standardization of phenotypic descriptions is needed before diseases can be compared. Several disease phenotype networks have been established in OMIM using different standardization methods. Two of these networks are important for phenotypic similarity analysis: the first and most commonly used network (mimMiner) is standardized by medical subject heading, and the other network (resnikHPO) is the first to be standardized by human phenotype ontology. This paper comprehensively evaluates for the first time the accuracy of these two networks in gene prioritization based on protein–protein interactions using large-scale, leave-one-out cross-validation experiments. The results show that both networks can effectively prioritize disease-causing genes, and the approach that relates two diseases using a logistic function improves prioritization performance. Tanimoto, one of four methods for normalizing resnikHPO, generates a symmetric network and it performs similarly to mimMiner. Furthermore, an integration of these two networks outperforms either network alone in gene prioritization, indicating that these two disease networks are complementary. [ABSTRACT FROM AUTHOR]

The maximal information coefficient (MIC) captures dependences between paired variables, including both functional and non-functional relationships. In this paper, we develop a new method, ChiMIC, to calculate the MIC values. The ChiMIC algorithm uses the chi-square test to terminate grid optimization and then removes the restriction of maximal grid size limitation of original ApproxMaxMI algorithm. Computational experiments show that ChiMIC algorithm can maintain same MIC values for noiseless functional relationships, but gives much smaller MIC values for independent variables. For noise functional relationship, the ChiMIC algorithm can reach the optimal partition much faster. Furthermore, the MCN values based on MIC calculated by ChiMIC can capture the complexity of functional relationships in a better way, and the statistical powers of MIC calculated by ChiMIC are higher than those calculated by ApproxMaxMI. Moreover, the computational costs of ChiMIC are much less than those of ApproxMaxMI. We apply the MIC values tofeature selection and obtain better classification accuracy using features selected by the MIC values from ChiMIC. [ABSTRACT FROM AUTHOR]