Accurate and Reliable Cancer Classification Based on Probabilistic Inference of Pathway Activity

With the advent of high-throughput technologies for measuring genome-wide expression profiles, a large number of methods have been proposed for discovering diagnostic markers that can accurately discriminate between different classes of a disease. However, factors such as the small sample size of typical clinical data, the inherent noise in high-throughput measurements, and the heterogeneity across different samples, often make it difficult to find reliable gene markers. To overcome this problem, several studies have proposed the use of pathway-based markers, instead of individual gene markers, for building the classifier. Given a set of known pathways, these methods estimate the activity level of each pathway by summarizing the expression values of its member genes, and use the pathway activities for classification. It has been shown that pathway-based classifiers typically yield more reliable results compared to traditional gene-based classifiers. In this paper, we propose a new classification method based on probabilistic inference of pathway activities. For a given sample, we compute the log-likelihood ratio between different disease phenotypes based on the expression level of each gene. The activity of a given pathway is then inferred by combining the log-likelihood ratios of the constituent genes. We apply the proposed method to the classification of breast cancer metastasis, and show that it achieves higher accuracy and identifies more reproducible pathway markers compared to several existing pathway activity inference methods.     

Pathway activity inference for multiclass disease classification through a mathematical programming optimisation framework

Background

Applying machine learning methods on microarray gene expression profiles for disease classification problems is a popular method to derive biomarkers, i.e. sets of genes that can predict disease state or outcome. Traditional approaches where expression of genes were treated independently suffer from low prediction accuracy and difficulty of biological interpretation. Current research efforts focus on integrating information on protein interactions through biochemical pathway datasets with expression profiles to propose pathway-based classifiers that can enhance disease diagnosis and prognosis. As most of the pathway activity inference methods in literature are either unsupervised or applied on two-class datasets, there is good scope to address such limitations by proposing novel methodologies.

Results

A supervised multiclass pathway activity inference method using optimisation techniques is reported. For each pathway expression dataset, patterns of its constituent genes are summarised into one composite feature, termed pathway activity, and a novel mathematical programming model is proposed to infer this feature as a weighted linear summation of expression of its constituent genes. Gene weights are determined by the optimisation model, in a way that the resulting pathway activity has the optimal discriminative power with regards to disease phenotypes. Classification is then performed on the resulting low-dimensional pathway activity profile.

Conclusions

The model was evaluated through a variety of published gene expression profiles that cover different types of disease. We show that not only does it improve classification accuracy, but it can also perform well in multiclass disease datasets, a limitation of other approaches from the literature. Desirable features of the model include the ability to control the maximum number of genes that may participate in determining pathway activity, which may be pre-specified by the user. Overall, this work highlights the potential of building pathway-based multi-phenotype classifiers for accurate disease diagnosis and prognosis problems.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-014-0390-2) contains supplementary material, which is available to authorized users.     

Exploring metabolic pathway disruption in the subchronic phencyclidine model of schizophrenia with the Generalized Singular Value Decomposition

Background

One of the major goals in gene and protein expression profiling of cancer is to identify biomarkers and build classification models for prediction of disease prognosis or treatment response. Many traditional statistical methods, based on microarray gene expression data alone and individual genes' discriminatory power, often fail to identify biologically meaningful biomarkers thus resulting in poor prediction performance across data sets. Nonetheless, the variables in multivariable classifiers should synergistically interact to produce more effective classifiers than individual biomarkers.

Results

We developed an integrated approach, namely network-constrained support vector machine (netSVM), for cancer biomarker identification with an improved prediction performance. The netSVM approach is specifically designed for network biomarker identification by integrating gene expression data and protein-protein interaction data. We first evaluated the effectiveness of netSVM using simulation studies, demonstrating its improved performance over state-of-the-art network-based methods and gene-based methods for network biomarker identification. We then applied the netSVM approach to two breast cancer data sets to identify prognostic signatures for prediction of breast cancer metastasis. The experimental results show that: (1) network biomarkers identified by netSVM are highly enriched in biological pathways associated with cancer progression; (2) prediction performance is much improved when tested across different data sets. Specifically, many genes related to apoptosis, cell cycle, and cell proliferation, which are hallmark signatures of breast cancer metastasis, were identified by the netSVM approach. More importantly, several novel hub genes, biologically important with many interactions in PPI network but often showing little change in expression as compared with their downstream genes, were also identified as network biomarkers; the genes were enriched in signaling pathways such as TGF-beta signaling pathway, MAPK signaling pathway, and JAK-STAT signaling pathway. These signaling pathways may provide new insight to the underlying mechanism of breast cancer metastasis.

Conclusions

We have developed a network-based approach for cancer biomarker identification, netSVM, resulting in an improved prediction performance with network biomarkers. We have applied the netSVM approach to breast cancer gene expression data to predict metastasis in patients. Network biomarkers identified by netSVM reveal potential signaling pathways associated with breast cancer metastasis, and help improve the prediction performance across independent data sets.     

Classification of early‐stage non‐small cell lung cancer by weighing gene expression profiles with connectivity information

Pathway‐based feature selection algorithms, which utilize biological information contained in pathways to guide which features/genes should be selected, have evolved quickly and become widespread in the field of bioinformatics. Based on how the pathway information is incorporated, we classify pathway‐based feature selection algorithms into three major categories—penalty, stepwise forward, and weighting. Compared to the first two categories, the weighting methods have been underutilized even though they are usually the simplest ones. In this article, we constructed three different genes’ connectivity information‐based weights for each gene and then conducted feature selection upon the resulting weighted gene expression profiles. Using both simulations and a real‐world application, we have demonstrated that when the data‐driven connectivity information constructed from the data of specific disease under study is considered, the resulting weighted gene expression profiles slightly outperform the original expression profiles. In summary, a big challenge faced by the weighting method is how to estimate pathway knowledge‐based weights more accurately and precisely. Only until the issue is conquered successfully will wide utilization of the weighting methods be impossible.     

Accurate cancer phenotype prediction with AKLIMATE,a stacked kernel learner integrating multimodal genomic data and pathway knowledge

Advancements in sequencing have led to the proliferation of multi-omic profiles of human cells under different conditions and perturbations. In addition, many databases have amassed information about pathways and gene “signatures”—patterns of gene expression associated with specific cellular and phenotypic contexts. An important current challenge in systems biology is to leverage such knowledge about gene coordination to maximize the predictive power and generalization of models applied to high-throughput datasets. However, few such integrative approaches exist that also provide interpretable results quantifying the importance of individual genes and pathways to model accuracy. We introduce AKLIMATE, a first kernel-based stacked learner that seamlessly incorporates multi-omics feature data with prior information in the form of pathways for either regression or classification tasks. AKLIMATE uses a novel multiple-kernel learning framework where individual kernels capture the prediction propensities recorded in random forests, each built from a specific pathway gene set that integrates all omics data for its member genes. AKLIMATE has comparable or improved performance relative to state-of-the-art methods on diverse phenotype learning tasks, including predicting microsatellite instability in endometrial and colorectal cancer, survival in breast cancer, and cell line response to gene knockdowns. We show how AKLIMATE is able to connect feature data across data platforms through their common pathways to identify examples of several known and novel contributors of cancer and synthetic lethality.     

Identifying causal genes and dysregulated pathways in complex diseases

In complex diseases, various combinations of genomic perturbations often lead to the same phenotype. On a molecular level, combinations of genomic perturbations are assumed to dys-regulate the same cellular pathways. Such a pathway-centric perspective is fundamental to understanding the mechanisms of complex diseases and the identification of potential drug targets. In order to provide an integrated perspective on complex disease mechanisms, we developed a novel computational method to simultaneously identify causal genes and dys-regulated pathways. First, we identified a representative set of genes that are differentially expressed in cancer compared to non-tumor control cases. Assuming that disease-associated gene expression changes are caused by genomic alterations, we determined potential paths from such genomic causes to target genes through a network of molecular interactions. Applying our method to sets of genomic alterations and gene expression profiles of 158 Glioblastoma multiforme (GBM) patients we uncovered candidate causal genes and causal paths that are potentially responsible for the altered expression of disease genes. We discovered a set of putative causal genes that potentially play a role in the disease. Combining an expression Quantitative Trait Loci (eQTL) analysis with pathway information, our approach allowed us not only to identify potential causal genes but also to find intermediate nodes and pathways mediating the information flow between causal and target genes. Our results indicate that different genomic perturbations indeed dys-regulate the same functional pathways, supporting a pathway-centric perspective of cancer. While copy number alterations and gene expression data of glioblastoma patients provided opportunities to test our approach, our method can be applied to any disease system where genetic variations play a fundamental causal role.     

New algorithms for multi-class cancer diagnosis using tumor gene expression signatures

MOTIVATION: The increasing use of DNA microarray-based tumor gene expression profiles for cancer diagnosis requires mathematical methods with high accuracy for solving clustering, feature selection and classification problems of gene expression data. RESULTS: New algorithms are developed for solving clustering, feature selection and classification problems of gene expression data. The clustering algorithm is based on optimization techniques and allows the calculation of clusters step-by-step. This approach allows us to find as many clusters as a data set contains with respect to some tolerance. Feature selection is crucial for a gene expression database. Our feature selection algorithm is based on calculating overlaps of different genes. The database used, contains over 16 000 genes and this number is considerably reduced by feature selection. We propose a classification algorithm where each tissue sample is considered as the center of a cluster which is a ball. The results of numerical experiments confirm that the classification algorithm in combination with the feature selection algorithm perform slightly better than the published results for multi-class classifiers based on support vector machines for this data set. AVAILABILITY: Available on request from the authors.     

Knowledge-based gene expression classification via matrix factorization

MOTIVATION: Modern machine learning methods based on matrix decomposition techniques, like independent component analysis (ICA) or non-negative matrix factorization (NMF), provide new and efficient analysis tools which are currently explored to analyze gene expression profiles. These exploratory feature extraction techniques yield expression modes (ICA) or metagenes (NMF). These extracted features are considered indicative of underlying regulatory processes. They can as well be applied to the classification of gene expression datasets by grouping samples into different categories for diagnostic purposes or group genes into functional categories for further investigation of related metabolic pathways and regulatory networks. RESULTS: In this study we focus on unsupervised matrix factorization techniques and apply ICA and sparse NMF to microarray datasets. The latter monitor the gene expression levels of human peripheral blood cells during differentiation from monocytes to macrophages. We show that these tools are able to identify relevant signatures in the deduced component matrices and extract informative sets of marker genes from these gene expression profiles. The methods rely on the joint discriminative power of a set of marker genes rather than on single marker genes. With these sets of marker genes, corroborated by leave-one-out or random forest cross-validation, the datasets could easily be classified into related diagnostic categories. The latter correspond to either monocytes versus macrophages or healthy vs Niemann Pick C disease patients.     

Finding minimum gene subsets with heuristic breadth-first search algorithm for robust tumor classification

ABSTRACT: BACKGROUND: Previous studies on tumor classification based on gene expression profiles suggest that gene selection plays a key role in improving the classification performance. Moreover, finding important tumor-related genes with the highest accuracy is a very important task because these genes might serve as tumor biomarkers, which is of great benefit to not only tumor molecular diagnosis but also drug development. RESULTS: This paper proposes a novel gene selection method with rich biomedical meaning based on Heuristic Breadth-first Search Algorithm (HBSA) to find as many optimal gene subsets as possible. Due to the curse of dimensionality, this type of method could suffer from over-fitting and selection bias problems. To address these potential problems, a HBSA-based ensemble classifier is constructed using majority voting strategy from individual classifiers constructed by the selected gene subsets, and a novel HBSA-based gene ranking method is designed to find important tumor-related genes by measuring the significance of genes using their occurrence frequencies in the selected gene subsets. The experimental results on nine tumor datasets including three pairs of cross-platform datasets indicate that the proposed method can not only obtain better generalization performance but also find many important tumor-related genes. CONCLUSIONS: It is found that the frequencies of the selected genes follow a power-law distribution, indicating that only a few top-ranked genes can be used as potential diagnosis biomarkers. Moreover, the top-ranked genes leading to very high prediction accuracy are closely related to specific tumor subtype and even hub genes. Compared with other related methods, the proposed method can achieve higher prediction accuracy with fewer genes. Moreover, they are further justified by analyzing the top-ranked genes in the context of individual gene function, biological pathway, and protein-protein interaction network.     

The role of phenotype in gene discovery in the whole genome sequencing era

As whole genome sequence becomes a routine component of gene discovery studies in humans, we will have an exhaustive catalog of genetic variation and the challenge becomes understanding the phenotypic consequences of these variants. Statistical genetic methods and analytical approaches that are concerned with optimizing phenotypes for gene discovery for complex traits offer two general categories of advantages. They may increase power to localize genes of interest and also aid in interpreting associations between genetic variants and disease outcomes by suggesting potential mechanisms and pathways through which genes may affect outcomes. Such phenotype optimization approaches include use of allied phenotypes such as symptoms or ages of onset to reduce genetic heterogeneity within a set of cases, study of quantitative risk factors or endophenotypes, joint analyses of related phenotypes, and derivation of new phenotypes designed to extract independent measures underlying the correlations among a set of related phenotypes through approaches such as principal components. New opportunities are also presented by technological advances that permit efficient collection of hundreds or thousands of phenotypes on an individual, including phenotypes more proximal to the level of gene action such as levels of gene expression, microRNAs, or metabolic and proteomic profiles.     

Match/X,A Gene Expression Pattern Recognition Algorithm Used to Identify Genes Which May be Related to CDC2 Function and Cell Cycle Regulation

Large-scale microarray gene expression studies can provide insight into complex genetic networks and biological pathways. A comprehensive gene expression database was constructed using Affymetrix GeneChip microarrays and RNA isolated from more than 6,400 distinct normal and diseased human tissues. These individual patient samples were grouped into over 700 sample sets based on common tissue and disease morphologies, and each set contained averaged expression data for over 45,000 gene probe sets representing more than 33,000 known human genes. Sample sets were compared to each other in more than 750 normal vs. disease pairwise comparisons. Relative up or down-regulation patterns of genes across these pairwise comparisons provided unique expression fingerprints that could be compared and matched to a gene of interest using the Match/X algorithm. This algorithm uses the kappa statistic to compute correlations between genes and calculate a distance score between a gene of interest and all other genes in the database. Using cdc2 as a query gene, we identified several hundred genes that had similar expression patterns and highly correlated distance scores. Most of these genes were known components of the cell cycle involved in G2/M progression, spindle function or chromosome arrangement. Some of the identified genes had unknown biological functions but may be related to cdc2 mediated mechanism based on their closely correlated distance scores. This algorithm may provide novel insights into unknown gene function based on correlation to expression profiles of known genes and can identify elements of cellular pathways and gene interactions in a high throughput fashion.     

Pathways related to mitochondrial dysfunction in cartilage of endemic osteoarthritis patients in China

In this paper, we present the first evidence of differences in the mitochondria-related gene expression profiles of adult articular cartilage derived from patients with Kashin-Beck disease and normal controls. The expression of 705 mitochondria-related genes was analyzed by mitochondria-related gene expression analysis and ingenuity pathways analysis. Mitochondria-related gene expression analysis identified 9 up-regulated genes in Kashin-Beck disease based on their average expression ratio. Three canonical pathways involved in oxidative phosphorylation, apoptosis signaling and pyruvate metabolism were identified, which indicate the involvement of mitochondrial dysfunction in the pathogenesis of Kashin-Beck disease.     

Biological and functional analysis of statistically significant pathways deregulated in colon cancer by using gene expression profiles

Gene expression profiling offers a great opportunity for studying multi-factor diseases and for understanding the key role of genes in mechanisms which drive a normal cell to a cancer state. Single gene analysis is insufficient to describe the complex perturbations responsible for cancer onset, progression and invasion. A deeper understanding of the mechanisms of tumorigenesis can be reached focusing on deregulation of gene sets or pathways rather than on individual genes. We apply two known and statistically well founded methods for finding pathways and biological processes deregulated in pathological conditions by analyzing gene expression profiles. In particular, we measure the amount of deregulation and assess the statistical significance of predefined pathways belonging to a curated collection (Molecular Signature Database) in a colon cancer data set. We find that pathways strongly involved in different tumors are strictly connected with colon cancer. Moreover, our experimental results show that the study of complex diseases through pathway analysis is able to highlight genes weakly connected to the phenotype which may be difficult to detect by using classical univariate statistics. Our study shows the importance of using gene sets rather than single genes for understanding the main biological processes and pathways involved in colorectal cancer. Our analysis evidences that many of the genes involved in these pathways are strongly associated to colorectal tumorigenesis. In this new perspective, the focus shifts from finding differentially expressed genes to identifying biological processes, cellular functions and pathways perturbed in the phenotypic conditions by analyzing genes co-expressed in a given pathway as a whole, taking into account the possible interactions among them and, more importantly, the correlation of their expression with the phenotypical conditions.     

Functional impact of missense variants in BRCA1 predicted by supervised learning

Many individuals tested for inherited cancer susceptibility at the BRCA1 gene locus are discovered to have variants of unknown clinical significance (UCVs). Most UCVs cause a single amino acid residue (missense) change in the BRCA1 protein. They can be biochemically assayed, but such evaluations are time-consuming and labor-intensive. Computational methods that classify and suggest explanations for UCV impact on protein function can complement functional tests. Here we describe a supervised learning approach to classification of BRCA1 UCVs. Using a novel combination of 16 predictive features, the algorithms were applied to retrospectively classify the impact of 36 BRCA1 C-terminal (BRCT) domain UCVs biochemically assayed to measure transactivation function and to blindly classify 54 documented UCVs. Majority vote of three supervised learning algorithms is in agreement with the assay for more than 94% of the UCVs. Two UCVs found deleterious by both the assay and the classifiers reveal a previously uncharacterized putative binding site. Clinicians may soon be able to use computational classifiers such as those described here to better inform patients. These classifiers can be adapted to other cancer susceptibility genes and systematically applied to prioritize the growing number of potential causative loci and variants found by large-scale disease association studies.     

A critical evaluation of network and pathway-based classifiers for outcome prediction in breast cancer

Recently, several classifiers that combine primary tumor data, like gene expression data, and secondary data sources, such as protein-protein interaction networks, have been proposed for predicting outcome in breast cancer. In these approaches, new composite features are typically constructed by aggregating the expression levels of several genes. The secondary data sources are employed to guide this aggregation. Although many studies claim that these approaches improve classification performance over single genes classifiers, the gain in performance is difficult to assess. This stems mainly from the fact that different breast cancer data sets and validation procedures are employed to assess the performance. Here we address these issues by employing a large cohort of six breast cancer data sets as benchmark set and by performing an unbiased evaluation of the classification accuracies of the different approaches. Contrary to previous claims, we find that composite feature classifiers do not outperform simple single genes classifiers. We investigate the effect of (1) the number of selected features; (2) the specific gene set from which features are selected; (3) the size of the training set and (4) the heterogeneity of the data set on the performance of composite feature and single genes classifiers. Strikingly, we find that randomization of secondary data sources, which destroys all biological information in these sources, does not result in a deterioration in performance of composite feature classifiers. Finally, we show that when a proper correction for gene set size is performed, the stability of single genes sets is similar to the stability of composite feature sets. Based on these results there is currently no reason to prefer prognostic classifiers based on composite features over single genes classifiers for predicting outcome in breast cancer.     

Gene selection for classification of cancers using probabilistic model building genetic algorithm

Recently, DNA microarray-based gene expression profiles have been used to correlate the clinical behavior of cancers with the differential gene expression levels in cancerous and normal tissues. To this end, after selection of some predictive genes based on signal-to-noise (S2N) ratio, unsupervised learning like clustering and supervised learning like k-nearest neighbor (k NN) classifier are widely used. Instead of S2N ratio, adaptive searches like Probabilistic Model Building Genetic Algorithm (PMBGA) can be applied for selection of a smaller size gene subset that would classify patient samples more accurately. In this paper, we propose a new PMBGA-based method for identification of informative genes from microarray data. By applying our proposed method to classification of three microarray data sets of binary and multi-type tumors, we demonstrate that the gene subsets selected with our technique yield better classification accuracy.     

Host gene expression signatures discriminate between ferrets infected with genetically similar H1N1 strains

Different respiratory viruses induce virus-specific gene expression in the host. Recent evidence, including those presented here, suggests that genetically related isolates of influenza virus induce strain-specific host gene regulation in several animal models. Here, we identified systemic strain-specific gene expression signatures in ferrets infected with pandemic influenza A/California/07/2009, A/Mexico/4482/2009 or seasonal influenza A/Brisbane/59/2007. Using uncorrelated shrunken centroid classification, we were able to accurately identify the infecting influenza strain with a combined gene expression profile of 10 selected genes, independent of the severity of disease. Another gene signature, consisting of 7 genes, could classify samples based on lung pathology. Furthermore, we identified a gene expression profile consisting of 31 probes that could classify samples based on both strain and severity of disease. Thus, we show that expression-based analysis of non-infected tissue enables distinction between genetically related influenza viruses as well as lung pathology. These results open for development of alternative tools for influenza diagnostics.     

Gene expression profiling in a renal cell carcinoma cell line: dissecting VHL and hypoxia-dependent pathways

The von Hippel-Lindau tumor suppressor, pVHL, is a key player in one of the best characterized hypoxia signaling pathways, the VHL-hypoxia-inducible factor (VHL-HIF) pathway. To better understand the role of VHL in the hypoxia signaling pathways of tumor cells, we used serial analysis of gene expression (SAGE) to investigate hypoxia-regulated gene expression in renal carcinoma cells (786-0), with and without VHL. The gene expression profiles of the cancer cells were compared to SAGE profiles from normal renal proximal tubule cells grown under both normoxia and hypoxia. The data suggest that the role of VHL as a tumor suppressor may be more complex than previously thought. Further, the data reveal that renal carcinoma cells have evolved an alternative hypoxia signaling pathway(s) compared with normal renal cells. These alternative hypoxia pathways demonstrate VHL-dependent and VHL-independent regulation. The genes involved in such pathways include those with potential importance in the physiological and pathological regulation of tumor growth and angiogenesis. Some of the genes identified as showing overexpression in the cancer cells, particularly those encoding secreted or membrane-bound proteins, could be potential biomarkers for tumors or targets for rational therapeutics that are dependent on VHL status.