A Machine Learned Classifier That Uses Gene Expression Data to Accurately Predict Estrogen Receptor Status

Background

Selecting the appropriate treatment for breast cancer requires accurately determining the estrogen receptor (ER) status of the tumor. However, the standard for determining this status, immunohistochemical analysis of formalin-fixed paraffin embedded samples, suffers from numerous technical and reproducibility issues. Assessment of ER-status based on RNA expression can provide more objective, quantitative and reproducible test results.

Methods

To learn a parsimonious RNA-based classifier of hormone receptor status, we applied a machine learning tool to a training dataset of gene expression microarray data obtained from 176 frozen breast tumors, whose ER-status was determined by applying ASCO-CAP guidelines to standardized immunohistochemical testing of formalin fixed tumor.

Results

This produced a three-gene classifier that can predict the ER-status of a novel tumor, with a cross-validation accuracy of 93.17±2.44%. When applied to an independent validation set and to four other public databases, some on different platforms, this classifier obtained over 90% accuracy in each. In addition, we found that this prediction rule separated the patients'' recurrence-free survival curves with a hazard ratio lower than the one based on the IHC analysis of ER-status.

Conclusions

Our efficient and parsimonious classifier lends itself to high throughput, highly accurate and low-cost RNA-based assessments of ER-status, suitable for routine high-throughput clinical use. This analytic method provides a proof-of-principle that may be applicable to developing effective RNA-based tests for other biomarkers and conditions.     

Optimization based tumor classification from microarray gene expression data

Background

An important use of data obtained from microarray measurements is the classification of tumor types with respect to genes that are either up or down regulated in specific cancer types. A number of algorithms have been proposed to obtain such classifications. These algorithms usually require parameter optimization to obtain accurate results depending on the type of data. Additionally, it is highly critical to find an optimal set of markers among those up or down regulated genes that can be clinically utilized to build assays for the diagnosis or to follow progression of specific cancer types. In this paper, we employ a mixed integer programming based classification algorithm named hyper-box enclosure method (HBE) for the classification of some cancer types with a minimal set of predictor genes. This optimization based method which is a user friendly and efficient classifier may allow the clinicians to diagnose and follow progression of certain cancer types.

Methodology/Principal Findings

We apply HBE algorithm to some well known data sets such as leukemia, prostate cancer, diffuse large B-cell lymphoma (DLBCL), small round blue cell tumors (SRBCT) to find some predictor genes that can be utilized for diagnosis and prognosis in a robust manner with a high accuracy. Our approach does not require any modification or parameter optimization for each data set. Additionally, information gain attribute evaluator, relief attribute evaluator and correlation-based feature selection methods are employed for the gene selection. The results are compared with those from other studies and biological roles of selected genes in corresponding cancer type are described.

Conclusions/Significance

The performance of our algorithm overall was better than the other algorithms reported in the literature and classifiers found in WEKA data-mining package. Since it does not require a parameter optimization and it performs consistently very high prediction rate on different type of data sets, HBE method is an effective and consistent tool for cancer type prediction with a small number of gene markers.     

Discovery and Validation of a Prostate Cancer Genomic Classifier that Predicts Early Metastasis Following Radical Prostatectomy

Purpose

Clinicopathologic features and biochemical recurrence are sensitive, but not specific, predictors of metastatic disease and lethal prostate cancer. We hypothesize that a genomic expression signature detected in the primary tumor represents true biological potential of aggressive disease and provides improved prediction of early prostate cancer metastasis.

Methods

A nested case-control design was used to select 639 patients from the Mayo Clinic tumor registry who underwent radical prostatectomy between 1987 and 2001. A genomic classifier (GC) was developed by modeling differential RNA expression using 1.4 million feature high-density expression arrays of men enriched for rising PSA after prostatectomy, including 213 who experienced early clinical metastasis after biochemical recurrence. A training set was used to develop a random forest classifier of 22 markers to predict for cases - men with early clinical metastasis after rising PSA. Performance of GC was compared to prognostic factors such as Gleason score and previous gene expression signatures in a withheld validation set.

Results

Expression profiles were generated from 545 unique patient samples, with median follow-up of 16.9 years. GC achieved an area under the receiver operating characteristic curve of 0.75 (0.67–0.83) in validation, outperforming clinical variables and gene signatures. GC was the only significant prognostic factor in multivariable analyses. Within Gleason score groups, cases with high GC scores experienced earlier death from prostate cancer and reduced overall survival. The markers in the classifier were found to be associated with a number of key biological processes in prostate cancer metastatic disease progression.

Conclusion

A genomic classifier was developed and validated in a large patient cohort enriched with prostate cancer metastasis patients and a rising PSA that went on to experience metastatic disease. This early metastasis prediction model based on genomic expression in the primary tumor may be useful for identification of aggressive prostate cancer.     

Clinical Value of Prognosis Gene Expression Signatures in Colorectal Cancer: A Systematic Review

Introduction

The traditional staging system is inadequate to identify those patients with stage II colorectal cancer (CRC) at high risk of recurrence or with stage III CRC at low risk. A number of gene expression signatures to predict CRC prognosis have been proposed, but none is routinely used in the clinic. The aim of this work was to assess the prediction ability and potential clinical usefulness of these signatures in a series of independent datasets.

Methods

A literature review identified 31 gene expression signatures that used gene expression data to predict prognosis in CRC tissue. The search was based on the PubMed database and was restricted to papers published from January 2004 to December 2011. Eleven CRC gene expression datasets with outcome information were identified and downloaded from public repositories. Random Forest classifier was used to build predictors from the gene lists. Matthews correlation coefficient was chosen as a measure of classification accuracy and its associated p-value was used to assess association with prognosis. For clinical usefulness evaluation, positive and negative post-tests probabilities were computed in stage II and III samples.

Results

Five gene signatures showed significant association with prognosis and provided reasonable prediction accuracy in their own training datasets. Nevertheless, all signatures showed low reproducibility in independent data. Stratified analyses by stage or microsatellite instability status showed significant association but limited discrimination ability, especially in stage II tumors. From a clinical perspective, the most predictive signatures showed a minor but significant improvement over the classical staging system.

Conclusions

The published signatures show low prediction accuracy but moderate clinical usefulness. Although gene expression data may inform prognosis, better strategies for signature validation are needed to encourage their widespread use in the clinic.     

Visualization-Aided Classification Ensembles Discriminate Lung Adenocarcinoma and Squamous Cell Carcinoma Samples Using Their Gene Expression Profiles

Introduction

The widespread application of microarray experiments to cancer research is astounding including lung cancer, one of the most common fatal human tumors. Among non-small cell lung carcinoma (NSCLC), there are two major histological types of NSCLC, adenocarcinoma (AC) and squamous cell carcinoma (SCC).

Results

In this paper, we proposed to integrate a visualization method called Radial Coordinate Visualization (Radviz) with a suitable classifier, aiming at discriminating two NSCLC subtypes using patients'' gene expression profiles. Our analyses on simulated data and a real microarray dataset show that combining with a classification method, Radviz may play a role in selecting relevant features and ameliorating parsimony, while the final model suffers no or least loss of accuracy. Most importantly, a graphic representation is more easily understandable and implementable for a clinician than statistical methods and/or mathematic equations.

Conclusion

To conclude, using the NSCLC microarray data presented here as a benchmark, the comprehensive understanding of the underlying mechanism associated with NSCLC and of the mechanisms with its subtypes and respective stages will become reality in the near future.     

Microarray Meta-Analysis Identifies Acute Lung Injury Biomarkers in Donor Lungs That Predict Development of Primary Graft Failure in Recipients

Objectives

To perform a meta-analysis of gene expression microarray data from animal studies of lung injury, and to identify an injury-specific gene expression signature capable of predicting the development of lung injury in humans.

Methods

We performed a microarray meta-analysis using 77 microarray chips across six platforms, two species and different animal lung injury models exposed to lung injury with or/and without mechanical ventilation. Individual gene chips were classified and grouped based on the strategy used to induce lung injury. Effect size (change in gene expression) was calculated between non-injurious and injurious conditions comparing two main strategies to pool chips: (1) one-hit and (2) two-hit lung injury models. A random effects model was used to integrate individual effect sizes calculated from each experiment. Classification models were built using the gene expression signatures generated by the meta-analysis to predict the development of lung injury in human lung transplant recipients.

Results

Two injury-specific lists of differentially expressed genes generated from our meta-analysis of lung injury models were validated using external data sets and prospective data from animal models of ventilator-induced lung injury (VILI). Pathway analysis of gene sets revealed that both new and previously implicated VILI-related pathways are enriched with differentially regulated genes. Classification model based on gene expression signatures identified in animal models of lung injury predicted development of primary graft failure (PGF) in lung transplant recipients with larger than 80% accuracy based upon injury profiles from transplant donors. We also found that better classifier performance can be achieved by using meta-analysis to identify differentially-expressed genes than using single study-based differential analysis.

Conclusion

Taken together, our data suggests that microarray analysis of gene expression data allows for the detection of “injury" gene predictors that can classify lung injury samples and identify patients at risk for clinically relevant lung injury complications.     

Transcriptomic Meta-Analysis of Multiple Sclerosis and Its Experimental Models

Background

Multiple microarray analyses of multiple sclerosis (MS) and its experimental models have been published in the last years.

Objective

Meta-analyses integrate the information from multiple studies and are suggested to be a powerful approach in detecting highly relevant and commonly affected pathways.

Data sources

ArrayExpress, Gene Expression Omnibus and PubMed databases were screened for microarray gene expression profiling studies of MS and its experimental animal models.

Study eligibility criteria

Studies comparing central nervous system (CNS) samples of diseased versus healthy individuals with n >1 per group and publically available raw data were selected.

Material and Methods

Included conditions for re-analysis of differentially expressed genes (DEGs) were MS, myelin oligodendrocyte glycoprotein-induced experimental autoimmune encephalomyelitis (EAE) in rats, proteolipid protein-induced EAE in mice, Theiler’s murine encephalomyelitis virus-induced demyelinating disease (TMEV-IDD), and a transgenic tumor necrosis factor-overexpressing mouse model (TNFtg). Since solely a single MS raw data set fulfilled the inclusion criteria, a merged list containing the DEGs from two MS-studies was additionally included. Cross-study analysis was performed employing list comparisons of DEGs and alternatively Gene Set Enrichment Analysis (GSEA).

Results

The intersection of DEGs in MS, EAE, TMEV-IDD, and TNFtg contained 12 genes related to macrophage functions. The intersection of EAE, TMEV-IDD and TNFtg comprised 40 DEGs, functionally related to positive regulation of immune response. Over and above, GSEA identified substantially more differentially regulated pathways including coagulation and JAK/STAT-signaling.

Conclusion

A meta-analysis based on a simple comparison of DEGs is over-conservative. In contrast, the more experimental GSEA approach identified both, a priori anticipated as well as promising new candidate pathways.     

Can Survival Prediction Be Improved By Merging Gene Expression Data Sets?

Background

High-throughput gene expression profiling technologies generating a wealth of data, are increasingly used for characterization of tumor biopsies for clinical trials. By applying machine learning algorithms to such clinically documented data sets, one hopes to improve tumor diagnosis, prognosis, as well as prediction of treatment response. However, the limited number of patients enrolled in a single trial study limits the power of machine learning approaches due to over-fitting. One could partially overcome this limitation by merging data from different studies. Nevertheless, such data sets differ from each other with regard to technical biases, patient selection criteria and follow-up treatment. It is therefore not clear at all whether the advantage of increased sample size outweighs the disadvantage of higher heterogeneity of merged data sets. Here, we present a systematic study to answer this question specifically for breast cancer data sets. We use survival prediction based on Cox regression as an assay to measure the added value of merged data sets.

Results

Using time-dependent Receiver Operating Characteristic-Area Under the Curve (ROC-AUC) and hazard ratio as performance measures, we see in overall no significant improvement or deterioration of survival prediction with merged data sets as compared to individual data sets. This apparently was due to the fact that a few genes with strong prognostic power were not available on all microarray platforms and thus were not retained in the merged data sets. Surprisingly, we found that the overall best performance was achieved with a single-gene predictor consisting of CYB5D1.

Conclusions

Merging did not deteriorate performance on average despite (a) The diversity of microarray platforms used. (b) The heterogeneity of patients cohorts. (c) The heterogeneity of breast cancer disease. (d) Substantial variation of time to death or relapse. (e) The reduced number of genes in the merged data sets. Predictors derived from the merged data sets were more robust, consistent and reproducible across microarray platforms. Moreover, merging data sets from different studies helps to better understand the biases of individual studies and can lead to the identification of strong survival factors like CYB5D1 expression.     

Clinical potential of DNA methylation in gastric cancer: a meta-analysis

Background

Accumulating evidence indicates aberrant DNA methylation is involved in gastric tumourigenesis, suggesting it may be a useful clinical biomarker for the disease. The aim of this study was to consolidate and summarize published data on the potential of methylation in gastric cancer (GC) risk prediction, prognostication and prediction of treatment response.

Methods

Relevant studies were identified from PubMed using a systematic search approach. Results were summarized by meta-analysis. Mantel-Haenszel odds ratios were computed for each methylation event assuming the random-effects model.

Results

A review of 589 retrieved publications identified 415 relevant articles, including 143 case-control studies on gene methylation of 142 individual genes in GC clinical samples. A total of 77 genes were significantly differentially methylated between tumour and normal gastric tissue from GC subjects, of which data on 62 was derived from single studies. Methylation of 15, 4 and 7 genes in normal gastric tissue, plasma and serum respectively was significantly different in frequency between GC and non-cancer subjects. A prognostic significance was reported for 18 genes and predictive significance was reported for p16 methylation, although many inconsistent findings were also observed. No bias due to assay, use of fixed tissue or CpG sites analysed was detected, however a slight bias towards publication of positive findings was observed.

Conclusions

DNA methylation is a promising biomarker for GC risk prediction and prognostication. Further focused validation of candidate methylation markers in independent cohorts is required to develop its clinical potential.     

A systems biology-based classifier for hepatocellular carcinoma diagnosis

Aim

The diagnosis of hepatocellular carcinoma (HCC) in the early stage is crucial to the application of curative treatments which are the only hope for increasing the life expectancy of patients. Recently, several large-scale studies have shed light on this problem through analysis of gene expression profiles to identify markers correlated with HCC progression. However, those marker sets shared few genes in common and were poorly validated using independent data. Therefore, we developed a systems biology based classifier by combining the differential gene expression with topological features of human protein interaction networks to enhance the ability of HCC diagnosis.

Methods and Results

In the Oncomine platform, genes differentially expressed in HCC tissues relative to their corresponding normal tissues were filtered by a corrected Q value cut-off and Concept filters. The identified genes that are common to different microarray datasets were chosen as the candidate markers. Then, their networks were analyzed by GeneGO Meta-Core software and the hub genes were chosen. After that, an HCC diagnostic classifier was constructed by Partial Least Squares modeling based on the microarray gene expression data of the hub genes. Validations of diagnostic performance showed that this classifier had high predictive accuracy (85.88∼92.71%) and area under ROC curve (approximating 1.0), and that the network topological features integrated into this classifier contribute greatly to improving the predictive performance. Furthermore, it has been demonstrated that this modeling strategy is not only applicable to HCC, but also to other cancers.

Conclusion

Our analysis suggests that the systems biology-based classifier that combines the differential gene expression and topological features of human protein interaction network may enhance the diagnostic performance of HCC classifier.     

Using Highly Detailed Administrative Data to Predict Pneumonia Mortality

Background

Mortality prediction models generally require clinical data or are derived from information coded at discharge, limiting adjustment for presenting severity of illness in observational studies using administrative data.

Objectives

To develop and validate a mortality prediction model using administrative data available in the first 2 hospital days.

Research Design

After dividing the dataset into derivation and validation sets, we created a hierarchical generalized linear mortality model that included patient demographics, comorbidities, medications, therapies, and diagnostic tests administered in the first 2 hospital days. We then applied the model to the validation set.

Subjects

Patients aged ≥18 years admitted with pneumonia between July 2007 and June 2010 to 347 hospitals in Premier, Inc.’s Perspective database.

Measures

In hospital mortality.

Results

The derivation cohort included 200,870 patients and the validation cohort had 50,037. Mortality was 7.2%. In the multivariable model, 3 demographic factors, 25 comorbidities, 41 medications, 7 diagnostic tests, and 9 treatments were associated with mortality. Factors that were most strongly associated with mortality included receipt of vasopressors, non-invasive ventilation, and bicarbonate. The model had a c-statistic of 0.85 in both cohorts. In the validation cohort, deciles of predicted risk ranged from 0.3% to 34.3% with observed risk over the same deciles from 0.1% to 33.7%.

Conclusions

A mortality model based on detailed administrative data available in the first 2 hospital days had good discrimination and calibration. The model compares favorably to clinically based prediction models and may be useful in observational studies when clinical data are not available.     

Circulating microRNAs as a Fingerprint for Liver Cirrhosis

Background

Sensitive and specific detection of liver cirrhosis is an urgent need for optimal individualized management of disease activity. Substantial studies have identified circulation miRNAs as biomarkers for diverse diseases including chronic liver diseases. In this study, we investigated the plasma miRNA signature to serve as a potential diagnostic biomarker for silent liver cirrhosis.

Methods

A genome-wide miRNA microarray was first performed in 80 plasma specimens. Six candidate miRNAs were selected and then trained in CHB-related cirrhosis and controls by qPCR. A classifier, miR-106b and miR-181b, was validated finally in two independent cohorts including CHB-related silent cirrhosis and controls, as well as non−CHB-related cirrhosis and controls as validation sets, respectively.

Results

A profile of 2 miRNAs (miR-106b and miR-181b) was identified as liver cirrhosis biomarkers irrespective of etiology. The classifier constructed by the two miRNAs provided a high diagnostic accuracy for cirrhosis (AUC = 0.882 for CHB-related cirrhosis in the training set, 0.774 for CHB-related silent cirrhosis in one validation set, and 0.915 for non−CHB-related cirrhosis in another validation set).

Conclusion

Our study demonstrated that the combined detection of miR-106b and miR-181b has a considerable clinical value to diagnose patients with liver cirrhosis, especially those at early stage.     

A Composite Model for Subgroup Identification and Prediction via Bicluster Analysis

Background

A major challenges in the analysis of large and complex biomedical data is to develop an approach for 1) identifying distinct subgroups in the sampled populations, 2) characterizing their relationships among subgroups, and 3) developing a prediction model to classify subgroup memberships of new samples by finding a set of predictors. Each subgroup can represent different pathogen serotypes of microorganisms, different tumor subtypes in cancer patients, or different genetic makeups of patients related to treatment response.

Methods

This paper proposes a composite model for subgroup identification and prediction using biclusters. A biclustering technique is first used to identify a set of biclusters from the sampled data. For each bicluster, a subgroup-specific binary classifier is built to determine if a particular sample is either inside or outside the bicluster. A composite model, which consists of all binary classifiers, is constructed to classify samples into several disjoint subgroups. The proposed composite model neither depends on any specific biclustering algorithm or patterns of biclusters, nor on any classification algorithms.

Results

The composite model was shown to have an overall accuracy of 97.4% for a synthetic dataset consisting of four subgroups. The model was applied to two datasets where the sample’s subgroup memberships were known. The procedure showed 83.7% accuracy in discriminating lung cancer adenocarcinoma and squamous carcinoma subtypes, and was able to identify 5 serotypes and several subtypes with about 94% accuracy in a pathogen dataset.

Conclusion

The composite model presents a novel approach to developing a biclustering-based classification model from unlabeled sampled data. The proposed approach combines unsupervised biclustering and supervised classification techniques to classify samples into disjoint subgroups based on their associated attributes, such as genotypic factors, phenotypic outcomes, efficacy/safety measures, or responses to treatments. The procedure is useful for identification of unknown species or new biomarkers for targeted therapy.     

A meta-learning approach for B-cell conformational epitope prediction

Background

One of the major challenges in the field of vaccine design is identifying B-cell epitopes in continuously evolving viruses. Various tools have been developed to predict linear or conformational epitopes, each relying on different physicochemical properties and adopting distinct search strategies. We propose a meta-learning approach for epitope prediction based on stacked and cascade generalizations. Through meta learning, we expect a meta learner to be able integrate multiple prediction models, and outperform the single best-performing model. The objective of this study is twofold: (1) to analyze the complementary predictive strengths in different prediction tools, and (2) to introduce a generic computational model to exploit the synergy among various prediction tools. Our primary goal is not to develop any particular classifier for B-cell epitope prediction, but to advocate the feasibility of meta learning to epitope prediction. With the flexibility of meta learning, the researcher can construct various meta classification hierarchies that are applicable to epitope prediction in different protein domains.

Results

We developed the hierarchical meta-learning architectures based on stacked and cascade generalizations. The bottom level of the hierarchy consisted of four conformational and four linear epitope prediction tools that served as the base learners. To perform consistent and unbiased comparisons, we tested the meta-learning method on an independent set of antigen proteins that were not used previously to train the base epitope prediction tools. In addition, we conducted correlation and ablation studies of the base learners in the meta-learning model. Low correlation among the predictions of the base learners suggested that the eight base learners had complementary predictive capabilities. The ablation analysis indicated that the eight base learners differentially interacted and contributed to the final meta model. The results of the independent test demonstrated that the meta-learning approach markedly outperformed the single best-performing epitope predictor.

Conclusions

Computational B-cell epitope prediction tools exhibit several differences that affect their performances when predicting epitopic regions in protein antigens. The proposed meta-learning approach for epitope prediction combines multiple prediction tools by integrating their complementary predictive strengths. Our experimental results demonstrate the superior performance of the combined approach in comparison with single epitope predictors.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-014-0378-y) contains supplementary material, which is available to authorized users.     

IPred - integrating ab initio and evidence based gene predictions to improve prediction accuracy

Background

Gene prediction is a challenging but crucial part in most genome analysis pipelines. Various methods have evolved that predict genes ab initio on reference sequences or evidence based with the help of additional information, such as RNA-Seq reads or EST libraries. However, none of these strategies is bias-free and one method alone does not necessarily provide a complete set of accurate predictions.

Results

We present IPred (Integrative gene Prediction), a method to integrate ab initio and evidence based gene identifications to complement the advantages of different prediction strategies. IPred builds on the output of gene finders and generates a new combined set of gene identifications, representing the integrated evidence of the single method predictions.

Conclusion

We evaluate IPred in simulations and real data experiments on Escherichia Coli and human data. We show that IPred improves the prediction accuracy in comparison to single method predictions and to existing methods for prediction combination.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1315-9) contains supplementary material, which is available to authorized users.     

Plant microRNA-Target Interaction Identification Model Based on the Integration of Prediction Tools and Support Vector Machine

Background

Confident identification of microRNA-target interactions is significant for studying the function of microRNA (miRNA). Although some computational miRNA target prediction methods have been proposed for plants, results of various methods tend to be inconsistent and usually lead to more false positive. To address these issues, we developed an integrated model for identifying plant miRNA–target interactions.

Results

Three online miRNA target prediction toolkits and machine learning algorithms were integrated to identify and analyze Arabidopsis thaliana miRNA-target interactions. Principle component analysis (PCA) feature extraction and self-training technology were introduced to improve the performance. Results showed that the proposed model outperformed the previously existing methods. The results were validated by using degradome sequencing supported Arabidopsis thaliana miRNA-target interactions. The proposed model constructed on Arabidopsis thaliana was run over Oryza sativa and Vitis vinifera to demonstrate that our model is effective for other plant species.

Conclusions

The integrated model of online predictors and local PCA-SVM classifier gained credible and high quality miRNA-target interactions. The supervised learning algorithm of PCA-SVM classifier was employed in plant miRNA target identification for the first time. Its performance can be substantially improved if more experimentally proved training samples are provided.