A scalable method for integration and functional analysis of multiple microarray datasets

BackgroundGene expression profiles based on microarray data have been suggested by many studies as potential molecular prognostic indexes of breast cancer. However, due to the confounding effect of clinical background, independent studies often obtained inconsistent results. The current study investigated the possibility to improve the quality and generality of expression profiles by integrated analysis of multiple datasets. Profiles of recurrence outcome were derived from two independent datasets and validated by a third dataset.ResultsThe clinical background of patients significantly influenced the content and performance of expression profiles when the training samples were unbalanced. The integrated profiling of two independent datasets lead to higher classification accuracy (71.11% vs. 70.59%) and larger ROC curve area (0.789 vs. 0.767) of the testing samples. Cell cycle, especially M phase mitosis, was significantly overrepresented by the 60-gene profile obtained from integrated analysis (p < 0.0001). This profiles significantly differentiated poor and good prognosis in a third patient cohort (p = 0.003). Simulation procedures demonstrated that the change of profile specificity had more instant influence on the performance of expression profiles than the change of profile sensitivity.ConclusionThe current study confirmed that the gene expression profile generated by integrated analysis of multiple datasets achieved better prediction of breast cancer recurrence. However, the content and performance of profiles was confounded by clinical background of training patients. In future studies, prognostic profile applicable to the general population should be derived from more diversified and balanced patient cohorts in larger scale.

Recent high-throughput sequencing advances have expanded the breadth of available omics datasets and the integrated analysis of multiple datasets obtained on the same samples has allowed to gain important insights in a wide range of applications. However, the integration of various sources of information remains a challenge for systems biology since produced datasets are often of heterogeneous types, with the need of developing generic methods to take their different specificities into account. We propose a multiple kernel framework that allows to integrate multiple datasets of various types into a single exploratory analysis. Several solutions are provided to learn either a consensus meta-kernel or a meta-kernel that preserves the original topology of the datasets. We applied our framework to analyse two public multi-omics datasets. First, the multiple metagenomic datasets, collected during the TARA Oceans expedition, was explored to demonstrate that our method is able to retrieve previous findings in a single kernel PCA as well as to provide a new image of the sample structures when a larger number of datasets are included in the analysis. To perform this analysis, a generic procedure is also proposed to improve the interpretability of the kernel PCA in regards with the original data. Second, the multi-omics breast cancer datasets, provided by The Cancer Genome Atlas, is analysed using a kernel Self-Organizing Maps with both single and multi-omics strategies. The comparison of these two approaches demonstrates the benefit of our integration method to improve the representation of the studied biological system. Proposed methods are available in the R package mixKernel, released on CRAN. It is fully compatible with the mixOmics package and a tutorial describing the approach can be found on mixOmics web site http://mixomics.org/mixkernel/. jerome.mariette@inra.fr or nathalie.villa-vialaneix@inra.fr. Supplementary data are available at Bioinformatics online.

Long non-coding RNAs (lncRNAs) are ncRNA transcripts >200 nucleotides that are important genetic regulators. LncRNAs can directly regulate mRNA through a lncRNA-mRNA regulatory mode and can also regulate mRNA through competitive binding to micro (mi)RNA, which is generally known as the competitive endogenous RNA (ceRNA) network. The present study evaluated the functional roles and regulatory networks of lncRNAs in chronic glomerulonephritis (CGN). The proliferative ability of mouse glomerular mesangial cells (GMCs) induced by different concentrations of lipopolysaccharide (LPS) was assessed using the Cell Counting Kit-8 assay, and RNA sequencing (RNA-seq) was performed to identify differentially expressed lncRNAs in LPS-induced GMCs. Based on the sequencing results, six lncRNAs were selected for validation using reverse transcription-quantitative PCR (RT-qPCR). Furthermore, the lncRNA-mRNA regulatory network and the lncRNA-miRNA-mRNA ceRNA network were constructed to assess the role and mechanism of CGN-related lncRNAs. To elucidate the biological functions of lncRNAs, Gene Ontology (GO) biological process term enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed on all mRNAs involved in the lncRNA-mRNA regulatory network and in the ceRNA network. A total of 1,532 differentially expressed lncRNAs, including 594 upregulated lncRNAs and 938 downregulated lncRNAs, were identified using RNA-seq. The results of RT-qPCR validation were consistent with RNA-seq results. An lncRNA-mRNA regulatory network, including 236 lncRNAs and 556 mRNAs, and a ceRNA network, including 6 lncRNAs, 18 miRNAs and 419 mRNAs, were successfully constructed. The GO biological process term enrichment and KEGG pathway enrichment analyses demonstrated that those lncRNAs were often related to inflammatory response and substance metabolism. The present study identified key CGN-related lncRNAs in LPS-induced GMCs, and further demonstrated a global view of the lncRNA-mRNA regulatory network and ceRNA network involved in CGN. These results offered novel insights into the roles of lncRNAs in the pathogenesis of CGN and identified potential diagnostic biomarkers.

The role of MAP infection on the epigenetic modeling of gut immunity during the progression of Johne’s Disease (JD) is still unknown. This study investigated the DNA methylation patterns in ileal (IL) and ileal lymph node (ILLN) tissues from cows diagnosed with persistent subclinical MAP infection over a one to four years period. DNA samples from IL and ILLN tissues from cows negative (MAPneg) (n=3) or positive for MAP infection (MAPinf) (n=4) were subjected to whole genome bisulfite sequencing. A total of 11,263 and 62,459 differentially methylated cytosines (DMCs), and 1259 and 8086 differentially methylated regions (DMRs) (FDR<0.1) were found between MAPinf and MAPneg IL and ILLN tissues, respectively. The DMRs were found on 394 genes (denoted DMR genes) in the IL and on 1305 genes in the ILLN. DMR genes with hypermethylated promoters/5’UTR (43 (IL) and 88 (ILLN)) or hypomethylated promoters/5’UTR (10 (IL) and 25 (ILLN)) and having multiple functions including response to stimulus/immune response (BLK, BTC, CCL21, AVPR1A, CHRNG, GABRA4, TDGF1), cellular processes (H2AC20, TEX101, GLA, NCKAP5L) and metabolic processes (GSTO2, H2AC20, SUV39H1, PPA1, UBE2D2) are potential DNA methylation candidate genes of MAP infection. The ILLN DMR genes were enriched for more biological process (BP) gene ontology (GO) terms (n=374), most of which were related to cellular processes (27.6%), biological regulation (16.6%), metabolic processes (15.4%) and response to stimulus/immune response (8.2%) compared to 75 BP GO terms (related to cellular processes, metabolic processes and transport, and system development) enriched for IL DMR genes. ILLN DMR genes were enriched for more pathways (n=47) including 13 disease pathways compared with 36 enriched pathways, including 7 disease/immune pathways for IL DMR genes. In conclusion, the results show tissue specific responses to MAP infection with more epigenetic changes (DMCs and DMRs) in the ILLN than in the IL tissue, suggesting that the ILLN and immune processes were more responsive to regulation by methylation of DNA relative to IL tissue. Our data is the first to demonstrate a potential role for DNA methylation in the pathogenesis of MAP infection in dairy cattle.

BackgroundThis study aimed to screen the feature modules and characteristic genes related to ulcerative colitis (UC) and construct a support vector machine (SVM) classifier to distinguish UC patients.MethodsFour datasets that contained UC and control samples were obtained from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) with consistency were screened via the MetaDE method. The weighted gene coexpression network (WGCNA) was used to distinguish significant modules based on the four datasets. The protein–protein interaction network was established based on intersection genes. Enrichment analysis of Gene Ontology (GO) biological processes (BPs) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment were established based on DAVID. An SVM combined with recursive feature elimination was also applied to construct a disease classifier for the disease diagnosis of UC patients. The efficacy of the SVM classifier was evaluated through receiver operating characteristic curves.ResultsTwelve highly preserved modules were obtained using the WGCNA, and 2009 DEGs with significant consistency were selected using the MetaDE method. Sixteen significantly related GO BPs and 12 KEGG pathways were obtained, such as cytokine-cytokine receptor interaction, cell adhesion molecules, and leukocyte transendothelial migration. Subsequently, 41 genes were used to construct an SVM classifier, such as CXCL1, CCR2, IL1B, and IL1A. The area under the curve (AUC) was 0.999 in the training dataset, whereas the AUC was 0.886, 0.790, and 0.819 in the validation set (GSE65114, GSE37283, and GSE36807, respectively).ConclusionsAn SVM classifier based on feature genes might correctly identify healthy people or UC patients.

Bioinformatic analysis of neuropeptide related genes in patients diagnosed with invasive breast carcinoma

&lt;p&gt;Johne’s Disease (JD), caused by Mycobacterium avium subsp paratuberculosis (MAP), is an incurable disease of ruminants and other animal species and is characterized by an imbalance of gut immunity. The role of MAP infection on the epigenetic modeling of gut immunity during the progression of JD is still unknown. This study investigated the DNA methylation patterns in ileal (IL) and ileal lymph node (ILLN) tissues from cows diagnosed with persistent subclinical MAP infection over a one to 4 years period. DNA samples from IL and ILLN tissues from cows negative (MAPneg) (n = 3) or positive for MAP infection (MAPinf) (n = 4) were subjected to whole genome bisulfite sequencing. A total of 11,263 and 62,459 differentially methylated cytosines (DMCs), and 1259 and 8086 differentially methylated regions (DMRs) (FDR&lt;0.1) were found between MAPinf and MAPneg IL and ILLN tissues, respectively. The DMRs were found on 394 genes (denoted DMR genes) in the IL and on 1305 genes in the ILLN. DMR genes with hypermethylated promoters/5′UTR [3 (IL) and 88 (ILLN)] or hypomethylated promoters/5′UTR [10 (IL) and 25 (ILLN)] and having multiple functions including response to stimulus/immune response (BLK, BTC, CCL21, AVPR1A, CHRNG, GABRA4, TDGF1), cellular processes (H2AC20, TEX101, GLA, NCKAP5L, RBM27, SLC18A1, H2AC20BARHL2, NLGN3, SUV39H1, GABRA4, PPA1, UBE2D2) and metabolic processes (GSTO2, H2AC20, SUV39H1, PPA1, UBE2D2) are potential DNA methylation candidate genes of MAP infection. The ILLN DMR genes were enriched for more biological process (BP) gene ontology (GO) terms (n = 374), most of which were related to cellular processes (27.6%), biological regulation (16.6%), metabolic processes (15.4%) and response to stimulus/immune response (8.2%) compared to 75 BP GO terms (related to cellular processes, metabolic processes and transport, and system development) enriched for IL DMR genes. ILLN DMR genes were enriched for more pathways (n = 47) including 13 disease pathways compared with 36 enriched pathways, including 7 disease/immune pathways for IL DMR genes. In conclusion, the results show tissue specific responses to MAP infection with more epigenetic changes (DMCs and DMRs) in the ILLN than in the IL tissue, suggesting that the ILLN and immune processes were more responsive to regulation by methylation of DNA relative to IL tissue. Our data is the first to demonstrate a potential role for DNA methylation in the pathogenesis of MAP infection in dairy cattle.&lt;/p&gt;

BackgroundCells dynamically adapt their gene expression patterns in response to various stimuli. This response is orchestrated into a number of gene expression modules consisting of co-regulated genes. A growing pool of publicly available microarray datasets allows the identification of modules by monitoring expression changes over time. These time-series datasets can be searched for gene expression modules by one of the many clustering methods published to date. For an integrative analysis, several time-series datasets can be joined into a three-dimensional gene-condition-time dataset, to which standard clustering or biclustering methods are, however, not applicable. We thus devise a probabilistic clustering algorithm for gene-condition-time datasets.ResultsIn this work, we present the EDISA (Extended Dimension Iterative Signature Algorithm), a novel probabilistic clustering approach for 3D gene-condition-time datasets. Based on mathematical definitions of gene expression modules, the EDISA samples initial modules from the dataset which are then refined by removing genes and conditions until they comply with the module definition. A subsequent extension step ensures gene and condition maximality. We applied the algorithm to a synthetic dataset and were able to successfully recover the implanted modules over a range of background noise intensities. Analysis of microarray datasets has lead us to define three biologically relevant module types: 1) We found modules with independent response profiles to be the most prevalent ones. These modules comprise genes which are co-regulated under several conditions, yet with a different response pattern under each condition. 2) Coherent modules with similar responses under all conditions occurred frequently, too, and were often contained within these modules. 3) A third module type, which covers a response specific to a single condition was also detected, but rarely. All of these modules are essentially different types of biclusters.ConclusionWe successfully applied the EDISA to different 3D datasets. While previous studies were mostly aimed at detecting coherent modules only, our results show that coherent responses are often part of a more general module type with independent response profiles under different conditions. Our approach thus allows for a more comprehensive view of the gene expression response. After subsequent analysis of the resulting modules, the EDISA helped to shed light on the global organization of transcriptional control. An implementation of the algorithm is available at http://www-ra.informatik.uni-tuebingen.de/software/IAGEN/.

Current research in biomedical informatics involves analysis of multiple heterogeneous data sets. This includes patient demographics, clinical and pathology data, treatment history, patient outcomes as well as gene expression, DNA sequences and other information sources such as gene ontology. Analysis of these data sets could lead to better disease diagnosis, prognosis, treatment and drug discovery. In this report, we present a novel machine learning framework for brain tumor classification based on heterogeneous data fusion of metabolic and molecular datasets, including state-of-the-art high-resolution magic angle spinning (HRMAS) proton (1H) magnetic resonance spectroscopy and gene transcriptome profiling, obtained from intact brain tumor biopsies. Our experimental results show that our novel framework outperforms any analysis using individual dataset.

Molecular fingerprinting of the podocyte reveals novel gene and protein regulatory networks

Introduction: Circular RNAs (circRNAs) regulatory network is important in human cancer. We, therefore, mapped the regulatory networks driven by circRNA in luminal-subtype breast cancer. Methods: Breast cancer-related microarray datasets from GEO database were analyzed for the differentially expressed circRNAs, miRNAs, and mRNAs. The potential downstream RNAs were collected using Circular RNA Interactome or Targetscan database. Protein-protein interaction (PPI) analysis was performed for the filtered genes to identify hub genes. The functions were annotated by the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. CircRNA-miRNA-mRNA networks were mapped using Cytoscape software. Hsa_circ_0086735-miR-1296-5p-STAT1 axis was used for verification. The expression levels of hsa_circ_0086735, miR-1296-5p, and STAT1 mRNA were confirmed by qRT-PCR in luminal-subtype tissues and cell lines. The interactions among them were verified by Luciferase reporter assay and RNA pull-down assay. Cell proliferation and apoptosis were assayed. Overall and distant metastasis-free survival was analyzed. Results: A total of 70 genes were finally targeted and enriched in multi-process and multi-pathway. Networks containing 96 circRNA-miRNA-mRNA axes were constructed. Hsa_circ_0086735 and STAT1 mRNA was upregulated in luminal breast cancer, while miR-1296-5p was downregulated. Hsa_circ_0086735-miR-1296-5p-STAT1 axis promotes breast cancer progression and contributes to tamoxifen resistance. High hsa_circ_0086735 was associated with poor overall and distant metastasis-free survival. Discussion: This study identified the hsa_circ_0086735-miR-1296-5p-STAT1 as an important regulatory axis in luminal-subtype breast cancer, aiding to determine potential therapeutic targets.

To date, a range of ion channels have been identified in chondrocytes using a number of different techniques, predominantly electrophysiological and/or biomolecular; each of these has its advantages and disadvantages. Here we aim to compare and contrast the data available from biophysical and microarray experiments. This letter analyses recent transcriptomics datasets from chondrocytes, accessible from the European Bioinformatics Institute (EBI). We discuss whether such bioinformatic analysis of microarray datasets can potentially accelerate identification and discovery of ion channels in chondrocytes. The ion channels which appear most frequently across these microarray datasets are discussed, along with their possible functions. We discuss whether functional or protein data exist which support the microarray data. A microarray experiment comparing gene expression in osteoarthritis and healthy cartilage is also discussed and we verify the differential expression of 2 of these genes, namely the genes encoding large calcium-activated potassium (BK) and aquaporin channels.

Following nicotine, cannabis is the most used drug during pregnancy. In utero cannabis exposure is linked to adverse outcomes, including fetal growth restriction. Cannabidiol (CBD), the non-hallucinogenic constituent of cannabis, is perceived as safe for use during pregnancy by ~25% of women. While studies assessing the effects of Δ-9-THC, the other principal cannabinoid in cannabis, have been initiated, research examining the impact of CBD on pregnancy is scarce. The placenta facilitates the exchange of gas, nutrients and waste between the fetus and mother. Our group has recently identified altered placental morphology and reduced fetal growth after in utero CBD exposure in rats. However, the influence of CBD on genes, transcription factors and biological processes in the placenta has not been assessed. This study aimed to identify biological pathways in the placenta that were significantly altered by CBD-exposure and correlate the findings with previously performed histological analysis. Pregnant rats were administered a daily dose of 3 mg/kg CBD or a vehicle control from embryonic day (E) 6.5-18.5. On E19.5, placentae were harvested for transcriptomic analysis. RNA extraction, library construction and bulk RNA sequencing was performed by Genome Quebec. The top differentially expressed genes were analysed using Metascape software, revealing the most up- and down-regulated Gene Ontology (GO) biological process pathways. Down-regulated GO biological processes included tube morphogenesis, angiogenesis, and blood vessel morphogenesis, which supported the angiogenic complications identified by histology. Further, the most significantly up-regulated GO biological processes were metabolic, including peptide, glycoprotein, glycosaminoglycan, and tetrahydrofolate. This is significant, as the histological analysis identified elevated Glut3 expression in the placentae from CBD-exposed pregnancies. Elevated Glut3 in fetal growth restriction is associated with changes to metabolic pathways. Consequently, additional studies will include analysis of CBD-exposed rat trophoblast cells to establish whether metabolic changes may underlie the altered morphology identified in CBD-exposed placentae.

Microarray technology aims to discover the relationship between genes and cancers. But the analysis of multiclass microarray datasets is a difficult problem in considering the small sample size along with the class imbalance problem. In this paper, we propose a Genetic Programing (GP) based Error Correcting Output Codes (ECOC) algorithm to tackle this problem. In our GP framework, each individual represents a codematrix, and a legality checking mechanism is applied to avoid the production of illegal codematrices. So the algorithm evolves towards optimum ECOC codematrices. In experiments, our algorithm is compared with other methods based on four famous microarray datasets. Experimental results prove that our algorithm can achieve better results in most cases.

Several approaches have been proposed for the analysis of DNA microarray datasets, focusing on the performance and robustness of the final feature subsets. The novelty of this paper arises in the use of q-values to pre-filter the features of a DNA microarray dataset identifying the most significant ones and including this information into a genetic algorithm for further feature selection. This method is applied to a lung cancer microarray dataset resulting in similar performance rates and greater robustness in terms of selected features (on average a 36.21% of robustness improvement) when compared to results of the standard algorithm.