Clustering Of Microarray Data Research Articles

Comparative Study of three Clustering Algorithms for Microarray Data

High throughput genomic data analysis is becoming an increasingly integral part of biomedical research. The information derived from gene expression analysis helps in diagnosing the treatment modality given to the patient. However, the amount of data is humongous and becomes complex to examine manually. Unsupervised machine learning algorithms perform complex tasks on an unlabelled data by clustering to comprehend the underlying structure and behaviour of the pattern. Clustering microarray data, examines the differential expressed genes found by grouping the genes based on the similarity of the expression values. In this study, we propose to elucidate the best clustering algorithm for gene expression data on various clinical conditions. The proposed study was carried on three gene expression datasets of Severe acute respiratory syndrome, Amyotrophic lateral sclerosis and Parkinson’s disease. Differentially expressed genes were found at three p-values 0.01, 0.05, 0.001 and the most significant number of genes were retrieved at p-value 0.05. We experimented the differential expressed genes on three clustering algorithms, namely Hierarchical clustering, k-means clustering and fuzzy clustering of the three diseases. The performance of the three clustering algorithms was evaluated using the internal validity index, wherein Hierarchical clustering was found to be best for gene expression data.

ILRC: a hybrid biomarker discovery algorithm based on improved L1 regularization and clustering in microarray data

BackgroundFinding significant genes or proteins from gene chip data for disease diagnosis and drug development is an important task. However, the challenge comes from the curse of the data dimension. It is of great significance to use machine learning methods to find important features from the data and build an accurate classification model.ResultsThe proposed method has proved superior to the published advanced hybrid feature selection method and traditional feature selection method on different public microarray data sets. In addition, the biomarkers selected using our method show a match to those provided by the cooperative hospital in a set of clinical cleft lip and palate data.MethodIn this paper, a feature selection algorithm ILRC based on clustering and improved L1 regularization is proposed. The features are firstly clustered, and the redundant features in the sub-clusters are deleted. Then all the remaining features are iteratively evaluated using ILR. The final result is given according to the cumulative weight reordering.ConclusionThe proposed method can effectively remove redundant features. The algorithm’s output has high stability and classification accuracy, which can potentially select potential biomarkers.

TriRNSC: triclustering of gene expression microarray data using restricted neighbourhood search.

Computational analysis of microarray data is crucial for understanding the gene behaviours and deriving meaningful results. Clustering and biclustering of gene expression microarray data in the unsupervised domain are extremely important as their outcomes directly dominate healthcare research in many aspects. However, these approaches fail when the time factor is added as the third dimension to the microarray datasets. This three-dimensional data set can be analysed using triclustering that discovers similar gene sets that pursue identical behaviour under a subset of conditions at a specific time point. A novel triclustering algorithm (TriRNSC) is proposed in this manuscript to discover meaningful triclusters in gene expression profiles. TriRNSC is based on restricted neighbourhood search clustering (RNSC), a popular graph-based clustering approach considering the genes, the experimental conditions and the time points at an instance. The performance of the proposed algorithm is evaluated in terms of volume and some performance measures. Gene Ontology and KEGG pathway analysis are used to validate the TriRNSC results biologically. The efficiency of TriRNSC indicates its capability and reliability and also demonstrates its usability over other state-of-art schemes. The proposed framework initiates the application of the RNSC algorithm in the triclustering of gene expression profiles.

Genome-wide analysis of maize MBD gene family and expression profiling under abiotic stress treatment at the seedling stage

Methyl-CpG-binding domain (MBD) proteins in plants are important trans-acting factors specifically recognizing methylated DNA. The MBD proteins can compact chromatins to repress transcription by recruiting chromatin-modifying complexes that contain histone deacetylase activities and chromatin remodeling factors, and play a crucial biological role in the growth and development in plants. Currently, very little is known regarding the structure and function of MBD genes in plants. In this study, we performed a genome-wide identification and expression profile analysis of maize MBD genes (ZmMBDs) from the latest version of the maize (B73) genome. By analyzing phylogenetic relationship of MBD gene families from Arabidopsis, rice, wheat, and maize, all 14 MBD proteins in maize were categorized into four subclasses. Furthermore, chromosome location and schematic structure revealed an unevenly distribution on chromosomes and structure features of MBD genes in maize, respectively. Eventually, EST expression data mining, microarray data clustering analysis, and semi-quantitative and quantitative expression profile analyses detected in seedling leaves and stems by heat, drought, and salt-stress treatments have demonstrated that these genes had temporal and spatial expression pattern and exhibited different expression levels under heat, drought, and salt-stress conditions, suggesting that functional diversification of the MBD gene family in maize. In addition, through electrophoretic mobility shift assay (EMSA) a representative MBD protein, ZmMBD11, exhibited in vitro DNA-binding activity, indicating that that the MBD proteins in maize might play a role in reading cytosine methylation. Taken together, these results would provide an important theoretical basis for future functional verification of ZmMBD genes and also facilitate future experimental research to further unravel the mechanisms of epigenetic regulation in plants.

K-means Clustering and Principal Components Analysis of Microarray Data of L1000 Landmark Genes

Abstract Dimensionality reduction methods such as principal component analysis (PCA) are used to select relevant features, and k-means clustering performs well when applied to data with low effective dimensionality. This study integrated PCA and k-means clustering using the L1000 dataset, containing gene microarray data from 978 landmark genes, which have been previously shown to predict expression of ~81% of the remaining 21,290 target genes with low error. Groups within the L1000 dataset were characterized using both microarray and clinical metadata to assess whether 978 landmark genes would improve clustering results, compared to a random set of 978 genes. The role of clinical variables, including morphological diagnosis, were assessed across k-means clustering groups within homogeneous tissue samples in the L1000 dataset. Results show that the 978 landmark genes better differentiated k-means clusters, relative to 978 randomly selected non-landmark genes. K-means clusters generated from the landmark genes showed more separation of cluster groups when plotted against the first two principal components, which capture a greater proportion of variation for the 978 landmark genes. These results suggest that the 978 landmark genes better represent the overall genetic profile of these heterogeneous samples. Future studies will implement predictive analytics techniques to further investigate the interaction of microarray data and clinical variables such as cancer stage.

Weighted Cluster Ensemble Based on Partition Relevance Analysis With Reduction Step

Over the last decade, the advent of the cluster ensemble framework has enabled more accurate and robust data analysis than traditional single clustering algorithms. The improved clustering of microarray data has had a particularly strong impact in the fields of genomics and medicine. However, when we bring several ensemble members together to form a consensus, low-quality data partitions can seriously compromise the final solution. One way to overcome this problem is the weighted cluster ensemble approach based on Partition Relevance Analysis (PRA), which uses internal cluster validity indices to evaluate and weight the ensemble members before the fusion. Unfortunately, the selection of appropriate validation indices for given data is far from trivial. In this paper, we propose an additional step in PRA that reduces the size of the committee of cluster validation indices. It does so by eliminating redundant and noisy indices using data dimensionality reduction methods. Our extension works in an unsupervised way, minimizing the amount of user intervention and required expert knowledge. We adapted three conventional consensus functions based on the principle of evidence accumulation to work with PRA weights. We demonstrate the advantages of the proposed reduction step of PRA based on extensive experiments with 25 gene expression and 15 non-genetic real-world datasets, where we compared 15 consensus functions. The source code is available at https://github.com/nejci/PRAr.

On the Role of Clustering and Visualization Techniques in Gene Microarray Data

As of today, bioinformatics is one of the most exciting fields of scientific research. There is a wide-ranging list of challenging problems to face, i.e., pairwise and multiple alignments, motif detection/discrimination/classification, phylogenetic tree reconstruction, protein secondary and tertiary structure prediction, protein function prediction, DNA microarray analysis, gene regulation/regulatory networks, just to mention a few, and an army of researchers, coming from several scientific backgrounds, focus their efforts on developing models to properly address these problems. In this paper, we aim to briefly review some of the huge amount of machine learning methods, developed in the last two decades, suited for the analysis of gene microarray data that have a strong impact on molecular biology. In particular, we focus on the wide-ranging list of data clustering and visualization techniques able to find homogeneous data groupings, and also provide the possibility to discover its connections in terms of structure, function and evolution.

Improving Clustering of MicroRNA Microarray Data by Incorporating Functional Similarity

Improving Clustering of MicroRNA Microarray Data by Incorporating Functional Similarity

Informatively clustering longitudinal microarrays using binary or survival outcome data

ABSTRACTThe goal of this research is to discover what groups of genes are associated with the disease process. We use binary and failure time outcomes to inform the clustering of longitudinally-collected microarray data. We propose a linear model with normally distributed cluster-specific random effects for the longitudinal gene expression trajectory. The random effects are linearly related to a latent continuous representation of the outcome, where the probability or hazard of the outcome depends on these latent variables. We apply our method to microarray data collected from trauma patients in the inflammation and host response to Injury project.

Impaired Response to Influenza Vaccination in AML Patients Post-Chemotherapy Associated with a Highly Atypical B-Cell Profile

The immunotherapy of cancer is arguably the most promising therapy under development, and vaccination against cancer antigens is a promising strategy to stimulate adaptive immune responses against malignant clones. However, the ability of patients treated with chemotherapy to respond appropriately to immune challenge may be suboptimal. This study was designed to determine the functional capacity of the immune system in adult acute myeloid leukemia (AML) patients who have completed chemotherapy and are potential candidates for immunotherapy. We used the response to influenza vaccination as a surrogate for the health of the immune system in AML patients in a complete remission (CR) post-chemotherapy.Ten adult AML patients in CR after completion of intensive chemotherapy were recruited to the clinical protocol J1293. They were on average 37 weeks post-treatment (range 4-148) when receiving the 2012-2013 inactivated seasonal influenza vaccine. Peripheral blood samples were collected at baseline and 30 days post-vaccination. Ten age and sex matched healthy donors (HD) served as baseline controls. Serological response to vaccination was assessed via microneutralization assays; multi-parameter flow cytometry was used to characterize lymphocyte subsets. ELISPOT assays were used to evaluate lymphocyte function, microarrays were used to assess gene expression, and deep sequencing of the B-cell receptor heavy chain (IGH) was performed to determine expansion and clonality of B-cells. Wilcoxon rank-sum tests were used to assess statistical significance.Only 2/10 (20%) patients seroconverted (AML responders, or AML-R) with a four-fold increase or greater in influenza-specific antibody. One responder was 148 weeks post-treatment; the other had acute promyelocytic leukemia (APL) and was 4 weeks post-treatment. Deep immunophenotyping revealed no striking differences in T-cell compartments between AML at baseline and HD, indicating rapid T-cell recovery after chemotherapy. In contrast, we observed a highly atypical B-cell profile. AML non-responders (AML-NR) at day 0 versus HD had significantly reduced frequencies of mature IgA+ (4.5% vs. 11%) and IgG+ (2.4% vs. 5.9%) B-cells (as a percentage of CD19+CD20+). Further dissection exposed markedly higher frequencies of CD10+CD27- transitional B-cells (36% vs. 16%, p<0.005) and dramatically fewer memory (resting, tissue-like, activated) B-cells (5% vs. 25%, p<0.005) (as a percentage of CD19+CD20+) in AML-NR at day 0 versus HD. There were no significant changes in any B-cell population at day 30 over baseline in any patient. Examining frequencies of transitional, naive, and memory B-cells in AML patients at day 0 when ranked by time since treatment showed a decrease in transitional B-cells with a corresponding increase in naive B-cells over time but no concurrent increase in memory B-cell frequencies (Fig 1). These data suggest B-cell deficiencies of several types: loss of the memory B-cell compartment due to chemotherapy, a subsequent excess of transitional B-cells, and a lack of naive B-cell development into specific, class-switched effectors of the antibody response, which together likely lead to humoral immune incompetence.Interestingly, functional T-cell assays revealed that of 5 evaluable patients, including the 2 AML-R and 3 AML-NR, 5/5 (100%) patients had an increase in influenza-specific cytokine production (1.24 - 4.40x higher on day 30 over baseline), suggesting a functional T-cell response even with deficient influenza-specific antibody production. Supervised clustering of microarray data identified many upregulated genes in AML-NR related to apoptosis, BCR, IL2, IL-4, IL-8, and IL-12 signaling pathways, indicative of developing B-cells. IGH sequencing demonstrated AML-NR had greater variability in CDR3 length than seen in HD, consistent with an antigen inexperienced B-cell repertoire.These data suggest that while some aspects of cellular immunity recover comparatively quickly, the humoral immune system is incompletely reconstituted in the year following intensive cytotoxic chemotherapy for AML. Abnormal frequencies of transitional and memory B-cells may explain the poor response to vaccination often seen in patients after chemotherapy. Furthermore, the uncoupled recovery of B-cell and T-cell immune capacity observed here might have implications for the success of immunotherapies based on vaccination. [Display omitted] DisclosuresNoonan:Celgene: Speakers Bureau. Borrello:Celgene: Research Funding.

OverDBC: A new density-based clustering method with the ability of detecting overlapped clusters from gene expression data

Using microarray techniques, it is possible to measure the expression levels of thousands of genes under several experimental conditions. Extracting information frommicroarray data is an important problem in Bioinformatics scope. Producing overlapping clusters is a major issue in clustering methods. While most of the research in this area has focused on clustering using disjoint cluster, many real microarray datasets and as a result many gene regulatory networks have inherently overlapping partitions. Genes have more than one function by coding for proteins that participate in multiple metabolic pathways. So, Overlapped clusters have an important role in discovering the relationship between genes and finding overlap gene regulatory networks. Recent proposed clustering methods rely on the search of optimal disjoint clusters. In this paper, we propose a new density based clustering (OverDBC) with a bound on the number of overlap clusters. OverDBC allows genes membership in a restricted number of clusters where the total number of clusters is unbounded. We define closeness as a new concept for finding core genes along with the density concept. We compare OverDBC with DBscan (a non-overlapping density-based clustering) algorithm. We prove that OverDBC may be significantly better than non-overlapping clustering in microarray data.

Genome Wide Expression Profiling of Cancer Cell Lines Cultured in Microgravity Reveals Significant Dysregulation of Cell Cycle and MicroRNA Gene Networks.

Zero gravity causes several changes in metabolic and functional aspects of the human body and experiments in space flight have demonstrated alterations in cancer growth and progression. This study reports the genome wide expression profiling of a colorectal cancer cell line-DLD-1, and a lymphoblast leukemic cell line-MOLT-4, under simulated microgravity in an effort to understand central processes and cellular functions that are dysregulated among both cell lines. Altered cell morphology, reduced cell viability and an aberrant cell cycle profile in comparison to their static controls were observed in both cell lines under microgravity. The process of cell cycle in DLD-1 cells was markedly affected with reduced viability, reduced colony forming ability, an apoptotic population and dysregulation of cell cycle genes, oncogenes, and cancer progression and prognostic markers. DNA microarray analysis revealed 1801 (upregulated) and 2542 (downregulated) genes (>2 fold) in DLD-1 cultures under microgravity while MOLT-4 cultures differentially expressed 349 (upregulated) and 444 (downregulated) genes (>2 fold) under microgravity. The loss in cell proliferative capacity was corroborated with the downregulation of the cell cycle process as demonstrated by functional clustering of DNA microarray data using gene ontology terms. The genome wide expression profile also showed significant dysregulation of post transcriptional gene silencing machinery and multiple microRNA host genes that are potential tumor suppressors and proto-oncogenes including MIR22HG, MIR17HG and MIR21HG. The MIR22HG, a tumor-suppressor gene was one of the highest upregulated genes in the microarray data showing a 4.4 log fold upregulation under microgravity. Real time PCR validated the dysregulation in the host gene by demonstrating a 4.18 log fold upregulation of the miR-22 microRNA. Microarray data also showed dysregulation of direct targets of miR-22, SP1, CDK6 and CCNA2.

A Granular Self-Organizing Map for Clustering and Gene Selection in Microarray Data.

A new granular self-organizing map (GSOM) is developed by integrating the concept of a fuzzy rough set with the SOM. While training the GSOM, the weights of a winning neuron and the neighborhood neurons are updated through a modified learning procedure. The neighborhood is newly defined using the fuzzy rough sets. The clusters (granules) evolved by the GSOM are presented to a decision table as its decision classes. Based on the decision table, a method of gene selection is developed. The effectiveness of the GSOM is shown in both clustering samples and developing an unsupervised fuzzy rough feature selection (UFRFS) method for gene selection in microarray data. While the superior results of the GSOM, as compared with the related clustering methods, are provided in terms of β -index, DB-index, Dunn-index, and fuzzy rough entropy, the genes selected by the UFRFS are not only better in terms of classification accuracy and a feature evaluation index, but also statistically more significant than the related unsupervised methods. The C-codes of the GSOM and UFRFS are available online at http://avatharamg.webs.com/software-code.

Rough Based Symmetrical Clustering for Gene Expression Profile Analysis.

Identification of coexpressed genes is the central goal in microarray gene expression data analysis. Point symmetry-based clustering is an important unsupervised learning technique for recognizing symmetrical convex or non-convex shaped clusters. To enable fast automatic clustering of large microarray data, in this article, a distributed time-efficient scalable parallel rough set based hybrid approach for point symmetry-based clustering algorithm has been proposed. A natural basis for analyzing gene expression data using the symmetry-based algorithm, is to group together genes with similar symmetrical patterns of expression. Rough-set theory helps in faster convergence and initial automatic optimal classification, thereby solving the problem of unknown knowledge of number of clusters in microarray data. This new parallel implementation with K-means algorithm also satisfies the linear speedup in timing on large microarray datasets. This proposed algorithm is compared with another parallel symmetry-based K-means and parallel version of existing K-means over four artificial and benchmark microarray datasets. We also have experimented over three skewed cancer gene expression datasets. The statistical analysis are also performed to establish the significance of this new implementation. The biological relevance of the clustering solutions are also analyzed.

English

The countenance level measurement for thousands of genes is allowed in a parallel fashion by microarray technology. Clustering is one of the first stage accepted to reveal information from gene expression data. Choosing an appropriate proximity measure (similarity or distance) is having great significance in addition to selecting a clustering algorithm for attaining reasonable clustering results. Til today, there are no inclusive guidelines regarding how to elect proximity measures for clustering microarray data. The choice of proximity measures is studied for the clustering of microarray data by estimating the performance of twelve proximity measures in some data sets from time course and cancer experiments. Given that different measures hoisted out for time course and cancer data evaluations, their choice should be specific to each scenario. To estimate measures on time-course data, the pre-processed and collected data sets from the microarray literature in a benchmark is used along with a new methodology, called Intrinsic Biological Separation Ability (IBSA). Both can be employed in future research to assess the effectiveness of new measures for gene time-course data.

Embryonic stem cells derived from in vivo or in vitro-generated murine blastocysts display similar transcriptome and differentiation potential.

The use of assisted reproductive technologies (ART) such as in vitro fertilization (IVF) has resulted in the birth of more than 5 million children. While children conceived by these technologies are generally healthy, there is conflicting evidence suggesting an increase in adult-onset complications like glucose intolerance and high blood pressure in IVF children. Animal models indicate similar potential risks. It remains unclear what molecular mechanisms may be operating during in vitro culture to predispose the embryo to these diseases. One of the limitations faced by investigators is the paucity of the material in the preimplantation embryo to test for molecular analysis. To address this problem, we generated mouse embryonic stem cells (mESC) from blastocysts conceived after natural mating (mESCFB) or after IVF, using optimal (KSOM + 5% O2; mESCKAA) and suboptimal (Whitten’s Medium, + 20% O2, mESCWM) conditions. All three groups of embryos showed similar behavior during both derivation and differentiation into their respective mESC lines. Unsupervised hierarchical clustering of microarray data showed that blastocyst culture does not affect the transcriptome of derived mESCs. Transcriptomic changes previously observed in the inner cell mass (ICM) of embryos derived in the same conditions were not present in mESCs, regardless of method of conception or culture medium, suggesting that mESC do not fully maintain a memory of the events occurring prior to their derivation. We conclude that the fertilization method or culture media used to generate blastocysts does not affect differentiation potential, morphology and transcriptome of mESCs.

Cluster analysis of cancer data using semantic similarity, sequence similarity and biological measures

Clustering of genes on the basis of expression profiles is generally the first step in understanding how a class of genes behaves in a biological process. A number of supervised and unsupervised algorithms are available in statistics and machine learning literature for clustering microarray data, but the algorithms are restricted in their ability to evaluate the results of a clustering algorithm in the light of biologically meaningful clusters. If two gene sequences are similar, then we would expect that their genetic expressions are similar and that they are similarly annotated in the Gene Ontology (GO) databases. Hence a comparison of the expression level similarity of two gene sequences against their corresponding similarity of annotation in the GO can establish this fact. Semantic similarity has now become a valuable tool for validating the results drawn from biomedical studies such as gene clustering and gene expression data analysis. This paper borrows from our previous work on meta-ensembles using cancer datasets where the output of several clustering algorithms are subsequently fed to a consensus building process to generate a stable set of cluster results. Next, these cluster results are further refined through a sequence of biological validation process for each gene pair of a given cluster using semantic similarity and sequence similarity. We have tested our approach on several benchmark cancer datasets in an attempt to provide a more accurate biological analysis of the clusters and the results have been found to be satisfactory.

Bioinformatics multivariate analysis determined a set of phase-specific biomarker candidates in a novel mouse model for viral myocarditis.

Myocarditis is an inflammatory disease of the cardiac muscle and is mainly caused by viral infections. Viral myocarditis has been proposed to be divided into 3 phases: the acute viral phase, the subacute immune phase, and the chronic cardiac remodeling phase. Although individualized therapy should be applied depending on the phase, no clinical or experimental studies have found biomarkers that distinguish between the 3 phases. Theiler's murine encephalomyelitis virus belongs to the genus Cardiovirus and can cause myocarditis in susceptible mouse strains. Using this novel model for viral myocarditis induced with Theiler's murine encephalomyelitis virus, we conducted multivariate analysis including echocardiography, serum troponin and viral RNA titration, and microarray to identify the biomarker candidates that can discriminate the 3 phases. Using C3H mice infected with Theiler's murine encephalomyelitis virus on 4, 7, and 60 days post infection, we conducted bioinformatics analyses, including principal component analysis and k-means clustering of microarray data, because our traditional cardiac and serum assays, including 2-way comparison of microarray data, did not lead to the identification of a single biomarker. Principal component analysis separated heart samples clearly between the groups of 4, 7, and 60 days post infection. Representative genes contributing to the separation were as follows: 4 and 7 days post infection, innate immunity-related genes, such as Irf7 and Cxcl9; 7 and 60 days post infection, acquired immunity-related genes, such as Cd3g and H2-Aa; and cardiac remodeling-related genes, such as Mmp12 and Gpnmb. Sets of molecules, not single molecules, identified by unsupervised principal component analysis, were found to be useful as phase-specific biomarkers.

MicroRNAs define distinct human neuroblastoma cell phenotypes and regulate their differentiation and tumorigenicity.

BackgroundNeuroblastoma (NB) is the most common extracranial solid tumor in children. NB tumors and derived cell lines are phenotypically heterogeneous. Cell lines are classified by phenotype, each having distinct differentiation and tumorigenic properties. The neuroblastic phenotype is tumorigenic, has neuronal features and includes stem cells (I-cells) and neuronal cells (N-cells). The non-neuronal phenotype (S-cell) comprises cells that are non-tumorigenic with features of glial/smooth muscle precursor cells. This study identified miRNAs associated with each distinct cell phenotypes and investigated their role in regulating associated differentiation and tumorigenic properties.MethodsA miRNA microarray was performed on the three cell phenotypes and expression verified by qRT-PCR. miRNAs specific for certain cell phenotypes were modulated using miRNA inhibitors or stable transfection. Neuronal differentiation was induced by RA; non-neuronal differentiation by BrdU. Changes in tumorigenicity were assayed by soft agar colony forming ability. N-myc binding to miR-375 promoter was assayed by chromatin-immunoprecipitation.ResultsUnsupervised hierarchical clustering of miRNA microarray data segregated neuroblastic and non-neuronal cell lines and showed that specific miRNAs define each phenotype. qRT-PCR validation confirmed that increased levels of miR-21, miR-221 and miR-335 are associated with the non-neuronal phenotype, whereas increased levels of miR-124 and miR-375 are exclusive to neuroblastic cells. Downregulation of miR-335 in non-neuronal cells modulates expression levels of HAND1 and JAG1, known modulators of neuronal differentiation. Overexpression of miR-124 in stem cells induces terminal neuronal differentiation with reduced malignancy. Expression of miR-375 is exclusive for N-myc-expressing neuroblastic cells and is regulated by N-myc. Moreover, miR-375 downregulates expression of the neuronal-specific RNA binding protein HuD.ConclusionsThus, miRNAs define distinct NB cell phenotypes. Increased levels of miR-21, miR-221 and miR-335 characterize the non-neuronal, non-malignant phenotype and miR-335 maintains the non-neuronal features possibly by blocking neuronal differentiation. miR-124 induces terminal neuronal differentiation with reduction in malignancy. Data suggest N-myc inhibits neuronal differentiation of neuroblastic cells possibly by upregulating miR-375 which, in turn, suppresses HuD. As tumor differentiation state is highly predictive of patient survival, the involvement of these miRNAs with NB differentiation and tumorigenic state could be exploited in the development of novel therapeutic strategies for this enigmatic childhood cancer.

Non-Negative Factorization for Clustering of Microarray Data

Typically, gene expression data are formed by thousands of genes associated to tens or hundreds of samples. Gene expression data comprise relevant (discriminant)information as well as irrelevant information often interpreted as noise. The irrelevant information usually affects the efficiency of discovering and grouping meaningful latent information correlated to biological significance, process closely related to data clustering. Class discovery through clustering may help in identifying latent features that reflect molecular signatures, ultimately leading to class forming. One solution for improving the class discovery efficiency is provided by data dimensionality reduction, where data is decomposed into lower dimensional factors, so that those factors approximate original data.