RNA sequencing (RNA-seq) is a New Generation Sequencing (NGS) method used for the analysis of transcripts and differential gene expression profiles. MicroRNA (miRNAs, 22--25 nt long) are, among small non coding RNAs (sncRNA) obtained through RNA-seq, key regulators in multiple cellular functions, as they play a crucial role in different physiological processes. miRNAs are in fact differentially expressed in several types of cancer, in specific tissues and during specific cell status. Clustering algorithms have been applied to microarray data in order to discover groups of genes (clusters) that are co-regulated with respect to certain experimental conditions. Because many regulation mechanisms involve only set of genes and limited set of experimental conditions, a new approach is needed. In this context, biclustering techniques represent suitable approaches because they allow to separate, in a data matrix, groups of rows and columns, standing for genes and samples , that exhibits similar values or similar characteristics. We present a biclustering approach in order to identify some patterns of miRNA expression deregulation in human breast cancer versus healthy controls. We applied the Iterative Signature Algorithm (ISA) tool, which has proved one of the most efficient when applied to gene expression datasets. Considering a real word breast cancer dataset, composed of 185 samples, we identified 12 miRNA biclusters, each of them involving different types of tumor samples and miRNA families. We showed the association between specific sub-class of tumor samples having the same immuno-histo-chemical (IHC) and/or histological features. Biclusters have been validated in the current scientific using the MetaMirClust and UCSC Genome Browser online tools, as well as another biclustering algorithm (SAMBA). The proposed biclustering led to the identification of different groups of miRNAs and patient conditions, that eventually have to be validated by in-vitro experiment.
A biclustering approach for the analysis of miRNA expression profiles
Antonino Fiannaca;Laura La Paglia;Massimo La Rosa;Antonio Messina;Riccardo Rizzo;Pietro Storniolo;Mario Tripiciano;
2015
Abstract
RNA sequencing (RNA-seq) is a New Generation Sequencing (NGS) method used for the analysis of transcripts and differential gene expression profiles. MicroRNA (miRNAs, 22--25 nt long) are, among small non coding RNAs (sncRNA) obtained through RNA-seq, key regulators in multiple cellular functions, as they play a crucial role in different physiological processes. miRNAs are in fact differentially expressed in several types of cancer, in specific tissues and during specific cell status. Clustering algorithms have been applied to microarray data in order to discover groups of genes (clusters) that are co-regulated with respect to certain experimental conditions. Because many regulation mechanisms involve only set of genes and limited set of experimental conditions, a new approach is needed. In this context, biclustering techniques represent suitable approaches because they allow to separate, in a data matrix, groups of rows and columns, standing for genes and samples , that exhibits similar values or similar characteristics. We present a biclustering approach in order to identify some patterns of miRNA expression deregulation in human breast cancer versus healthy controls. We applied the Iterative Signature Algorithm (ISA) tool, which has proved one of the most efficient when applied to gene expression datasets. Considering a real word breast cancer dataset, composed of 185 samples, we identified 12 miRNA biclusters, each of them involving different types of tumor samples and miRNA families. We showed the association between specific sub-class of tumor samples having the same immuno-histo-chemical (IHC) and/or histological features. Biclusters have been validated in the current scientific using the MetaMirClust and UCSC Genome Browser online tools, as well as another biclustering algorithm (SAMBA). The proposed biclustering led to the identification of different groups of miRNAs and patient conditions, that eventually have to be validated by in-vitro experiment.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
