META-ANALYSIS OF GENE EXRESSION DATA FOR IDENTIFICATION OF COMMON ALTERED GENES IN BREAST AND OVARIAN CANCERS

Abstract

A technology that has been used extensively for analyzing breast and ovarian cancer malignancies is microarray technology. The combined investigation for breast and ovarian cancers across multiple gene expression studies are not well reported. In preliminary phase of study, data analysis was performed by combining gene expression profiles of eight different published microarray studies based on breast and ovarian cancers. Breast and ovarian cancers’ genetic makeup are very similar and heritable mutations in the tumor suppressor genes BRCA1 and BRCA2 incline individuals to breast and ovarian cancers. Microarray data of both cancers was screened and downloaded from NCBI GEO. The raw data files were extracted and only low, high and normal grade tumor samples were included for the metaanalysis. After combining all the eight microarray data, normalization, and pre-processing, differential gene expression analysis (DEGs) were carried out. The statistical test that was used for identifying DEGs was one way ANOVA and this was followed by clustering the top DEGs. The clustering analysis explored the common genes and sample expression pattern including co-expressed gene sets across two types of cancers (breast and ovarian). This metaanalysis unified eight results of previous gene expression studies in breast and ovarian cancers. This analysis was performed using two different softwares viz. Robina and Genespring. The combined microarray data analysis result revealed the connection between common expressed genes in breast and ovarian cancers. It was found that the common DEGs and subsequently the co-expressed genes have strong enrichment from cell proliferation, ER signaling, actin cytoskeleton and Mitogen Activated Protein Kinase (MAPK) pathways. The research was continued by further pathway analysis of DEGs and co-expressed genes which then explored the common molecular basis, signatures and potential important regulatory pathways in these two cancer developments. The common up-regulated genes deduced after performing all the steps were IRF5, IKZF2 and CCNL1 and the common down-regulated genes included ATF3, HMGA1 and NRIP3. The identified common altered genes in breast and ovarian expression data, which can serve as potential biomarkers, were validated using in silico method.