Identifying the optimal target genes associated with multiple myeloma by a novel bioinformatical analysis

Abstract

Multiple myeloma (MM) is one of the most frequent malignant hematopoietic diseases, the pathogenesis of which remains unclear. It is well known that miRNAs are aberrantly expressed in many tumors, thus, investigating the target genes of miRNAs contributes to understanding the functional effect of miRNAs on MM. In this study, plasma samples of 147 patients with MM and 15 normal donors were collected. Using high-throughout microarray and limma package to screen the differentially expressed genes. Furthermore, to accurately predict the optimal target genes of MM, the logFC, targetScanCS and targetScanPCT values of known genes in four miRNAs (i.e. has-miR-21, has-miR-20a, has-miR-148a and has-miR-99b) were used to compute the targetScore values. As a result, 171 genes with larger difference were screened out using t-test, F-test and eBayes statistics analysis. Furthermore, 34 potential target genes associated with MM were selected by integrating the differentially expressed genes (DEGs) and the genes obtained by targetScore algorithm. Additionally, combining with the mutated genes in MM and the obtained DEGs, 41 consistently expressed genes were obtained. Finally, 5 optimal target genes, including SYK, LCP1, HIF1A, ALDH1A1 and MAFB, were screened out by the intersection of 34 DEGs and 41 mutated genes. In a word, this novel target gene prediction algorithm may contribute to improve our understanding on the pathogenesis of miRNAs in MM, which open up a new approach for future study.