Whole Genome microRNA Expression Data in Childhood Acute Lymphoblastic Leukemia and Evaluation of microRNA Pathways Using Fuzzy C-means

Objective: Hard clustering approaches may cause some of the relationships to be overlooked due to their nature of algorithms especially in genetic datasets. But hidden relationships can be revealed by fuzzy approaches. Purpose of this study was evaluating effect of microRNAs (miRNA) on children with acute lymphoblastic leukaemia (ALL) by using miRNA expression data obtained from bone marrow samples with sets containing different numbers of elements of fuzzy Cmeans (FCM). Material and Methods: miRNA expression levels of 43 newly diagnosed ALL patients and 14 healthy subjects were analysed via FCM. Clusters containing different numbers of miRNAs were evaluated, common properties in messenger RNA (mRNA) pathways were investigated and new pathways associated with ALL and cancer were described via miRNA target prediction tools. Results: Significant miRNA profile was compared to control cases. Only 46 out of 108 miRNAs were found to be significantly upregulated or downregulated. Of forty six miRNAs: 8 miRNAs were labelled as tumour suppressor (17.4%), 17 miRNAs were labelled as onco-miR (37.0%) and 21 miRNAs could not be labelled (45.6%) for hematological malignancy. Fourteen (%30.4) miRNAs were found to be apoptosis-related, 27 miRNAs were in leukemia-related (58.7%) and 15 labelled miRNAs were related with cancer pathways (32.6%). hsa-miR-181b, hsa-miR- 146a, hsa-miR-155, hsa-miR-181c-5p, hsa-miR-7-1-3p, hsa-miR-708- 5p onco-miRs constituted a set. These miRNAs targeted 801 common mRNAs (p<0.05). When this sub-cluster was searched in the literature and miRNA target prediction tools system, it was found to be involved in cancer-related pathways except ALL. Conclusion: Hidden relationships can be defined by fuzzy approaches and those pathways may provide guidance to open up new horizons in the field of miRNA studies.