Abstract
BACKGROUND: Large-scale "omics" datasets have not been leveraged and integrated with functional analyses to discover potential drivers of cardiomyopathy. This study addresses the knowledge gap. METHODS: We coupled RNA sequence (RNA-Seq) variant detection and transcriptome profiling with pathway analysis to model drug refractory dilated cardiomyopathy (drDCM) using the BaseSpace sequencing hub and Ingenuity Pathway Analysis. We used RNA-Seq case-control datasets (n = 6 cases, n = 4 controls), exome sequence familial DCM datasets (n = 3 Italians, n = 5 Italians, n = 5 Chinese), and controls from the HapMap project (n = 5 Caucasians, and n = 5 Asians) for disease modeling and putative mutation discovery. Variant replication datasets: n = 128 cases and n = 15 controls. Source of datasets: NCBI Sequence Read Archive. STATISTICS: Pairwise differential expression analyses to determine differentially expressed genes and t-tests to calculate p-values. We adjusted for false discovery rates and reported q-values. We used chi-square tests to assess independence among variables, the Fisher's Exact Tests and overlap p-values for the pathways and p-scores to rank network. RESULTS: Data revealed that ECHS1(enoyl-CoA hydratase, short chain 1(log(2)(foldchange) = 1.63329) hosts a mirtron, MIR3944 expressed in drDCM (FPKM = 5.2857) and not in controls (FPKM = 0). Has-miR3944-3p is a putative target of BAG1 (BCL2 associated athanogene 1(log(2)(foldchange) = 1.31978) and has-miR3944-5p of ITGAV (integrin subunit alpha V(log(2)(foldchange) = 1.46107) and RHOD (ras homolog family member D(log(2)(foldchange) = 1.28851). There is an association between ECHS1:11 V/A(rs10466126) and drDCM (p = 0.02496). The interaction (p = 2.82E-07) between ECHS1:75 T/I(rs1049951) and ECHS1:rs10466126 is associated with drDCM (p < 2.2e-16). ECHS1:rs10466126 and ECHS1:rs1049951 are in linkage disequilibrium (D' = 1). The interaction (p = 7.84E-08) between ECHS1:rs1049951 and the novel ECHS1:c.41insT variant is associated with drDCM (p < 2.2e-16). The interaction (p = 0.001096) between DBT (Dihydrolipoamide branched chain transacylase E2):384G/S(rs12021720) and ECHS1:rs10466126 is associated with drDCM (p < 2.2e-16). At the mRNA level, there is an association between ECHS1 (log(2)(foldchange) = 1.63329; q = 0.013927) and DBT (log(2)(foldchange) = 0.955072; q = 0.0368792) with drDCM. ECHS1 is involved in valine (-log (p = 3.39E00)), isoleucine degradation (p = 0.00457), fatty acid β-oxidation (-log(p) = 2.83E00), and drug metabolism:cytochrome P450 (z-score = 2.07985196) pathways. The mitochondria (-log(p) = 8.73E00), oxidative phosphorylation (-log(p) = 5.35E00) and TCA-cycle II (-log(p) = 2.70E00) are dysfunctional. CONCLUSIONS: We introduce an integrative data strategy that considers the interplay between the DNA, mRNA, and associated pathways, which represents a possible diagnostic, prognostic, biomarker, and personalized treatment discovery approach in genomically heterogeneous diseases.