**1. Introduction**

Dilated cardiomyopathy (DCM) is the predominant heart muscle disease, characterized by a dilatated left ventricle (LV) and reduced contractility. It is associated with high hospitalization rates, sudden cardiac death risk, and substantial demand for heart failure therapies. The prevalence of all forms of DCM is estimated to be as high as 1:250 [1]. In the patients with DCM, approximately 30%–50% are assumed to have familial predisposition of the disease [2], and 40% of the familial DCM patients possess currently identifiable genetic variations [3], with most of them having an autosomal-dominant

transmission pattern [4]. On the other side, some sporadic DCM patients have de novo genetic mutations. As the technology of next generation sequencing grows expeditiously, causative genetic variants of DCM have been detected in over 30 genes, with a great number of them encoding sarcomere proteins, such as *TTN, MYH6, MYH7,* and *TNNT2*, and others encoding proteins constituting calcium or potassium channels, such as *SCN5A*, proteins essential in the nuclear membrane, such as *LMNA*, as well as others such as *BAG3* or *TAZ (G 4.5)* [1,5]. Truncating variants in *TTN* (TTNtv) account for 15%–25% of familial DCM and 10%–18% of sporadic DCM [6]. Otherwise, most DCM-related genetic variants are reported to be nonsynonymous missense, while other types of mutations, such as frameshifts, insertions, deletions, and splice-site-mutations, were also detected [1].

Recently, a drastically increased number of disease-causing variants have been pinpointed with the help of genome-wide association studies. In the era of precision medicine, an in-depth understanding of the disease-causing mechanism of these detected variants has proved to be challenging, but it is the key to the development of novel personalized therapies [7]. Aside from traditional single-point gene variants, new evidence suggests a possible role of genetic–environmental interactions in human cardiomyopathies, for example, through alterations of DNA methylation [8,9] and through changes of histone modifications [10]. In a recent study by us on human DCM, the DNA methylation level within the promoter region was found to correlate negatively with gene expression. It was observed that the pattern of DNA methylation in promoter regions is significantly changed when comparing DCM patients with healthy controls. Numerous hot spots with statistically significant phenotype–epigenotype correlations were identified in the genome of DCM patients [8].

On the other side, post-transcriptional modifications of sarcomeric proteins were also reported to play a critical role in the pathogenesis of DCM. Aberrantly spliced sarcomere proteins, including titin, troponin T, tropomyosin, and LDB3 protein, were detected in patients with DCM, generating abnormal protein products predisposing people to heart failure. *TTN*, encoding titin, is the most commonly mutated gene in DCM. An RNA-binding splicing factor RBM20 can bind directly to the intronic parts of titin and affect its splicing. Genetic variations in *RBM20* in 3%–5% of genetic DCM cases cause the expression of a dysregulated isoform of titin, N2BA-G, generating reduced passive tension of the muscle fiber in DCM [11–13].

We sought here to identify a role of DNA methylation changes in the somatic tissue of DCM patients and the impact of such alterations on mRNA splicing. By using an epigenome-wide and whole transcriptome approach, we found a strong association of intragenic DNA methylation and exon usage. Particularly interesting are changes of the Titin locus, where we found reciprocal alterations in an encoded Titin-antisense non-coding RNA in DCM compared to control.
