Data of methylome and transcriptome derived from human dilated cardiomyopathy

Alterations in DNA methylation and gene expression have been implicated in the development of human dilated cardiomyopathy (DCM). Differentially methylated probes (DMPs) and differentially expressed genes (DEGs) were identified between the left ventricle (LV, a pathological locus for DCM) and the right ventricle (RV, a proxy for normal hearts). The data in this DiB are for supporting our report entitled “Methylome analysis reveals alterations in DNA methylation in the regulatory regions of left ventricle development genes in human dilated cardiomyopathy” (Bong-Seok Jo, In-Uk Koh, Jae-Bum Bae, Ho-Yeong Yu, Eun-Seok Jeon, Hae-Young Lee, Jae-Joong Kim, Murim Choi, Sun Shim Choi, 2016) [1].

Alterations in DNA methylation and gene expression have been implicated in the development of human dilated cardiomyopathy (DCM). Differentially methylated probes (DMPs) and differentially expressed genes (DEGs) were identified between the left ventricle (LV, a pathological locus for DCM) and the right ventricle (RV, a proxy for normal hearts). The data in this DiB are for supporting our report entitled "Methylome analysis reveals alterations in DNA methylation in the regulatory regions of left ventricle development genes in human dilated cardiomyopathy" (Bong-Seok Jo, In-Uk Koh, Jae-Bum Bae, Ho-Yeong Yu, Eun-Seok Jeon, Hae-Young Lee, Jae-Joong Kim, Murim Choi, Sun Shim Choi, 2016) [1].
& Provide new insights on the interaction network constructed by genes of DMP-DEG pairs.

Data
DCM samples where methylome and transcriptome data were produced used in the present DiB were listed in Table S1. The present data contain as followings: cleaning and normalization procedures (Figs. S1 and S2), global DNA methylation pattern ( (Table 1), protein-protein interactions for the 45 genes (Fig. S12), and the relationship between methylation and expression of genes (i.e., TBX5 and HAND1) (Fig. S13).

Ethics statement
The data were prepared in accordance with principles (the Helsinki Declaration). It was approved by the Institutional Review Board (IRB) of The Samsung Medical Center (South Korea) (No. 2012-02-065). All participants have provided written informed consent and obtained the IRB approval for the consent procedure.

Tissue sample and chip data preparation from human DCM patients
Please refer to 'Materials and Methods' section of our original article published in Genomics [1] for the detailed procedures about where tissue samples originated from, how to extract DNAs and RNAs, and what kinds of chip technologies were used for data productions.

Finding DMPs and DEGs between LV and RV
One of the Bioconductor packages named RnBeads [2] was used for parsing raw intensity data generated from the Illumina 450 K IDAT files [3]. A total of 13,170 DMPs were chosen by a rank implemented by 'combinedRank' function of RnBeads program [2]. Please refer to our original paper [1] and the RnBeads program manual for the detailed protocols [2]. To identify DEGs, we first removed probes of detection p-value of over 0.01 in any sample and performed a quantile normalization [1]. Then, the filtered microarray data were compared between the LV and RV samples. A twosample t-test was applied for selecting DEGs between the two samples at a FDR adjusted p o0.05 using R version 3.2.2 [4], from which a total of 3347 DEGs were identified. The eight top-ranked functional sub-networks with many nodes (genes) were selected from the 984 genes included in the input genes list. The most significant functional term in each subnetwork was selected to indicate the pathway of each sub-network. The down-and up-regulation of expression in each subnetwork was determined by the log2-transformed average fold change of the mean expression of the genes in a particular subnetwork and colored blue and red, respectively. TSS1500 or TSS200 in a sub-network indicates that the DMP positions were relatively enriched in that sub-network. The bold black border of the nodes indicates hypermethylation in LV, whereas the gray border of nodes indicates hypomethylation in LV; the red area indicates grouped genes with up-regulated expression, whereas the blue area indicates grouped genes with down-regulated expression.

Matching DMP-DEG pairs
Matching the 13,170 DMPs produced by the combinedRank cutoff (72,880) of RnBeads to the 3347 DEGs resulted in a total of 984 DMP-DEG pairs. This matching experiment was performed with home-built Python scripts. Odd-skipped related transcription factor 2 PPP1R13L Protein phosphatase 1, regulatory subunit 13 like PTCD2 Pentatricopeptide repeat domain 2 RAPGEF3 Rap guanine nucleotide exchange factor (GEF) 3

Functional characterization of DMP-containing genes
Function of genes located at the nearest DMPs was estimated by a freely available web-tool called GREAT (http://bejerano.stanford.edu/great/public/html) [5]. Significance test for gene ontology (GO) enrichment was performed with the binomial test in the GREAT analysis. The protein-protein interaction network analysis for the selected 45 genes was performed with GeneMANIA (ver. 3.4.0) through the Cytoscape (ver. 3.2.0) [6].