Cis-Cardio: A comprehensive analysis platform for cardiovascular-relavant cis-regulation in human and mouse

Cis-regulatory elements are important molecular switches in controlling gene expression and are regarded as determinant hubs in the transcriptional regulatory network. Collection and processing of large-scale cis-regulatory data are urgent to decipher the potential mechanisms of cardiovascular diseases from a cis-regulatory element aspect. Here, we developed a novel web server, Cis-Cardio, which aims to document a large number of available cardiovascular-related cis-regulatory data and to provide analysis for unveiling the comprehensive mechanisms at a cis-regulation level. The current version of Cis-Cardio catalogs a total of 45,382,361 genomic regions from 1,013 human and mouse epigenetic datasets, including ATAC-seq, DNase-seq, Histone ChIP-seq, TF/TcoF ChIP-seq, RNA polymerase ChIP-seq, and Cohesin ChIP-seq. Importantly, Cis-Cardio provides six analysis tools, including region overlap analysis, element upstream/downstream analysis, transcription regulator enrichment analysis, variant interpretation, and protein-protein interaction-based co-regulatory analysis. Additionally, Cis-Cardio provides detailed and abundant (epi-) genetic annotations in cis-regulatory regions, such as super-enhancers, enhancers, transcription factor binding sites (TFBSs), methylation sites, common SNPs, risk SNPs, expression quantitative trait loci (eQTLs), motifs, DNase I hypersensitive sites (DHSs), and 3D chromatin interactions. In summary, Cis-Cardio is a valuable resource for elucidating and analyzing regulatory cues of cardiovascular-specific cis-regulatory elements. The platform is freely available at http://www.licpathway.net/Cis-Cardio/index.html.


INTRODUCTION
The cardiovascular system plays a vital role in transporting blood and nutrients around the body. Dysfunction or injury of the cardiovascular system can lead to poor prognosis of cardiovascular diseases, such as myocardial infarction, heart failure, and atherosclerosis. However, the molecular regulatory mechanism of these diseases is unclear, and it is urgent to unveil it to maintain normal physiological functions. Abnormal gene expression is a risk factor in the development of complex diseases and is essential for understanding the pathological mechanisms. Thus, our study aims to dissect the comprehensive gene regulation patterns in the cardiovascular system.
Gene expression programs are complex and are driven by transcription regulators that occupy at cis-regulatory elements, such as promoters and distal enhancers, thereby supervising the expression activity of downstream genes. [1][2][3] Mechanically, the mediator complex links signals from multiple regulators, such as transcription factor (TF) and transcription co-factor (TcoF), and recruits cohesin complexes to bind to RNA polymerase II and initiate gene transcription. 4,5 As the crucial commanders of gene expression, it is important to illustrate the downstream regulatory atlas of cis-regulatory regions. 6,7 Recently, Zheng et al. constructed the Cistrome Database that contains approximately 47,000 human and mouse samples from about 24,000 collected epigenetics datasets to dissect the global gene expression programs. 8 cis-regulation of gene expression also provides an important perspective to understand disease etiology. 9 In the field of cardiovascular diseases, some studies have also revealed core regulatory networks based on cis-regulation. For instance, Hocker et al. found >280,000 cis-regulatory regions of heart failure and annotated two variants that affect cis-regulatory regions controlling KCNH2/ HERG expression and action potential repolarization in single-cell resolution. 10 Huang et al. found that super-enhancer-driven circRNA Nfix could promote cardiac regenerative repair by inhibiting Ybx1 ubiquitin-dependent degradation and activating miR-214 expression after myocardial infarction. 11 Galang et al. revealed an Isl1 enhancer that regulates pacemaker cells development and sinoatrial node function via cis-regulation. 12 Cis-regulatory regions have the strong ability to recruit transcriptional regulators. Genome variations in TFs, cofactors, and chromatin regulator binding sites are also major causes of cardiovascular dysfunction, such as mutations in GATA4 13 and TBX5. 14 Therefore, identification and annotation of cis-regulatory regions are the central topics in transcription regulation. Moreover, investigating the downstream regulatory cues of cis-regulatory regions is also crucial to dissect the cardiovascular-specific gene expression pattern. These studies demonstrate the importance and widespread utility of cis-regulatory regions for addressing key regulatory cues associated with cardiovascular physiological and pathological processes.
Technologically, several high-throughput sequencing techniques, such as chromatin immunoprecipitation sequencing (ChIP-seq), assay for transposase-accessible chromatin sequencing (ATAC-seq), and DNaseI sequencing (DNase-seq) have been developed for identifying genome-wide cis-regulatory regions. 15 Based on these publicly available epigenomics datasets, some databases or web tools have also been developed to focus on understanding the regulatory potentials and biological functions of cis-regulatory regions, such as ENCODE, Cistrome, ReMap, ChIP-Atlas, SEdb2.0, and GREAT. 8,[16][17][18][19][20] These resources have provided valuable data for cisregulation studies. Moreover, single-cell transcription regulation data have also been released, such as single-cell ATAC-seq (scA-TAC-seq) from scEnhancer. 21 However, all these resources have paid more attention to provide genome-wide cis-regulatory regions and basic functions but have not focused on regulatory annotations, including comprehensive upstream and downstream regulatory annotations. Especially, the barrier to understanding the genetic and molecular basis of cardiovascular diseases is the paucity of resources to mark the cardiovascular-specific gene regulatory programs. Thus, it is highly desirable to construct an integrated resource and analysis tools of cardiovascular-related cis-regulatory regions, which provides comprehensive annotations of cis-regulatory regions and enables biologists to annotate, analyze, and understand these cardiovascular-related cis-regulatory regions.
To investigate the cis-regulatory mechanisms of the cardiovascular system, we developed the Cis-Cardio platform (http://www. licpathway.net/Cis-Cardio/index.html), which is a comprehensive server for analyzing human and mouse cardiovascular-related cis-regulatory elements. Cis-Cardio is designed to document and annotate a large number of cardiovascular-specific cis-regulatory elements and to uncover the comprehensive mechanisms in cis-regulation level. The current version of Cis-Cardio catalogs a total of 45,382,361 candidate cis-regulatory elements from over 1,013 human and mouse epigenetic datasets, including ATAC-seq, scA-TAC-seq, DNase-seq, Histone ChIP-seq, TF/TcoF ChIP-seq, RNA polymerase ChIP-seq, and Cohesin ChIP-seq. These datasets were manually curated from numerous epigenetic databases and almost covered all samples of cardiovascular systems, such as tissues, primary cells, and induced pluripotent stem cells. Emphatically, Cis-Cardio provides detailed and abundant (epi-) genetic annotations in cis-regulatory regions, such as super-enhancers, enhancers, transcription factor binding sites (TFBSs), methylation sites, common SNPs, risk SNPs, expression quantitative trait loci (eQTLs), motifs, DNase I hypersensitive sites (DHSs), and 3D chromatin interactions. Cis-Cardio also provides cis-element downstream target genes by mapping binding regions into genomes in three methods. Furthermore, Cis-Cardio provides various annotations for ciselement target genes, including pathways, Gene Ontology (GO) terms, and expression changes in major cardiovascular diseases. Especially, Cis-Cardio provides six types of cis-regulatory analyses for users, including genome region overlap analysis, upstream/ downstream regulatory axes analysis, transcription regulator enrichment analysis, variant interpretation, and transcription co-regulatory analysis. Cis-Cardio is a user-friendly platform to analyze, query, browse, and visualize information associated with cis-regulatory elements. We believe that Cis-Cardio could become a useful and effective platform for exploring potential functions and cis-regulation in cardiovascular diseases.

Overview and characteristic of Cis-Cardio
The main framework and functions of Cis-Cardio are illustrated in Figure 1, including the collection of cis-regulatory regions from epigenetic datasets, (epi-) genetic annotations, disease gene expression, and the six analysis panels. Briefly, the current version of Cis-Cardio cataloged a total of 45,382,361 genomic regions from over 1,013 human and mouse epigenetic datasets (including ATAC-seq, The top area contains the data scope of the server, which includes collection of ATAC-seq, DNase-seq, Histone ChIP-seq, TF/TcoF ChIP-seq, RNA polymerase ChIP-seq, and Cohesin ChIP-seq. The middle area contains (epi-) genetic annotations in cis-regulatory regions. The bottom area contains six analysis tools for cardiovascular-related cis-regulation. Cis-Cardio is a user-friendly platform to analyze, query, browse, and visualize information associated with cis-regulatory elements.
scATAC-seq, DNase-seq, Histone ChIP-seq, TF/TcoF ChIP-seq, RNA polymerase ChIP-seq, and Cohesin ChIP-seq) of five resources. Modifications and variations of cis-regulatory elements play key roles in guiding gene expression. For example, super-enhancers or enhancers can recruit multiple transcription regulators, such as TFs and TcoFs, to form a transcription complex to participate in gene regulation. [22][23][24] Variations in these regulatory elements can disrupt the high binding affinity between transcription regulators and enhancers, leading to the dysfunction of downstream regulatory axes. 25 Chromatin interaction information between distal elements and proximal elements provides more evidence to identify the potential cis-regulatory elements. Additionally, DNA methylation level also determines the downstream gene expression. In the field of cardiovascular disease, previous studies have also demonstrated the significance of cis-regulatory patterns. 26 Therefore, for each cis-regulatory region, we integrated a large number of (epi-) genetic annotations and provided downstream gene annotations from multiple resources and strategies ( Figure 2A).
Additionally, we performed peak annotation for all candidate cis-regulatory elements via ChIPseeker. Results showed that a majority of candidate cis-regulatory elements were located at the promoter and distal intergenic regions ( Figure 2B). Hence, Cis-Cardio provided the detailed and abundant (epi-) genetic annotations of cis-regulatory regions, such as distal super-enhancers, distal enhancers, TFBSs, methylation sites, common SNPs, risk SNPs, eQTLs, motifs, DHSs, and 3D chromatin interactions. Downstream target genes, pathways, GO terms, and expression changes in major cardiovascular diseases were also provided. Importantly, Cis-Cardio provided six analysis tools to help users decipher the multi-omics regulatory networks of cardiovascular diseases. Moreover, Cis-Cardio covered almost transcription regulators and cell/tissue types in the field of cardiovascular disease ( Figures 2C and 2D). Summarily, Cis-Cardio is a user-friendly web server to analyze, browse, and visualize information associated with cis-regulatory elements.

Case study of Cis-Cardio
Cis-Cardio provided six analysis tools to dissect the candidate cis-regulatory element-mediated transcription regulation mechanism of cardiovascular diseases. To illustrate the use and analysis performances of Cis-Cardio, we performed four case studies by integrating cardiac-specific genes to enrich upstream transcription regulators, inputting cardiac-specific super-enhancers to locate the superenhancer-related regulatory axis and locating regulatory information for cardiovascular-related SNP sites.  The transcription regulator is the key supervisor of transcription regulation, which can regulate downstream gene expression by binding DNA regulatory elements and formatting chromatin loops. The activity of epigenetics factors also plays an important role in the regulatory processes. Notably, compared with gene expression or geneset-based TF enrichment analysis, ChIP-seq-based methods have the advantages in prediction accuracy and tissue specificity. Here, we integrated all the regulatory factor-gene pairs of all epigenetics data and used a hypergeometric test to identify the potential upstream regulators. As an example, we first collected the 55 human cardiacspecific genes from CellMarker, 27 which included the typical markers GATA4, NKX2-5, TNNT2, MYH6, and MYH7 ( Figure 3A, left). Then we set the analysis parameters as default and clicked "Analyze" to perform enrichment analysis. On the results page, Cis-Cardio will list all the ranked regulators based on hypergeometric test. Detailed enrichment analysis statistics and annotated genes are also provided for users to optimize the transcription regulator-target gene relationships ( Figure 3B).
As a result, some potential upstream regulators were identified with top-ranking statistical p value, such as MED1, GATA4, and TBX5 ( Figure 3A, right), which was in coincidence with previous studies. Ang et al. demonstrated that GATA4 broadly co-occupied cardiac super-enhancers with TBX5. 13 Mutation with GATA4 could decrease TBX5 recruitment to cardiac super-enhancers, leading to the dysregulation of downstream genes and phenotypic abnormalities. MED1 is a typical chromatin mark of super-enhancers. 28 These results suggested that MED1-GATA4-TBX5 complex could occupy in super-enhancers to maintain cardiomyocyte identity. Moreover, results showed that GATA4 overlapped more target genes than other regulators, including validated targets NPPA and NPPB, suggesting the capacity of GATA4 in regulating cardiac marker genes. Above all, Cis-Cardio could identify key transcription regulators of a gene set, suggesting the usefulness in exploring the cardiovascular-specific transcription regulation mechanisms.
Identification of the overlapped candidate cis-regulatory elements for Nppa/Nppb super-enhancer Cis-Cardio processed and curated a large number of cis-regulatory regions from diverse epigenetics data, which covered the vast majority of cardiovascular-specific active chromatins. We provided the genomic region intersection analysis to identify the overlapped region of interest from background epigenetics annotation data. Super-enhancers, which are marked by H3K27ac, EP300, MED1, and BRD4, are the representative regulatory elements in maintaining cell identity. Studies revealed that super-enhancers exert functions by recruiting numerous transcription regulators. Here, we collected and analyzed the mouse super-enhancer region of heart failure disease genes Nppa/Nppb 29 ( Figure 3C, left). Cis-Cardio listed all the information of each overlapped region, including genomic locus information, epigenetics data type, and tissue/cell information ( Figure 3D, left). Users can click the Peak_ID to view the detailed information of regions of interest.
As a result, 577 cis-regulatory regions were identified to overlap with the input super-enhancer, indicating the regulatory importance of the super-enhancer in cardiovascular diseases. Based on the cell fractions pie chart of all overlapped regions, we found that cis-regulatory regions from cardiovascular disease samples were extracted, such as cardiac hypertrophy and myocardial infarction ( Figure 3C, right).
Overlapped regions of embryonic heart validated the marker effects of Nppa/Nppb in heart development. Additionally, factor histogram of overlapped regions validated the high chromatin activity and strong ability of super-enhancers in recruiting transcription regulators.

Identification of upstream regulators for heart failure genes
Previous studies have demonstrated that dysregulation of upstream transcription regulators of disease genes is the major cause of cardiovascular diseases. For instance, fetal genes in cardiac development have been demonstrated to regulate heart failure via mediating key biological processes, such as calcium handling, oxidative phosphorylation, switch from fatty acid to glucose metabolism, and mitochondrial dysfunction. All these fetal genes are driven in part by TFs, such as the MEF family, GATA4, NTAT, SRF, and NKX2.5. 30,31 Cis-Cardio integrated comprehensive transcription regulator/cis-regulatory element/gene pairs, and we then collected the top 20 heart failure disease genes from DisGeNET with ranked gene-disease association score ( Figure S1A). As a result, Cis-Cardio could locate the upstream candidate cis-regulatory elements and TFs of these disease genes (Figure S1B). For instance, HIF1A is the potential target gene of heart left ventricle H3K4me3 ChIP-seq peak "Cardio_01_1032818447," which is regulated by 53 TFs, including the known heart failure regulators SP1 and YY1. 32,33 In addition, we also listed the regulatory details of the HIF1A-related cis-regulatory region ( Figure S1C). Annotation results showed that this region had the enormous potential to regulate gene expression as super-enhancers and enhancers in multiple cells. Target genes of this region were also upregulated in cardiovascular diseases, especially in coronary artery disease, which is one of major causes of heart failure.

Interpretation of variations of coronary heart diseases
Previous studies have demonstrated that non-coding genetics variations could modify TF binding affinities on cis-regulatory elements and disturb distal-proximal element connections to participate in the processes of cardiovascular diseases. 34 Cis-Cardio collected a large set of variants and integrated abundant annotations to interpret the potential functions of these variants. SMAD3 is the crucial regulator of coronary heart disease. We mapped the SMAD3-related top variant "rs17293632" of coronary heart disease genome-wide association study (GWAS) data into Cis-Cardio, and the results showed that this variant was enriched in multiple cis-regulatory elements of coronary artery smooth muscle cell, which was consistent with previous research 35 (Figures S1D and S1E). In the element "Car-dio_01_1019927866" of JUND ChIP-seq data, some TFs located in the variant were identified, such as EGR1 and FOXD3. Furthermore, the peak detail page also presented the comprehensive TF binding information and peak annotation information for users to unveil the regulatory cues between variants and cis-regulatory elements.
User-friendly interface for browsing, searching, and downloading cis-regulatory data Cis-Cardio provides six analysis tools and a quick browse for retrieving the cis-regulatory data ( Figures 4A and 4B). On the browse page, users can query the factors of interest via a search box or filter the samples via a checkbox. Here, we used human samples as the filter and selected a sample of cardiac-specific TF GATA4 as an example to introduce the performance of Cis-Cardio. In the browse table, users can obtain the metadata of samples, such as cell type, cell name, data source, and dataset accession number. Users can click the data ID links to the data detail page. On the detail page, users can view the comprehensive annotations of GATA4, including GATA4 ChIPseq overview, GATA4 ChIP-seq peaks, peak annotations, and genomic peak annotation information, which are useful to help investigate the regulatory patterns of factors. For instance, the results showed that GATA4 preferred to bind distal regions, which is coincident with previous studies. 13 For the samples of transcription regulators, Cis-Cardio also provided the basic annotations for transcription regulators; here, we displayed the basic annotations for GATA4, including GATA4-related gene ontology/pathway annotation, expression, and disease information ( Figure 4C).
Importantly, users can click "Peak id" to obtain the details about each cis-regulatory region of interest. In the peak detail page, Cis-Cardio provided abundant detailed annotations for the cis-regulatory region of interest, including super-enhancer, enhancer, TF binding site, eQTL, common SNP, risk SNP, LD SNPs, DNase I hypersensitive sites, binding TFs predicted by motif, methylation sites of 450k array, methylation sites of whole-genome bisulfite sequencing, and 3D chromatin interactions. All the annotation information could help users to understand the primary causes of diseases and identify the potential synergetic regulatory axis for clinical therapeutics. Cis-Cardio pro-vided the predictive downstream genes of the cis-regulatory peak based on three methods. Target gene differential expression in cardiovascular diseases and function annotation were also embedded in the server, which could help users to screen the disease gene efficiently. Moreover, users can view the cis-regulatory region by genome browser and download the data of interest in the "Download" section (file descriptions are provided in the supplemental information).

DISCUSSION
Recently, more and more attention has been paid to investigate cardiovascular-specific gene transcription programs based on dissection of the communicating cues between cis-regulatory elements and transcription regulators. 10,36 With the development of high-throughput techniques, the volume of cardiovascular-related omics data has accumulated rapidly, especially epigenomics and transcriptome data. However, it remains a challenge to integrate and process the data from multiple perspectives. And it is necessary to develop web tools that comprised the regulation data and provided diverse analysis functions to meet the needs. Here, we developed the Cis-Cardio platform, which aims to document a large number of available resources of cardiovascular-related cis-regulatory data and to annotate and uncover the comprehensive mechanisms at the cis-regulation level. The current version of Cis-Cardio cataloged a total of 45,382,361 genomic regions from 1,013 human and mouse epigenetic datasets, including ATAC-seq, DNase-seq, Histone ChIP-seq, TF/TcoF ChIP-seq, RNA polymerase ChIP-seq, and Cohesin ChIP-seq. Moreover, to summarize the potential cardiovascular-related cis-regulatory regions, we merged all the cis-regulatory regions of each sample into an integrated region set with peak frequency using BEDTools ( Figure S2). As a result, 1,395,462 unique regions were merged and provided in the "Download" page of the server. Users can find the potential hot regions of interest based on the high peak frequencies. Cis-Cardio is the first resource for investigating the cardiovascular-related candidate cis-regulatory elements, with the largest human and mouse samples and the most comprehensive annotation information. We provide a convenient platform for researchers to explore regulated information about candidate cis-regulatory elements and candidate cis-regulatory element-associated regulatory analyses.  on the "Search" page. Cis-Cardio provides four query search methods for users to obtain the cis-regulatory data. (5) Cis-Cardio provides a user-friendly "Data-Browse" page. Users may further click on the "ID" to view candidate cis-regulatory elements for a given sample. (6) Cis-Cardio embeds a personalized genome browser with intuitive data visualization. Moreover, we also validated that Cis-Cardio has the ability to identify upstream regulators of cardiac marker genes and to identify important genome regions based on integration of epigenetics data.

Cis-Cardio
The current version of Cis-Cardio processed a large number of cardiovascular-specific candidate cis-regulatory elements. However, we also found that the elements from the transcription regulator were sparse. With the data increased, in the future updates, we will extend the scale of the candidate cis-regulatory elements from transcription regulator datasets and add more annotations. And we will develop a novel analysis tool to infer more transcription regulators. Cis-Cardio is a user-friendly platform to analyze, query, browse, and visualize information associated with cis-regulatory elements. We believe that Cis-Cardio could become a useful and effective platform for exploring potential functions and analyzing regulation of candidate cis-regulatory elements in cardiovascular diseases.

Cardiovascular-related cis-regulation datasets
Previous studies have released abundant high-throughput data to investigate cis-regulation in the field of cardiovascular disease, such as Histone ChIP-seq, ATAC-seq, scATAC-seq, DNase-seq, TF ChIP-seq, TcoF ChIP-seq, Cohesin ChIP-seq, and RNA polymerase ChIP-seq. In this study, we focused on dissecting these cis-regulation data and manually collected 1,013 cardiovascular-related cis-regulation datasets with binding peaks of human and mouse from ENCODE, ChIP-Atlas scEnhancer, and Cistrome. Next, liftOver (http://genome.ucsc.edu/cgi-bin/hgLiftOver) software was used to normalize and covert all the peaks into hg19 (human) and mm10 (mouse) genome versions. We also provided an online liftOver tool to convert genome coordinates from hg19 to hg38. All cis-regulatory regions from each sample were merged into the unique cis-regulatory regions by BEDTools with default parameters. All the biosample information is provided in Table S1.

Upstream annotations of cis-regulatory regions
In the current version of Cis-Cardio, we embedded a large number of (epi-) genetic annotations in cis-regulatory regions, such as superenhancer, enhancer, TFBS, methylation sites, common SNPs, risk SNPs, eQTLs, histone modifications, and 3D chromatin interactions. All the data sources are listed in Table S2.

Super-enhancer and enhancer
Cis-regulatory regions contain several types of DNA regulatory elements, including proximal promoters and distal enhancers. Notably, distal enhancers, especially super-enhancers, are considered to play prominent roles in driving cell-specific gene expression programs.
Here, to annotate the enhancers and super-enhancers within cardio-vascular cis-regulatory regions, we firstly collected the enhancer data from EnhancerAtlas, 37 HACER, 38 ENCODE, 16 FANTOM5, 39 DENDB, 40 and ENdb, 41 including 14,797,266 human enhancers and 439,092 mouse enhancers. Secondly, for super-enhancer annotations, we manually processed H3K27ac ChIP-seq data from ENCODE, Roadmap, NCBI GEO/SRA, and Genomics of Gene Regulation Project. In brief, we executed Bowtie software for each H3K27ac ChIP-seq profile and ran MACS software to call all active peaks. ROSE was used to identify super-enhancer regions. Moreover, we also collected super-enhancers of human and mouse from SEA 42 and dbSuper. 43 As a result, 2,678,273 human super-enhancers and 11,609 mouse super-enhancers were embedded in the current Cis-Cardio.

TF binding sites
Cis-regulatory regions have a strong ability to recruit TFs to exert regulatory functions for downstream genes. To uncover the specific TF binding events of each cis-regulatory region, we performed Find Individual Motif Occurrences pipeline to call the motif binding sites in these regions for $700 TFs. 44 In detail, we curated more than 3,000 DNA binding motifs from the TRANSFAC and MEME suite, which collected from JASPAR CORE 2020 vertebrates, Homeodomains, Jolma2013, UniPROBE, and Wei2010. Significant TF binding sites within regions were defined with an optimized p value threshold of 1e-6 from motif analysis. 19 As supplementary, we collected 5,547,656 human TFBSs and 2,858,356 mouse TFBSs from UCSC and performed BEDTools to reserve the intersected TFBSs within cis-regulatory regions.

SNPs/linkage disequilibrium SNPs/risk SNPs/eQTLs
Variants of the cis-regulatory regions determine the TF binding affinities and participate in the downstream gene transcription program.

Chromatin interaction/DHS/methylation
Emerging evidence has demonstrated that chromatin marks can help to uncover the regulatory effects and mechanisms between cis-regulatory regions, and downstream genes mark interaction data, such as DNA chromatin interaction, DNase activity, and DNA methylation states. In the current server, we downloaded the chromatin interaction data from 4DGenome and Oncobase, which included data from ChIA-PET 3C, 4C, 5C, and Hi-C. DHS annotation data of cisregulatory regions were downloaded from UCSC and ENCODE. In total, 69,860,705 human DHSs of 293 samples and 9,802,229 mouse DHSs of 56 samples were obtained. In addition, we also obtained DNA methylation states of 30,392,523 methylation sites of 450k array and 166,855,665 methylation sites of whole-genome shotgun bisulfite sequencing from ENCODE.

Functional annotations of TFs and TcoFs
To characterize the biological functions of TFs and TcoFs, we provided more annotation information, including TF/TcoF-mediated pathway, gene ontology, expression, and disease from multiple sources. In brief, we obtained TF/TcoF expression profiles from NCBI, GTEx, ENCODE, and FANTOM5. The experimentally validated TF/TcoF-disease relationships were obtained from DisGeNET, GAD, and MGI. Moreover, TF/TcoF-related pathways were downloaded from our previous study ComPAT, which curated 2,169 human and mouse pathways from 10 resources, including KEGG, Reactome, NetPath, WikiPathways, PANTHER, PID, HumanCyc, CTD, SMPDB, and INOH. The pathway gene set is provided in Table S3.
Intersection analysis between genomic regions was performed by BEDTools with the command: bedtools intersect -a query.bed -b annotation.bed -f 1E-9 -wa -wb -bed -u > result.bed.

Cis-regulatory region-related downstream target genes
To comprehensively characterize the regulatory details of cis-regulatory regions, we embedded multiple methods to identify the downstream target genes of each region. For all cis-regulatory data in bed format files, we firstly used a python script (ROSE gen-eMapper.py) to annotate cis-regulatory region-related target genes.
Notably, target genes of three strategies of ROSE (overlap, proximal, and closest) were merged. Secondly, we also used binding and expression target analysis minus (BETA minus) to identify cis-regulatory region downstream genes. 48 Importantly, we also embedded the Activity-by-Contact model to optimize the target genes of cis-regulatory elements, which is a high-confidence method by integrating histone modification and chromatin contact. 49 Target genes from the above methods were integrated and used for further analysis.

Differentially expressed gene annotations of cis-regulatory regions
To investigate the regulatory axis of cis-regulatory regions in multiple cardiovascular diseases, we have collected and processed gene expression profiles and identified gene differential expression information of eight major cardiovascular diseases, including heart failure, hypertrophic/dilated cardiomyopathy, myocardial infarction, coronary artery disease, cardiac hypertrophy, and pulmonary arterial hypertension. Briefly, we obtained the original expression profiles from GEO supplementary tables. Then we divided the samples into two groups (control group and disease group) based on the sample label. Expression data of microarrays were performed with log2 transform function, and then we identified gene differential information via SAM test, which was a non-parameter test for differential gene analysis. RNA-seq data were processed by DEseq2 with raw count matrix. p values were adjusted via false discovery rate method.

Novel online analysis tools for deciphering regulatory cues of cardiovascular diseases
Dysfunction of gene-mediated downstream regulatory axes is considered the cause and potential therapeutic target of multiple cardiovascular diseases. Gene expression is determined by upstream transcription regulation programs, such as SNP, transcription regulator activity, and DNA regulatory elements states. To help biologists investigate the regulatory mechanisms of cardiovascular diseases in the aspect of transcriptional regulation, we integrated multi-omics data and developed six online analysis tools as follows.

Genomic region intersection analysis
Genomic region intersection analysis tool could identify the cis-regulatory regions that overlapped with the user's genomic regions of interest. Briefly, users can upload a "bed" format file of genomic regions or a region list, choose the species (human or mouse), and set overlap size (intersection size ratio between input regions and background regions) to identify potential cis-regulatory regions that locate at the similar genomic regions of input regions. Cis-Cardio will display all the cis-regulatory regions that overlap with the input regions, and users can also obtain detailed annotations of the overlapped regions, such as cis-regulation type, peak information, target gene, and cell type information.

Upstream regulatory axes analysis
Upstream regulatory axes analysis tool was developed for identification of upstream regulatory mechanisms for genes of interest, and it can list the input genes related cis-regulatory regions and binding TFs. Users can submit a gene list to Cis-Cardio to map the target genes of all candidate cis-regulatory elements. If the submitted genes are the target genes of a cis-regulatory element, Cis-Cardio will extract the upstream regulatory TFs of the cis-regulatory element and form the TFs/candidate cis-regulatory elements/submitted genes regulatory axes.

Downstream regulatory pathway analysis
TFs are usually located at the terminal of the signal pathways. Users can submit a gene list, and Cis-Cardio will identify enriched pathways in up to 10 pathway databases via a hypergeometric test. The p value of the enriched pathway was measured as follows: Here, m represents the total number of genes in all pathways, t represents the number of input genes, n represents the number of genes of each pathway, and r represents the number of overlap genes between input genes and each pathway gene.
In each pathway, Cis-Cardio will locate the terminal TFs and extract the terminal TF-bound downstream candidate cis-regulatory elements. Therefore, users can find the submitted genes/pathway/ TFs/candidate cis-regulatory element regulatory axes using this analysis tool.

Transcription regulator enrichment analysis
Transcription regulator enrichment analysis tool provides an enrichment function to find the upstream regulators of input genes based on epigenomics data. Users can submit a gene list to Cis-Cardio to identify the upstream transcription regulators and epigenetics factors based on a hypergeometric test. The background TR-target gene pairs were constructed from the data based on two target gene assignment strategies. Cis-Cardio will display all the datasets that overlap with the input genes via the hypergeometric test and Jaccard index. Furthermore, we also provided the genome region-based transcription regulator enrichment analysis via LOLA. Users can submit regions of interest to enrich the transcription regulators.

Cis-regulatory variants interpretation analysis
Variants that locate at the cis-regulatory regions determine the binding affinity of TFs and participate in the gene expression programs.
Cis-regulatory variants interpretation analysis can quickly map cardiovascular-related candidate cis-regulatory elements and TF binding sites that contain variants of interest. Users can submit a variant name (such as rs10817286) to Cis-Cardio to extract the downstream peaks that locate the variant. Previous studies revealed that variants determine the TF binding ability in DNA regulatory elements. 34 Here, users can obtain the TF binding sites in cell-specific peaks via motif analysis.

PPI-based co-regulatory analysis
It is important to decipher the regulatory axes of the regulatory proteins that are located in the cell nucleus, as they can exert functions by regulating transcription regulators. Here, users can submit a gene list (nuclear protein of interest) to find the direct TFs based on a proteinprotein interaction (PPI) network. Users can also obtain the network topological importance score of the input genes and the downstream regulatory data of the interactive transcription regulators.  50 We recommend using a modern web browser that supports the HTML5 standard, such as Firefox, Google Chrome, Safari, Opera, or IE 9.0+, for the best display.

DATA AND CODE AVAILABILITY
All data supporting the findings of this study are available within the paper and online Cis-Cardio server (http://www.licpathway.net/Cis-Cardio/index.html).