The dark genome in cardiovascular medicine Authors: Despoina Kesidou, Simon D Brown, Lars Maegdefessel, Igor Ulitsky, Andrew H Baker Published: 2026 ABSTRACT Graphical Abstract Graphical AbstractThe dark genome landscape and therapeutic targeting in cardiovascular disease. Schematic overview of functional components within the dark genome. (Left) Non-coding RNAs, transposable elements, regulatory DNA regions (promoters, enhancers), introns and intergenic regions, pseudogenes and structural elements (telomeres, centromeres) modulate cardiovascular homeostasis. (Centre) Non-coding RNA species, including microRNAs, long non-coding RNAs (lncRNAs), circular RNAs (circRNAs), and other non-coding RNA species, serve as regulators modulating molecular processes at the transcriptional, post-transcriptional, translational and post-translational level. (Right) Pathological dysregulation of these dark genome elements drives cardiovascular disease progression, which can be therapeutically targeted to restore cardiovascular function. Ultimately, as a greater understanding of the dark genome emerges, novel target identification will enable a new class of advanced cardiovascular therapeutics to arrest or reverse disease progression as well as creating a wealth of new information.For image description, please refer to the figure legend and surrounding text. Keywords: Dark genome, Non-coding RNAs (ncRNAs), Cardiovascular disease, Transcriptomics, RNA therapeutics, Translational cardiovascular medicine Introduction—What is the ‘dark genome’? The genome contains the genetic blueprint required for an organism’s development and physiology. Protein-coding genes were traditionally considered the principal cellular mediators, and their importance in cardiovascular health is well-established.1 However, the genetic sequence corresponding to protein-coding genes accounts for only ∼1%–2% of the human genome.2 Previously, the remaining 98%–99% was considered non-functional or ‘junk’ DNA.3 Modern genomic studies have since revealed that biological complexity is instead encoded within non-coding regulatory elements and their transcriptional output.4–6 The non-protein-coding DNA sequence of the genome consists of overlapping genomic elements collectively termed the ‘dark genome’ (Figure 1): Make-up of the dark genome within the mammalian nucleus. Schematic overview of the major components of the dark genome and their organization within the mammalian nucleus. The dark genome comprises regulatory DNA elements (enhancers and promoters), structural elements (centromeres and telomeres), intergenic regions, introns and untranslated regions, transposable elements, and non-protein-coding genes that give rise to regulatory non-coding RNAs Regulatory DNA elements (enhancers, silencers, and promoters) Regulatory DNA elements (enhancers, silencers, and promoters) Structural elements (centromeres and telomeres) Structural elements (centromeres and telomeres) Introns Introns Transposable elements Transposable elements Pseudogenes Pseudogenes Intergenic regions Intergenic regions Non-protein-coding genes Non-protein-coding genes In this review, we discuss these components of the dark genome with emerging relevance to cardiovascular health and disease, summarizing experimental and computational approaches used for characterization, key methodological challenges, and future priorities for exploiting dark genome (patho)biology in cardiovascular research and medicine. A glossary of key terms, with definitions and relevance to the dark genome, is provided in Supplementary data online, Table S1. Dark genome regulatory layers in cardiovascular disease In this section, we discuss major regulatory layers of the dark genome that are increasingly implicated in cardiovascular biology, including transposable elements, regulatory DNA elements, chromatin organization and structural features, pseudogenes, and RNA processing and splicing. While cardiovascular evidence varies, collective evidence illustrates the diverse ways in which these components can influence cardiovascular homeostasis and disease. Regulatory DNA elements Enhancers and promoters are DNA elements that govern spatiotemporal control of gene expression. Promoters are DNA regions close to a gene’s start site that facilitate transcription initiation. Enhancers are regulatory DNA elements located distal to the transcription start site and increase gene transcription by bringing the enhancer and promoter into contact and recruiting transcription factors. In the healthy heart and vasculature, precise combinations of active enhancers and promoters define the gene expression programmes that establish and maintain cell identity.7 Variants in these regulatory regions are now recognized as central contributors to human cardiovascular disease. Genome-wide profiling of human heart tissue revealed thousands of active promoters and enhancers with region-specific activity patterns, many of which differ between healthy and failing hearts.8 These observations highlight that changes in gene regulation are a feature of cardiac disease. In heart failure, ventricular tissue shows changes in gene regulation, including shifts in promoter use and enhancer activity, which are linked to altered expression of genes involved in contractile function and metabolism.9 Consistent with this regulatory mechanism contributing to inter-individual variation, common genetic variants in regulatory DNA regions have been linked directly to cardiomyopathy phenotypes. Gacita et al. (2021) reported that non-coding variants within enhancer elements controlling cardiomyopathy genes, including MYH7 and LMNA, modified cardiomyocyte gene expression and correlated with disease progression in human cohorts and engineered human cardiomyocytes (Figure 2A).11 This study illustrates that enhancer–promoter variation influences gene regulation in cardiovascular tissues and contributes to disease susceptibility, progression, and phenotypic heterogeneity. Regulatory elements of the dark genome: enhancers and higher-order chromatin structure. (A) Enhancer–promoter interactions. Transcription of the MYH6/MYH7 tandem is regulated through the action of multiple enhancer sequences (C1–C6). Genetic variation in these enhancer sequences modifies MYH6/MYH7 gene expression, leading to a dose-dependent increase in MYH6, a faster contractile rate in engineered cardiac tissues, and is associated with cardiomyopathy progression.9 (B) Higher-order, 3D chromatin organization is governed via topologically associated domains (TADs), which ensure correct gene expression patterns within chromosome territories. Mutations within TAD sequences can result in aberrant enhancer–promoter interactions and altered gene expression patterns, leading to diseases, such as cardiac laminopathies.10 Methods to identify regulatory elements and assess their functionality Several complementary approaches are used to identify and quantify regulatory DNA element activity. Chromatin accessibility assays, such as assay for transposase-accessible chromatin sequencing, identify genomic regions that are open and transcriptionally active, highlighting likely enhancers and promoters.12 Chromatin immunoprecipitation sequencing can then help distinguish these elements based on characteristic histone marks associated with transcriptional activity.13 In parallel, cap analysis of gene expression sequencing maps transcription start sites and promoter usage at high resolution.14 These methods enable construction of regulatory maps; these techniques are now being applied to cardiovascular tissues to help define how enhancer and promoter activity is altered in disease. Chromatin organization and structural features Beyond discrete regulatory elements, the dark genome also includes structural and organizational features that shape how the genome functions in space and time. The physical organization of the genome within the nucleus adds a further layer of regulatory complexity as enhancers and promoters communicate through three-dimensional chromatin looping, and this spatial organization is constrained by topologically associating domains (TADs): regions of the genome that preferentially interact with each other and are insulated from neighbouring domains.15 TAD boundaries help ensure that enhancers activate the correct target genes; when these boundaries are disrupted, enhancers can aberrantly activate genes in neighbouring domains, with potential pathological consequences.15 Alterations in three-dimensional chromatin organization have been described in cardiovascular diseases, where widespread changes in enhancer–promoter contacts accompany transcriptional reprogramming (Figure 2B).10 Other structural components of the dark genome, including centromeres, telomeres, satellite DNA, and tandem repeats, are now increasingly recognized as functionally important to cell function. Telomeres are repetitive DNA sequences at chromosome ends that protect chromosomes from degradation and fusion, and telomere shortening has well-established links to vascular ageing, cellular senescence, atherosclerosis, and cardiovascular risk (Figure 3A).16,19 In contrast, centromeres, which are specialized chromosomal regions required for accurate chromosome segregation during cell division, and satellite DNA, which consists of repetitive sequences found in centromeric and nearby pericentromeric regions, remain less well studied in cardiovascular disease.20 These regions are sources of active transcription rather than inert repeats.20 In cancer, e.g. dysregulated satellite transcription has been linked to chromatin instability, abnormal chromosome segregation, and stress responses,21 raising the possibility that similar mechanisms may also be relevant in cardiovascular pathology. This remains an important area for future investigation. Regulatory element of the dark genome: chromosome structure and alternative splicing. (A) Overview of human chromosome structure, with key structural elements such as centromeres and telomeres annotated. Various factors such as age, oxidative stress, and chronic inflammation can result in progressive telomere shortening. Cells with shortened telomeres exit the cell cycle and enter senescence, which is associated with the secretion of a variety of molecules, collectively known as the senescence-associated secretory phenotype, and can become apoptotic. While conclusive studies remain to be published, some studies suggest telomere shortening may lead to an increased risk of cardiovascular disease.16 (B) Alternative splicing is a process by which different protein isoforms can be generated through the differential splicing of exons. This is governed, in part, by splice regulatory elements present within introns. Mutations in TTN introns have been shown to lead to splicing errors resulting in truncated Titin isoform expression.17 These can lead to reduced sarcomere elasticity and dilated cardiomyopathy.18 RNA processing: alternative splicing Alternative splicing joins different exon combinations from a single precursor RNA, allowing one gene to generate multiple transcripts and protein isoforms. This process is regulated in part by non-coding intronic sequences containing splicing-regulatory elements, which help control exon inclusion or exclusion and thereby shape isoform expression in human tissues.22–24 Mutations in intronic sequences can generate pathological splice variants and have been implicated in conditions including dilated cardiomyopathy,18 heart arrhythmia in myotonic dystrophy patients,25 hypertrophic cardiomyopathy,26 and hereditary long QT syndrome.27 This principle is illustrated by intronic variants in the TTN gene, which can disrupt healthy, physiological splicing, creating truncation variants of the Titin protein, causing reduced sarcomere elasticity and dilated cardiomyopathy (Figure 3B),17 as well as variants in LMNA and MYBPC3, which have been linked to aberrant splicing in cardiomyopathy and cardiac hypertrophy/heart failure.28,29 These examples highlight that the dark genome contributes to cardiovascular disease not only through regulatory DNA elements, but also through non-coding sequence motifs that govern RNA maturation. Transposable elements Transposable elements represent the single largest category of genomic sequence in humans (∼45%).30 Transposable elements are mobile DNA sequences, mostly derived from ancient viral insertions or related to them, that can propagate throughout the genome via a ‘copy-and-paste’ mechanism, called retrotransposition. Through this action, they can influence gene regulation, chromatin structure, and genomic stability by introducing or disrupting regulatory elements.31 Transposable elements are broadly sub-classified based on the presence or absence of long terminal repeats (LTRs), which resemble retroviruses, and non-LTR elements. The latter includes long interspersed nuclear elements (LINEs), which are autonomous, and short interspersed nuclear elements (SINEs), such as Alu elements, which rely on LINE machinery for mobility.32 The expression of most transposable elements is tightly constrained by multiple molecular processes that often differ between germline and somatic cells. Despite this high level of transcriptional control, transposable element expression and function are increasingly recognized in cardiovascular disease (Figure 4A). Vascular calcification is a key feature of atherosclerosis and is associated with common cardiovascular conditions. It has been shown that expression of the largest and most prevalent retrotransposon, LINE-1, is increased in calcified human arteries,33 and that silencing LINE-1 or inhibiting its activity can attenuate calcification and pro-inflammatory signalling.33 These results suggest that transposable element activation contributes to calcific pathology in cardiovascular disease. Transposable element-derived sequences can also be assimilated into cardiovascular gene networks; e.g. a transposable element-derived exon within the CARMN locus that contains miRNAs and a host gene long non-coding RNA (lncRNA), has been reported to contribute to differentiation of cardiac precursor cells into vascular smooth muscle cells.34 Increased expression of another class of transposable elements, endogenous retroviruses, has also been reported in patients with dilated cardiomyopathy.38 These observations suggest that transposable elements can influence cardiovascular biology both through pathological activation of transposition and through evolutionary incorporation into regulatory programmes. Transposable elements and pseudogenes as regulators of cardiovascular disease. (A) Transposable elements. Top, schematic overview showing how transposable element sequences can influence genomic organization and gene regulation. Bottom left, increased LINE-1 activation promotes transcription, cDNA generation, and retrotransposition, and has been linked to vascular calcification and vascular remodelling.33 Bottom right, transposable element-derived exon incorporation into the CARMN locus illustrates evolutionary co-option into a cardiovascular regulatory programme linked to differentiation from cardiac precursor cells to vascular smooth muscle cells.34 (B) Pseudogenes. Top: defining features of pseudogenes. Bottom: representative examples of cardiovascular relevance. Left: variation at the APOOP1 pseudogene locus has been associated with altered LDL cholesterol and atherosclerosis.35 Middle: ZNF542P expression has been linked to intracellular cholesterol levels and variable LDL-cholesterol response to simvastatin.36 Right: PTENP1 acts as a competing endogenous RNA that reduces miRNA-mediated repression of PTEN, thereby influencing pathways relevant to atherosclerosis and vascular fibrosis.37 Pseudogenes Pseudogenes are genomic sequences derived from ancient protein-coding genes that generally no longer encode functional proteins, most commonly due to disruptive mutations or loss of regulatory elements,39 and long considered non-functional relics.40 However, increasing evidence now shows that some pseudogenes are transcribed and can regulate gene expression in multiple disease contexts.35,36 Cardiovascular examples for functional pseudogenes are beginning to emerge (Figure 4B). In an Amish population, variation at the pseudogene APOOP1 was associated with increased circulating low-density lipoprotein (LDL) cholesterol, and functional studies in zebrafish supported a role in hypercholesterolaemia and vascular plaque formation.35 In a separate study, expression of the pseudogene ZNF542P was associated with intracellular cholesterol levels and variability in LDL-cholesterol response to simvastatin, suggesting a role in treatment response.36 Another example is PTENP1, a pseudogene of PTEN gene, which has been implicated in smooth muscle cell proliferation,37 a process relevant to atherosclerosis41 and vascular fibrosis.42 Together, these observations suggest that pseudogenes may contribute to cardiovascular phenotypes through regulatory, rather than protein-coding, functions. Non-coding RNAs: essential regulators of cell function and dysfunction Non-coding RNAs (ncRNAs) represent a major transcriptional output of the dark genome and an increasingly tractable regulatory layer that is now a major research focus. Significant advances in transcriptomic profiling and functional genomics have expanded ncRNA annotation across cardiovascular tissues and disease states, strengthening links between ncRNA regulation and cardiovascular phenotypes. Because many ncRNAs show marked cell type- and context-specific expression, they are attractive candidates for mechanistic investigation and therapeutic modulation. Together, these RNAs form a major interface through which the dark genome shapes gene-regulatory networks relevant to cardiovascular health and disease. Classes of non-coding RNAs and their genomic prevalence Non-coding RNA classification usually combines transcript length with the biogenesis/processing to produce such transcripts. There are two major subtypes: small non-coding RNAs (typically <500 nucleotides) and long non-coding RNAs (>500 nucleotides).43 In the current GENCODE human genome annotation,44 78 691 distinct genomic locations or ‘gene loci’ are catalogued. Of these, 20 097 are protein–coding, whereas the remainder comprise non-coding RNA genes (35 899 long non-coding RNA genes and 7563 small non-coding RNA genes), pseudogenes (14 701), and immunoglobulin/T-cell receptor gene segments (649) (Figure 5A).44 Genome annotation remains an active area, and continued improvements in sequencing technologies and functional genomics are expected to refine gene loci and non-coding RNA catalogues over time.77 Non-coding RNAs are broadly categorized into small non-coding RNAs, including miRNAs, piRNAs, snoRNAs, Y RNAs, and tsRNAs, and long non-coding RNAs, including lncRNAs and circRNAs, based on transcript length and biogenesis (Figure 5B). Genomically, non-coding RNA loci can be located between protein-coding genes (intergenic), within introns or exons of other genes (intragenic), or on the opposite strand to a protein-coding gene (antisense), each configuration with distinct implications for their regulation and function (Figure 5C). While the translational evidence remains variable across non-coding RNA classes, several have already been implicated in major cardiovascular disease processes (Figure 5D), and representative therapeutic strategies currently under preclinical or clinical evaluation are summarized in Supplementary data online, Table S2. Genomic organization and structure of non-coding RNAs in the human genome. (A) Composition of human gene loci based on current GENCODE annotation (version 49). (B) Major classes of non-coding RNAs are broadly categorized into long non-coding RNAs (lncRNAs and circRNAs) and small non-coding RNAs (including miRNAs, snoRNAs, yRNAs, tsRNAs, and piRNAs). (C) Genomic organization of non-coding RNA loci, illustrating intergenic, intragenic (intronic and exonic), and antisense transcription relative to protein-coding genes. (D) Overview of key non-coding RNA classes and exemplars associated with major cardiovascular disease and disease-relevant processes, including atherosclerosis and cardiac diseases,45–51 cardiomyocyte hypertrophy and survival,52–59 vascular remodelling,60–63 apoptosis/senescence,64–67 inflammation,68–71 extracellular matrix regulation and fibrosis.72–76 Small non-coding RNAs in cardiovascular disease Small non-coding RNAs comprise several classes, including miRNAs, small nucleolar RNAs (snoRNAs), small nuclear RNAs (snRNAs), tRNA-derived small RNAs (tsRNAs), and Y RNAs. MicroRNAs in cardiovascular disease MiRNAs are short (∼22-nucleotide) RNAs that bind to target messenger RNAs and typically repress gene expression by reducing their stability and translation.78 Their biogenesis, expression patterns, and cardiovascular roles have been comprehensively reviewed elsewhere79,80; an overview of canonical miRNA biogenesis and therapeutic modalities is summarized in Figure 6. Although several miRNAs have shown promise in preclinical cardiovascular studies, few have advanced to clinical trials (see Supplementary data online, Table S2). Here we briefly focus on the most clinically advanced examples, and readers are referred to a recent comprehensive review in the field for in-depth detail of miRNAs in cardiovascular research.84 The role and mechanism of function of microRNAs in cardiovascular disease. (A) Canonical miRNA biogenesis and mechanism of action. (B) Clinical translation of miRNA-based therapies, illustrated by CDR132L (anti-miR-132-3p) for heart failure with reduced ejection fraction81,82 and MRG110 (anti-miR-92a-3p) for ischaemia-associated angiogenic indications.83 Modulation of miRNA function has progressed through two main strategies: inhibition using antisense oligonucleotides (ASOs, anti-miRs) and restoration using synthetic mimics. As an example, CDR132L is an anti-miR targeting miR-132-3p, which acts by de-repressing FOXO3 and SERCA2A, promoting improved contractility and reduced cardiomyocyte hypertrophy (Figure 6B).52,81 In a Phase Ib trial, CDR132L demonstrated good tolerability and a median 23.3% N-terminal pro-B-type natriuretic peptide (NT-proBNP) reduction,82 and is currently under evaluation in the HF-REVERT trial (NCT05350969) in patients with reduced ejection fraction after myocardial infarction.81 MRG110 targets the anti-angiogenic miR-92a-3p, de-repressing targets including ITGA5, SIRT1, and KLF2/KLF4 to promote angiogenesis and wound healing (Figure 6B).83 A first-in-human study demonstrated sustained reduction of circulating miR-92a-3p,83 with a further Phase I trial ongoing (NCT03494712); given that miR-92a-3p inhibition reduces atherosclerosis in preclinical models,85 these results are of cardiovascular interest. MiRNA mimics offer a complementary approach to restore miRNA activity. Unbiased mimic screens have identified miRNAs capable of promoting cardiac regeneration,86 and mimic delivery has shown cardioprotective effects in preclinical ischaemia models.87 Emerging delivery strategies using lipid nanoparticle and modified RNA-based approaches are resolving delivery challenges, yet this is an area of key focus to elevate safety and efficacy.88,89 These early-phase studies support the feasibility of miRNA-targeted therapeutics, though definitive evaluation of clinical benefit will require adequately powered cardiovascular trials. Beyond microRNAs: emerging roles for SnoRNAs, Y RNA fragments, TsRNAs, and PiRNAs Although miRNAs are the best characterized small non-coding RNAs, evidence indicates that other small non-coding RNA classes contribute to cardiovascular phenotypes. A greater understanding of these is crucial to reveal novel pathways and to identify therapeutic targets. Other types of small non-coding RNAs, include snoRNAs, Y RNA-derived fragments, tsRNAs, and piRNAs (see Supplementary data online, Table S2, Figure 7). Evidence for these in cardiovascular disease is less mature; however, for some classes, including snoRNAs, mechanistic links are emerging. SnoRNAs guide the modification of other RNAs (most commonly ribosomal RNAs and snRNAs) (Figure 7A). In the context of vascular remodelling, snoRNAs, including SNORD113-6/AF357425 have been shown to modulate signalling pathways which are important for arterial fibroblast function.72 In parallel, broader human studies report co-ordinated regulation of multiple snoRNAs in diseased vascular tissue, supporting a wider contribution to cardiovascular pathology.90 Mechanisms of action of selected small non-coding RNAs in cardiovascular biology. (A) Example of small nucleolar RNA (snoRNA) mechanism of function. SnoRNAs guide chemical modification of ribosomal and small nuclear RNAs (snRNAs) within snoRNA–protein complexes and can generate snoRNA-derived fragments with regulatory functions in disease contexts. (B) Example of y RNA mechanism of action. Y RNAs can be packaged within extracellular vesicles (EVs) and transferred between cells, modulating immune responses and cardioprotection following injury. The figure also represents an alternative therapeutic modality in which yRNAs can be loaded into lipid nanoparticles and delivered independently to EVs.68,69 C. Example of PIWI-interacting RNA (piRNA) mechanism of function. PiRNAs bind PIWI proteins to regulate gene expression through post-transcriptional silencing and chromatin-associated repression, with emerging roles in cardiomyocyte hypertrophy and survival Y-RNA fragments are another class of small non-coding RNAs. Although their functions were unclear for years, they have been linked to RNA quality control, proliferation, and DNA replication (Figure 7B).91 In the context of cardiac injury, Marbán and colleagues identified a short Y RNA fragment (EV-YF1) within extracellular vesicles (EVs) from cardiac cells.68 EVs are small particles released by cells that carry biological cargo, such as small non-coding RNAs, and their composition can change in disease. In ischaemia–reperfusion models, EVs containing EV-YF1 were reported to signal to immune cells and increase the expression of anti-inflammatory cytokines, conferring cardioprotection.68 Importantly, this observation has suggested a novel therapeutic approach: synthetic Y RNAs can be incorporated within lipid nanoparticles, enabling mass production of advanced therapeutics for cardiovascular disease.69 Recent work has extended this paradigm by identifying additional EV-enriched Y RNA species with therapeutic potential, including yREX3, which exerts cardioprotection.69 Another class of small ncRNAs, tsRNAs are generated from transfer RNAs (tRNAs) that normally deliver amino acids to the ribosome during protein synthesis.92 TsRNAs are increasingly recognized as stress-responsive regulators in the cardiovascular system, with multiple fragments detected in models of cardiac hypertrophy, ischaemia–reperfusion injury, pulmonary vascular disease, and mitochondrial cardiomyopathy.93 However, functional studies remain relatively limited and confined to experimental settings. Finally, piRNAs are a class of small RNAs that bind PIWI proteins (Argonaute-family RNA-binding proteins that partner with small RNAs) to regulate gene expression (Figure 7C). They are mainly found in reproductive tissues and are best known for silencing transposable elements, but can also affect RNA stability and gene regulation at the chromatin level.94 In cardiovascular disease, a small number of studies have suggested roles for specific cardiac piRNAs, including CHAPIR (linked to pathological hypertrophy)53 and HAAPIR (linked to cardiomyocyte apoptosis).64 However, the evidence remains limited. More broadly, other small non-coding RNA species, including vault RNAs,95 7SL/SRP RNA-derived fragments96 and 7SK RNA,97 as well as rRNA-derived fragments, remain poorly characterized in cardiovascular health and disease. Improved annotation, cell type-resolved profiling, and functional perturbation studies are therefore needed to define which of these RNAs are causal drivers, biomarkers, or therapeutic targets in cardiovascular disease. Such efforts have the potential to expand the repertoire of regulatory pathways implicated in cardiovascular disease and to identify new candidates for biomarker development and therapeutic modulation. Long non-coding RNAs in cardiovascular disease Whereas small non-coding RNAs primarily act through post-transcriptional mechanisms, long non-coding RNAs add an additional regulatory layer that can shape transcriptional programmes, chromatin state, and cell phenotype (Figure 8A). The long non-coding RNA category comprises long non-coding RNAs (lncRNAs) and circular RNAs (circRNAs), two abundant classes with recognized roles in cardiovascular health and disease (see Supplementary data online, Table S2).43,98–100 Although the functions of cardiovascular lncRNAs have been comprehensively reviewed elsewhere,101,102 here we highlight several well-defined lncRNAs and briefly discuss how these contribute to cardiovascular disease-relevant pathways (Figure 8B and C). In vascular smooth muscle cells (vSMCs), CARMN promotes a contractile identity and stabilizes the vSMC phenotype,45 while the lncRNA SMILR becomes expressed in response to proliferative stimuli preferentially in vSMCs, where it acts to accelerate cell proliferation, a pathogenic phenotype of several cardiovascular diseases (Figure 8B).60–62 LncRNAs involved in matrix remodelling, such as the cardiac fibroblast-enriched Wisper, orchestrate myofibroblast activation and extracellular matrix deposition (Figure 8C).73 Together, these representative lncRNAs highlight how long non-coding transcription integrates multiple regulatory modalities to shape cardiovascular disease progression. Regulatory mechanisms of long non-coding RNAs in cardiovascular disease. (A) Canonical Long non-coding RNA (lncRNA) biogenesis and mechanism of action. (B) The lncRNA SMILR becomes expressed in vascular smooth muscle cells (vSMCs) in response to vascular damage, pro-inflammatory stimuli, or growth factors. SMILR binds to the mRNA of the mitotic gene CENPF, where it acts to shield CENPF from Stauffen1 (Stau1)-mediated degradation, thereby promoting vSMC proliferation.60–62 (C) The lncRNA WISPER is upregulated in cardiac fibroblasts from infarcted hearts and promotes cardiac fibrosis and is very selective for cardiac fibroblasts. An antisense oligonucleotide (ASO), HTX-001, decreases WISPER levels, reduces cardiac fibrosis, and improves cardiac function.73 Therapeutic targeting of lncRNAs Substantial experimental evidence now demonstrates that blocking or enhancing the activity of lncRNAs may offer therapeutic benefit across many cardiovascular pathologies. Short interfering RNA (siRNA) and short hairpin RNA (shRNA) have been developed to reduce gene function primarily based on the naturally occurring mechanisms of RNA interference (RNAi), and these same technologies are now also being applied to target lncRNAs. A discussion of the mechanisms of RNAi is beyond the scope of this review and has been reviewed comprehensively.103 There is currently a scarcity of lncRNA-targeting siRNA in clinical studies. This is partly due to the relatively recent appreciation of lncRNAs and their role in disease, and due to poor cross-species conservation compared to protein-coding genes, which impedes pre-clinical studies in animal models. Despite these challenges, a greater cell-specific expression pattern of lncRNAs makes them attractive therapeutic targets. Circular RNAs in cardiovascular disease Circular RNAs (circRNAs) are covalently closed, single-stranded RNAs that can be generated from both protein-coding and non-coding genes and form stable RNA circles. CircRNAs have been implicated in regulatory roles, including interaction with RNA-binding proteins, modulation of transcription, and, in some cases, regulation of miRNA activity (Figure 9A). Their stability and tissue-specific expression patterns have made them attractive candidates as cardiovascular biomarkers and potential therapeutic targets.105 Regulatory mechanisms of circular RNAs in cardiovascular disease. (A) Canonical circular RNA (circRNA) biogenesis and mechanisms of action, including acting as microRNA (miRNA) sponge, RNA binding protein (RBP) sponge, and roles in translation process and protein recruitment. (B) CircANRIL is reported to bind to PES1, a component of the PeBow complex, essential for ribosome biogenesis. By binding to PES1, circANRIL prevents ribosomal RNA (rRNA) processing, leading to reduced ribosome biogenesis. Reduced ribosome biogenesis triggers the increased expression of p53, a master regulator of the DNA damage and apoptosis response. This leads to increased apoptosis and reduced proliferation in vascular smooth muscle cells (vSMCs).104 C. Overexpression of the circRNA CircSlc8a1 accelerates cardiac hypertrophy and heart failure in a pressure overload heart model. Adeno-associated virus (AAV9)-mediated RNAi knockdown of CircSlc8a1 using a short hairpin RNA (shRNA) successfully attenuates cardiac hypertrophy.54 Within cardiomyocytes, several circRNAs have been shown to influence stress adaptation and pathological growth. CircRNAs participate in metabolic and inflammatory injury. For example, reduced expression of circRNA DICAR drives cardiomyocyte pyroptosis in diabetic cardiomyopathy,65 and vascular circRNAs such as circANRIL shape proliferative and inflammatory responses linked to atherosclerosis (Figure 9B).104 CircSlc8a1 modulates hypertrophic signalling (Figure 9C)54 while circFOXO3 promotes cellular senescence through interaction with proteins that normally restrain stress responses and senescence pathways.66 CircRNAs also operate in the vascular compartment, where endothelial transcripts such as circHIPK3 support angiogenic competence and vascular repair.106 Together, these mechanistically defined examples illustrate how circular transcript architecture enables diverse regulatory roles across cardiac and vascular compartments. Network-level regulation by non-coding RNAs in cardiovascular disease Non-coding RNAs act within interconnected regulatory networks in which multiple RNA classes converge to shape gene-expression programmes relevant to cardiovascular disease. Rather than functioning as isolated regulators, miRNAs, lncRNAs, and circRNAs frequently participate in shared circuits, including competing endogenous RNA networks in which lncRNAs or circRNAs modulate miRNA availability and thereby influence downstream mRNA targets, though the mechanistic underpinnings of such networks that can account for the relative stoichiometries of messenger RNAs and lncRNAs remain unclear.79,107 Such lncRNA–miRNA–mRNA and circRNA–miRNA–mRNA interactions have been described in myocardial infarction, heart failure, and atherosclerosis, where they affect processes such as cardiomyocyte survival, fibroblast activation, and vascular inflammation.108 Beyond these networks, snoRNA-derived fragments and tRNA-derived small RNAs also intersect with miRNA- and lncRNA-centred regulatory modules,109–112 although the overall prevalence and impact of such interactions in cardiovascular disease remain poorly characterized.113 At the tissue level, EVs provide a further layer of integration by packaging distinct non-coding RNA species, enabling co-ordinated transfer of regulatory signals between cardiomyocytes, fibroblasts, endothelial cells, and immune cells during injury and remodelling.114 Together, these observations support a view of the cardiovascular dark genome as a distributed complex regulatory network in which multiple non-coding RNA classes intersect to orchestrate cardiovascular disease-relevant transcriptional programmes. Defining this network will require sustained research that integrates perturbation experiments with bioinformatics approaches with the aim of identifying causal regulators that could be prioritized as therapeutic targets for cardiovascular disease.115,116 Will there be other non-coding RNAs that emerge? Despite rapid progress in annotation, a subset of non-coding transcripts remains difficult to classify. One example is non-coding RNAs derived from 3′ untranslated regions (3′ UTRs), which can exhibit subcellular localizations distinct from their parent messenger RNAs,117,118 and may therefore have distinct regulatory functions.119 However, their biogenesis, mechanism, and regulation are still largely unknown.120 Ongoing advances in transcriptome profiling and RNA-focused functional genomics are now enabling more systematic research of such atypical RNA species. Whether these non-canonical RNAs contribute to cardiovascular physiology or disease remains to be determined. Tools, technologies, and methodological challenges in dark genome research A range of computational approaches is now enabling discovery, functional annotation, and therapeutic targeting of dark genome elements across cardiovascular tissues and disease context. While methods for regulatory DNA elements have been discussed in Section 2.2, here we focus on approaches for non-coding RNA discovery and functional characterization (Figure 10), before addressing broader methodological challenges that apply across dark genome research. Computational and experimental approaches for discovery and functional annotation of non-coding RNAs. Overview of key technologies used to identify and characterize non-coding RNAs, including RNA sequencing strategies (poly(A)-selected, total RNA and small RNA sequencing), methods to capture transient regulatory transcription, short- and long-read sequencing for isoform resolution, and approaches for functional perturbation: RNA interference (RNAi), CRISPR-based modulation and locus editing RNA sequencing strategies for non-coding RNA discovery High-throughput RNA sequencing profiles RNA species and their abundance across tissues, cell types, disease states, or in response to stimuli. The discovery and characterization of non-coding RNAs (ncRNAs) rely on high-throughput sequencing strategies tailored to specific RNA populations. While standard poly(A)-selected RNA-sequencing captures many lncRNAs, total RNA library preparations utilizing ribosomal RNA depletion are necessary to identify circular and non-polyadenylated transcripts. Small ncRNAs, such as miRNAs, tRNA fragments, and Y RNA fragments, often require dedicated size-selection workflows. Additionally, short-read sequencing often struggles with complex or repetitive isoforms; however, long-read sequencing and targeted capture can provide more accurate full-length transcript reconstruction.121,122 Interpretation requires care, as library preparation can introduce sequence- and modification-dependent biases (notably for tRNA-derived fragments), and complementary validation is often needed.123 Moreover, a substantial fraction of transcription occurs near regulatory regions such as promoters and enhancers, producing RNAs that are often unstable and therefore difficult to detect in steady-state RNA sequencing. Methods that profile nascent RNA can reveal this transient transcriptional activity and help map where transcription initiates and how it is regulated.124,125 Decoding non-coding function: perturbation readouts and AI-guided prediction Beyond identification, establishing function of RNA species necessitates integrating perturbation assays, such as CRISPR-based repression or activation and quantitative readouts like nascent RNA profiling and chromatin profiling, which captures transient transcriptional activity at enhancers and promoters. To manage the scale of these data, machine learning and AI-guided frameworks are employed to prioritize candidate regulatory elements and translate non-coding genetic risk into causal cardiovascular mechanisms.126,127 Methodological challenges Despite rapid progress, several challenges limit the reliable identification and functional characterization of dark genome elements in cardiovascular research. First, the lack of standardized protocols and analytical pipelines across studies makes direct comparison of results difficult, and community-agreed minimum reporting standards are needed to improve reproducibility.128 Second, current genome annotations remain incomplete as functional evidence exists for only a small fraction of the tens of thousands of annotated non-coding loci, while many genuinely functional elements likely remain unannotated.77 Third, because the heart and vasculature contain multiple distinct cell types, studies using bulk tissue average signals across cardiomyocytes, fibroblasts, endothelial cells, and immune cells, potentially masking cell-type-specific regulatory activity.129 Single-cell and single-nucleus approaches are resolving this heterogeneity; however, most widely used single-cell transcriptomic datasets still rely on poly(A)-based workflows and therefore provide incomplete coverage of many non-coding RNA classes, although more recent total-RNA methods are starting to address this limitation.130 Fourth, demonstrating that a non-coding locus is functionally important, rather than merely associated or transcribed, remains technically demanding. For many annotated non-coding loci implicated in cardiovascular disease, causal roles, target genes, and underlying mechanisms have yet to be established.131 Fifth, many dark genome elements, particularly lncRNAs, show limited conservation between humans and the animal models used in preclinical research, complicating experimental validation and translation.132 Finally, the majority of genetic variants associated with cardiovascular disease in large population studies map to non-coding regions, yet linking these variants to specific regulatory mechanisms and target genes remains a major unsolved challenge.133 Current limitations and next steps Three priorities stand out for the field. First, expanded profiling of cardiovascular cell types using long-read and single-cell sequencing will broaden the catalogue of functionally relevant dark genome elements. Second, scalable experimental tools, including CRISPR-based screens115 across cardiovascular cell types, are needed to systematically identify which dark genome elements are causally important, and improved computational methods will help integrate these findings with existing genomic and clinical data. Third, translating dark genome discoveries into clinical benefit will require both advances in targeted delivery of nucleic acid therapeutics134 and integration of non-coding regulatory information into risk prediction frameworks.135 Addressing these challenges will ultimately determine whether the dark genome can be fully exploited for cardiovascular diagnosis and therapy. Emerging concepts in dark genome biology The dark genome is increasingly recognized as a source of functional outputs beyond canonical RNAs. In addition to regulatory RNAs, some loci generate small bioactive peptides, and others appear to act through multiple co-existing mechanisms at the DNA, RNA, and peptide levels. In this section, we highlight these emerging concepts and their relevance to cardiovascular biology. The dark proteome and small open Reading frames Advances in transcriptomics, ribosome profiling, and proteogenomics have revealed that some loci previously annotated as non-coding generate functional micropeptides from small open reading frames, indicating that part of the dark genome encodes peptides that were historically excluded from genome annotations. These micropeptides are typically <100 codons in length but can exert meaningful regulatory effects, influencing cellular metabolism, stress responses, and calcium handling.136 While these micropeptides remain poorly annotated, several have already been shown to exert direct, mechanistically defined effects in cardiovascular tissues. Micropeptides such as DWORF,137 and MOXI/mitoregulin138 have demonstrated roles in modulating cardiomyocyte contractility, mitochondrial function, and metabolic adaptation, while others, such as SPAAR, influence endothelial signalling and angiogenic responses.139 While micropeptide function is better defined in cardiomyocytes and endothelial cells, emerging data now implicate micropeptides in fibroblast activation and vascular smooth muscle phenotypic switching; however, these roles are supported by a relatively small number of recent studies and remain to be explored in depth.140,141 Collectively, these discoveries illustrate that the dark genome contributes not only regulatory RNAs but also peptide-coding potential that broadens the scope of gene expression in cardiovascular biology. Multifunctional genetic loci Since the discovery of the first non-coding genes, it was acknowledged that some regions of the dark genome can exert regulatory effects through multiple, co-existing modalities, challenging the traditional separation of coding, non-coding, and regulatory DNA. At such loci, the DNA sequence can function as an enhancer or chromatin regulatory element, while the ‘non-coding’ RNA transcript mediates additional effects through transcriptional or post-transcriptional mechanisms.142 In some cases, as discussed in the previous section, these same transcripts also harbour small open reading frames that encode biologically active micropeptides, creating multimodal loci in which RNA and peptide outputs can mediate distinct, and in some contexts divergent, regulatory effects. The LINC00961/SPAAR locus exemplifies RNA-peptide multimodality, as the non-coding transcript and the encoded micropeptide exert distinct and, in angiogenic assays, opposing effects on endothelial behaviour.139 A further exemplar is the chromosome 9p21.3 coronary artery disease risk locus, where a non-coding, enhancer-rich interval contains risk alleles that alter inflammatory transcription factor binding and cis-regulatory activity, while the same region also produces ANRIL isoforms with demonstrable effects on vascular cell phenotypes (including circANRIL effects on ribosomal RNA maturation and cell proliferation and apoptosis), illustrating how a single locus can integrate DNA-level regulatory and RNA-mediated mechanisms.143 More broadly, cardiac enhancer-associated lncRNA loci illustrate how a single genomic region can integrate DNA- and RNA-based regulatory functions, underscoring the multimodal nature of dark-genome regulation in cardiovascular biology.73 Taken together, these examples highlight that cardiovascular risk loci and disease-associated transcripts may encode overlapping DNA-, RNA-, and peptide-level functions, and that a locus-centred view is often necessary to define causal mechanisms and prioritize translational targets. Conclusions The study of the dark genome has now reached a critical juncture. Two decades of rapid advances in sequencing technologies have extensively characterized the different DNA and RNA elements that our genome can produce. In addition to the discussed emerging concepts, key challenges remain. While progress has been made to solve long-standing DNA sequencing challenges, including the complete sequencing of the human genome (telomere-to-telomere),144 understanding the dynamic interplay between the various elements comprising this sequence remains an enormous challenge. This includes understanding how the 3D spatial architecture is governed through thousands of chromatin loops, how this varies between cell types, and if and how this architecture changes in response to disease or stimuli,145 and understanding how specific enhancers shape gene expression patterns hundreds of thousands of base pairs away. Understanding these processes will clarify our understanding of the dark genome. Clinical translation of key pre-clinical findings is also a critical next step. As discussed, a number of non-coding elements emerge now as either targets or therapeutic entities in various phases of clinical studies in cardiovascular medicine. The number of clinically evaluated non-coding elements is expected to increase substantially as these elements become fully functionally characterized. Supplementary Material Acknowledgements The authors thank Neil Craig and Craig Balmforth for providing a clinical perspective and for their helpful comments on the manuscript, which helped strengthen the translational relevance and clarify key clinical considerations. The authors also thank Alan Monziani for early input and for suggesting references that informed the review, and Susan Carpenter and Abdelaziz Beqqali for reviewing the manuscript and providing helpful suggestions. Contributor Information Despoina Kesidou, Institute for Neuroscience and Cardiovascular Research, Queens Medical Research Institute, University of Edinburgh, 47 Little France Crescent, Edinburgh EH16 4TJ, UK. Simon D Brown, Institute for Neuroscience and Cardiovascular Research, Queens Medical Research Institute, University of Edinburgh, 47 Little France Crescent, Edinburgh EH16 4TJ, UK. Lars Maegdefessel, Institute of Molecular Vascular Medicine, TUM University Hospital, Munich 80802, Germany; Cluster for Nucleic Acid Sciences and Technologies—NUCLEATE, Institute of Molecular Vascular Medicine, TUM University Hospital, Munich 80802, Germany. Igor Ulitsky, Department of Immunology and Regenerative Biology and Department of Molecular Neuroscience, Weizmann Institute of Science, Rehovot 7610001, Israel. Andrew H Baker, Institute for Neuroscience and Cardiovascular Research, Queens Medical Research Institute, University of Edinburgh, 47 Little France Crescent, Edinburgh EH16 4TJ, UK; Cardiovascular Research Institute Maastricht (CARIM) and Department of Pathology, University of Maastricht, Maastricht 6229HX, The Netherlands; MRC/BHF Centre of Research Excellence in Advanced Cardiac Therapies (REACT), Institute for Neuroscience and Cardiovascular Research, University of Edinburgh, 47 Little France Crescent, Edinburgh EH16 4TJ, UK. Supplementary data Supplementary data are available at European Heart Journal online. Declarations Disclosure of Interest D.K., I.U., and A.H.B. are named inventors on patents relating to non-coding RNA therapeutics (D.K.: PCT/GB2023/053178, PCT/EP2023/084938; I.U.: PCT/IL2021/051503; A.B.: PCT/GB2023/053178, PCT/EP2023/084938, PCT/GB2016/050708). D.K. and A.H.B. have also received research funding from Bristol Myers Squibb and Evotec in support of microRNA therapeutic research (beLAB1407). L.M. has received research funding from Novo Nordisk (Måløv, Denmark), Roche Diagnostics (Rotkreuz, Switzerland), and Bitterroot Bio (Palo Alto, CA, USA), and serves as a scientific advisor to Novo Nordisk (Måløv, Denmark), DrugFarm (Guilford, CT, USA), and Angiolutions (Hannover, Germany). I.U. has received funding from Mazhi Therapeutics (San Francisco, CA, USA) and serves as a scientific advisor to Renasis Bio (Rehovot, Israel), MetaboCure (Yokneam, Israel), and Mazhi Therapeutics (San Francisco, CA, USA). Data Availability No data were generated or analysed for or in support of this paper. Funding D.K is supported by beLAB1407 (Evotec–Bristol Myers Squibb BRIDGE partnership). S.D.B is supported by a British Heart Foundation Chair of Translational Sciences (CH/11/2/28733) awarded to A.H.B. A.H.B is supported by a British Heart Foundation Chair of Translational Sciences (CH/11/2/28733), Programme Grant RG/F/25/110149, UKRI/ERC Advanced Grant EP/Z533373/1, the UKRI Medical Research Council and British Heart Foundation Research Excellence (CoRE) in Advanced Cardiac Therapies (REACT) funding (MR/Z504865/1) and European Research Council Grant ‘PolymiRize’ funded by the UK Guarantee. L.M. is supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—Project-ID 533767322—EXC 3113/1, Cluster for Nucleic Acid Sciences and Technologies—NUCLEATE, the Deutsche Forschungsgemeinschaft-funded TRR267, a European Research Council (ERC) Consolidator grant with the acronym LongTx and grant number 101088370, BMBF in the framework of the Cluster4Future program (Cluster for Nucleic Acid Therapeutics Munich, CNATM—Project ID: 03ZU1201BM). I.U. is supported by the European Research Council (Consolidator Grant lncIMPACT), the Abisch-Frenkel RNA Therapeutics Center at the Weizmann Institute, and the Israeli Ministry of Health (IMOH) as part of the ERA4Health joint call, RECREATE consortium, grant number 3-19415. References Associated Data Supplementary Materials Data Availability Statement No data were generated or analysed for or in support of this paper.