4. Discussion

This study applied an integrative multiomic framework to a small cohort of paediatric NDD and CHD probands to characterise individual‐level genomic, transcriptomic and epigenetic findings. These analyses demonstrate how genomic, transcriptomic and epigenomic data can provide complementary diagnostic and mechanistic information at the individual‐patient level. These include rare variants in developmental and chromatin remodelling genes, heterogeneous methylation patterns across probands and epigenetic age estimates that, although variable and limited by small sample size, suggest that this framework may identify molecular subgroups within the broader NDD/CHD population. We interpret these findings as hypothesis‐generating, acknowledging that the small cohort size and overlapping confidence intervals across groups preclude definitive conclusions.

Whole genome sequencing identified 33 rare potentially deleterious variants and one microdeletion across 14 probands. The most significant finding was a de novo splice‐site variant in the chromatin remodelling gene ARID1B in Proband 19, functionally confirmed by RNA sequencing, which demonstrated aberrant splicing absent in both parents. This variant likely underlies the widespread methylation dysregulation observed in this proband. The SWI/SNF complex, of which ARID1B is a component, regulates chromatin accessibility during heart and brain development, and loss of function can produce genome‐wide epigenetic consequences [40, 41]. Chromatin‐disrupting variants have also been implicated in other neuropsychiatric conditions including OCD [42], supporting dysregulation of this machinery as a mechanism relevant across neurodevelopmental and psychiatric phenotypes.

The remaining probands with extreme methylation dysregulation showed distinct and heterogeneous patterns. Proband 13 (NDD), despite no confirmed causal variant, exhibited 72,355 differentially methylated sites with a spatially distinct signature of hypermethylation enriched at CpG islands and hypomethylation in open sea regions, suggesting disruption of methylation maintenance rather than global loss. Probands 19 and 20, despite having different clinical presentations (NDD and CHD, respectively), showed strikingly similar methylation profiles across genomic contexts, consistent with shared underlying genetic factors. In contrast, Proband 15 (NDD) showed enriched hypermethylation across all genomic contexts, suggesting global dysregulation of methylation machinery with potential for broad transcriptional repression in developmental networks.

Our results highlight both shared and distinct genetic factors contributing to NDD and CHD. Several genes identified in our cohort demonstrate pleiotropic effects across cardiac and neural tissues. Proband 16 has an MYH6 variant which may be relevant to this individual′s cardiac phenotype (although most reported patients with pathogenic variants in this gene have had much less severe cardiac manifestations than HLHS). This individual also has an inherited truncating variant in RERE; this is a variant of uncertain significance but could possibly contribute to the neurodevelopmental phenotype. This observation aligns with recent large‐scale sequencing studies suggesting that patients with both CHD and NDD carry a greater burden of damaging variants compared with those with isolated conditions [9, 43].

The epigenetic analysis revealed a substantial biological age acceleration in HLHS patients, significantly greater than both NDD‐only patients (mean: 9.5 years greater) and controls (mean: 7.4 years greater). This magnitude exceeds previous reports in paediatric disease cohorts and approaches levels typically observed in severe adult‐onset conditions [44–47], suggesting CHD is associated with broader biological effects beyond the cardiovascular system, though replication in larger cohorts is needed. This acceleration may reflect chronic hypoxia and surgical stress, though formal investigation of these mechanisms is needed. Our group recently demonstrated that CHD patients carry significantly higher polygenic risk scores for hypertension, Type 2 diabetes and obesity compared with healthy controls [48], suggesting both genetic predisposition and accelerated biological aging may contribute to the increased comorbidity burden.

Several limitations should be considered when interpreting these findings. The small sample size (15 probands) limits statistical power for detecting associations and may not capture the full spectrum of genetic variation relevant to both conditions. The study was designed to investigate shared pathways across patients with NDD and/or CHD; however, only two probands had both diagnoses, which limits direct comparison between isolated and comorbid presentations. Future studies with larger coaffected cohorts would better enable such comparisons. The absence of a matched control group for genomic analyses limits interpretation of methylation patterns, although this was taken into consideration for the analyses performed in this study.

The heterogeneity of both cardiac and neurodevelopmental phenotypes within our cohort presents challenges for identifying convergent mechanisms. Although ADHD was the predominant NDD diagnosis, the inclusion of ASD, intellectual disability, and other conditions may obscure condition‐specific genetic factors. Future studies with larger, phenotypically stratified cohorts will be essential to validate and extend these findings. Additionally, functional validation of identified variants through cellular or animal models would strengthen causal inferences, particularly for variants of uncertain significance.

Our integrated multiomic approach identified molecular findings relevant to both CHD and NDD, including variants affecting chromatin remodelling and heterogeneous patterns of epigenetic dysregulation. These observations support further investigation of shared developmental mechanisms in larger cohorts. The pathway enrichment analysis, which identified involvement of both muscle development and neuronal guidance, suggests possible molecular links between cardiac and neural development. This is in line with recent single‐cell transcriptomic studies indicating some shared progenitor populations during early embryogenesis [49]. Collectively, these findings highlight the value of integrative genomic and epigenomic profiling for capturing the complexity of these disorders and improving diagnostic precision for affected children. Larger, phenotypically stratified cohorts and functional validation studies will be necessary to clarify causality and determine the full clinical implications of these observations.