IBI-DT as a complementary framework for prioritizing candidate genes, interactions, and pathways in congenital heart disease

Congenital heart disease (CHD) has a complex and heterogeneous genetic architecture. Conventional population-level association analyses can identify average genetic effects across a cohort but may miss subgroup-specific, individualized, and interaction-related signals. To address this, we applied Individualized Bayesian Inference and Decision Tree (IBI-DT), a Bayesian marginal-likelihood-guided decision-tree framework, to CHD genomic data from the Pediatric Cardiac Genomics Consortium available through the Kids First Data Portal, with a primary focus on coding variation. After quality control, we restricted the analysis to exonic variants to focus on coding regions, yielding 90,298 variants from 2,260 subjects (713 CHD cases and 1,547 reference subjects). As a baseline population-level association analysis, Fisher’s exact test (FET) identified 22 Bonferroni-significant variants, whereas IBI-DT prioritized 42 variants exceeding the empirical recurrence threshold. IBI-DT recovered more known CHD genes among top-ranked variants than the baseline analysis, including four known CHD genes in the top 10 variants versus one, and seven in the top 42 variants versus three. In patient-specific analysis, IBI-DT provided broader case coverage, with top-ranked variants covering all 713 CHD cases by rank 43, whereas the baseline FET ranking covered only 10 CHD cases. Top IBI-DT variants were also less redundant than baseline FET variants (mean absolute Spearman correlation 0.04 vs. 0.29). Beyond single-variant prioritization, IBI-DT identified 163 variant pairs above the empirical recurrence threshold, reflecting conditional interaction structure derived from the hierarchical tree framework, and recovered more CHD-related pathways than the baseline FET analysis. IBI-DT provides complementary information beyond a baseline population-level association analysis for studying the heterogeneous genetic basis of CHD. By prioritizing subgroup-specific, individualized, and interaction-related signals from coding variants, IBI-DT may improve the identification of candidate genes, interaction patterns, and pathways for follow-up CHD studies.