SHARED GENETIC ARCHITECTURE AND DIRECTIONAL ASSOCIATIONS BETWEEN ATTENTION-DEFICIT/HYPERACTIVITY AND BORDERLINE PERSONALITY DISORDERS: AN INTEGRATIVE GENOMIC ANALYSIS

Background: attention-deficit/hyperactivity disorder (ADHD) and borderline personality disorder (BPD) frequently co-occur, but their shared genetic susceptibility and directional associations remain unclear. Methods: we analyzed European-ancestry genome-wide association study (GWAS) data for ADHD (38,691 cases; 186,843 controls) and BPD (12,339 cases; 1,041,717 controls). Linkage disequilibrium score regression, conjunctional false discovery rate (conjFDR), bidirectional Mendelian randomization (MR), MR accounting for linkage disequilibrium and sample overlap (MR-LAP), Causal Analysis Using Summary Effect estimates (CAUSE), colocalization, and Sum of Single Effects (SuSiE) fine-mapping assessed genetic sharing and directionality. Results: genetic correlation was substantial (r_g = 0.733). ConjFDR identified 48 shared loci; 36 (75%) showed concordant directions in both phenotype-specific comparisons. Functional annotation implicated axon guidance, oxidative phosphorylation, and developing midbrain cells. ADHD-to-BPD MR was positive (22 instruments; inverse-variance weighted beta = 0.487, SE = 0.082, p = 3.27 x 10^-9), persisting after outlier removal, MR-LAP correction, and exclusion of two instruments at colocalized loci. Reverse MR was positive (five instruments; beta = 0.303, SE = 0.112, p = 6.83 x 10^-3) but heterogeneous and variant-sensitive. CAUSE favored directional over sharing-only models in both directions (delta ELPD = 6.98 and 5.92, respectively). SuSiE colocalization supported shared signals at rs17576773, rs9969232, and rs4727799. Conclusions: ADHD and BPD exhibit extensive genetic overlap involving 48 shared loci and convergent evidence of directional associations. Findings across complementary methods more consistently support an ADHD-to-BPD relationship, suggesting that shared genetic susceptibility alone may not fully explain their co-occurrence.