Stochastic Genomic Instability as an Upstream Modifier of Pruning Dosage Distinguishes Schizophrenia from Cognitive Variation

Background: Schizophrenia and intelligence represent opposing poles of neurodevelopmental genetic architecture, yet both implicate synaptic pruning. Prior work has linked pruning excess to SCZ and calibrated refinement to cognition, but the upstream role of genomic instability—particularly germline predisposition to mitotic errors and low-level mosaicism—remains unexplored. Methods: We re‐analysed the PGC3 SCZ GWAS (effective N ≈ 58,749) and Savage et al. IQ GWAS (N = 269,867) using identical pipelines: MAGMA gene-based and gene-set analysis, LDSC partitioned heritability, GSEA/DGSEA pathway enrichment, and GTEx v8 brain TWAS. Eight overlapping gene sets were curated for SCZ (housekeeping, monoaminergic, glutamatergic, synaptic pruning, mCA-related, mitotic stress, clonal hematopoiesis, and negative controls). For IQ, 10 gene sets were tested (the same 7 as SCZ, plus shortened glutamatergic, shortened pruning, and pruning-specific). Seven sets were drawn in comparative tables (kindling excluded, as it was tested only in SCZ). Results: In SCZ, synaptic pruning dominated (MAGMA Bonferroni p = 4.5 × 10⁻⁵, 9/76 GSEA sets significant, LDSC 1.41‐fold enrichment), amplified by immune (MHC/complement) and instability signals (mCA 2.00‐fold, mitotic 2.10‐fold LDSC; MAD1L1/PPP1R16B up‐regulated in limbic TWAS). Glutamatergic enrichment was secondary. In IQ, glutamatergic pathways led (DGSEA pruning vs. glutamatergic dES = –0.459, p = 0.002), with pruning remaining significant only after removing glutamatergic overlap. DGSEA directly contrasted the traits: pruning > glutamatergic/mCA in SCZ; glutamatergic > pruning in IQ. Conclusions: We propose the Stochastic Genomic Instability Amplifies Pruning Dysregulation (SGI-APD) model: inherited mitotic/CH-like variants create mosaic neuronal substrates that exacerbate immune-mediated pruning in SCZ, while genomic stability permits calibrated refinement and glutamatergic optimisation in high-IQ cognition. This framework unifies polygenic and emerging somatic mechanisms, offers mechanistic explanations for heterogeneity, and highlights convergent therapeutic targets (complement inhibition + DNA-repair modulation).