Abstract
Incomplete penetrance and variable expressivity complicate interpretation of germline GATOR1 variants and individual-level phenotype prediction. We conducted a family-based systematic review and meta-analysis, reconstructing individual-level data from published pedigrees and reassessing variants using GATOR1-specific classification criteria. Sixty-three studies were included; 59 contributed 734 individuals heterozygous for germline pathogenic or likely pathogenic (P/LP) variants from 227 families. The pooled proportion of affected individuals was 65.7%. Recurrent variant status, variant consequence, DEPDC5 protein domain, transmitting-parent phenotype, and overall parental origin did not consistently distinguish affected from unaffected individuals, although paternal DEPDC5 transmission showed an exploratory association with higher disease expression (90.4% vs. 82.4%, p = 0.032). Among affected individuals, missense variants were associated with earlier seizure onset than loss-of-function (LoF) variants (0.67 vs. 4.50 years, p = 0.002). LoF variants in the DEPDC5 LassoA domain showed exploratory associations with cognitive impairment, MRI abnormalities, and epilepsy surgery (FDR-adjusted p ≤ 0.046). An expanded synthesis identified 14 published same-gene germline–somatic second-hit cases, of which 85.7% had focal cortical dysplasia (FCD) IIa; quantifiable somatic variant allele fractions (VAFs) for sequence-variant second hits were below 5% in all six single-region cases, whereas multiregion sampling in three cases showed spatial variation, including values below 1%. These findings suggest that disease expression and post-onset phenotype represent partly distinct aspects of clinical heterogeneity in GATOR1-related epilepsy.