NM_001412.4:c.338-1G>C is a canonical ±1 splice acceptor variant in intron 5 of EIF1AX, a gene in which loss of function is supported as a germline disease mechanism in uveal melanoma predisposition. Under ClinGen SVI PVS1 recommendations (PMC6185798), this qualifies for PVS1 at strong strength.¹ The variant is absent from gnomAD v2.1, v4.1, and gnomAD-Canada population databases, satisfying PM2 at moderate strength.² SpliceAI predicts a splice-altering effect (max delta score 0.68, acceptor loss). This computational evidence is already accounted for in PVS1 and is not separately applied as PP3 per PMC6185798 guidance against double-counting splice prediction evidence.³ No de novo data (PS2/PM6), no functional studies (PS3/BS3), no case-control data (PS4), no segregation data (PP1/BS4), and no patient phenotype information (PP4) are available for this variant. These criteria remain unmet. ClinVar variation ID 4484109 has zero submissions, no classification, and no review status. PP5 and BP6 cannot be applied.⁴ The variant has been observed in 3 somatic tumor samples in COSMIC (COSV65450922). Somatic occurrence does not independently satisfy germline criteria but is consistent with a role in tumorigenesis. The only associated publication (PMID:31275557) describes a pan-cancer splicing mutation repository but does not include NM_001412.4:c.338-1G>C in its dataset. No publication provides variant-specific evidence for any criterion.⁵ Overall classification: Likely Pathogenic. The combination of PVS1 (strong) and PM2 (moderate) satisfies the generic ACMG/AMP 2015 threshold for Likely Pathogenic (1 strong + 1 moderate).⁶