NM_000314.8:c.253+5G>T is an intronic variant at the +5 position of the PTEN donor splice site. This variant is absent from gnomAD v2.1 and v4.1 population databases (PM2_Supporting).

SpliceAI strongly predicts a deleterious splicing effect (max delta score 0.91; acceptor loss 0.91, donor loss 0.85), consistent with disruption of normal splicing at the exon 4-intron 4 junction.

The PTEN VCEP PVS1 decision tree applies to canonical GT-AG ±1,2 splice site disruptions. c.253+5G>T at the +5 position does not fall within the canonical splice site branch of the decision tree, and PVS1 is not met under the VCEP framework.

This variant is reported in ClinVar (VariationID 427617) as Pathogenic by four clinical laboratories and Likely pathogenic by one, with review status 'criteria provided, single submitter.' No expert panel review is available.

The variant has been observed in somatic cancers (COSMIC COSV64298366, n=2), consistent with a role in tumorigenesis, though somatic observations are not directly applicable to germline ACMG/AMP criteria.

Key functional evidence may exist in PMID:28677221 (Chen et al. 2017), which characterized cryptic splicing in 34 germline PTEN intronic variants from Cowden syndrome patients, but full-text confirmation that c.253+5G>T was specifically studied is unavailable. Until splicing assay data are confirmed, PS3 cannot be applied.

Multiple criteria require additional evidence to be fully assessed: PS3 (RNA/mini-gene splicing assay), PS4 (proband specificity scores), PP3 (VarSeak concordance with SpliceAI), PS1 (comparison to other pathogenic variants at c.253+5), and PP1/BS4 (segregation data).