| Gene / variant | Evidence type and exact reported observation | Interpretation / limitation | Primary source DOI and publication year |
|---|---|---|---|
| **KRT4** c.438_440delCAA; in-frame deletion of one aspartic-acid residue | Familial human case report: heterozygous variant detected by trio exome sequencing in the affected 2-year-old proband and affected father, confirmed by paternal Sanger sequencing; absent from the unaffected mother. Two monozygotic twin sisters had similar lesions but were not sequenced. (pqac-00000024, pqac-00000025) | Phenotype co-segregation in the tested father–son pair supports pathogenicity, but the family is small and no functional assay was reported. | [10.3390/genes13122184](https://doi.org/10.3390/genes13122184), **2022** |
| **KRT13** c.1023+23_1024-39del; 32-bp intronic deletion causing aberrant splicing and **p.Lys342_Gln359del** | Four-generation human pedigree: RT-PCR demonstrated deletion of 54 bases from the transcript and an in-frame loss of 18 amino acids. The variant was present in all **9/9 tested affected carriers** and absent from **11/11 tested unaffected relatives**; it arose de novo in the affected ancestral branch. (pqac-00000049, pqac-00000050) | Strong segregation plus RNA-level functional evidence supports pathogenicity. Full penetrance was observed in this pedigree, but this does not establish universal penetrance or prove that the variant caused the two oral cancers reported in the family. | [10.1002/ccr3.1073](https://doi.org/10.1002/ccr3.1073), **2017** |
| **KRT13** c.340C>T (**p.Arg114Cys**), rs545085703 | Four-patient diagnostic case series: inherited maternally by patient P-2; listed at **1/251,496 alleles** in gnomAD and classified “likely pathogenic” using VarSome. (pqac-00000027, pqac-00000034) | Rare missense candidate, but the transmitting mother was unaffected, no functional assay was reported, and the classification was an author-reported VarSome output rather than an independently curated ClinGen conclusion. Incomplete penetrance or non-causality remains possible. | [10.3390/bioengineering10020154](https://doi.org/10.3390/bioengineering10020154), **2023** |
| **KRT4** c.244_245insTTGGTGGCTTTGGTGCCGGCGGCTTCGGAGCTGGTTTCGGCA; **p.Gly81_Thr82insIleGlyGlyPheGlyAlaGlyGlyPheGlyAlaGlyPheGly** | Diagnostic case series: inherited from an unaffected parent in multiple patients; reported at **2/47,694 alleles** in gnomAD and classified as a variant of uncertain significance. (pqac-00000027, pqac-00000034) | **VUS**: rarity and in-frame location do not establish causality. Presence in unaffected parents and lack of functional evidence require cautious interpretation; common benign variants reported in the same series should not be treated as causal. | [10.3390/bioengineering10020154](https://doi.org/10.3390/bioengineering10020154), **2023** |
| **KRT4** c.259_260insCCGGCGGCTTCGGAGCTGGTTTCGGCACTGGTGGCTTTGGTG; **p.G87delinsAGGFGAGFGTGGFGG** | Single human case: found by exome sequencing in a 70-year-old man whose WSN lesion developed focal moderate-to-severe epithelial dysplasia over one year; absent from his healthy daughter. Single-cell RNA sequencing showed altered epithelial proliferation/differentiation and increased epithelial–fibroblast communication. (pqac-00000023, pqac-00000031, pqac-00000032) | Novel candidate reported in one patient. The study provided no direct variant-specific functional assay; smoking history and single-case observational design confound attribution. Association with dysplasia does **not** prove that the variant causes WSN or malignant transformation. | [10.1186/s12903-024-04300-y](https://doi.org/10.1186/s12903-024-04300-y), **2024** |


*Table: Variant-level evidence for KRT4- and KRT13-associated white sponge nevus, separating segregation and functional observations from uncertain or single-case interpretations. Population frequencies and classification limitations are retained to prevent overcalling causality.*