AXIN2 c.2013_2024del, p.(Thr672_Arg675del)

Reasoning

Met under the generic ACMG/AMP BA1 rule. The gnomAD v4.1 African/African American group maximum allele frequency is 7.33307% (5,502/75,030 alleles), with grpmax FAF 7.17121%; this exceeds the generic 5% BA1 threshold. Concordant gnomAD v2.1 data show African/African American AF 7.37246% (1,789/24,266) and grpmax FAF 7.17742%.

Rule

Generic ACMG/AMP 2015 BA1: allele frequency >5%.

Evidence considered

• No AXIN2 ClinGen VCEP specification or local gene-specific framework was available; generic ACMG/AMP was therefore used.
• gnomAD v4.1 reports 220 African/African American homozygotes, and gnomAD v2.1 reports 66, supporting that the high ancestry-specific frequency is a reproducible population observation rather than a single-dataset artifact.
• The BA1 threshold used is from Richards et al. (PMID:25741868): allele frequency above 5% in a population reference dataset is benign stand-alone evidence.

Sources

Reasoning

BP3 (Richards et al. 2015, PMID:25741868) applies to in-frame deletions/insertions in a repeat region without a known function. All three elements are met. (1) In-frame deletion: c.2013_2024del removes 12 nt = 4 codons, producing p.(Thr672_Arg675del) with no frameshift, no premature termination codon (no NMD), and no splice impact (SpliceAI max delta 0.019); the variant lies entirely within exon 8 (c.1908-2141) of the MANE Select transcript NM_004655.4. (2) Repeat region: the deleted interval ACGGGGGGTGGTG contains a 6-G homopolymer run and Mutalyzer normalization reports shift 6 (boundary ambiguity in a repetitive sequence). (3) Without a known function: residues 672-675 fall outside all four annotated AXIN2 conserved domains (TNKS-binding, RGS, beta-catenin-binding, DIX) in the C-terminal region between the beta-catenin-binding and DIX domains, removing only ~0.5% of the 843-residue protein, and the deletion is common and homozygous in population cohorts (gnomAD v4.1 AFR AF 7.33% with 220 AFR homozygotes; v2.1 AFR AF 7.37% with 66 homozygotes), indicating the region tolerates loss of these residues. BP3 met at supporting strength.

Rule

ACMG/AMP 2015 BP3 (Richards et al., PMID:25741868): in-frame deletions/insertions in a repeat region without a known function

Evidence considered

• Richards et al. 2015 (PMID:25741868) Table 3 defines BP3 as in-frame deletions/insertions in a repeat region without a known function; the companion PM4 passage retrieved in full text (publications/25741868.txt) confirms the in-frame deletion/insertion framework, though the BP3 sentence itself lies beyond the 40KB truncation of the retrieved text.
• Consequence: Mutalyzer/VariantValidator (prefetch.json) confirm a 12-nt in-frame deletion, NP_004646.3:p.(Thr672_Arg675del), entirely within exon 8 (VariantValidator start_exon/end_exon = 8; Mutalyzer exon 8 = c.1908-2141) of the MANE Select transcript NM_004655.4; no frameshift, so no NMD is expected.
• Repeat-region context: the deleted interval ACGGGGGGTGGTG contains a 6-G homopolymer run, and Mutalyzer normalization reports shift 6, indicating the deletion boundaries are ambiguous within a repetitive sequence.
• No known function: residues 672-675 lie outside AXIN2's four conserved domains (Tankyrase-binding, RGS, beta-catenin-binding, DIX), per the AXIN2 gene literature summarized in the case bundle (PMID:41361419 fulltext excerpt in case_summary.json), in the region between the beta-catenin-binding and DIX domains; only 4 of 843 amino acids (~0.5%) are removed.
• Population tolerance: gnomAD v4.1 overall AF 0.393%, AFR AF 7.33% with 220 AFR homozygotes (229 total homozygotes); gnomAD v2.1 AFR AF 7.37% with 66 homozygotes; gnomAD-Canada AF 0.46% - homozygous carriers are common, consistent with a nonessential repeat-region deletion.
• SpliceAI max delta 0.019 (< 0.1) indicates no splice impact, so the consequence is confined to the in-frame protein deletion.
• ClinVar variation 182018 (15 submissions, all Benign/benign) used only as consistency context; no criterion was inferred from another source's final classification.

Still missing

• No formal RepeatMasker/low-complexity track was retrieved for this locus; repeat-region determination rests on the 6-G homopolymer within the deleted interval and the Mutalyzer shift-6 normalization ambiguity.
• No dedicated functional study of this specific in-frame deletion was identified; 'without a known function' is supported by absence of domain annotation and population tolerance (homozygotes) rather than a direct assay.

Sources

Reasoning

SpliceAI, the only applicable calibrated computational tool for this in-frame deletion, predicts no significant splicing impact: max delta score = 0.019, well below the ~0.2 threshold used as a high-precision cutoff for splice-altering variants in the original SpliceAI publication (Jaganathan et al. 2019, PMID 30661751). No gene-specific VCEP/CSPEC PP3/BP4 lookup table exists for AXIN2 to override this generic threshold. Because this is an in-frame deletion rather than a missense or synonymous change, missense predictors (REVEL/BayesDel/aGVGD) and BP7 are not applicable; SpliceAI is the sole relevant multiple-lines-of-evidence source, so BP4 is applied at supporting strength per generic ACMG/AMP rules.

Rule

Generic ACMG/AMP BP4: multiple lines of computational evidence suggest no impact on gene/gene product. Applied at supporting strength based on SpliceAI max delta 0.019 << 0.2 threshold (PMID 30661751); no AXIN2 VCEP-specific BP4 rule found.

Evidence considered

• SpliceAI direct query for NM_004655.4:c.2013_2024del returned max_delta_score=0.019 (DS_AG=0.002, DS_AL=0.013, DS_DG=0.019, DS_DL=0.011), evidence_sentence: 'SpliceAI predicts no significant splice impact for this variant (max delta score = 0.02).'
• SpliceAI's ~0.2 delta-score threshold for high-precision splice-altering prediction comes from Jaganathan KK et al., 'Predicting Splicing from Primary Sequence with Deep Learning,' Cell 2019, PMID 30661751.
• The variant is a 12-bp in-frame deletion (p.Thr672_Arg675del), not a missense or synonymous substitution, so REVEL/BayesDel/aGVGD missense predictors and BP7 do not apply.

Still missing

• No gene-specific VCEP/CSPEC PP3/BP4 lookup table available for AXIN2 to confirm or override the generic SpliceAI threshold.

Sources

Reasoning

Not met. The variant is common in population reference datasets, including African/African American AF 7.33307% in gnomAD v4.1 (grpmax FAF 7.17121%) and AF 7.37246% in gnomAD v2.1 (grpmax FAF 7.17742%); it is not absent or rare in population databases.

Rule

Generic ACMG/AMP PM2 requires absence or rarity in population databases; this requirement is not satisfied.

Evidence considered

• gnomAD-Canada v1.0 independently reports overall AF 0.456% and African ancestry AF 6.582% (67/1,018 alleles), consistent with the gnomAD ancestry-specific frequency pattern.

Sources

Reasoning

NM_004655.4:c.2013_2024del is an in-frame 12-nt deletion (4 codons) in exon 8 of AXIN2 producing NP_004646.3:p.(Thr672_Arg675del), i.e., a protein length change of 4 of 843 residues, so the in-frame-deletion component of PM4 is present. However, PM4 (Richards et al. 2015, PMID:25741868) is scoped to in-frame deletions/insertions in a NONREPEAT region; the deleted 12-bp interval (ACGGGGGGTGGTG) contains a 6-G homopolymer run and Mutalyzer normalization reports a shift of 6, indicating the deletion boundaries are ambiguous within a repetitive sequence. The variant therefore falls in a repeat/low-complexity region, which PM4 explicitly excludes and which is the domain of BP3 instead. Consistent with a non-pathogenic length change, the region has no known function (residues 672-675 lie outside the TNKS-binding, RGS, beta-catenin-binding, and DIX domains) and the deletion is common in population cohorts (gnomAD v4.1 AFR AF 7.33%, 220 AFR homozygotes; v2.1 AFR AF 7.37%, 66 homozygotes), with no splice impact (SpliceAI max delta 0.019) and no NMD expected (no premature termination codon). PM4 not met.

Rule

ACMG/AMP 2015 PM4 (Richards et al., PMID:25741868): protein length changes due to in-frame deletions/insertions in a nonrepeat region or stop-loss variants

Evidence considered

• Richards et al. 2015 (PMID:25741868) defines PM4 as protein length changes due to in-frame deletions/insertions in a NONREPEAT region or stop-loss variants; retrieved full-text excerpt (publications/25741868.txt) states: 'in-frame deletions/insertions and stop losses are considered moderate evidence of pathogenicity.'
• Mutalyzer normalization (prefetch.json) classifies NM_004655.4:c.2013_2024del as an in-frame 12-nt deletion with protein consequence NP_004646.3:p.(Thr672_Arg675del) (4 amino acids removed) and reports shift 6 in the normalized model, i.e., the deletion can be represented at multiple positions within a repetitive sequence.
• The deleted 12-bp interval (ACGGGGGGTGGTG per g.63532561_63532572del) contains a 6-G homopolymer run (GGGGGG), placing the deletion in a homopolymer/low-complexity repeat context that PM4's nonrepeat-region scope excludes.
• SpliceAI (max delta score 0.019; DS_DG 0.019, all four delta scores < 0.1) predicts no significant splice impact, so the sole consequence is the in-frame protein length change; no premature termination codon is created, so NMD is not expected.
• gnomAD v4.1: overall AF 0.393% (6349/1613670 alleles), AFR AF 7.33% with 220 AFR homozygotes (229 total homozygotes); gnomAD v2.1: AFR AF 7.37% with 66 homozygotes - a common polymorphism with many homozygous carriers, inconsistent with a pathogenic protein-length change.
• ClinVar variation 182018 (15 submissions, all Benign/benign, 'criteria provided, multiple submitters, no conflicts') used only as consistency context; no criterion was inferred from another source's final classification.

Still missing

• No formal RepeatMasker/low-complexity track was retrieved for this locus; the repeat-region determination rests on the 6-G homopolymer within the deleted interval and the Mutalyzer shift-6 normalization ambiguity, which are considered sufficient.
• No functional assay specifically interrogating loss of residues 672-675 was identified; this is not required for PM4 not_met but would strengthen the benign-consequence picture.

Sources

Reasoning

No AXIN2-specific VCEP/CSPEC was found for this case (cspec.found=false; vcep_materials empty), so generic ACMG/AMP computational-evidence rules apply. This variant, NM_004655.4:c.2013_2024del (NP_004646.3:p.(Thr672_Arg675del)), is a 12-bp in-frame deletion removing 4 amino acids, not a missense or synonymous substitution, so missense in-silico predictors (REVEL, BayesDel, aGVGD) do not apply to this variant type and are not_available in the prefetch data regardless (revel.score=null, bayesdel.score=null). The only relevant computational tool is SpliceAI, which was run directly on this variant and returned a maximum delta score of 0.019 (DS_AG=0.002, DS_AL=0.013, DS_DG=0.019, DS_DL=0.011), far below the ~0.2 threshold established as a high-precision cutoff for predicted splice-altering effect in the original SpliceAI publication (Jaganathan et al. 2019, PMID 30661751). Pangolin scores (SG=0.011, SL=-0.054) are likewise near zero. This computational prediction therefore does not support a deleterious splicing effect, so PP3 is not met.

Rule

Generic ACMG/AMP PP3: computational evidence supports a deleterious effect on the gene/gene product. No AXIN2 VCEP-specific PP3/BP4 SpliceAI threshold was found; the well-established ~0.2 SpliceAI delta-score cutoff (PMID 30661751) was applied instead.

Evidence considered

• SpliceAI direct query for NM_004655.4:c.2013_2024del returned max_delta_score=0.019 (DS_AG=0.002, DS_AL=0.013, DS_DG=0.019, DS_DL=0.011), evidence_sentence: 'SpliceAI predicts no significant splice impact for this variant (max delta score = 0.02).'
• Pangolin scores from the same query (pangolin_SG=0.011, pangolin_SL=-0.054) are also near zero, consistent with no predicted splice gain or loss.
• SpliceAI's ~0.2 delta-score threshold for high-precision splice-altering prediction comes from Jaganathan KK et al., 'Predicting Splicing from Primary Sequence with Deep Learning,' Cell 2019, PMID 30661751.
• The variant is a 12-bp in-frame deletion (p.Thr672_Arg675del), not a missense or synonymous substitution, so REVEL/BayesDel/aGVGD missense predictors and BP7 do not apply to this variant.

Still missing

• No gene-specific VCEP/CSPEC PP3/BP4 lookup table available for AXIN2 to confirm or override the generic SpliceAI threshold.
• REVEL and BayesDel scores were not available in prefetch (not applicable in any case, as this is an in-frame deletion, not a missense variant).

Sources

Reasoning

The exact ClinVar record has no pathogenic or likely pathogenic expert-panel assertion; therefore PP5 is not applicable.

Rule

Apply PP5 only to an exact-variant ClinVar expert-panel Pathogenic or Likely pathogenic assertion.

Evidence considered

• The exact-variant ClinVar audit identified 15 submissions and zero expert-panel submissions; all recorded classifications are benign. Ordinary laboratory submissions were not used for PP5.

Sources

Reasoning

Not applied because BA1 is met; BA1 is stand-alone benign evidence and supersedes use of the lower-strength BS1 frequency criterion for the same population-frequency observation.

Rule

Generic ACMG/AMP hierarchy: do not co-apply BS1 when BA1 is met from the same frequency evidence.

Evidence considered

• The same high, replicated African/African American frequency that supports BA1 would also be incompatible with a rare fully penetrant Mendelian disorder, but BS1 is not double-counted once BA1 is assigned.

Sources

Reasoning

The exact ClinVar record has no benign or likely benign expert-panel assertion; therefore BP6 is not applicable.

Rule

Apply BP6 only to an exact-variant ClinVar expert-panel Benign or Likely benign assertion.

Evidence considered

• The exact-variant ClinVar audit identified 15 benign laboratory submissions but zero expert-panel submissions. Ordinary laboratory submissions and the aggregate ClinVar label were not used for BP6.

Sources

Reasoning

No proband-level phenotype or parental genotype data are provided; confirmed de novo occurrence cannot be assessed.

Rule

Generic ACMG/AMP 2015 PS2: confirmed de novo occurrence in an affected individual with phenotype consistent with the gene-associated disorder.

Evidence considered

• No AXIN2 c.2013_2024del proband, parental testing, or parentage information is present in the consolidated case-evidence bundle.
• No AXIN2-specific VCEP specification was retrieved; generic ACMG/AMP 2015 criteria apply. PS2 requires a de novo variant in a patient with the disease and no family history, with maternity and paternity confirmed.

Still missing

• Proband phenotype and molecular result context.
• Parental testing results confirming absence of the variant in both biological parents, including parentage confirmation.

Sources

Reasoning

No AXIN2-specific VCEP/ClinGen expert panel specification was found in this case (framework_mode: generic_acmg_fallback), so generic ACMG applies. No validated functional/biochemical assay data (e.g. Wnt/beta-catenin reporter assays, degradation/complex-binding assays) addressing NM_004655.4:c.2013_2024del (p.Thr672_Arg675del) were present anywhere in the case evidence bundle (prefetch, evidence, literature_pass, vcep_materials, fulltext_map). None of the three linked full-text publications (PMID:25741868 ACMG/AMP guideline, PMID:26467025 Sherloc-related scoring paper, PMID:25394175 cancer predisposition referral guideline) report a functional assay of this variant; they are methodological/guideline references, not primary functional studies.

Still missing

• No validated in vitro or in vivo functional assay (e.g. TOPflash/Wnt reporter, AXIN2 protein stability/degradation assay) for this specific in-frame deletion was located in the case bundle.
• No AXIN2 ClinGen VCEP specification available to define assay validation criteria or applicable evidence strength.

Reasoning

No variant-specific case-control or cohort enrichment data were provided for NM_004655.4:c.2013_2024del (p.Thr672_Arg675del).

Rule

Generic ACMG/AMP PS4: increased prevalence in affected individuals compared with controls must be demonstrated.

Evidence considered

• No retrieved publication reports this exact AXIN2 variant in affected cases or supplies case-control enrichment data.

Still missing

• A variant-specific study reporting affected-case and ancestry-matched control counts, with an enrichment analysis, is required.

Reasoning

Not assessed. No affected-proband observation documents this variant with a pathogenic/likely pathogenic AXIN2 variant, and no phase result establishes trans configuration. AXIN2 has no retrieved gene-specific VCEP specification, so generic ACMG/AMP cis/trans guidance applies.

Rule

Generic ACMG/AMP PM3: in a recessive disorder, a variant established in trans with a pathogenic variant supports PM3.

Evidence considered

• The supplied ClinVar record is an exact match for NM_004655.4:c.2013_2024del (p.Thr672_Arg675del), but the supplied record contains no co-occurring AXIN2 variant, affected-proband genotype, or phase result.
• Generic ACMG/AMP guidance (PMID:25741868) specifies PM3 when a variant in a recessive-disorder gene is established in trans with a pathogenic variant; no such observation is supplied here.

Still missing

• Affected-proband genotype data identifying a pathogenic/likely pathogenic second AXIN2 variant.
• Validated parental, segregation, read-based, or other phase evidence showing the two variants are in trans.
• Evidence that the relevant AXIN2 disease mechanism is recessive for the assessed phenotype.

Sources

Reasoning

No proband-level phenotype or parental genotype data are provided; an assumed de novo occurrence cannot be assessed.

Rule

Generic ACMG/AMP 2015 PM6: assumed de novo occurrence in an affected individual, without confirmed parentage.

Evidence considered

• No AXIN2 c.2013_2024del proband or parental testing information is present in the consolidated case-evidence bundle.
• No AXIN2-specific VCEP specification was retrieved; generic ACMG/AMP 2015 criteria apply. PM6 is used for a de novo occurrence when maternity and paternity are not confirmed.

Still missing

• Proband phenotype and molecular result context.
• Parental testing results showing the variant is absent in both parents when parentage has not been confirmed.

Sources

Reasoning

No genotyped relatives, phenotype assignments, or informative meioses are provided; co-segregation cannot be evaluated.

Rule

Generic ACMG/AMP 2015 PP1: co-segregation with disease in multiple affected family members, with strength dependent on the segregations available.

Evidence considered

• The consolidated case-evidence bundle contains no pedigree, family genotypes, affected-relative data, or meiosis count for AXIN2 c.2013_2024del.
• No AXIN2-specific VCEP specification was retrieved; generic ACMG/AMP 2015 criteria apply. PP1 requires co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease.

Still missing

• Pedigree with affected/unaffected status and ages or clinical evaluations.
• Genotypes for informative relatives and the number of independent informative meioses.

Sources

Reasoning

No individual-level phenotype, diagnostic phenotype, or phenotype-to-gene specificity evidence was provided for the tested individual or for carriers of this exact variant.

Rule

Generic ACMG/AMP PP4: the patient's phenotype or family history must be highly specific for a disease with a single genetic etiology.

Evidence considered

• The available records contain no patient phenotype or exact-variant phenotype series suitable for PP4 assessment.

Still missing

• A documented phenotype that is highly specific for AXIN2-related disease, together with evidence that alternative causes have been considered, is required.

Reasoning

Not assessed. Although population databases report numerous homozygotes, the available records do not document individual phenotypes, ages, or disease status. In AXIN2-associated oligodontia-cancer predisposition, these aggregate counts alone cannot establish observation in adequately phenotyped unaffected individuals or resolve age-related cancer risk and penetrance.

Rule

Generic ACMG/AMP BS2 requires observation in healthy adult individuals for a fully penetrant early-onset disorder; aggregate population homozygote counts without phenotype and age data are insufficient here.

Evidence considered

• gnomAD v4.1 reports 229 total homozygotes (220 African/African American); gnomAD v2.1 reports 66 total homozygotes; gnomAD-Canada v1.0 reports three homozygotes. None of the supplied population records establishes that these individuals are unaffected for the relevant AXIN2-associated phenotype at an informative age.

Still missing

• Phenotype- and age-annotated observations of homozygous individuals, with disease-specific penetrance context, are needed for BS2.

Sources

Reasoning

Same as PS3: no AXIN2 VCEP specification found (generic ACMG fallback), and no validated functional assay data of any kind (pathogenic-consistent or benign-consistent) for this variant were present in the case evidence bundle. Absence of PS3-supporting functional data does not itself establish BS3; a benign-consistent functional assay result was not located.

Still missing

• No validated functional assay result (of any direction) for NM_004655.4:c.2013_2024del was found in the case bundle.
• No AXIN2 ClinGen VCEP specification available to define assay validation criteria or applicable evidence strength.

Reasoning

No informative family phenotypes and genotypes are provided; lack of segregation cannot be established.

Rule

Generic ACMG/AMP 2015 BS4: lack of segregation in affected family members, assessed using informative family data.

Evidence considered

• The consolidated case-evidence bundle contains no family genotype-phenotype data for AXIN2 c.2013_2024del.
• No AXIN2-specific VCEP specification was retrieved; generic ACMG/AMP 2015 criteria apply. BS4 requires lack of segregation in affected family members.

Still missing

• Pedigree with reliably assessed disease status and ages for variant-positive relatives.
• Genotypes demonstrating the variant in informative unaffected relatives or its absence in affected relatives, with phenocopy and penetrance considerations addressed.

Sources

Reasoning

Not assessed. No pathogenic/likely pathogenic second AXIN2 variant and no cis/trans phase result are supplied. Therefore neither a cis observation in a recessive setting nor a trans observation in a dominant setting can be evaluated under generic ACMG/AMP guidance.

Rule

Generic ACMG/AMP BP2: assess only when phase is known with a pathogenic variant and the relevant recessive or dominant disease model is established.

Evidence considered

• The supplied ClinVar exact-variant record contains no co-occurring AXIN2 variant or phase information for this allele.
• Generic ACMG/AMP guidance (PMID:25741868) describes BP2 from cis co-occurrence in a recessive setting or trans co-occurrence with a pathogenic variant in a dominant setting; neither observation is available in the supplied evidence.

Still missing

• A documented pathogenic/likely pathogenic second AXIN2 variant in the same individual.
• Validated phase evidence establishing cis or trans configuration.
• Phenotype-specific inheritance-mode and penetrance information needed to determine whether the applicable BP2 scenario is relevant.

Sources

Reasoning

No individual-level evidence was provided showing that an alternate molecular diagnosis fully explains the phenotype in a carrier of this exact variant.

Rule

Generic ACMG/AMP BP5: the variant is found in a case with an alternate molecular basis for disease.

Evidence considered

• No co-occurring molecular diagnosis or phenotype attribution data were retrieved for this exact variant.

Still missing

• A documented affected individual with an independently established alternate molecular diagnosis sufficient to explain the phenotype is required.