APC c.221-29G>C, p.?

Reasoning

The governing ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for APC (v2.1) defines BA1 as stand-alone benign when the gnomAD Popmax Filtering Allele Frequency (AF) is >= 0.1% (0.001), and its instructions for this criterion state: 'General recommendation: Use the non-cancer dataset from gnomAD (v2.1.1)'. In the gnomAD v2.1 non-cancer dataset (same variant record page as registry key gnomad_v2) the variant is carried on 252/233,222 non-cancer exome alleles (AF 0.00108052, 0.10805%), and the non-cancer exome popmax (grpmax) filtering AF is 0.00476693 (0.4767%) - approximately 4.8-fold above the 0.1% cutoff. The South Asian subpopulation drives the popmax value with 165/30,298 alleles (AF 0.5446%); the highest single-population figure in the v3.1 non-cancer genomes (HGDP:BRAHUI 2/44, AF 4.5%) is a small-sample artifact and is not the basis of this call. Independent datasets agree in direction and magnitude: gnomAD v2.1 all-comers popmax FAF 0.4784% (total AF 0.09651%, 269/278,728, 3 homozygotes), gnomAD v4.1 popmax FAF 0.5074% (total AF 0.07822%, 1,191/1,522,624, 9 homozygotes), gnomAD v3.1 non-cancer genomes popmax FAF 0.4118% (AF 0.04261%, 63/147,868, 0 homozygotes), and gnomAD-Canada v1.0 genomes AF 0.10859% (20/18,418, 0 homozygotes). The BA1 verdict rests on the popmax filtering AF arm named by the specification, not on the combined non-cancer exome+genome total, which is 260/264,558 alleles (AF 0.09827%) and sits marginally below the 0.1% figure - a threshold artifact of dataset composition rather than evidence against BA1. Disease-expectation context reinforces the call: FAP is autosomal dominant, close to 100% penetrant and of childhood/young-adult onset (PMID:25645574), so an allele at roughly 1 in 926 in a non-cancer adult population is far more common than this disorder permits. Because BA1 is met at stand-alone strength, the classification contribution of this criterion alone is Benign.

Rule

BA1 (Benign Stand Alone), InSiGHT APC specification v2.1: gnomAD Popmax Filtering Allele Frequency (AF) >= 0.1% (0.001), evaluated on the gnomAD v2.1.1 non-cancer dataset.

Evidence considered

• Governing specification: ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel Specifications for APC, version 2.1 (cspec). BA1 rule text: 'GnomAD Popmax Filtering Allele Frequency (AF) >= 0.1% (0.001)'; default strength 'Benign Stand Alone'; instructionsToUse: 'General recommendation: Use the non-cancer dataset from gnomAD (v2.1.1)'.
• gnomAD v2.1 non-cancer subset, exomes (same variant record page as registry key gnomad_v2, build GRCh37, variant id 5-112102857-G-C): 252/233,222 alleles, AF 0.00108052 (0.10805%), homozygotes = 3; South Asian 165/30,298 alleles, AF 0.00544590 (0.54459%), homozygotes = 3; non-cancer exome popmax (grpmax) filtering AF 0.00476693 (0.4767%). This popmax filtering AF exceeds the 0.1% BA1 cutoff.
• gnomAD v2.1 all-comers (gnomad_v2) for comparison: total AF 0.00096510 (0.09651%, 269/278,728 alleles, 3 homozygotes), exome AF 0.00105501 (0.10550%, 261/247,392 alleles), popmax filtering AF 0.00478413 (0.4784%), South Asian AF 0.00546268 (0.54627%).
• gnomAD v4.1 all-comers (gnomad_v4, build GRCh38, variant id 5-112767160-G-C): total AF 0.00078220 (0.07822%, 1,191/1,522,624 alleles, 9 homozygotes), joint popmax filtering AF 0.00507372 (0.5074%), South Asian AF 0.00547589 (0.54759%, 488/89,118 alleles, 7 homozygotes).
• gnomAD v2.1 non-cancer combined exome+genome total: 260/264,558 alleles, AF 0.00098277 (0.09827%), homozygotes = 3; the genome-only non-cancer component has popmax filtering AF 0.00016915 (0.0169%). Flagged as a dataset-composition artifact: the combined total falls marginally below the 0.1% cutoff while both the exome-only figures and the popmax filtering AF arm exceed it.
• gnomAD v3.1 non-cancer genomes (genome-only dataset; not the source specified by the VCEP, used only as corroboration): total AF 0.00042606 (0.04261%, 63/147,868 alleles, 0 homozygotes), popmax filtering AF 0.00411844 (0.4118%); the highest-frequency population reported is HGDP:BRAHUI at 2/44 alleles (AF 4.55%), a small-sample artifact not used in this call.
• gnomAD-Canada v1.0 (HostSeq genomes, registry key gnomad_canada): AF 0.00108589 (0.10859%, 20/18,418 alleles, 0 homozygotes).
• Disease-expectation context for a frequency-based benign call: FAP is dominantly inherited with penetrance close to 100% (PMID:25645574), so an allele frequency near 0.1% (about 1 in 926 alleles) in a non-cancer adult cohort vastly exceeds the frequency compatible with this disorder.
• Generic ACMG stand-alone threshold (>= 0.05, PMID:25741868) was not applied because the APC VCEP specification takes precedence and supplies its own lower, gene-specific BA1 cutoff.

Sources

Reasoning

The governing InSiGHT APC specification (v2.1) defines BS1 at strong strength when the gnomAD Popmax Filtering Allele Frequency is >= 0.001% (0.00001), and directs use of the non-cancer dataset from gnomAD v2.1.1 for this criterion. The same measurement used for BA1 satisfies it by a wide margin: the non-cancer exome popmax (grpmax) filtering AF is 0.00476693 (0.4767%), roughly 477-fold above the 0.001% cutoff, and the non-cancer exome allele frequency itself is 0.10805% (252/233,222 alleles), which is also above the BS1 threshold on its own. All-comers data are concordant (gnomAD v2.1 popmax FAF 0.4784%; gnomAD v4.1 popmax FAF 0.5074%). The allele frequency in a non-cancer population is therefore far greater than expected for FAP, an autosomal dominant disorder with penetrance close to 100% (PMID:25645574). An important limitation is recorded rather than hidden: BA1 is met from the identical frequency observation, so BS1 adds nothing to the evidence tally and the two criteria must not be summed - they are one measurement applied twice. BS1 is reported as met (threshold satisfied) with the subsumption flagged for downstream combination logic.

Rule

BS1 (Benign Strong), InSiGHT APC specification v2.1: gnomAD Popmax Filtering Allele Frequency (AF) >= 0.001% (0.00001), evaluated on the gnomAD v2.1.1 non-cancer dataset.

Evidence considered

• Governing specification: ClinGen InSiGHT APC specification v2.1 (cspec). BS1 rule text: 'GnomAD Popmax Filtering Allele Frequency (AF) >= 0.001% (0.00001)'; default strength 'Benign Strong'; instructionsToUse: 'General recommendation: Use the non-cancer dataset from gnomAD (v2.1.1)'.
• gnomAD v2.1 non-cancer exomes (gnomad_v2 record page): popmax (grpmax) filtering AF 0.00476693 (0.4767%) versus the 0.001% BS1 cutoff; allele frequency 0.10805% (252/233,222 alleles).
• gnomAD v2.1 all-comers popmax filtering AF 0.00478413 (0.4784%), total AF 0.09651% (269/278,728 alleles); gnomAD v4.1 popmax filtering AF 0.00507372 (0.5074%), total AF 0.07822% (1,191/1,522,624 alleles); gnomAD-Canada v1.0 AF 0.10859% (20/18,418 alleles). All exceed the 0.001% threshold.
• gnomAD v3.1 non-cancer genomes popmax filtering AF 0.00411844 (0.4118%), total AF 0.04261% (63/147,868 alleles), used only as corroboration.
• Subsumption note: BA1 (>= 0.1% popmax filtering AF) is also met from this same measurement, so BS1 is not additive with BA1.
• Generic ACMG BS1 defaults (>= 0.01, PMID:28518168; PMID:25741868) were not applied because the APC VCEP specification takes precedence.

Still missing

• BS1 and BA1 derive from the same popmax filtering AF measurement; BS1 must not be counted additively with BA1 in the combination step.

Sources

Reasoning

The InSiGHT APC specification (v2.1) defines BS2 at strong strength as '>= 10 points for healthy individuals OR >= 2 times in homozygous state', and its instructions for this criterion state that the gnomAD v2.1.1 non-cancer dataset cannot be used for heterozygous healthy individuals (limited phenotype information, and it is normally consumed by BA1/BS1) but 'can be used to search for homozygous individuals'. That permitted route is satisfied here: the gnomAD v2.1 non-cancer exomes contain 3 homozygotes for this allele (all 3 in the South Asian subpopulation, 165/30,298 alleles), which is above the >= 2 homozygous observation requirement, and gnomAD v4.1 independently reports 9 homozygotes (7 South Asian, 1 Middle Eastern, 1 non-Finnish European). The penetrance premise that makes a homozygous observation in an unselected adult cohort strong benign evidence is met: FAP is dominantly inherited and close to 100% penetrant with childhood/young-adult onset (PMID:25645574), so homozygotes for a truly causal APC allele should not be present, let alone recurrent, in a non-cancer population cohort. The second arm (>= 10 healthy-individual points) is not reachable from this evidence, because the non-cancer dataset is explicitly barred for heterozygous healthy individuals and no individually phenotyped healthy carriers were available; BS2 therefore rests on the homozygote arm alone. The residual uncertainty is that gnomAD's non-cancer subset carries limited phenotype annotation and these homozygotes are not individually phenotype-verified, so the call is reported at strong rather than very strong strength, which is the maximum the specification allows for this criterion in any case.

Rule

BS2 (Benign Strong), InSiGHT APC specification v2.1: >= 10 points for healthy individuals OR >= 2 times in homozygous state, with homozygote ascertainment permitted from the gnomAD v2.1.1 non-cancer dataset.

Evidence considered

• Governing specification: ClinGen InSiGHT APC specification v2.1 (cspec). BS2 strong rule text: '>= 10 points for healthy individuals OR >= 2 times in homozygous state'; instructionsToUse: 'The non-cancer dataset from gnomAD (v2.1.1) cannot be used for heterozygous healthy individuals, because of the limited phenotype information and since it is usually already used for BA1/BS1. However, the non-cancer dataset from gnomAD (v2.1.1) can be used to search for homozygous individuals.'
• gnomAD v2.1 non-cancer exomes (gnomad_v2 record page): homozygote count = 3 (233,222 alleles; 252 allele carriers), with all 3 homozygotes in the South Asian subpopulation (165/30,298 alleles, homozygote count 3). Three homozygous observations satisfy the specification's >= 2 requirement.
• gnomAD v4.1 (gnomad_v4) independent corroboration: 9 homozygotes in total (7 South Asian, 1 Middle Eastern, 1 non-Finnish European) among 1,522,624 alleles.
• gnomAD v3.1 non-cancer genomes: 0 homozygotes among 147,868 alleles - noted for completeness, but this genome-only dataset is not the source specified by the VCEP and its lower homozygote count does not offset the v2.1.1 homozygotes.
• gnomAD-Canada v1.0: 0 homozygotes among 18,418 alleles.
• Penetrance premise for interpreting homozygotes as benign evidence: FAP is dominantly inherited and close to 100% penetrant with childhood/young-adult onset (PMID:25645574), so recurrent homozygotes for a causal allele should not occur in a non-cancer adult cohort.
• The case-level evidence contains no individually phenotyped healthy homozygous carrier, so the specimen-level 'healthy individual' points route (1 point per age >= 50 with fewer than 5 adenomas, 0.5 point for control/non-cancer keywords) could not be scored; this does not affect the homozygote arm.

Still missing

• gnomAD's non-cancer subset has limited phenotype information; the 3 homozygotes are not individually confirmed as unaffected adults with colonoscopy data, so homozygote ascertainment is population-level rather than phenotype-verified.
• The '>= 10 points for healthy individuals' arm of BS2 was not assessable (the specification bars using the non-cancer dataset for heterozygous healthy individuals), so BS2 rests on the homozygous-state arm alone.

Sources

Reasoning

Scope first: BP4 is evaluated on the SpliceAI path because NM_001127510.3:c.221-29G>C is an intronic change (29 nucleotides upstream of the exon 3 acceptor of APC transcript variant 2), not a missense change; the APC VCEP explicitly states BP4 is not applicable to missense variants and is reserved for synonymous (silent) or intronic variants, so this variant is inside the criterion's scope. The governing rule (ClinGen InSiGHT APC v2.1) is satisfied when 'multiple in silico splicing predictors suggest no impact on gene or gene product'. SpliceAI Lookup returns a max delta score of 0.02 (DS_AG 0.001, DS_AL 0.013, DS_DG 0.000, DS_DL 0.020), i.e. no native acceptor or donor site is weakened and no cryptic site is created, and the same lookup reports Pangolin SG 0.001 / Pangolin SL -0.007 - two independent algorithms concurring on no impact. Against the published SpliceAI benign threshold of <=0.1 (Jaganathan et al. 2019, Cell, PMID:30661751), the 0.02 max delta is comfortably below the cutoff, and it also sits below the VCEP's own 'moderate cryptic gain' band of 0.2-0.8. BP4 is met at supporting strength. The APC VCEP permits BP7 together with BP4 ('The use of BP7 with BP4 is allowed'), so the shared predictor evidence is not treated as an impermissible double count.

Rule

APC VCEP v2.1 BP4 (supporting): 'Synonymous (silent) or intronic variants: Multiple in silico splicing predictors suggest no impact on gene or gene product.' (Missense variants: BP4 is not applicable.)

Evidence considered

• APC VCEP v2.1 BP4 (benign supporting) rule: 'Missense variants: BP4 is not applicable. Synonymous (silent) or intronic variants: Multiple in silico splicing predictors suggest no impact on gene or gene product.' Recommended programs are referenced to the panel's PS1 instructions (SpliceAI, MaxEntScan, VarSeak).
• Variant NM_001127510.3:c.221-29G>C is intronic (29 nucleotides upstream of the exon 3 acceptor; NC_000005.10:g.112767160G>C), which places it in BP4's synonymous/intronic scope rather than the inapplicable missense scope.
• SpliceAI Lookup result: max delta 0.02 (DS_AL 0.013, DS_DL 0.020, DS_AG 0.001, DS_DG 0.000) with Pangolin SG 0.001 and Pangolin SL -0.007 - two independent algorithms agree that the native splice sites are unaffected and no cryptic site is created.
• Generic calibrated SpliceAI BP4 threshold of <=0.1 (Jaganathan et al. 2019, Cell, PMID:30661751) applied because the APC VCEP supplies no separate numeric cutoff; the observed max delta of 0.02 is well below it, so BP4 supporting is met.
• The APC VCEP explicitly allows BP4 to be combined with BP7 ('The use of BP7 with BP4 is allowed'), so the identical predictor set underpinning both codes is applied as the specification intends rather than counted twice against an unstated prohibition.

Still missing

• MaxEntScan and VarSeak (also recommended in the APC VCEP specification) were not retrieved; the two available algorithms both predict no splice impact, so the 'multiple predictors' condition of the VCEP BP4 rule is already satisfied.

Sources

Reasoning

Scope first: the generic ACMG BP7 wording is limited to synonymous variants, but the governing gene-specific specification expressly widens it - the ClinGen InSiGHT APC v2.1 BP7 supporting rule reads: 'A synonymous (silent) or intronic variant at or beyond +7/-21 for which multiple splicing prediction algorithms predict no impact to the splice consensus sequence nor the creation of a new splice site.' The gene-specific rule governs, and NM_001127510.3:c.221-29G>C is an intronic variant 29 nucleotides from the exon 3 acceptor, i.e. beyond the -21 boundary of the exon-proximal region, so it is within BP7's scope rather than outside it. Two algorithms in the SpliceAI Lookup output agree there is no splice impact: SpliceAI max delta 0.02 (DS_AG 0.001, DS_AL 0.013, DS_DG 0.000, DS_DL 0.020 - no consensus sequence affected and no new site created, since both gain scores are ~0) and Pangolin SG 0.001 / SL -0.007. This satisfies the 'multiple splicing prediction algorithms predict no impact' requirement, and the result is far below the closest published benign benchmark for splice predictors (SpliceAI <=0.1; Jaganathan et al. 2019, Cell, PMID:30661751) as well as below the VCEP's 0.2-0.8 moderate cryptic-gain band. BP7 is met at supporting strength; the APC VCEP states that use of BP7 with BP4 is allowed, so this does not double-count the shared prediction with BP4.

Rule

APC VCEP v2.1 BP7 (supporting): 'A synonymous (silent) or intronic variant at or beyond +7/-21 for which multiple splicing prediction algorithms predict no impact to the splice consensus sequence nor the creation of a new splice site.'

Evidence considered

• APC VCEP v2.1 BP7 (benign supporting) rule: 'A synonymous (silent) or intronic variant at or beyond +7/-21 for which multiple splicing prediction algorithms predict no impact to the splice consensus sequence nor the creation of a new splice site.' This gene-specific wording extends BP7 beyond the generic synonymous-only definition and therefore covers the intronic variant assessed here.
• NM_001127510.3:c.221-29G>C sits 29 nucleotides upstream of the exon 3 acceptor site (NC_000005.10:g.112767160G>C), i.e. beyond the -21 cutoff, so its position is inside the VCEP's BP7 scope; it is not a canonical +/-1 or 2 splice site (the APC VCEP's PVS1 decision-tree lists cover only those positions).
• The APC VCEP supplementary material records that the APC BP7 rule was modified from the ClinGen CDH1 specification and that use of BP7 together with BP4 is allowed.
• SpliceAI Lookup result for this variant: max delta 0.02, with DS_DL 0.020 and DS_AL 0.013 (no native donor/acceptor site lost), DS_AG 0.001 and DS_DG 0.000 (no new splice site created), plus Pangolin SG 0.001 and Pangolin SL -0.007. Two algorithms predict no impact on the splice consensus sequence and no creation of a cryptic site.
• The published SpliceAI benign benchmark of <=0.1 (Jaganathan et al. 2019, Cell, PMID:30661751) is far exceeded by the observed 0.02, and the VCEP's own guidance (native site loss at 0.8-1.0; strong cryptic gain 0.8-1.0, moderate 0.2-0.8) is likewise not approached.

Still missing

• MaxEntScan and VarSeak were not retrieved, but SpliceAI and Pangolin (two independent algorithms) both predict no impact, satisfying the VCEP requirement for multiple algorithms.

Sources

Reasoning

The governing APC VCEP specification retains PS1 for exactly two situations: (1) a missense variant producing the same amino-acid change as a previously established (Likely) Pathogenic APC missense variant - currently only c.3077A>G p.(Asn1026Ser) and c.3084T>A p.(Ser1028Arg), both Likely Pathogenic, so other variants at those residues reach only PS1_Moderate - and (2) a splice variant that affects splicing at the same nucleotide as a previously established (Likely) Pathogenic variant, provided splice prediction is above the defined thresholds or similar to the established variant by multiple in silico predictors. NM_001127510.3:c.221-29G>C is intronic, 29 bp upstream of the acceptor site of the exon beginning at c.221 (NM_000038.6 exon 4 / NM_001127510.2 exon 5), with NP_001120982.1:p.? - it encodes no amino-acid substitution, so the same-amino-acid arm is inapplicable and only the same-nucleotide splice arm is testable. That arm fails: no previously established Pathogenic or Likely Pathogenic APC variant at this nucleotide exists. The only ClinVar record at this position is this variant itself (VCV000217948 / ClinVarID 217948), asserted Likely benign by four submitters including two clinical laboratories, with criteria provided by multiple submitters and no conflicts (2 stars, no expert-panel review, no criterion leads), COSMIC returns no entry, and no fetched publication (PMID:21368914, PMID:25645574, PMID:25741868, PMID:27854360) mentions c.221-29 or c.221 at all. The assessed variant's own splice effect also would not qualify, since SpliceAI reports a maximum delta score of 0.02 (DS_AG 0.001, DS_AL 0.013, DS_DG 0.000, DS_DL 0.020), far below the VCEP's SpliceAI ranges used for splice consequence evaluation (cryptic site gain considered moderate at 0.2-0.8 and strong at 0.8-1.0) and below the generic supporting PP3 cutoff of 0.2. Finally, the variant is present in gnomAD v2.1 at total AF 0.0965% (269/278,728 alleles, 3 homozygotes) with a South Asian frequency of 0.546% (166/30,388 alleles), which is not compatible with a fully penetrant pathogenic splice allele. PS1 is therefore not met.

Rule

APC VCEP PS1: the variant under assessment must either result in the same amino-acid change as a previously established Pathogenic (PS1_Strong) or Likely Pathogenic (PS1_Moderate) APC missense variant, or - for splice variants - affect splicing at the same nucleotide as a previously established (Likely) Pathogenic variant with splice prediction above the defined thresholds or similar to the established variant by multiple in silico predictors.

Evidence considered

• The APC VCEP PS1 rule (ClinGen InSiGHT specification v2/v2.1, PS1 row, supplementary material page 2) requires for missense variants the same amino-acid change as a previously established Pathogenic/Likely Pathogenic variant, and for splice variants that the variant affect splicing at the same nucleotide as a previously established (Likely) Pathogenic variant, with splice prediction above the defined thresholds or similar to that variant by multiple in silico predictors.
• The VCEP states that only two amino-acid positions (p.Asn1026 and p.Ser1028, i.e. c.3077A>G and c.3084T>A) currently host Likely Pathogenic APC missense variants, and that other variants leading to the same missense change at those positions meet PS1_Moderate.
• NM_001127510.3:c.221-29G>C is intronic: normalization gives no amino-acid change (NP_001120982.1:p.?), so the same-amino-acid arm of PS1 cannot be satisfied and no residue is available for comparison.
• ClinVar (variation 217948) lists this exact variant as Likely benign with criteria provided by multiple submitters and no conflicts (2 stars; 4 submissions, 2 from clinical laboratories, no expert-panel submission), so the same nucleotide carries no previously established pathogenic assertion; no criterion leads were extracted from those submissions.
• SpliceAI lookup for NM_001127510.3:c.221-29G>C reports max delta score 0.02 (DS_AG 0.001, DS_AL 0.013, DS_DG 0.000, DS_DL 0.020; Pangolin SG 0.001, SL -0.007), an indeterminate/no-impact result that is below the VCEP SpliceAI ranges applied to splice-consequence evaluation (moderate 0.2-0.8 and strong 0.8-1.0 for cryptic site gain) and below the generic 0.2 supporting splice cutoff.
• gnomAD v2.1 exomes/genomes for 5-112102857-G-C report total AF 0.0965% (269/278,728 alleles, 3 homozygotes; grpmax FAF 0.478%), with the highest population frequency in South Asians (AF 0.546%, 166/30,388, 3 homozygotes); this stands against a pathogenic same-nucleotide splice designation.
• COSMIC returned no result for APC c.221-29G>C, and a text search of all fetched full texts (PMID:21368914, PMID:25645574, PMID:25741868, PMID:27854360) found no occurrence of c.221-29 or c.221, so no published same-nucleotide (or same-residue) comparator variant exists for PS1.

Sources

Reasoning

The InSiGHT APC specification (v2.1) defines PM2 as 'Rare in controls, defined by an allele frequency <= 0.0003% (0.000003) if the allele count is > 1 OR by an allele frequency < 0.001% (0.00001) if the allele count is <= 1', and directs use of the total population from the non-cancer dataset of gnomAD v2.1.1. The allele count in the non-cancer exomes is 252, which is greater than 1, so the applicable ceiling is 0.000003 (0.0003%). The observed non-cancer exome frequency is 0.00108052 (0.10805%, 252/233,222 alleles), approximately 360-fold above that ceiling; the combined non-cancer exome+genome total is 0.00098277 (0.09827%, 260/264,558 alleles), also far above it, and the non-cancer genome-only component is 0.00025530 (0.02553%). The allele is likewise not rare in all-comers gnomAD: v2.1 total AF 0.09651% (269/278,728 alleles) and v4.1 total AF 0.07822% (1,191/1,522,624 alleles). Every available frequency source, including the lowest (v3.1 non-cancer genomes at 0.04261%, 63/147,868 alleles), exceeds the PM2 rarity ceiling by one to two orders of magnitude, so no dataset choice alters the outcome. The variant is a common, recurrent population allele rather than a rare one, and PM2 (rare in controls) is therefore not met; because it fails on frequency alone, no strength was assigned.

Rule

PM2 (rare in controls), InSiGHT APC specification v2.1: AF <= 0.0003% (0.000003) when allele count > 1, or < 0.001% (0.00001) when allele count <= 1, evaluated on the gnomAD v2.1.1 non-cancer total population.

Evidence considered

• Governing specification: ClinGen InSiGHT APC specification v2.1 (cspec). PM2 rule text: 'Rare in controls, defined by an allele frequency <= 0.0003% (0.000003) if the allele count is > 1 OR by an allele frequency < 0.001% (0.00001) if the allele count is <= 1'; default strength 'Pathogenic Moderate' with a Supporting-strength rule; instructionsToUse: 'General recommendation: Use the total population from the non-cancer dataset from gnomAD (v2.1.1)'.
• gnomAD v2.1 non-cancer exomes (gnomad_v2 record page): 252/233,222 alleles, AF 0.00108052 (0.10805%) with allele count > 1, versus the applicable 0.000003 (0.0003%) PM2 ceiling - not rare in controls.
• gnomAD v2.1 non-cancer combined exome+genome total population: 260/264,558 alleles, AF 0.00098277 (0.09827%); the non-cancer genome-only component is 8/31,336 alleles, AF 0.00025530 (0.02553%). Both remain far above the PM2 ceiling.
• gnomAD v2.1 all-comers (gnomad_v2): total AF 0.00096510 (0.09651%, 269/278,728 alleles, 3 homozygotes); South Asian subpopulation AF 0.00546268 (0.54627%, 166/30,388 alleles).
• gnomAD v4.1 all-comers (gnomad_v4): total AF 0.00078220 (0.07822%, 1,191/1,522,624 alleles, 9 homozygotes).
• gnomAD v3.1 non-cancer genomes, the lowest frequency source identified: AF 0.00042606 (0.04261%, 63/147,868 alleles, 0 homozygotes), still orders of magnitude above the PM2 rarity ceiling.
• gnomAD-Canada v1.0: AF 0.00108589 (0.10859%, 20/18,418 alleles, 0 homozygotes).
• The generic ACMG/ClinGen-SVI PM2 supporting threshold (<= 0.0001, PMID:25741868) is also clearly exceeded, so the conclusion is unchanged whether the VCEP or generic threshold is used; the VCEP value is quoted here because the APC specification takes precedence.

Sources

Reasoning

Scope first: NM_001127510.3:c.221-29G>C (NC_000005.10:g.112767160G>C; GRCh37 5-112102857-G-C) is an intronic change 29 nucleotides upstream of the exon 3 acceptor site of APC transcript variant 2 (protein consequence p.?; no amino-acid substitution). PP3/BP4 evidence therefore comes from the SpliceAI path only; REVEL is neither applicable nor available (no REVEL score was retrieved, and it would say nothing about a non-missense change). The governing framework is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for APC v2.1, whose PP3 supporting rule requires that 'multiple in silico splicing predictors support a deleterious effect' for a non-canonical splicing variant (for presumed missense variants it directs that only splicing predictors be used). The specification names SpliceAI, MaxEntScan and VarSeak as recommended programs but supplies no separate numeric PP3 cutoff, so the published calibrated SpliceAI thresholds apply: supporting >=0.2 and moderate >=0.5 (Jaganathan et al. 2019, Cell, PMID:30661751). SpliceAI Lookup for this variant returns a max delta score of 0.02 (DS_AG 0.001, DS_AL 0.013, DS_DG 0.000, DS_DL 0.020) - no loss or gain of an acceptor or donor site - and the same output reports Pangolin SG 0.001 / SL -0.007, likewise indicating no splice effect. Machine-readable consequence is 'sequence_variant', consistent with an intronic change with no coding effect. Not a single available predictor supports a deleterious splicing effect, and 0.02 is an order of magnitude below the 0.2 supporting threshold and far below the VCEP's 0.8-1.0 splice-loss/strong-cryptic-gain band; PP3 is not met.

Rule

APC VCEP v2.1 PP3 (supporting): 'Missense variants: Do not use computational prediction models for conservation, evolution, etc. In silico splicing predictors should be used for presumed missense variants to reveal possible splicing effects. Non-canonical splicing variants: Multiple in silico splicing predictors support a deleterious effect.'

Evidence considered

• APC VCEP v2.1 (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel, doc 1529292791) PP3 supporting rule: for non-canonical splicing variants, multiple in silico splicing predictors must support a deleterious effect; for presumed missense variants only splicing predictors are to be used. Recommended programs are SpliceAI, MaxEntScan and VarSeak, with numeric guidance given in the panel's PS1 instructions: SpliceAI splice-site loss is considered for 0.8-1.0, cryptic-site gain strong for 0.8-1.0 and moderate for 0.2-0.8; MaxEntScan requires >3 for a credible native site and -15% for native site loss (Houdayer et al. 2012, PMID:22505045).
• Variant NM_001127510.3:c.221-29G>C (GRCh38 NC_000005.10:g.112767160G>C; GRCh37 5-112102857-G-C) is intronic, 29 nucleotides upstream of the exon 3 acceptor site of APC transcript variant 2, with predicted protein consequence p.?. The consequence type is therefore non-canonical intronic/splice-region, which selects the SpliceAI evidence path for PP3/BP4 and makes REVEL not applicable.
• SpliceAI Lookup output for this variant: max delta score 0.02, with DS_AL 0.013, DS_DL 0.020, DS_AG 0.001 and DS_DG 0.000 (no acceptor or donor site lost or gained); the same lookup reports Pangolin SG 0.001 and Pangolin SL -0.007, also indicating no splice effect. Scratch/delta positions (DP_AL 29, DP_DL 230) do not correspond to any realigned site use.
• Generic calibrated SpliceAI thresholds were applied because the APC VCEP provides no separate numeric PP3 cutoff: PP3 supporting >=0.2 and PP3 moderate >=0.5 (Jaganathan et al. 2019, Cell, PMID:30661751). The observed max delta 0.02 is below both, and no second predictor supports a deleterious splicing effect, so PP3 supporting is not met.
• The APC VCEP supplementary material records the PP3 modification (based on Lee et al. 2018) and adds no gene-specific numeric splice threshold; it confirms BP4 is not applicable to missense variants, consistent with restricting the in silico splice path used here to the intronic/synonymous scope.

Still missing

• MaxEntScan and VarSeak (the other splice programs recommended in the APC VCEP specification) were not retrieved for this variant; both available algorithms (SpliceAI, Pangolin) predict no splice effect, so their absence cannot create PP3 support and does not change the verdict.

Sources

Reasoning

PP5 can be triggered only by an exact-variant ClinVar expert-panel classification of Pathogenic/Likely pathogenic. No such assertion exists for APC c.221-29G>C. ClinVar variation 217948 is a Likely benign record assembled from four submissions, every one of which is an ordinary laboratory assertion (is_expert_panel = false for all four: Mayo Clinic Laboratories [no assertion criteria provided, research method], KCCC/NGS Laboratory [clinical testing, single submitter], Center for Genomic Medicine, Rigshospitalet [clinical testing, single submitter], Breakthrough Genomics [method not provided]); the submission audit reports expert_panel_submissions = 0, usable_for_criterion_leads = 0, validated_pmids = 0 and an empty criterion_leads list, and the scrape records expert_panels = {} (zero expert-panel assertions). The aggregate record is at best 'criteria provided, multiple submitters, no conflicts' (2-star) - an aggregate laboratory label, which the criterion explicitly excludes - and it is a benign-direction label in any case, not the Pathogenic/Likely pathogenic direction PP5 requires. Independently, the governing APC VCEP v2.1 does not use PP5 at all: 'Reputable source recently reports variant as pathogenic but the evidence is not available to the laboratory to perform an independent evaluation - Not used as recommended by Biesecker et al. 2018 (PMID: 29543229): Expert opinions should only be considered if accompanied by the primary evidence used.' With no expert-panel assertion and no usable primary evidence trail in any submission, PP5 is not met; no numeric threshold or comparison operator applies to this criterion.

Rule

PP5 requires an exact-variant ClinVar expert-panel Pathogenic/Likely pathogenic classification; APC VCEP v2.1 additionally does not use PP5 (Biesecker et al. 2018, PMID:29543229).

Evidence considered

• ClinVar variation 217948 (key 'clinvar'): classification 'Likely benign', review status 'criteria provided, single submitter' on the record page and 'criteria provided, multiple submitters, no conflicts' (2 stars) via the API; source_data.expert_panels = {} and clinvar submission audit is_expert_panel = false for all four submissions, expert_panel_submissions = 0.
• Submission audit detail: accession SCV000256947 (Mayo Clinic Laboratories, research method, no assertion criteria provided, no valid PMIDs), SCV004015881 (KCCC/NGS Laboratory, clinical testing, single submitter, no validated PMID trail), SCV002550551 (Rigshospitalet, clinical testing, single submitter, no validated PMID trail), SCV005226734 (Breakthrough Genomics, method not provided, criteria provided); none is an expert-panel submission and none supplies usable primary evidence.
• APC VCEP v2.1 PP5 row: 'Not used as recommended by Biesecker et al. 2018 (PMID: 29543229): Expert opinions should only be considered if accompanied by the primary evidence used.' - so PP5 is also excluded by the gene-specific specification.
• Triaged PP5 candidate sources (PMID:20301519, PMID:21368914, PMID:27854360, PMID:26389258) were reviewed; none reports the variant as pathogenic or presents primary evidence naming c.221-29G>C, and one of them (PMID:26389258) could not be retrieved in full text.

Still missing

• No ClinGen expert-panel assertion (InSiGHT AP VCEP or equivalent) exists for this exact variant in ClinVar, so PP5 can never be triggered for this variant as currently reported.
• All four ClinVar submissions lack validated primary evidence (validated_pmids = 0; rejected_pmids includes a submission with no valid PMIDs and one with no assertion criteria), so no underlying evidence could be independently reviewed either.

Sources

Reasoning

BP6 may be triggered only by an exact-variant ClinVar expert-panel classification of Benign/Likely benign. The Likely benign label on ClinVar variation 217948 is not that, and must not be used for BP6. The record is an aggregate of four ordinary laboratory submissions - Mayo Clinic Laboratories (research method, no assertion criteria provided), KCCC/NGS Laboratory (clinical testing, single submitter, criteria provided), Center for Genomic Medicine, Rigshospitalet (clinical testing, single submitter, criteria provided) and Breakthrough Genomics (method not provided, criteria provided) - and every one has is_expert_panel = false; the submission audit reports expert_panel_submissions = 0, and the scrape records expert_panels = {}. Two of the four submissions were rejected as criterion leads (no valid PMIDs / no assertion criteria / non-clinical or unclear method) and none has a validated PMID trail (validated_pmids = 0), so there is no expert-panel re-evaluation behind the benign label, only the same laboratory assertions the criterion excludes. The record's aggregate review status ('criteria provided, multiple submitters, no conflicts', 2 stars; 'criteria provided, single submitter' on the record page) is exactly the ordinary-submission category that must never trigger BP6, and worse, the underlying submissions provide no evidence the laboratory could not independently evaluate. Independently, the APC VCEP v2.1 does not use BP6: 'Reputable source recently reports variant as benign but the evidence is not available to the laboratory to perform an independent evaluation - Not used as recommended by Biesecker et al. 2018 (PMID: 29543229): Expert opinions should only be considered if accompanied by the primary evidence used.' BP6 is therefore not met; no numeric threshold or comparison operator applies to this criterion, and its direction must not be inferred from other criteria in this group.

Rule

BP6 requires an exact-variant ClinVar expert-panel Benign/Likely benign classification; APC VCEP v2.1 additionally does not use BP6 (Biesecker et al. 2018, PMID:29543229).

Evidence considered

• ClinVar variation 217948 (key 'clinvar'): aggregate classification 'Likely benign' from four submissions (2 clinical laboratories shown in the record summary), review status 'criteria provided, single submitter' (record page) / 'criteria provided, multiple submitters, no conflicts' (2 stars, API); source_data.expert_panels = {}.
• Submission audit: is_expert_panel = false for all four submissions; expert_panel_submissions = 0; SCV000256947 rejected for no_valid_pmids, no_assertion_criteria and non_clinical_or_unclear_method; validated_pmids = 0 overall.
• APC VCEP v2.1 BP6 row: 'Not used as recommended by Biesecker et al. 2018 (PMID: 29543229): Expert opinions should only be considered if accompanied by the primary evidence used.'
• No retrieved publication (PMID:20301519, PMID:21368914, PMID:25645574, PMID:25741868, PMID:27854360) reports c.221-29G>C at all, so no independent literature-based benign assertion exists either.

Still missing

• No ClinGen expert-panel assertion (Benign/Likely benign) exists for this exact variant, so the only permitted BP6 trigger is absent; the benign direction is carried only by ordinary laboratory submissions.
• No submission provides the underlying primary evidence for its benign call (no validated PMID trail, criteria lists not exposed), so the benign assertion could not be independently evaluated.

Sources

Reasoning

The governing InSiGHT/ClinGen APC VCEP specification makes PS2 applicable and scores it with de novo points: >=4 de novo scores = Very Strong, 2-3.5 = Strong, 1-1.5 = Moderate. The de novo score per proband is assigned from the APC phenotype/parental-relationship tables: a proband with >=1 phenotype point scores 2 if parental relationships are confirmed and 1 if unconfirmed; a proband with 0.5 phenotype points scores 1 or 0.5 respectively. This case contains no proband record, no parental genotypes, no maternity/paternity confirmation, and no phenotype-point data; the four ClinVar submissions on this variant are variant-level assertions only, with no family members, de novo observations, or parental testing reported and no criterion leads. No proband-level de novo score can therefore be computed and PS2 cannot be assessed.

Rule

APC VCEP PS2: >=4 de novo scores = Very Strong; 2-3.5 = Strong; 1-1.5 = Moderate; de novo score per proband from APC Table 2 (2 or 1 when phenotype point >=1, 1 or 0.5 when phenotype point = 0.5, depending on confirmed vs unconfirmed parental relationships).

Evidence considered

• The APC VCEP specification lists PS2 as applicable with de novo score thresholds: >=4 = Very Strong, 2-3.5 = Strong, 1-1.5 = Moderate.
• The APC VCEP supplementary material states that PS2 and PM6 were modified based on the SVI recommendation for de novo criteria.
• APC Table 1/2 converts a proband's phenotype points and parental-relationship status into a de novo score (2 or 1 when phenotype points >=1; 1 or 0.5 when phenotype points = 0.5), so both a proband phenotype score and parental-relationship status are required inputs.
• The ClinVar record for this variant (ClinVarID 217948, 4 submissions, all variant-level) reports no family members, parental testing, de novo occurrence, or criterion leads; the submission audit found zero usable criterion leads.
• No fetched publication mentions APC c.221-29G>C, so no published de novo observation is available.

Still missing

• Proband clinical phenotype and confirmation of an affected individual carrying the variant
• Variant testing results for both biological parents
• Confirmation of maternity and paternity
• Phenotype-point score for the proband (APC Table 1) to convert a de novo event into a de novo score
• Documentation that the family history is negative for the APC phenotype

Sources

Reasoning

The governing framework is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for APC, version 2.1 (framework_mode vcep), which defines PS3 by assay type and magnitude rather than by in silico prediction: RNA assays showing a premature stop codon or in-frame skipping of exon 13 or 14 (Very Strong when full-length transcript is absent, Strong when <10% full-length transcript remains), RNA assays showing in-frame skipping of other exons or alternative-transcript over-expression (Moderate/Supporting), or protein assays showing increased beta-catenin-regulated transcription and/or decreased beta-catenin binding, with the protein route restricted to the beta-catenin binding domain (codons 959-2129). NM_001127510.3:c.221-29G>C is an intronic change 29 bp upstream of the exon 3 acceptor (NM_000038.6(APC):c.221-29G>C; 5:112102857G>C GRCh37), so it falls in the splice-region class the APC VCEP addresses with RNA assay rules. No RNA/transcript analysis, minigene assay, or protein assay for this variant was identified in any retrieved source. The four full-text publications available (PMID:25741868 ACMG/AMP 2015 standards, PMID:25645574 ACG clinical guideline, PMID:21368914 clinical utility gene card, PMID:27854360 ACMG SF v2.0 policy statement) are consensus or guideline documents; none names the variant or any intronic/splice-region APC variant and none reports variant-level functional data. ClinVar variation 217948 contains four submissions (two usable clinical-testing submissions), all classified Likely benign with no assertion criteria or functional evidence listed, no criterion leads, and no validated PMIDs, so it provides no assay-based lead in either direction. SpliceAI predicts no significant splice impact for this variant (max delta score 0.02), but that is an in silico prediction, not the RNA analysis the APC VCEP PS3 rules require, and ACMG/AMP 2015 (PMID:25741868) states that for splice-region variants 'confirmation of impact requires functional analysis by either RNA or protein analysis'. Because the assay evidence class PS3 requires does not exist for this variant, PS3 is not assessed rather than being asserted or refuted.

Rule

APC VCEP v2.1 PS3 (RNA assays showing a premature stop codon or in-frame skipping of exon 13/14, at Very Strong/Strong/Moderate/Supporting by transcript fraction; other in-frame skipping or alternative-transcript over-expression; or protein assay within codons 959-2129).

Evidence considered

• APC VCEP v2.1 (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel) defines PS3 by assay and magnitude: RNA assays showing a premature stop codon or in-frame skipping of exon 13 or 14 (Very Strong when full-length transcript is absent, Strong when <10% of full-length transcript), RNA assays showing other in-frame skipping or alternative-transcript over-expression (Moderate/Supporting), or protein assays showing increased beta-catenin-regulated transcription and/or decreased beta-catenin binding, the latter restricted to the beta-catenin binding domain (codons 959-2129).
• The APC VCEP supplementary materials add only the note that NMD considerations are handled with PVS1; they provide no variant-specific assay data for this variant.
• No RNA/transcript analysis, minigene assay, or protein assay performed on this variant was found in any retrieved source; the four full-text papers retrieved (PMID:25741868, PMID:25645574, PMID:21368914, PMID:27854360) are consensus or guideline documents and none mentions c.221-29G>C or any intronic/splice-region APC variant.
• ClinVar variation 217948 carries four submissions (two usable clinical-testing submissions, from Mayo Clinic Laboratories, Rigshospitalet and KCCC), all classified Likely benign; none lists assertion criteria, functional evidence, criterion leads, or validated PMIDs, so no assay-based signal for PS3 is available from that source.
• SpliceAI reports max delta score 0.02 (DS_AG 0.001, DS_AL 0.013, DS_DG 0.0, DS_DL 0.02) for NM_001127510.3:c.221-29G>C; this is an in silico prediction only and is not the RNA assay evidence the APC VCEP PS3 rules require, so it neither supports nor refutes PS3.
• ACMG/AMP 2015 (PMID:25741868) notes that splice-region variants must be confirmed by functional analysis using RNA or protein methods, and that in silico splice predictors are a single piece of evidence that cannot alone support a clinical assertion.

Still missing

• No RNA/transcript analysis (RT-PCR, transcript sequencing, or splice-junction quantification) of c.221-29G>C in a constitutional patient sample.
• No minigene or other splicing reporter assay for this variant.
• No protein assay (beta-catenin-regulated transcription or beta-catenin binding by surface plasmon resonance) for this variant; note this route is limited by the APC VCEP to codons 959-2129, which c.221-29 does not fall within.
• No functionally characterised comparator or control data at this splice region (c.221-30 to c.221-20) that could be transferred to this variant.
• No ClinVar submission or publication for this variant reports assertion criteria, an assay, or a criterion lead.

Sources

Reasoning

Governing framework: the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for APC v2.1 (registry key 'cspec', doc 1529292791; PDF set 'vcep_apc_specifications_supplementary_material_v2', 'vcep_table_1_262'), which redefines PS4 - 'The prevalence of the variant in affected individuals is significantly increased compared to the prevalence in controls' - via Fortuno et al. 2020 (PMID:33300245) as a phenotype-point score curated per affected proband from Table 1. Operator check for this criterion, quoted from its own VCEP source text (not from any other criterion in this group): '>= 16 phenotype points' = Very Strong; '4-15.5 phenotype points' = Strong; '2-3.5 phenotype points' = Moderate; '1-1.5 phenotype point' = Supporting. The ladder is monotonic upward: strength increases as the point total rises, so 'met' requires points >= the tier minimum; the opposite ('<=' / downward) direction does not apply to PS4. Substituted at the lowest tier: with 0 curated phenotype points, '0 >= 1? no' - i.e. even the Supporting tier fails. That point total is 0 because no affected individual carrying c.221-29G>C is reported with any Table 1 feature (colorectal adenoma number, desmoid, medulloblastoma, hepatoblastoma, CHRPE, gastric or duodenal adenomas, osteomas, typical FAP family history) in any retrieved source: ClinVar variation 217948 carries four ordinary-laboratory submissions with no proband phenotype documentation, 0 expert-panel submissions and 0 usable criterion leads; the GeneReviews chapter APC-Associated Polyposis Conditions (PMID:20301519) does not mention the variant (searches for 'c.221-29' and '29G' returned no hits, its Table 7 of notable APC pathogenic variants lists only the different variant c.221-1G>A); and the four retrieved full texts (PMID:21368914, PMID:25645574, PMID:25741868, PMID:27854360) do not mention the variant or any intronic APC variant. No case-control enrichment estimate (odds ratio or equivalent) for this variant exists in the retrieved evidence, and the only population data available point away from enrichment: the allele is carried in controls at 0.0782% (1191/1,522,624 alleles, 9 homozygotes, gnomAD v4.1) and 0.0965% (269/278,728 alleles, 3 homozygotes, gnomAD v2.1), with popmax 0.5476% in South Asians. Because the case supplies no proband phenotype and no affected carrier of this variant, the phenotype-point total cannot be curated from case data and no enrichment can be demonstrated, so PS4 is not assessed rather than scored; a reviewer should note that on any proband with a documented polyposis phenotype this criterion would have to be re-scored.

Rule

APC VCEP v2.1 PS4 (Fortuno et al. 2020, PMID:33300245): phenotype points per proband from Table 1 - >= 16 (Very Strong), 4-15.5 (Strong), 2-3.5 (Moderate), 1-1.5 (Supporting); higher points = stronger, met requires points >= tier minimum.

Evidence considered

• APC VCEP v2.1 (cspec doc 1529292791) redefines PS4 through the phenotype-point system of Table 1, with the tiers written as '>= 16 phenotype points' (Very Strong), '4-15.5 phenotype points' (Strong), '2-3.5 phenotype points' (Moderate) and '1-1.5 phenotype point' (Supporting); direction is upward, so met requires points >= the tier minimum.
• ClinVar variation 217948 (key 'clinvar'): four submissions, all is_expert_panel = false; submission audit reports expert_panel_submissions = 0, usable_for_criterion_leads = 0, criterion_leads = []; aggregate label 'Likely benign' derives from ordinary laboratory submissions with no proband phenotype documentation.
• GeneReviews 'APC-Associated Polyposis Conditions' (PMID:20301519, NBK1345) was read in full from NCBI Bookshelf and searched for the variant: no hit for 'c.221-29' or '29G'; Table 7 'Notable APC Pathogenic Variants' lists the different variant c.221-1G>A (founder in Newfoundland), confirming the variant under assessment is not a reported notable APC variant.
• Clinical utility gene card for FAP/AFAP (PMID:21368914) and the ACG hereditary GI cancer guideline (PMID:25645574): both are gene-level reviews of APC that never name c.221-29G>C or any intronic APC variant, so neither contributes affected-individual phenotype points.
• ACMG/AMP 2015 standards (PMID:25741868) and ACMG SF v2.0 (PMID:27854360): methodology/policy documents that define PS4 in the generic framework and list APC as a secondary-findings gene; neither reports this variant or any case observation.
• gnomAD v2.1 and v4.1 (keys 'gnomad_v2', 'gnomad_v4'): the variant is present in population controls (0.0965%, 269/278,728 alleles, 3 homozygotes; and 0.0782%, 1191/1,522,624 alleles, 9 homozygotes) with popmax 0.5476% in South Asians, which is evidence against increased prevalence in affected individuals rather than for it.

Still missing

• No proband-level phenotype data exist for this case (no adenoma counts, colonoscopy findings, desmoid/CHRPE/osteoma observations or family history), so the APC VCEP phenotype-point score underlying PS4 cannot be curated.
• No affected individual carrying c.221-29G>C is reported in any retrieved source (ClinVar criterion leads = 0; variant absent from PMID:20301519, PMID:21368914, PMID:25645574, PMID:25741868, PMID:27854360), so no case-control enrichment estimate exists.
• PMID:26389258 (PDQ Cancer Genetics Risk Assessment and Counseling) was triaged as a PS4/PP5 candidate source but no full text was retrieved (abstract page screenshot only); it would not be expected to carry variant-level phenotype data, but it was not checkable for this variant.

Sources

Reasoning

The governing InSiGHT/ClinGen APC VCEP specification makes PM6 applicable and scores an assumed de novo event with de novo points: 2-3.5 de novo scores = Strong, 1-1.5 = Moderate, 0.5 = Supporting. The de novo score per proband is taken from the APC phenotype/parental-relationship tables and requires at minimum a proband phenotype point of 0.5 (unconfirmed parental relationships) or 1 (parental relationships confirmed). No case record documents the variant in an affected proband, no parental testing (confirmed or absent) is reported, and no phenotype points are available; the ClinVar submissions are variant-level only and provide no assumed de novo observation or criterion leads. PM6 therefore cannot be assessed.

Rule

APC VCEP PM6: 2-3.5 de novo scores = Strong; 1-1.5 = Moderate; 0.5 = Supporting; de novo score per proband from APC Table 2 using phenotype points and parental-relationship status.

Evidence considered

• The APC VCEP specification lists PM6 as applicable with de novo score thresholds: 2-3.5 = Strong, 1-1.5 = Moderate, 0.5 = Supporting.
• The APC VCEP supplementary material states that PM6 (assumed de novo without confirmation of paternity and maternity) was modified based on the SVI recommendation for de novo criteria.
• APC Table 2 requires phenotype points plus parental-relationship status (confirmed vs unconfirmed) before any de novo score, including the 0.5 Supporting minimum, can be assigned.
• The ClinVar record for this variant reports only variant-level submissions (2 clinical laboratories, Likely benign), with no proband-level de novo statement or criterion leads from any submission.
• No fetched publication mentions APC c.221-29G>C, so no published assumed de novo observation exists for this variant.

Still missing

• Case record establishing the variant in an affected proband
• Parental testing status (tested negative, not tested, or unconfirmed relationships)
• Phenotype-point score for the proband (APC Table 1) needed for the 0.5-1.5 de novo score band
• Documentation of family history to distinguish assumed de novo from inherited occurrence

Sources

Reasoning

The governing InSiGHT/ClinGen APC VCEP specification makes PP1 applicable: segregation in 3-4 meioses in >=1 family = Supporting, 5-6 meioses in >=1 family = Moderate, >=7 meioses in >=2 families = Strong. The specification further requires that counted affected individuals exhibit at least 0.5 phenotype point (with relatives having >=10 or 'multiple' colorectal adenomas counted as 0.5 point) and that only genotype- and phenotype-positive individuals and obligate carriers with phenotype be counted. This case contains no pedigree or family structure, no affected relatives, no variant-positive relatives, no obligate carriers, no meiosis count, and no family-level phenotype scoring; the ClinVar submissions are variant-level only with no relative genotypes, and no fetched paper reports this variant in any family. Segregation therefore cannot be evaluated.

Rule

APC VCEP PP1: 3-4 segregating meioses in >=1 family = Supporting; 5-6 meioses in >=1 family = Moderate; >=7 meioses in >=2 families = Strong; counted individuals must have >=0.5 phenotype point (relatives with >=10 or 'multiple' colorectal adenomas count as 0.5 point).

Evidence considered

• The APC VCEP PP1 rules require 3-4 segregating meioses in >=1 family for Supporting, 5-6 meioses in >=1 family for Moderate, and >=7 meioses in >=2 families for Strong.
• The APC VCEP PP1 instructions require counted affected individuals to exhibit at least 0.5 phenotype point (relatives with >=10 or 'multiple' colorectal adenomas count as 0.5 point) and count only genotype- and phenotype-positive individuals and obligate carriers with phenotype.
• The APC VCEP supplementary material states that the PP1 modifications follow Luo et al., 2019.
• No family pedigree, affected relative, variant-positive relative, or meiosis count is reported for this variant; the ClinVar submissions provide no relative-level genotype data and no criterion leads.
• No fetched publication mentions APC c.221-29G>C, so no published family segregation data exist for this variant.

Still missing

• Pedigree or family structure
• Number of affected relatives and their variant-carrier status
• Number of informative meioses and number of families
• Phenotype-point scores for each counted relative (APC Table 1, including >=10 or 'multiple' colorectal adenomas as 0.5 point)
• Genotype and phenotype status of obligate carriers

Sources

Reasoning

The governing framework is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for APC, version 2.1, which defines BS3 by assay design and controls: an RNA assay of a synonymous or intronic variant in a constitutional patient sample demonstrating no mRNA aberration, at Strong strength only when biallelic expression is shown and/or NMD inhibition was used and at Supporting strength otherwise; or, alternatively for the Supporting tier, a protein assay showing retention of beta-catenin-regulated transcription activity comparable to wild-type, which the specification limits to variants within the beta-catenin binding domain (codons 959-2129). NM_001127510.3:c.221-29G>C is an intronic variant 29 bp upstream of the exon 3 acceptor (5:112102857G>C, GRCh37). No RNA assay, reporter assay, or protein assay result for this variant was identified in any retrieved source, so neither the Strong nor the Supporting tier of BS3 can be evaluated: the RNA route requires an actual transcript analysis with documented assay controls (biallelic expression and/or NMD inhibition), and the protein route is additionally precluded here because c.221-29 lies far upstream of codon 959 and therefore outside the beta-catenin binding domain the specification restricts it to. The available publications are consensus/guideline documents with no variant-level data, and ClinVar variation 217948 lists Likely benign classifications from clinical laboratories without assertion criteria, functional evidence, or criterion leads, so it cannot be used to infer a benign functional result. SpliceAI predicts no significant splice impact (max delta 0.02) and the variant is common in gnomAD (popmax AF 0.547% in South Asian, 9 homozygotes in gnomAD v4.1), but in silico prediction and population frequency are not RNA-assay evidence and belong to the PP3/BP4 and population criteria respectively; they cannot be substituted for the assay BS3 requires. BS3 is therefore not assessed.

Rule

APC VCEP v2.1 BS3 (RNA assay in a constitutional patient sample showing no mRNA aberration, Strong if biallelic expression and/or NMD inhibition demonstrated, else Supporting; or protein assay showing beta-catenin-regulated transcription comparable to wild-type, restricted to codons 959-2129).

Evidence considered

• APC VCEP v2.1 (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel) defines BS3 as an RNA assay of a synonymous or intronic variant in a constitutional patient sample demonstrating no mRNA aberration, Strong when biallelic expression is shown and/or NMD inhibition was used, Supporting otherwise, or alternatively a protein assay showing retention of beta-catenin-regulated transcription comparable to wild-type restricted to the beta-catenin binding domain (codons 959-2129).
• The APC VCEP supplementary materials record 'No comment' against BS3 and supply no variant-specific functional data for this variant.
• No RNA assay, transcript analysis, or reporter/protein assay result for c.221-29G>C was found in any retrieved source; the four full-text papers retrieved (PMID:25741868, PMID:25645574, PMID:21368914, PMID:27854360) are consensus or guideline documents that do not mention the variant.
• ClinVar variation 217948 holds four submissions classified Likely benign, none with assertion criteria, functional evidence, criterion leads, or validated PMIDs; a source's classification and a laboratory label are not assay results, so BS3 receives no evidence from this source.
• The BS3 protein route is not available for this variant on the specification's own terms: it is limited to the beta-catenin binding domain (codons 959-2129), whereas c.221-29 lies upstream of codon 221.
• SpliceAI max delta score 0.02 (no predicted splice impact) and the gnomAD population frequency (v4.1 popmax AF 0.547% in South Asian, 9 homozygotes) are not RNA-assay evidence; they are adjudicated under the in silico splicing criteria and the population criteria respectively and cannot satisfy the assay requirement of BS3.

Still missing

• No RNA assay of c.221-29G>C in a constitutional patient sample, with or without biallelic expression documentation or NMD inhibition, so the BS3 RNA route cannot be evaluated at either strength.
• No beta-catenin-regulated transcription or beta-catenin binding assay; the APC VCEP limits this BS3 route to codons 959-2129, which excludes c.221-29.
• No functionally characterised APC splice-region variant with matched controls that could be used as a comparator for this variant.
• No ClinVar submission or publication reports assay data or assertion criteria for this variant.

Sources

Reasoning

The governing InSiGHT/ClinGen APC VCEP specification makes BS4 applicable: Strong requires an affected member without the variant scoring at least 1 phenotype point, or at least two affected members without the variant each scoring at least 0.5 phenotype points; Supporting requires one affected member without the variant scoring at least 0.5 phenotype points. Assessing non-segregation therefore requires genotyped affected relatives who lack the variant together with phenotype-point scoring against APC Table 1. This case contains no pedigree, no affected noncarriers, no relative-level genotypes, and no phenotype scoring; the ClinVar submissions are variant-level only, and no fetched paper reports this variant in any family. Non-segregation cannot be assessed.

Rule

APC VCEP BS4: Strong requires one affected noncarrier with >=1 phenotype point or two affected noncarriers each with >=0.5 points; Supporting requires one affected noncarrier with >=0.5 points.

Evidence considered

• The APC VCEP BS4 Strong rule requires an affected member without the variant scoring at least 1 phenotype point, or at least two affected members without the variant each scoring at least 0.5 phenotype points.
• The APC VCEP BS4 Supporting rule requires one affected member without the variant scoring at least 0.5 phenotype points.
• Phenotype points for affected noncarriers must be assigned using APC Table 1, which is not populated for this case.
• No affected noncarrier, relative-level genotype, pedigree, or phenotype-point score is reported for this variant; the ClinVar submissions contain no relative-level data and no criterion leads.
• No fetched publication mentions APC c.221-29G>C, so no published non-segregation evidence exists for this variant.

Still missing

• Affected relatives who tested negative for the variant
• Variant testing results and relationships for those relatives
• Phenotype details and APC phenotype-point scores for affected noncarriers
• Family pedigree and number of non-segregating affected members
• Assessment for phenocopies or a second pathogenic variant that could explain apparent non-segregation

Sources

Reasoning

BP2 is applicable for APC under the governing ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification version 2.1, at Benign Supporting strength, with the rule 'Observed in trans with a (Likely) Pathogenic APC variant OR >= 3 times in an unknown phase with different (Likely) Pathogenic APC variants.' The VCEP modification narrows the generic ACMG BP2 definition (which also covers 'in cis with a pathogenic variant in any inheritance pattern') to trans-configuration and phase-unknown co-occurrence observations involving a (likely) pathogenic APC allele. Satisfying BP2 therefore requires a positive allelic observation: either a phase-confirmed trans configuration with a pathogenic APC variant, or at least three independent occurrences alongside different pathogenic APC variants when phase is unknown. No source consulted for this case provides any such observation. The four ClinVar submissions for variation 217948 (Mayo Clinic Laboratories, Kuwait Cancer Control Center, Rigshospitalet Copenhagen, Breakthrough Genomics) all report Likely benign with zero expert-panel submissions, and the submission audit records zero criterion-lead submissions and no usable criterion leads, meaning no submitter reported a second APC allele, a phase determination, or a co-occurrence count for this variant. The reviewed full-text papers (PMID:21368914, PMID:25645574, PMID:25741868, PMID:27854360) do not mention c.221-29G>C or any allelic configuration involving it; they contribute only gene-level context such as the autosomal dominant, near-completely penetrant inheritance of FAP, which is the reason a trans observation with a pathogenic APC allele would be informative. The case is a variant-level analysis with no affected-proband record, no parental or offspring genotypes, and no phasing assay, so neither branch of the VCEP BP2 rule can be evaluated. BP2 is therefore recorded as not assessed rather than met or not met: the required observation has neither been made nor excluded, and its absence must not be converted into benign evidence. The phase and co-occurrence gap is listed below.

Rule

APC VCEP v2.1 BP2 (Benign Supporting): Observed in trans with a (Likely) Pathogenic APC variant OR >= 3 times in an unknown phase with different (Likely) Pathogenic APC variants.

Evidence considered

• The APC InSiGHT VCEP v2.1 BP2 rule (Benign Supporting) requires the variant to be observed in trans with a (likely) pathogenic APC variant, or at least three times in an unknown phase with different (likely) pathogenic APC variants; unlike generic BP2 it does not include a cis-with-pathogenic configuration.
• ClinVar variation 217948 carries four submissions (Mayo Clinic Laboratories; Kuwait Cancer Control Center KCCC/NGS; Rigshospitalet, Copenhagen University Hospital; Breakthrough Genomics), all classified Likely benign, none from an expert panel, and the submission audit reports 0 usable criterion leads and 0 criterion-lead submissions, so no submitter described a second APC allele, phase, or co-occurrence count for this variant.
• No affected-proband record, parental or offspring genotype, or phasing assay is present in this case, so the trans configuration required by the first branch of the VCEP BP2 rule cannot be evaluated.
• The four full-text papers reviewed for this case do not mention c.221-29G>C or any in-trans partner variant; they supply only gene-level inheritance context (FAP is autosomal dominant with near-complete penetrance), which is why a trans observation would be informative but does not itself satisfy BP2.
• Absence of an in-trans observation was not treated as benign evidence, so BP2 is recorded as not assessed rather than met.

Still missing

• No affected-proband genotype, second APC allele, or phase determination (parental/offspring testing) is documented for APC c.221-29G>C in any consulted source.
• No observation of this variant in trans with a (likely) pathogenic APC variant is reported, and no count of >= 3 phase-unknown occurrences alongside different (likely) pathogenic APC variants is available.
• None of the four ClinVar submissions provided assertion criteria usable as criterion leads, so no submitter-supplied allelic or co-occurrence evidence could be recovered.

Sources

Reasoning

Governing APC VCEP v2.1 text for BP5 ('Variant found in a case with an alternate molecular basis for disease'): 'Can be used, if a (Likely) Pathogenic variant in another adenomatous polyposis gene was identified (heterozygous variants in POLD1 or POLE; biallelic variants in MUTYH, NTHL1 or MSH3; in patients with onset in childhood / adolescence: biallelic variants in MLH1, MSH2, MSH6 or PMS2). This rule is only applicable when a colorectal polyposis phenotype is present.' The VCEP definition carries no numeric or likelihood-ratio threshold, so no comparison operator or substituted inequality applies to BP5; the only numeric comparison in this group is PS4's phenotype-point ladder, whose operator is the upward '>=', and that direction must not be imported into BP5. Neither prerequisite for BP5 is documented in this case: there is no colorectal polyposis phenotype for any carrier of c.221-29G>C (the case is a variant-only curation with no adenoma counts, colonoscopy or clinical findings, and no reported affected carrier in ClinVar, GeneReviews PMID:20301519, the FAP/AFAP gene card PMID:21368914, the ACG guideline PMID:25645574, PMID:25741868 or PMID:27854360), and no (Likely) Pathogenic variant in MUTYH, POLD1, POLE, NTHL1, MSH3, MLH1, MSH2, MSH6 or PMS2 is reported alongside this variant. An absent alternate genetic basis cannot be asserted from an absence of testing, so BP5 is not assessed rather than scored.

Rule

APC VCEP v2.1 BP5: usable only when a (Likely) Pathogenic variant in another adenomatous polyposis gene (POLD1, POLE, MUTYH, NTHL1, MSH3, or MLH1/MSH2/MSH6/PMS2 for childhood/adolescent onset) is identified AND a colorectal polyposis phenotype is present; no numeric threshold defined.

Evidence considered

• APC VCEP v2.1 supplementary specifications, BP5 row: alternate-basis rule limited to (Likely) Pathogenic variants in POLD1, POLE, MUTYH, NTHL1, MSH3 (and, for childhood/adolescent onset, MLH1, MSH2, MSH6 or PMS2), explicitly 'only applicable when a colorectal polyposis phenotype is present'.
• No proband phenotype (adenoma counts, colonoscopy findings, age at onset) or other-gene panel result appears in any source available for this case, so both the phenotype precondition and the alternate-basis finding are undocumented.
• No retrieved source reports c.221-29G>C together with a variant in a second adenomatous polyposis gene.

Still missing

• No colorectal polyposis phenotype is documented for any carrier, and the VCEP restricts BP5 to this situation ('This rule is only applicable when a colorectal polyposis phenotype is present').
• No other-gene testing is recorded for this case; MUTYH (biallelic), POLD1, POLE, NTHL1, MSH3 and, for childhood/adolescent onset, MLH1/MSH2/MSH6/PMS2 results are unavailable, so an alternate molecular basis can be neither confirmed nor excluded.

Sources