POLD1 c.3205G>A, p.(Val1069Ile)

Reasoning

PM2, applied at the supporting strength recommended by the ClinGen SVI 2020 downgrade (Richards et al. 2015, PMID:25741868), requires an allele frequency at or below 0.0001 in population databases. NM_002691.4:c.3205G>A is far below that: gnomAD v4.1 joint AF 3.75897e-06 (6/1,596,184 alleles, 0 homozygotes) with a grpmax filtering allele frequency of 1.25e-06, and gnomAD v2.1 AF 8.88992e-06 (2/224,974 alleles, 0 homozygotes). The highest frequency observed in any ancestry group in either dataset is the Admixed American subpopulation of gnomAD v2.1 at 3.01696e-05 (1/33,146 alleles), which is still about three-fold below the 1e-4 cutoff (gnomAD v4.1 Admixed American 1.69033e-05, 1/59,160). Because no governing VCEP exists for POLD1, the default all-comers GNOMAD_V2_1 and GNOMAD_V4_1 totals are the primary source; the non-cancer subsets corroborate the result rather than contradict it (gnomAD v2.1 non-cancer exomes AF 4.6892e-06, 1/213,256; gnomAD v3.1 non-cancer absent; gnomAD-Canada v1.0 absent). One ClinVar submitter discloses PM2 as its population criterion for this variant but states that the frequency data are considered unreliable because of poor data-quality metrics at this position in gnomAD; that positional artifact is noted, yet with popmax 3.02e-05 there is still roughly three-fold headroom below the cutoff and the variant remains rare in every independent source, so the verdict is unchanged. For a dominantly inherited, incompletely penetrant, late-onset cancer predisposition gene, rarity in population databases is weak evidence, which is why supporting strength rather than moderate is appropriate. PM2 is met at supporting strength.

Rule

PM2 (supporting): allele frequency <= 0.0001 (ClinGen SVI 2020 supporting downgrade; Richards et al., PMID:25741868) - satisfied (max observed 3.02e-05 in any ancestry group; 3.76e-06 overall in gnomAD v4.1)

Evidence considered

• gnomAD v4.1 joint (exomes + genomes, all-comers default source): AF 3.75897e-06 (6/1,596,184 alleles), 0 homozygotes, grpmax filtering allele frequency 1.25e-06; exome component 3.46257e-06 (5/1,444,016) and genome component 6.57168e-06 (1/152,168).
• gnomAD v2.1 (exomes only, all-comers default source): AF 8.88992e-06 (2/224,974 alleles), 0 homozygotes; no grpmax filtering allele frequency reported for this variant.
• Highest observed ancestry-group frequency: Admixed American, gnomAD v2.1 AF 3.01696e-05 (1/33,146 alleles, 0 homozygotes) and gnomAD v4.1 AF 1.69033e-05 (1/59,160 alleles, 0 homozygotes).
• Non-cancer corroboration: gnomAD v2.1 non-cancer exomes AF 4.6892e-06 (1/213,256 alleles, 0 homozygotes); gnomAD v3.1 non-cancer absent; gnomAD-Canada v1.0 absent.
• PM2 supporting-strength threshold of allele frequency <= 0.0001 is the ClinGen SVI 2020 recommendation to downgrade PM2 (Richards et al. 2015, PMID:25741868).
• ClinVar variation 537093 (3 submissions: 2x Uncertain significance, 1x Likely benign; no expert-panel submissions): the Labcorp submission SCV000767598 discloses PM2 as its population criterion while stating 'the frequency data for this variant in the population databases is considered unreliable, as metrics indicate poor data quality at this position in the gnomad database'; this was treated as a positional data-quality caveat and the verdict rests on the frequency values above, not on any submitter's classification.
• No POLD1-specific ClinGen CSPEC, VCEP specification, or local gene framework was located, so no non-cancer or exome-only population source is mandated; the all-comers GNOMAD_V2_1 and GNOMAD_V4_1 totals were used as the primary sources with the non-cancer subsets reported as corroboration.

Still missing

• A ClinVar submitter reports poor data-quality metrics at this position in gnomAD (frequency data considered unreliable), so the true allele frequency could exceed the reported value; the reported popmax of 3.02e-05 leaves roughly three-fold headroom below the 1e-4 cutoff.
• PMID:28492532 (Sherloc), cited in the Labcorp ClinVar submission trail, was not retrieved as full text in this case; it is a criteria-refinement methodology paper that would not supply variant-specific population frequency data, so its absence does not affect this determination.

Sources

Reasoning

Scope: the variant is a missense substitution, so BP4 is evaluated on the REVEL missense path only; the SpliceAI/splice path is not applicable to a missense variant and 'no predicted splice impact' would not be BP4 support here. With no POLD1-specific ClinGen specification available, the generic ClinGen SVI REVEL calibration applies. The local REVEL v1.3 lookup returned 0.197, which is at or below the BP4 supporting threshold of <=0.29 (Pejaver et al. 2022, PMID:36413997), indicating the amino-acid substitution is predicted tolerated; it does not reach the moderate (<=0.183) or strong (<=0.016) benign tiers, so BP4 is met at supporting strength only. BP4 is not claimed at a higher strength, and the same REVEL score is not reused for any other code. BayesDel (-0.367908) was not used because no ACMG-strength calibration exists for it, and Pangolin splice scores were not used as they are not calibrated for PP3/BP4. SpliceAI produced no score (timeout), which is missing data and was not treated as benign evidence.

Rule

BP4 (missense path, REVEL): met at supporting if REVEL <= 0.29, moderate if <= 0.183, strong if <= 0.016 (ClinGen SVI REVEL calibration, Pejaver et al. 2022, PMID:36413997). REVEL 0.197 -> BP4 supporting.

Evidence considered

• Variant consequence: NM_002691.4:c.3205G>A encodes NP_002682.2:p.(Val1069Ile), a missense substitution; BP4 therefore draws on the REVEL missense path and not on SpliceAI.
• No POLD1 ClinGen VCEP specification or gene-specific PP3/BP4 lookup table was available, so generic ACMG/AMP computational rules were applied.
• REVEL v1.3 local predictor lookup (source_registry key 'revel') returned score 0.197; ClinGen SVI REVEL calibration places BP4 supporting at <=0.29, moderate at <=0.183, strong at <=0.016 (Pejaver et al. 2022, Am J Hum Genet, PMID:36413997). 0.197 satisfies supporting but not moderate or strong.
• BayesDel score -0.367908 was available but not used: no established ACMG-strength calibration exists for BayesDel and the applicable framework specifies no cutoff.
• Pangolin splice scores (SG 0.003, SL -0.001) are not calibrated for PP3/BP4 and were not used as benign support.
• SpliceAI returned no scores owing to a lookup timeout; this missing-data state was not converted into benign evidence, and a splice prediction would be out of scope for a missense variant's BP4 regardless.

Sources

Reasoning

PS1 (Richards et al. 2015, PMID:25741868) requires that the identical amino-acid change - here p.Val1069Ile (NM_002691.4:c.3205G>A, NP_002682.2) - has previously been established as pathogenic, regardless of the nucleotide change. That premise is not satisfied. The exact-match ClinVar record (ClinVarID 537093, NM_002691.4(POLD1):c.3205G>A, p.Val1069Ile) carries three germline submissions: two Uncertain significance (SCV000767598, Labcorp Genetics/Invitae; SCV001995559, GeneDx) and one Likely benign (SCV006051457, Ambry Genetics), giving an aggregate 'Conflicting classifications of pathogenicity' with review status 'criteria provided, conflicting classifications' (1 star). No submitter classifies the identical p.Val1069Ile change as pathogenic or likely pathogenic, and no expert-panel submission exists for it. The only validated PMID trail attached to that record is 28492532 (Sherloc: a comprehensive refinement of the ACMG-AMP variant classification criteria), a criteria-refinement framework paper rather than a case report establishing this substitution as pathogenic; its full text is not available in this case. A PubMed search restricted to POLD1 plus V1069 (title/abstract) returned 0 records, so no publication reports p.Val1069Ile as a pathogenic observation either. The single full-text paper fetched for this case (PMID:25394175, an ACMG/NSGC cancer-referral guideline) was searched for c.3205G>A, p.(V1069I) and V1069I and contains no POLD1 or V1069 content at all. Because the same amino-acid change has never been established as pathogenic in any source examined, PS1 is not met.

Rule

Same amino acid change as a previously established pathogenic variant regardless of nucleotide change.

Evidence considered

• ClinVar ClinVarID 537093 (NM_002691.4(POLD1):c.3205G>A, p.Val1069Ile): SCV000767598 Labcorp Genetics/Invitae - Uncertain significance, germline, criteria provided single submitter, cites PMID:28492532; SCV001995559 GeneDx - Uncertain significance, germline; SCV006051457 Ambry Genetics - Likely benign, germline. Aggregate classification 'Conflicting classifications of pathogenicity', review status 'criteria provided, conflicting classifications'. No submission reports the identical amino-acid change as pathogenic or likely pathogenic.
• The ClinVar submission summary for this record shows 2 clinical laboratories reporting Uncertain significance and 1 reporting Likely benign, with 0 expert-panel submissions and a single validated PMID (28492532).
• The only PMID trail on this record is 28492532 (Sherloc: a comprehensive refinement of the ACMG-AMP variant classification criteria, Nykamp et al. 2017) - a criteria-refinement framework paper, not a report establishing p.Val1069Ile as pathogenic; its full text is not present under this case's publications directory.
• PubMed search run for this criterion: 'POLD1[Title/Abstract] AND V1069' and 'POLD1 1069 missense' - both returned 0 records, so no published observation reports the p.Val1069Ile change.
• PMID:25394175 (fetched full text for this case) was searched for 'POLD1 c.3205G>A', 'c.3205G>A', 'p.(V1069I)', 'V1069I', 'NP_002682.2:p.(V1069I)'; variant_mentioned=false and the paper contains no POLD1 content, so it provides no same-amino-acid-change pathogenicity evidence.

Still missing

• The PMID cited by one ClinVar submitter (28492532, Sherloc) is not available as full text in this case; it is a methods/framework paper and carries no variant-level pathogenicity assertion for p.Val1069Ile.
• No ClinGen POLD1 VCEP/CSPEC specification exists for this gene, so the generic ACMG/AMP 2015 definition of PS1 was applied.

Sources

Reasoning

Governing framework: no ClinGen CSPEC and no gene-specific VCEP specification was retrieved for POLD1 (cspec.found=false; framework_mode=generic_acmg; framework_complete=false), so generic ACMG/AMP 2015 PS4 applies (Richards et al. 2015, PMID:25741868): the prevalence of the variant in affected individuals is significantly increased compared with the prevalence in controls (case-control odds ratio, relative risk, or penetrance data). PS4 was screened as a candidate criterion for the only high-priority paper in the literature packet (PMID:25394175, triaged from the ClinVar submission trail with candidate_criteria ['PP5','PS4']). That paper is an ACMG/NSGC practice guideline on referral indications for cancer predisposition assessment; its pre-computed full-text extraction records no mention of POLD1, c.3205G>A, or p.(Val1069Ile) anywhere in the paper and provides no case-control, cohort, or penetrance data for this variant (paper_extracts.json, PMID:25394175: variant_mentioned=false). No other case-control or cohort enrichment dataset for this exact variant was collected: the three ClinVar records for VCV000537093 are single-submitter clinical-testing submissions (Labcorp/Invitae Uncertain significance, GeneDx Uncertain significance, Ambry Genetics Likely benign) with no case or control denominators, the variant returned no records in the COSMIC somatic-cancer search, and it has no row in Cancer Hotspots. Because no affected-vs-control frequency comparison can be constructed at all, PS4's evidentiary requirement is unmet. Note on numeric direction: the generic specification consulted here supplies no numeric PS4 threshold, and neither does any VCEP specification for this gene, so no threshold inequality is written; the failure is the total absence of enrichment data rather than a directional comparison against a cutoff.

Rule

generic_acmg_2015: prevalence in affected individuals significantly increased vs controls (Richards et al. 2015, PMID:25741868)

Evidence considered

• PMID:25394175 (Hampel et al., ACMG/NSGC referral-indications practice guideline): pre-computed full-text extraction found no mention of POLD1, c.3205G>A, or p.(Val1069Ile), and no cohort, case-control, or penetrance data for this variant.
• ClinVar VCV000537093: 'Conflicting classifications of pathogenicity', 1 star, 3 single-submitter clinical-testing submissions (SCV000767598 Labcorp/Invitae Uncertain significance; SCV001995559 GeneDx Uncertain significance; SCV006051457 Ambry Genetics Likely benign); no case or control counts are reported, so no odds ratio or relative risk is derivable.
• COSMIC search 'POLD1 c.3205G>A' returned no records: 'This variant has not previously been reported in somatic cancers (COSMIC).'
• Cancer Hotspots query 'POLD1 V1069' returned no hotspot row: 'This variant does not lie in a statistically significant hotspot.'
• OncoKB: 'Unknown Oncogenic Effect'; no variant-specific curated evidence for V1069I, so no case-enrichment signal is contributed.
• gnomAD v2.1 non-cancer exomes report 1/213256 alleles (AF 4.69e-06) and gnomAD v2.1 total 2/224974 alleles (AF 8.89e-06); these are population-frequency observations, not case enrichment, and cannot support PS4.

Still missing

• No case-control, cohort, or penetrance dataset for POLD1 c.3205G>A (p.Val1069Ile) was located in the collected sources; a systematic literature search dedicated to case-control enrichment for this variant was not run beyond the single triaged paper.
• No numeric PS4 threshold exists in the governing generic ACMG/AMP specification or in any retrieved gene-specific specification, so a threshold-based comparison cannot be performed even if data later become available.

Sources

Reasoning

PM1 (Richards et al. 2015, PMID:25741868) requires the variant to lie in a mutational hotspot and/or a critical, well-established functional domain that is without benign variation. No gene-specific POLD1 specification and no domain table were available: vcep_materials.json has no domain_tables entry for POLD1 (domain_tables is an empty object), final_classification_framework records found=false with source 'generic_acmg_fallback' and scoring model ACMG/AMP 2015, and no ClinGen CSPEC exists for this gene. PM1 was therefore assessed from generic evidence, and two independent lines of evidence argue against it. (1) No hotspot: the case's CancerHotspots query for POLD1 V1069 returned no row ('This variant does not lie in a statistically significant hotspot'), and the COSMIC query for POLD1 c.3205G>A returned no somatic-cancer record. (2) Domain: the domain in which residue 1069 actually sits is the C-terminal CysB motif (UniProt P28340, residues 1058-1076), and residue 1069 is not one of that motif's metal-coordinating residues (1058, 1061, 1071, 1076). The functional domain established as critical for germline POLD1 cancer predisposition is instead the N-terminal proofreading exonuclease domain - OncoKB gene background states 'Germline heterozygous loss-of-function mutations in the exonuclease domain of POLD1 cause polyposis and predispose individuals to colorectal, endometrial, and possibly brain cancers' - and residue 1069 lies far outside it (the exonuclease domain is reported as approximately amino acids 268-471). The C-terminal region containing residue 1069 is also not a region free of benign variation: ClinVar lists Likely benign POLD1 missense variants in it (p.Val1025Leu c.3073G>T, p.Leu1044Val c.3130C>G, p.Arg1086Gly c.3256C>G) alongside Uncertain significance p.Val1041Ile and p.Val1069Asp, and gnomAD v4 carries many rare missense alleles across the CysB motif (for example p.His1066Arg at 39/1,447,732 and p.Arg1074Gln at 33/1,445,492 exome alleles). One counter-consideration is recorded for human review: the CysB region does harbour clustered pathogenic de novo missense variants at adjacent residues (p.Arg1060Cys causing combined immunodeficiency, and p.Ile1070Asn, p.Thr1072Pro, p.Ser1073Arg causing MDPL), so a reviewer who treats the CysB motif as the critical domain and its sub-1e-4-frequency variation as compatible with 'without benign variation' could support PM1 at supporting strength. None of those variants affects residue 1069 or establishes it as pathogenic. On the evidence assembled - no hotspot at this codon, residue outside the gene's established disease-critical domain, and documented benign/uncertain variation in the surrounding C-terminal region - PM1 is not met.

Rule

Located in a mutational hot spot and/or critical and well-established functional domain (e.g., active site of an enzyme) without benign variation.

Evidence considered

• No gene-specific framework was available for POLD1 (final_classification_framework: found=false, source='generic_acmg_fallback', framework_mode null, scoring_model 'ACMG/AMP 2015', notes unchanged), and vcep_materials.json contains an empty domain_tables object ('domain_tables': {}), so no governing VCEP domain table for this gene exists to check the residue against; PM1 was therefore assessed under generic ACMG/AMP 2015.
• UniProt P28340 (POLD1_HUMAN, 1107 aa, the same length as NP_002682.2) annotates a CysA-type zinc finger at residues 1012-1029 and a CysB motif at residues 1058-1076 with metal-coordinating residues 1058, 1061, 1071 and 1076. Residue 1069 lies inside the CysB motif but is not itself a coordinating residue.
• For germline POLD1 cancer predisposition the established critical domain is the N-terminal proofreading exonuclease domain. OncoKB gene-level background (oncokb source) states 'Germline heterozygous loss-of-function mutations in the exonuclease domain of POLD1 cause polyposis and predispose individuals to colorectal, endometrial, and possibly brain cancers', and the POLD1 gene summary likewise states 'Germline mutations in its exonuclease domain cause polyposis and predispose people to colorectal, endometrial, and possibly brain cancers.' Residue 1069 is far outside that domain (reported as approximately amino acids 268-471).
• The case's CancerHotspots query for POLD1 V1069 returned no hotspot row (evidence_sentence: 'This variant does not lie in a statistically significant hotspot'), and the COSMIC query for POLD1 c.3205G>A returned 'This variant has not previously been reported in somatic cancers (COSMIC).' Neither source documents a mutational hotspot at this codon.
• The C-terminal region containing residue 1069 is not free of benign/uncertain variation: ClinVar POLD1 records include Likely benign missense variants p.Val1025Leu (c.3073G>T), p.Leu1044Val (c.3130C>G) and p.Arg1086Gly (c.3256C>G) in the same C-terminal region, together with Uncertain significance p.Val1041Ile (c.3121G>A) and p.Val1069Asp (c.3206T>A).
• gnomAD v4 lists numerous rare missense alleles across the CysB motif and its flanks, e.g. p.His1066Arg 39/1,447,732 exome alleles, p.Arg1074Gln 33/1,445,492, p.Arg1060His 18/1,451,752, p.Val1069Ile 5/1,444,016 exome plus 1/152,168 genome alleles - all far below 1e-4, i.e. rare but not absent, showing the region tolerates some amino-acid substitution.
• Counter-evidence recorded for review: published de novo pathogenic missense variants cluster in the CysB region - p.Ile1070Asn and p.Ser1073Arg (PMID:35590056, 'Mild MDPL in a patient with a novel de novo missense variant in the Cys-B region of POLD1', which quotes the same conserved CysB motif as the region in which the previously reported p.(Ile1070Asn) lies and concludes 'Our analysis supports a central role for the CysB region in regulating POLD1 activity in health and disease'), p.Thr1072Pro (PMID:34517090), and p.Arg1060Cys (PMID:31629014: 'Molecular dynamics simulation revealed that the R1060C mutation disrupts the intramolecular interaction between the POLD1 CysB motif and the catalytic domain'). These are at different residues (1060, 1070, 1072, 1073) than the variant under assessment.

Still missing

• No ClinGen POLD1 VCEP specification or vcep_materials domain_tables entry was available, so no authoritative list of approved critical functional domains for POLD1 could be checked; domain boundaries were taken from UniProt P28340 and the gene-level literature.
• Human review of the PM1 domain call is advisable: the CysB motif (1058-1076) containing residue 1069 does contain clustered pathogenic de novo missense variants at residues 1060, 1070, 1072 and 1073 (MDPL / combined immunodeficiency), which a reviewer could weigh as supporting-strength domain evidence.
• The CancerHotspots result is a somatic-cancer hotspot database; POLD1 somatic variants are rare, so its negative result at V1069 is weak evidence against a germline hotspot.

Sources

Reasoning

PM5 (Richards et al. 2015, PMID:25741868) requires that a different missense change at the same amino-acid residue has previously been determined to be pathogenic. Residue 1069 of NP_002682.2 has exactly one alternate change on record in ClinVar: NM_002691.4:c.3206T>A, p.Val1069Asp (ClinVarID 3021003), whose aggregate germline classification is 'Uncertain significance' with review status 'criteria provided, multiple submitters, no conflicts'; there is no pathogenic or likely pathogenic same-residue comparator. The variant under assessment itself (c.3205G>A, p.Val1069Ile, ClinVarID 537093) is likewise only Uncertain significance/Likely benign with conflicting classifications. An independent ClinVar query for residue 1069 ('POLD1[gene] AND 1069[residue]') returned four records, of which only these two concern codon 1069 (the other two, c.1069C>G p.Leu357Val and c.1069C>A p.Leu357Met, are nucleotide-position matches at codon 357). A PubMed search for the residue returned no records. The clustered pathogenic POLD1 missense variants in the neighbouring CysB region (p.Arg1060Cys, p.Ile1070Asn, p.Thr1072Pro, p.Ser1073Arg) are at different residues and therefore cannot satisfy PM5, which is residue-specific by definition. Because no different missense change at residue 1069 has been determined to be pathogenic, PM5 is not met.

Rule

Missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before.

Evidence considered

• ClinVar residue-1069 records (from the ClinVar variation summaries): ClinVarID 3021003, NM_002691.4(POLD1):c.3206T>A (p.Val1069Asp) - aggregate 'Uncertain significance', review status 'criteria provided, multiple submitters, no conflicts', conditions 'Hereditary cancer-predisposing syndrome' and 'Colorectal cancer, susceptibility to, 10'; and ClinVarID 537093, NM_002691.4(POLD1):c.3205G>A (p.Val1069Ile) - aggregate 'Conflicting classifications of pathogenicity' (2 Uncertain significance, 1 Likely benign). No pathogenic or likely pathogenic missense change at residue 1069 exists in ClinVar.
• The case artifact pm5_candidates.json reports found=false with candidate_summary totals of 0 (total_candidates_seen 0, same_residue_candidates 0) and reason 'Unable to confirm classic same-residue PM5 semantics safely; skipping automatic candidate harvesting', with the instruction 'Review whether any collected same-residue comparator variants satisfy the governing PM5 rule; do not apply PM5 from this artifact alone'. That gap was closed by querying ClinVar directly for residue 1069.
• ClinVar search performed for this criterion: 'POLD1[gene] AND 1069[residue]' returned 4 records - ClinVarID 3021003 (c.3206T>A, p.Val1069Asp), 537093 (c.3205G>A, p.Val1069Ile), 1782705 (c.1069C>G, p.Leu357Val) and 1782683 (c.1069C>A, p.Leu357Met); only the first two are at codon 1069.
• PubMed search performed for this criterion ('POLD1[Title/Abstract] AND V1069') returned 0 records, so no publication reports a different pathogenic missense change at residue 1069.
• PMID:25394175 (the only full-text paper fetched for this case) was searched for c.3205G>A, p.(V1069I) and V1069I (variant_mentioned=false) and contains no POLD1 content, so it contributes no same-residue comparator.
• Adjacent-residue pathogenic POLD1 variants in the CysB motif - p.Arg1060Cys (PMID:31629014), p.Ile1070Asn and p.Ser1073Arg (PMID:35590056), p.Thr1072Pro (PMID:34517090) - are at residues other than 1069 and therefore cannot be used as PM5 comparators.

Still missing

• The automated PM5 candidate-harvesting artifact for this case declined to collect comparators (found=false, same_residue_candidates=0); that gap was closed by a direct ClinVar residue query, which identified p.Val1069Asp as the sole alternate residue-1069 change and confirmed it is not pathogenic.
• No ClinGen POLD1 VCEP/CSPEC specification exists, so the generic ACMG/AMP 2015 definition of PM5 was applied.

Sources

Reasoning

PP2 (Richards et al. 2015, PMID:25741868) requires both that the gene has a low rate of benign missense variation and that missense variants are a common mechanism of disease. For POLD1 the mechanistic half is satisfied: dominant germline missense variants are a well-established cause of disease - exonuclease-domain missense variants cause polymerase proofreading-associated polyposis (OncoKB gene background: 'Germline heterozygous loss-of-function mutations in the exonuclease domain of POLD1 cause polyposis and predispose individuals to colorectal, endometrial, and possibly brain cancers'), and CysB-region missense variants cause MDPL (p.Ile1070Asn, p.Ser1073Arg, PMID:35590056; p.Thr1072Pro, PMID:34517090) and combined immunodeficiency (p.Arg1060Cys, PMID:31629014). The first requirement, however, is not met. gnomAD v4 gene constraint for POLD1 (ENSG00000062822) gives observed/expected missense 0.838 (1338 observed versus 1595.8 expected), missense Z 2.75, and pLI 5.4e-15 (no constraint); the missense Z falls below the Z > 3.09 significance threshold that gnomAD uses to define a gene as constrained against missense variation (Lek et al. 2016, Nature, PMID:27535533), so benign missense variation in POLD1 is not demonstrably rare at the gene level. ClinVar reinforces this: the gene carries large numbers of Uncertain significance and Likely benign missense variants throughout its length, including in the C-terminal region (p.Val1025Leu and p.Leu1044Val and p.Arg1086Gly Likely benign; p.Val1041Ile and p.Val1069Asp Uncertain significance). Because the low-benign-missense-variation arm of PP2 is not supported by the gene-level constraint data or the ClinVar missense spectrum, PP2 is not met for this variant.

Rule

Missense variant in a gene that has a low rate of benign missense variation and where missense variants are a common mechanism of disease.

Evidence considered

• gnomAD v4 gene constraint for POLD1 (ENSG00000062822, retrieved from the gnomAD API): pLI 5.376e-15, oe_lof 0.609, oe_lof_upper 0.734, oe_mis 0.838, obs_mis 1338, exp_mis 1595.78, mis_z 2.752. The missense Z of 2.75 is below the Z > 3.09 threshold used by gnomAD to call significant missense constraint (Lek et al. 2016, Nature, PMID:27535533), so POLD1 does not show a low rate of benign missense variation by that measure.
• Gene-level mechanism context (oncokb / pvs1_gene_context): POLD1 encodes the catalytic subunit of DNA polymerase delta with both polymerase and 3'-5' exonuclease (proofreading) activity; dominant germline missense variants in the exonuclease domain cause polymerase proofreading-associated polyposis and colorectal/endometrial cancer predisposition, and pvs1_gene_context.json documents a supported germline disease mechanism for this gene. Missense variation is therefore a genuine mechanism of POLD1 disease - the mechanistic arm of PP2 is satisfied.
• Further evidence that missense is a POLD1 disease mechanism: p.Arg1060Cys in the CysB motif causes combined immunodeficiency (PMID:31629014), p.Ile1070Asn and p.Ser1073Arg cause MDPL (PMID:35590056), and p.Thr1072Pro causes MDPL (PMID:34517090).
• ClinVar POLD1 spectra consulted for the benign-missense-variation arm: Likely benign missense variants including p.Val1025Leu (c.3073G>T), p.Leu1044Val (c.3130C>G) and p.Arg1086Gly (c.3256C>G), and Uncertain significance variants including p.Val1041Ile (c.3121G>A) and p.Val1069Asp (c.3206T>A), showing that benign/uncertain missense variation is not rare in this gene.
• This variant's own in-silico scores (REVEL 0.197, BayesDel -0.368) were not used for PP2, which is a gene-level criterion; PP3/BP4 are assessed in the in-silico group.

Still missing

• No ClinGen POLD1 VCEP/CSPEC specification exists, so PP2 was applied from the generic ACMG/AMP 2015 definition; the 'low rate of benign missense variation' arm was operationalised with gnomAD gene-level constraint rather than a gene-specified threshold.
• The mechanistic arm of PP2 (missense as a common disease mechanism for POLD1) is satisfied; only the low-benign-missense-variation arm fails, which is why the verdict is recorded as not_met rather than not_applicable.

Sources

Reasoning

Scope: NM_002691.4:c.3205G>A is a missense variant (NP_002682.2:p.(Val1069Ile)), so PP3/BP4 must be taken from the missense protein-impact path (REVEL) only; the splice path is not applicable to this consequence type. No gene-specific ClinGen VCEP specification exists for POLD1 (cspec not found; framework_mode generic_acmg), so the generic ACMG/AMP computational rules and the ClinGen SVI REVEL calibration apply. The local REVEL v1.3 lookup returned 0.197 for this variant. That is below even the supporting-strength PP3 threshold of >=0.644 (Pejaver et al. 2022, PMID:36413997), so the substitution is not predicted damaging and PP3 is not met at any strength. No PP3-supporting in-silico evidence exists for this variant: BayesDel is present (-0.367908) but has no established ACMG-strength calibration and was not used, the HCI prior lookup found nothing, and Pangolin splice scores (SG 0.003, SL -0.001) are not calibrated for PP3/BP4. SpliceAI is not part of this variant's applicable path and returned no scores in any case (timeout), so it contributes nothing here in either direction.

Rule

PP3 (missense path, REVEL): met only if REVEL >= 0.644 supporting / >= 0.773 moderate / >= 0.932 strong (ClinGen SVI REVEL calibration, Pejaver et al. 2022, PMID:36413997). REVEL 0.197 does not meet any tier.

Evidence considered

• Variant consequence: NM_002691.4:c.3205G>A encodes NP_002682.2:p.(Val1069Ile), a single amino-acid substitution (missense), so the REVEL missense path governs PP3/BP4 and the SpliceAI splice path is out of scope for this criterion.
• No ClinGen VCEP specification or gene-specific PP3/BP4 lookup table for POLD1 was available (cspec.found=false; final_classification_framework.source=generic_acmg_fallback), so generic ACMG/AMP rules were used.
• REVEL v1.3 local predictor lookup (source_registry key 'revel') returned score 0.197 for this variant; the ClinGen SVI REVEL calibration assigns PP3 supporting at >=0.644, moderate at >=0.773, strong at >=0.932 (Pejaver et al. 2022, Am J Hum Genet, PMID:36413997). 0.197 is below all PP3 thresholds.
• BayesDel score -0.367908 was present but was not used: no generic ACMG-strength calibration for BayesDel is established, and the POLD1 framework specifies no cutoff.
• Pangolin splice scores (SG 0.003, SL -0.001) are not calibrated for PP3/BP4 and were not used.
• SpliceAI returned no scores because of a lookup timeout; this is missing data, not evidence of absent splice impact, and would not be usable as PP3 support in any case because this variant is missense.

Sources

Reasoning

PP5 applies only when a reputable source reports the variant as pathogenic and the laboratory cannot independently evaluate the underlying evidence; under the governing policy for this group it may be triggered only by an exact-variant ClinVar classification from an expert panel (ClinGen VCEP or equivalent expert-panel review). The ClinVar record for the exact variant NM_002691.4(POLD1):c.3205G>A (p.Val1069Ile), VCV000537093, carries review status 'criteria provided, conflicting classifications' at 1 star, comprising three single-submitter clinical-testing submissions: SCV000767598 Labcorp Genetics (formerly Invitae) Uncertain significance, SCV001995559 GeneDx Uncertain significance, and SCV006051457 Ambry Genetics Likely benign. The submission audit for this record reports expert_panel_submissions = 0 and is_expert_panel = false for every one of the three submissions, and the aggregate label 'Conflicting classifications of pathogenicity' is an aggregate rather than a submission-level expert assertion. The only submission criteria signals extracted are PM2 and PP3, both from uncertain-significance submissions, so no submission even asserts a pathogenic-strength classification. The single validated PMID in the submission trail (28492532, the Sherloc criteria-refinement methodology paper) is a framework publication, not variant-specific case evidence. The exact-variant expert-panel Pathogenic/Likely pathogenic classification PP5 requires therefore does not exist, and PP5 is not met. No numeric threshold is involved, so no inequality applies.

Rule

PP5 policy: requires an exact-variant ClinVar classification of Pathogenic/Likely pathogenic from an expert panel (VCEP); ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions never trigger PP5

Evidence considered

• ClinVar VCV000537093 (ClinVarID 537093, NM_002691.4(POLD1):c.3205G>A, p.Val1069Ile): overall classification 'Conflicting classifications of pathogenicity', review status 'criteria provided, conflicting classifications', 1 star, 3 submissions from 3 submitters.
• Submission-level detail: SCV000767598 (Labcorp Genetics/Invitae, Uncertain significance, criteria provided single submitter, germline), SCV001995559 (GeneDx, Uncertain significance, Not Provided condition), SCV006051457 (Ambry Genetics, Likely benign, criteria provided single submitter) - none is an expert-panel submission.
• ClinVar submission audit counts: submissions_total 3, expert_panel_submissions 0, criterion_lead_submissions 1 (SCV000767598, solely because it carries a validated PMID trail), validated_pmids ['28492532'].
• Extracted submission criteria signals are PM2 (SCV000767598) and PP3 (SCV001995559); no pathogenic-strength criterion is asserted by any submitter.
• PMID:28492532 (Sherloc, Nykamp et al. 2017) is the only validated citation attached to a submission and is a criteria-refinement methodology paper, not evidence about this variant.
• PMID:25394175 was triaged with candidate_criteria ['PP5','PS4'] from the ClinVar trail, but its full text does not mention POLD1 or this variant, so it reports no classification of this variant.

Still missing

• No ClinGen expert-panel (VCEP) classification exists for this variant; if a POLD1 (or hereditary cancer) VCEP later reviews VCV000537093, PP5 would need re-evaluation.
• The aggregate ClinVar label and the individual non-expert submissions remain available only as weak, non-triggering context.

Sources

Reasoning

BA1 is the stand-alone benign criterion requiring an allele frequency >= 0.05 (Richards et al. 2015, PMID:25741868). NM_002691.4:c.3205G>A (POLD1 p.Val1069Ile) is present only at vanishingly low frequency across every population resource examined. The gnomAD v2.1 all-comers exome total is AF 8.88992e-06 (2/224,974 alleles, 0 homozygotes), and the most frequent population in either dataset is the Admixed American subpopulation of gnomAD v2.1 at AF 3.01696e-05 (1/33,146 alleles, 0 homozygotes). gnomAD v4.1 joint (exomes + genomes) gives a total AF of 3.75897e-06 (6/1,596,184 alleles, 0 homozygotes) with a grpmax filtering allele frequency of 1.25e-06 and an Admixed American subpopulation maximum of 1.69033e-05 (1/59,160 alleles). The non-cancer subsets agree (v2.1 non-cancer exomes AF 4.6892e-06, 1/213,256), and the variant is absent from gnomAD-Canada v1.0. Every one of these values is more than three orders of magnitude below the 0.05 threshold, so BA1 is not met, and no gene-specific BA1 consideration could lower the bar into this frequency range for a rare cancer predisposition gene. No POLD1-specific ClinGen/VCEP specification or local gene framework was available for this case, so the generic ACMG/AMP 2015 stand-alone benign threshold was applied.

Rule

BA1 (stand-alone benign): allele frequency >= 0.05 (ACMG/AMP 2015, Richards et al., PMID:25741868) - not satisfied (max observed 3.02e-05)

Evidence considered

• gnomAD v2.1 all-comers (exomes only): total AF 8.88992e-06 (2/224,974 alleles), 0 homozygotes; highest subpopulation Admixed American AF 3.01696e-05 (1/33,146 alleles), 0 homozygotes; no grpmax filtering allele frequency reported for this variant.
• gnomAD v4.1 joint (exomes + genomes): total AF 3.75897e-06 (6/1,596,184 alleles), 0 homozygotes; grpmax filtering allele frequency 1.25e-06; highest subpopulation Admixed American AF 1.69033e-05 (1/59,160 alleles), 0 homozygotes.
• gnomAD v2.1 non-cancer exomes subset: AF 4.6892e-06 (1/213,256 alleles), 0 homozygotes. gnomAD v3.1 non-cancer and gnomAD-Canada v1.0 (HostSeq genomes): variant not present.
• The BA1 stand-alone benign allele-frequency threshold of >= 0.05 is the ACMG/AMP 2015 criterion (Richards et al. 2015, PMID:25741868).
• No POLD1 ClinGen CSPEC, VCEP specification, or local custom gene framework was located for this case (framework_mode = generic_acmg, gene_vcep_dir = null), so the generic ACMG/AMP 2015 population thresholds were used rather than a gene-specific or non-cancer-source-specific rule.

Sources

Reasoning

BS1 requires an allele frequency greater than that expected for the disorder, i.e. at or above the gene-specific threshold where a gene-specific maximum credible allele frequency can be derived from disease prevalence, allelic heterogeneity and penetrance, or the generic default of >= 0.01 otherwise (Whiffin et al. 2017, PMID:28518168). No POLD1-specific ClinGen/VCEP specification or local gene framework was available for this case, and the case evidence contains no prevalence or penetrance estimate for POLD1-associated polymerase proofreading-associated polyposis, so a gene-specific threshold could not be derived and the generic 0.01 default was applied. Against that threshold the variant is far too rare to qualify: gnomAD v2.1 total AF 8.88992e-06 (2/224,974), gnomAD v4.1 joint AF 3.75897e-06 (6/1,596,184) with grpmax filtering AF 1.25e-06, gnomAD v2.1 non-cancer exomes AF 4.6892e-06 (1/213,256), and the highest frequency seen in any subpopulation is 3.01696e-05 (Admixed American, gnomAD v2.1). The observations sit roughly 300-fold or more below the 1% threshold, and no plausible gene-specific derivation for a rare, incompletely penetrant cancer predisposition gene would place the benign threshold as low as 3e-05, so BS1 is not met.

Rule

BS1 (strong benign): allele frequency >= 0.01 generic SVI default; gene-specific derivation preferred (Whiffin et al. 2017, PMID:28518168) - not satisfied (max observed 3.02e-05)

Evidence considered

• gnomAD v2.1 all-comers (exomes only): total AF 8.88992e-06 (2/224,974 alleles), 0 homozygotes; highest subpopulation Admixed American AF 3.01696e-05 (1/33,146 alleles).
• gnomAD v4.1 joint (exomes + genomes): total AF 3.75897e-06 (6/1,596,184 alleles), 0 homozygotes; grpmax filtering allele frequency 1.25e-06; highest subpopulation Admixed American AF 1.69033e-05 (1/59,160 alleles).
• gnomAD v2.1 non-cancer exomes subset: AF 4.6892e-06 (1/213,256 alleles), 0 homozygotes; gnomAD v3.1 non-cancer genomes and gnomAD-Canada v1.0: variant absent.
• BS1 generic strength threshold of >= 0.01 allele frequency, with per-gene derivation preferred, is the SVI default derived from Whiffin et al. 2017 (PMID:28518168).
• No prevalence, penetrance, or allelic-heterogeneity figures for POLD1-associated polymerase proofreading-associated polyposis were available in the case evidence, so no gene-specific maximum credible allele frequency (and therefore no gene-specific BS1 threshold) could be derived.

Still missing

• No POLD1 disease prevalence or penetrance estimate was present in the case evidence, so the prevalence/penetrance-based gene-specific BS1 (maximum credible allele frequency) derivation described by Whiffin et al. 2017 (PMID:28518168) could not be computed and the generic 0.01 threshold was used; this does not change the verdict, because observed frequencies are ~300-fold or more lower.

Sources

Reasoning

BS2 requires the variant to have been observed in a healthy adult individual for a disorder in which full penetrance is expected at an early age (homozygous for a recessive disorder; heterozygous for a dominant disorder). Neither limb is satisfied here. The variant was observed only in the heterozygous state and no homozygotes were reported in any dataset: 0 homozygotes among 224,974 alleles in gnomAD v2.1, 0 among 213,256 alleles in the gnomAD v2.1 non-cancer exome subset, and 0 among 1,596,184 alleles in gnomAD v4.1 joint; the variant is entirely absent from gnomAD v3.1 non-cancer and gnomAD-Canada v1.0. For the dominant limb, the six documented gnomAD v4.1 heterozygous carriers carry no phenotype, age, or health-status annotation, so they cannot be counted as healthy adults, and gnomAD v4.1 aggregates participants regardless of cancer status, which is material for a cancer predisposition gene. Independently, the criterion's premise fails: germline POLD1 variants cause polymerase proofreading-associated polyposis, an autosomal dominant, adult-onset, incompletely penetrant colorectal/endometrial cancer predisposition, so full penetrance at an early age is not expected. BS2 is not met.

Rule

BS2 (strong benign): observed in a healthy adult individual for a disorder with full penetrance expected at an early age - not satisfied (0 homozygotes in all datasets; no phenotyped healthy adult carriers; POLD1 disease is adult-onset with incomplete penetrance)

Evidence considered

• Homozygote counts: gnomAD v2.1 total 0/224,974 alleles; gnomAD v2.1 non-cancer exomes 0/213,256 alleles; gnomAD v4.1 joint 0/1,596,184 alleles (exome 0/1,444,016; genome 0/152,168).
• gnomAD v3.1 non-cancer genomes and gnomAD-Canada v1.0: variant not present, so no carrier or homozygote observations from either source.
• Highest subpopulation frequencies remain heterozygous-only sightings: Admixed American AF 3.01696e-05 (1/33,146) in gnomAD v2.1 and 1.69033e-05 (1/59,160) in gnomAD v4.1, both with 0 homozygotes.
• POLD1 germline variants in the exonuclease domain cause polymerase proofreading-associated polyposis with autosomal dominant, adult-onset, incompletely penetrant colorectal, endometrial and possibly brain cancer risk (gene-level curated context: POLD1 gene summary and OncoKB gene-level background), so BS2's requirement for a disorder with full penetrance expected at an early age is not satisfied for this gene-disease pair.
• No POLD1-specific ClinGen/VCEP specification or local gene framework was available, so the generic ACMG/AMP 2015 BS2 definition was applied.

Still missing

• gnomAD does not publish per-carrier phenotype, age, or health status, and gnomAD v4.1 includes participants regardless of cancer status, so it cannot be established whether the six heterozygous carriers are unaffected adults; this only affects whether BS2 could ever be met, not the current verdict.

Sources

Reasoning

BP1 (Richards et al. 2015, PMID:25741868) applies only when truncating variants are the primary - effectively the only - mechanism by which variants cause disease in the gene. POLD1 does not fit that profile: missense variants are an established dominant disease mechanism alongside truncating ones. Exonuclease-domain missense variants cause polymerase proofreading-associated polyposis (OncoKB gene background: 'Germline heterozygous loss-of-function mutations in the exonuclease domain of POLD1 cause polyposis and predispose individuals to colorectal, endometrial, and possibly brain cancers'), CysB-region missense variants cause MDPL and combined immunodeficiency (p.Arg1060Cys, PMID:31629014; p.Ile1070Asn and p.Ser1073Arg, PMID:35590056; p.Thr1072Pro, PMID:34517090), the recurrent in-frame deletion p.Ser605del also causes MDPL, and ClinVar contains POLD1 missense variants classified Likely pathogenic (for example p.Ser478Arg, c.1432A>C). The case's own PVS1 context likewise records a supported germline loss-of-function mechanism (pvs1_gene_context.json, lof_mechanism_supported=true), which means truncating variants contribute but are explicitly not the only mechanism. Because missense variation is a well-documented cause of POLD1 disease, the premise that primarily truncating variants cause disease is false, and BP1 is not met.

Rule

Missense variant in a gene for which primarily truncating variants are known to cause disease.

Evidence considered

• OncoKB gene-level background (oncokb source) for POLD1: 'Germline heterozygous loss-of-function mutations in the exonuclease domain of POLD1 cause polyposis and predispose individuals to colorectal, endometrial, and possibly brain cancers' - i.e. the disease is caused by a specific class of missense/proofreading-domain alleles, not solely by truncating variants.
• CysB-region missense variants cause autosomal dominant disease: p.Arg1060Cys causes combined immunodeficiency (PMID:31629014: 'We identified a missense mutation (c.3178C>T; p.R1060C) in POLD1 in 3 related subjects...'), while p.Ile1070Asn and p.Ser1073Arg (PMID:35590056) and p.Thr1072Pro (PMID:34517090) cause MDPL syndrome; MDPL is also caused by the recurrent in-frame deletion p.Ser605del.
• ClinVar POLD1 records include missense variants classified Likely pathogenic (e.g. p.Ser478Arg, NM_002691.4:c.1432A>C), confirming missense pathogenicity in the germline setting for this gene.
• pvs1_gene_context.json records lof_mechanism_supported=true with mechanism_rationale 'Targeted germline literature review identified disease-focused publications supporting POLD1 loss of function as a germline disease mechanism' - truncating variants do contribute, but the same review set and the gene-level background show missense variants are also a mechanism, which is the situation BP1 excludes.

Still missing

• No ClinGen POLD1 VCEP/CSPEC specification exists, so BP1 was applied from the generic ACMG/AMP 2015 definition.
• The relative proportion of truncating versus missense POLD1 pathogenic variants was not quantified from a curated variant spectrum; the criterion was decided on the established existence of a missense disease mechanism, which is sufficient to exclude BP1.

Sources

Reasoning

BP2 (Richards et al. 2015, PMID:25741868) requires the variant to be observed in cis with a pathogenic variant in any inheritance pattern, or in trans with a pathogenic variant in a fully penetrant dominant gene/disorder. Nothing in the assembled evidence satisfies either arm. The exact-match ClinVar record (ClinVarID 537093) contains three germline submissions for this single variant - two Uncertain significance and one Likely benign - and none reports a co-occurring pathogenic POLD1 variant in cis or trans, nor any phased genotype. Population data show no genotype consistent with a second hit on the same allele: gnomAD v2.1 reports 2/224,974 alleles with 0 homozygotes, gnomAD v4.1 reports 6/1,596,184 alleles with 0 homozygotes, and gnomAD-Canada v1.0 reports the variant absent. The two triaged papers (PMID:25394175, PMID:28492532) are framework/guideline documents that do not mention this variant and report no genotype data. Because no pathogenic variant was observed in cis or trans with c.3205G>A in any source examined, BP2 is not met; the caveat is that phase was never directly determined in this case, so this rests on absence of the required observation rather than on a negative phasing experiment.

Evidence considered

• BP2 applied from the generic ACMG/AMP 2015 definition (Richards et al., PMID:25741868); no POLD1 VCEP/CSPEC criteria specification exists for this case.
• ClinVar ClinVarID 537093 exact transcript match NM_002691.4(POLD1):c.3205G>A: SCV000767598 (Labcorp/Invitae, Uncertain significance, germline, cites PMID:28492532), SCV001995559 (GeneDx, Uncertain significance, germline), SCV006051457 (Ambry Genetics, Likely benign, germline); no submission describes a second POLD1 variant in cis or trans, and no phase information is provided.
• ClinVar submitter SCV000767598 cites PMID:28492532 (Sherloc, an ACMG/AMP criteria-refinement framework paper); that paper reports no variant-level data, no phase data, and does not mention c.3205G>A.
• gnomAD v2.1: 2/224,974 alleles (AF 8.89e-06), 0 homozygotes; gnomAD v4.1: 6/1,596,184 alleles (AF 3.76e-06), 0 homozygotes; gnomAD-Canada v1.0: variant absent - no population genotype indicates a co-occurring pathogenic allele on either chromosome.
• Gene-level context: POLD1 germline disease is dominantly inherited (heterozygous exonuclease-domain loss of function causes polyposis/colorectal cancer predisposition), so the BP2 'in trans with a pathogenic variant in a fully penetrant dominant gene' arm is the operative one for this gene.
• Searched the triaged literature packet (2 PMIDs) and the full-text extraction for cis/trans or co-occurrence observations of POLD1 c.3205G>A / p.Val1069Ile; no such observation was found.

Still missing

• No phasing, zygosity, segregation or family-genotype data were available; BP2 was evaluated from the absence of any reported co-occurring pathogenic variant rather than from a directly phased genotype.
• No ClinGen POLD1 VCEP/CSPEC specification exists, so the generic ACMG/AMP 2015 definition of BP2 was applied.

Sources

Reasoning

BP6 is the benign-direction counterpart of PP5: it applies only when a reputable source reports the variant as benign and the laboratory cannot independently evaluate the evidence, and under the governing policy for this group it may be triggered only by an exact-variant ClinVar classification of Benign or Likely benign from an expert panel (ClinGen VCEP or equivalent expert-panel review). For NM_002691.4(POLD1):c.3205G>A (p.Val1069Ile), the only benign-direction record in VCV000537093 is SCV006051457 from Ambry Genetics (Likely benign, review status 'criteria provided, single submitter', method clinical testing, origin germline, is_expert_panel = false, extracted criteria_found = []). That is an ordinary laboratory submission of the kind the policy explicitly excludes from triggering BP6, and it carries no supporting criteria that could be adopted. The record as a whole is 1-star and conflicting (two Uncertain significance submissions versus one Likely benign), and the submission audit reports expert_panel_submissions = 0, so no ClinGen expert-panel Benign/Likely benign classification exists. Because the sole benign-direction input is a non-expert single-submitter assertion, BP6 is not met. Numeric direction check: BP6 as applied here carries no numeric likelihood-ratio or frequency threshold in the governing generic ACMG/AMP specification or in any retrieved gene-specific specification, so no comparison operator or inequality exists to evaluate; the verdict rests entirely on the absence of an expert-panel benign classification, not on any value comparison.

Rule

BP6 policy: requires an exact-variant ClinVar classification of Benign/Likely benign from an expert panel (VCEP); ordinary laboratory submissions and aggregate ClinVar labels never trigger BP6

Evidence considered

• ClinVar VCV000537093 benign-direction submission SCV006051457 (Ambry Genetics): classification Likely benign, review status 'criteria provided, single submitter', method clinical testing, origin germline, is_expert_panel false, trust_tier single_submitter, evidence_signals.criteria_found empty.
• ClinVar record-level status: 'Conflicting classifications of pathogenicity', 1 star, submission summary 2 clinical laboratories Uncertain significance and 1 clinical laboratory Likely benign.
• ClinVar submission audit counts for this exact variant: submissions_total 3, usable_for_classification 3, expert_panel_submissions 0, validated_pmids ['28492532'].
• PMID:28492532 is the only validated citation in the submission trail (a criteria-refinement methodology paper) and is not attached to the Likely benign submission at all, which lists no PMIDs.

Still missing

• No ClinGen expert-panel (VCEP) Benign/Likely benign classification exists for this variant, so BP6 cannot be applied regardless of the individual laboratory submissions.
• The Ambry Genetics Likely benign assertion provides no extracted supporting criteria and remains an unpublished, non-adoptable single-submitter opinion.

Sources

Reasoning

PS2 (Richards et al. 2015, PMID:25741868) requires a de novo occurrence of the variant in a patient with the disease, with both maternity and paternity confirmed and no family history. NM_002691.4:c.3205G>A (NP_002682.2:p.(Val1069Ile)) in POLD1 is presented here as a bare variant with no proband attached: the case carries no patient record, no parental genotypes, no statement of maternity/paternity confirmation, no phenotype or disease status for a carrier, and no family history. Screening of every case evidence file for de novo, proband, parental, pedigree, meiosis, inheritance, family and segregation content returned nothing, and direct reads of prefetch.json, vcep_materials.json, clinvar_submission_audit.json and gene_context.json closed the remaining files without adding any. The ClinVar record for this variant (VCV000537093, variation ID 537093) holds three germline single-submitter entries - Uncertain significance from Labcorp Genetics (formerly Invitae) and from GeneDx, and Likely benign from Ambry Genetics - and none of them states a de novo occurrence or reports parental testing; the only submission-level criterion signals extracted from the submitter text are PM2 (population) and PP3 (in silico), neither of which speaks to inheritance. The single fetched full text, PMID:25394175, is an ACMG/NSGC cancer-referral practice guideline that does not mention POLD1 or this variant, and the POLD1 gene-context papers retrieved for the PVS1 strand contain no occurrence of c.3205G>A or p.(V1069I), so no publication supplies a de novo observation for this variant. Because PS2's core requirement - a confirmed de novo event in an affected proband - is absent from the record rather than merely uncounted, this criterion cannot be evaluated on the available evidence. Generic ACMG/AMP 2015 governs, as no ClinGen CSPEC and no POLD1 gene-specific specification exists (framework_complete = false, source = generic_acmg_fallback).

Rule

ACMG/AMP 2015 PS2 (PMID:25741868): de novo (both maternity and paternity confirmed) in a patient with the disease and no family history

Evidence considered

• PS2 definition applied from the generic ACMG/AMP 2015 guideline (PMID:25741868), which is the governing text for this case because no ClinGen CSPEC and no POLD1 gene-specific specification exists (case_summary compact_evidence.cspec.found = false; vcep_materials.json states 'no cspec url available' with an empty summaries/files_found list; final_classification_framework.found = false, source = generic_acmg_fallback, scoring_model = ACMG/AMP 2015).
• Case-wide search for family-level evidence: the review manifest, case summary, source registry, evidence, literature pass, full-text map and paper extracts contain no occurrence of 'de novo', 'proband', 'parental', 'pedigree', 'meiosis', 'inheritance', 'family' or 'segregation', and direct reads of prefetch.json, vcep_materials.json, clinvar_submission_audit.json, gene_context.json and evidence_brief.json added none; the case has no patient record at all.
• ClinVar (source_registry key 'clinvar'), VCV000537093 / variation ID 537093: three germline clinical-testing submissions, all 'criteria provided, single submitter' - SCV000767598 Labcorp Genetics (formerly Invitae) Uncertain significance, last evaluated 2025-12-09; SCV001995559 GeneDx Uncertain significance, 2023-11-10; SCV006051457 Ambry Genetics Likely benign, 2025-01-10. Expert panels = {} (no VCEP submission). None of the three submissions states a de novo occurrence or reports parental/maternity/paternity testing.
• ClinVar submission-level criterion signals extracted from the submitter text are limited to PM2 (Labcorp: 'the frequency data for this variant in the population databases is considered unreliable...') and PP3 (GeneDx: 'in silico analysis supports that this missense variant does not alter protein structure/function'); neither submission reports de novo or inheritance information for this variant.
• Paper-extract review (paper_extracts.json, source_registry key 'PMID:25394175'): the only fetched full text is an ACMG/NSGC practice guideline on referral indications for cancer predisposition assessment; variant_mentioned = false, no POLD1 content and no family or de novo data. The ClinVar-cited Sherloc paper (PMID:28492532) was not retrieved as full text and reports no variant-specific inheritance data.
• Gene-context literature retrieved for the PVS1 strand (PMIDs 41673532, 42286372, 36291559, 37030525, 38749661) was screened for the exact strings 'c.3205G>A', '3205G' and 'V1069I' with no occurrence, so no publication provides a de novo observation for this variant.
• Variant context: NM_002691.4:c.3205G>A is a rare missense substitution (gnomAD v2.1 total AF 8.89e-06, 2/224974 alleles; ClinVar conflicting classifications), so inherited-versus-de-novo status cannot be inferred from frequency or from another submitter's classification.

Still missing

• No proband-level genotypes: whether c.3205G>A occurred de novo in an affected individual is unknown.
• No parental testing or confirmation of maternity and paternity for the carrier of this variant.
• No phenotype/disease status for the variant carrier and no family history, both of which PS2 requires.

Sources

Reasoning

PS3 (Richards et al. 2015, PMID:25741868) requires a well-established in vitro or in vivo functional study showing a deleterious effect on the gene or gene product, and is variant-specific by default. No POLD1 CSPEC or VCEP criterion specification exists for this case (no CSPEC URL available; framework_complete=false), so the generic ACMG/AMP 2015 criterion definition governs. NM_002691.4:c.3205G>A is a missense substitution, NP_002682.2:p.(Val1069Ile), so a functional-assay line of evidence is applicable in principle and was actively searched for. No functional-assay evidence for this variant exists in any consulted source. OncoKB returned 'Unknown Oncogenic Effect' with not_reviewed=true and no variant-specific reviewed functional evidence, only gene-level text ('POLD1, a DNA polymerase, is infrequently altered by mutation in cancer'). ClinVar VCV000537093 carries three submitter classifications (Labcorp/Invitae and GeneDx Uncertain significance, Ambry Likely benign, all 'criteria provided, single submitter') but no submitter reports functional assay data, and the only PMID trail behind it (28492532) is the Sherloc classification-framework paper, which reports no variant-level data. COSMIC has no record for this variant and Cancer Hotspots lists neither the exact variant nor a significant hotspot at residue V1069. The single full-text publication retrieved for this case, PMID:25394175 (ACMG/NSGC cancer-referral practice guideline), does not mention POLD1 or this variant anywhere. Independent PubMed and PMC searches (NCBI E-utilities, accessed 2026-09-18) for 'POLD1 V1069I', 'c.3205G>A' and 'Val1069Ile' returned zero records. POLD1 germline disease is mediated by loss of proofreading through exonuclease-domain variants, and published POLD1 functional assays characterise that N-terminal exonuclease domain, which does not encompass position 1069; no systematically characterised residue range covering position 1069 was identified in any source, so PS3 cannot be credited by range extrapolation either. PS3 is recorded as not_assessed rather than not_met because the negative finding is the absence of any functional study of this variant, not a functional study showing it to be tolerated; no PS3 strength (supporting through very strong) can be assigned.

Rule

generic_acmg_2015:PS3

Evidence considered

• Criterion definition: PS3 requires a well-established functional study showing a damaging effect on the gene or gene product and is variant-specific by default; adjudicated under generic ACMG/AMP 2015 (Richards et al., PMID:25741868) because no POLD1 CSPEC/VCEP criterion specification was available for this case (no CSPEC URL; framework_complete=false; gene_vcep_dir null).
• OncoKB variant page for POLD1 V1069I: classification 'Unknown Oncogenic Effect', is_variant_specific=false, not_reviewed=true, citations_pmids empty; summary is gene-level only ('POLD1, a DNA polymerase, is infrequently altered by mutation in cancer'). No variant-specific reviewed functional evidence.
• ClinVar VCV000537093 (NM_002691.4(POLD1):c.3205G>A, p.Val1069Ile): 3 submissions - SCV000767598 Labcorp Genetics (formerly Invitae) Uncertain significance, SCV001995559 GeneDx Uncertain significance, SCV006051457 Ambry Genetics Likely benign, each 'criteria provided, single submitter'; conflicting classifications, 1 star. No submission reports a functional assay, and the one validated PMID trail (28492532, Sherloc) is a classification-framework refinement with no data on this variant.
• COSMIC (query 'POLD1 c.3205G>A'): no result - variant not reported in somatic cancer data. Cancer Hotspots (query 'POLD1 V1069'): no hotspot row and no exact variant listed, so no residue-level functional or recurrence signal.
• PMID:25394175 (Hampel et al. 2015, ACMG/NSGC referral-indications practice guideline): full text searched for c.3205G>A, p.(V1069I), V1069I and NM_002691.4 - neither the variant nor POLD1 is mentioned; the paper contains no functional assay data, case observations or gene-level mechanistic discussion of POLD1.
• Independent literature search (NCBI E-utilities, PubMed and PMC, accessed 2026-09-18): PubMed 'POLD1 V1069I' = 0 records, 'POLD1 "c.3205G>A"' = 0 records, '"Val1069Ile"' = 0 records in both PubMed and PMC. No indexed publication reports this variant, so the PS3 gap cannot be closed from the retrievable literature.
• Gene/mechanism context: POLD1 germline disease is mediated by loss of proofreading via exonuclease-domain variants, and the published POLD1 functional assays reported in the gene-level literature characterise that N-terminal exonuclease domain rather than the C-terminal region containing residue 1069; no study with systematic range characterisation covering position 1069 was identified.

Still missing

• No variant-specific functional assay (POLD1 3'->5' exonuclease/proofreading activity assay, polymerase-activity assay, yeast or mammalian cell mutagenesis/reversion assay, or patient-derived cell mismatch-accumulation assay) reports results for NM_002691.4:c.3205G>A p.(Val1069Ile); such an assay would be required to assess PS3 at any strength.
• No systematic mutagenesis, saturation or tiling dataset spanning POLD1 residue 1069 was found, so PS3 cannot be credited from a systematically characterised range containing the variant.
• OncoKB, ClinVar, COSMIC and Cancer Hotspots contain no variant-level functional annotation that could substitute for a primary assay; no published ClinVar submission or citing paper supplies unpublished functional data.
• Whether any functional data exist outside the indexed literature (e.g. an unpublished commercial-laboratory assay or a variant-specific report in a non-indexed source) was not resolvable from the sources available for this case.

Sources

Reasoning

PM6 (Richards et al. 2015, PMID:25741868) requires an assumed de novo occurrence without confirmation of paternity and maternity: a proband in whom the variant is present and for whom parental testing is absent or unconfirmed but whose family structure nevertheless indicates a new event. This case contains no proband and no parental data of any kind. Screening of the review manifest, case summary, source registry, evidence, literature pass, full-text map and paper extracts for proband, parental, pedigree, inheritance and de novo content returned nothing, and direct reads of prefetch.json, vcep_materials.json, clinvar_submission_audit.json and gene_context.json added no carrier-level information. The ClinVar record (VCV000537093) comprises three single-submitter germline entries - two Uncertain significance and one Likely benign - that give no inheritance statement, and its only extracted criterion signals are PM2 and PP3, both unrelated to assumed de novo status. The single fetched full text (PMID:25394175) does not mention POLD1 or this variant, and the POLD1 gene-context papers screened for the variant show no occurrence of it. Because no assumed de novo event is described anywhere in the record, the premise PM6 addresses is absent rather than merely unconfirmed, and the criterion cannot be applied. Generic ACMG/AMP 2015 governs this variant, no ClinGen CSPEC or POLD1-specific specification being available.

Rule

ACMG/AMP 2015 PM6 (PMID:25741868): assumed de novo, but without confirmation of paternity and maternity

Evidence considered

• PM6 definition applied from the generic ACMG/AMP 2015 guideline (PMID:25741868); no ClinGen CSPEC or POLD1 gene-specific specification governs this case (case_summary compact_evidence.cspec.found = false; vcep_materials.json 'no cspec url available'; final_classification_framework.found = false, source = generic_acmg_fallback).
• Case-wide search for proband/parental content ('de novo', 'assumed de novo', 'proband', 'parental', 'pedigree', 'meiosis', 'segregation', 'family history') found no occurrence in any of the bundled case evidence files, and prefetch.json, vcep_materials.json, clinvar_submission_audit.json and gene_context.json read directly also contain none.
• ClinVar (source_registry key 'clinvar'), VCV000537093: three germline submissions from three clinical laboratories (Labcorp/Invitae and GeneDx Uncertain significance; Ambry Genetics Likely benign), each 'criteria provided, single submitter', no expert-panel review; no submission asserts or implies a de novo origin and none reports parental testing.
• Paper-extract review (paper_extracts.json, source_registry key 'PMID:25394175'): the sole fetched full text is an ACMG/NSGC referral-indications practice guideline, variant_mentioned = false; it defines no case for POLD1 c.3205G>A and reports no assumed de novo event.
• Gene-context publications PMIDs 41673532, 42286372, 36291559, 37030525 and 38749661 were searched for 'c.3205G>A', '3205G' and 'V1069I' and do not mention the variant, so none describes this variant in a proband.
• Variant context: this is a rare missense variant (gnomAD v2.1 AF 8.89e-06; a single non-Finnish European and a single Admixed American allele, no homozygotes), so no de novo inference is available from allele data.

Still missing

• No proband record tying this variant to an affected individual for whom inheritance could be inferred.
• No parental testing information, even unconfirmed, that would allow the assumed-de-novo path of PM6 to be considered.
• No family structure or disease status for relatives that could indicate a new event.

Sources

Reasoning

PP1 (Richards et al. 2015, PMID:25741868) requires co-segregation of the variant with disease in multiple affected family members of a gene definitively known to cause the disease, which presupposes at least one informative pedigree with genotype data on affected relatives. POLD1 does satisfy the gene-level half of that premise - the gene is an established colorectal cancer predisposition gene (ClinVar conditions for this variant include 'Colorectal cancer, susceptibility to, 10' and 'Hereditary cancer-predisposing syndrome'; the case's gene context describes germline exonuclease-domain mutations causing polyposis with colorectal, endometrial and possibly brain cancer predisposition) - but the family half is entirely absent. This case provides no pedigree, no proband, no relatives and no meioses: screening of the review manifest, case summary, source registry, evidence, literature pass, full-text map and paper extracts for segregation, pedigree, proband, affected-relative and family-history content returned nothing, and direct reads of prefetch.json, vcep_materials.json, clinvar_submission_audit.json and gene_context.json added none. The ClinVar record (VCV000537093) is three single-submitter germline entries with no segregation or family data, and the only retrieved full text (PMID:25394175) is an ACMG/NSGC referral guideline with no POLD1 content; the gene-context papers screened contain no occurrence of c.3205G>A or p.(V1069I). With zero informative meioses available, no co-segregation can be counted at any strength, so PP1 is not evaluable. Generic ACMG/AMP 2015 governs the case, as no ClinGen CSPEC or POLD1 gene-specific specification exists.

Rule

ACMG/AMP 2015 PP1 (PMID:25741868): co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease

Evidence considered

• PP1 definition applied from the generic ACMG/AMP 2015 guideline (PMID:25741868); no ClinGen CSPEC or POLD1 gene-specific framework is available for this case (case_summary compact_evidence.cspec.found = false; vcep_materials.json 'no cspec url available', framework source = generic_acmg_fallback).
• Search of all case evidence files for segregation content ('segregation', 'co-segregation', 'pedigree', 'proband', 'meioses', 'family history', 'affected relative') returned no matches, including prefetch.json, vcep_materials.json, clinvar_submission_audit.json and gene_context.json read directly.
• Gene-disease premise (the only part of PP1 that could be satisfied here): POLD1 is a recognized hereditary colorectal cancer predisposition gene, and the ClinVar record for this variant lists conditions 'Colorectal cancer, susceptibility to, 10' and 'Hereditary cancer-predisposing syndrome'; the case gene-context summary describes germline POLD1 exonuclease-domain mutations causing polyposis with colorectal, endometrial and possibly brain cancer predisposition. This establishes the gene-disease relationship but provides no family genotypes.
• ClinVar (source_registry key 'clinvar'), VCV000537093: three germline single-submitter submissions (Labcorp/Invitae and GeneDx Uncertain significance; Ambry Genetics Likely benign); none reports a pedigree, affected relative, or co-segregation observation for c.3205G>A.
• Paper-extract review (paper_extracts.json, source_registry key 'PMID:25394175'): the only fetched full text is a general ACMG/NSGC cancer-referral guideline with variant_mentioned = false; it reports no POLD1 co-segregation observations.
• Gene-context publications (PMIDs 41673532, 42286372, 36291559, 37030525, 38749661) were screened for the variant strings 'c.3205G>A', '3205G' and 'V1069I'; the variant does not appear in them, and no family study of this variant was identified.

Still missing

• No pedigree or family structure linked to the variant carrier.
• No genotype data on affected or unaffected relatives; zero informative meioses available for co-segregation counting.
• No variant-specific family study in the retrieved literature (PubMed/ClinVar citations for this variant report no segregation).

Sources

Reasoning

Generic ACMG/AMP 2015 PP4 (Richards et al. 2015, PMID:25741868) requires the patient's phenotype or family history to be highly specific for a disease with a single genetic etiology. This case record contains no proband-level phenotype, no HPO term list, no family-history description, and no clinical indication beyond the variant itself; the only phenotype-adjacent labels available are the ClinVar condition labels attached to VCV000537093, namely 'Colorectal cancer, susceptibility to, 10', 'Hereditary cancer-predisposing syndrome', and 'not provided'. Those are broad cancer-predisposition categories that are shared across many genes and are not a phenotype specific to a single-gene etiology, so they cannot satisfy PP4 even if a phenotype were attributable to this variant. No case report describing an affected carrier of p.(Val1069Ile) was retrieved: the single triaged paper (PMID:25394175) does not mention the variant or POLD1. With neither a proband phenotype/family history nor variant-level phenotype-specificity evidence, PP4's prerequisite is missing rather than contradicted, so the criterion is recorded as not_assessed. No numeric threshold is involved in PP4, so no inequality applies.

Rule

generic_acmg_2015: patient's phenotype or family history highly specific for a disease with a single genetic etiology (Richards et al. 2015, PMID:25741868)

Evidence considered

• ClinVar VCV000537093 condition labels: 'Colorectal cancer, susceptibility to, 10', 'Hereditary cancer-predisposing syndrome', 'not provided' - broad predisposition categories rather than a phenotype highly specific for a single-gene disorder.
• Case-level phenotype inputs are absent: review_manifest.json, case_summary.json, evidence.json, and prefetch.json contain no proband phenotype, HPO terms, or family-history field (a text search of prefetch.json returned only a paper title containing the word 'indications').
• PMID:25394175 was the only paper with a phenotype- or case-observation theme in the literature packet, and its extraction records no POLD1 or variant mention, so it supplies no phenotype-specificity evidence.

Still missing

• No proband phenotype, HPO term set, or family history is present anywhere in the collected case evidence, so phenotype specificity for this variant cannot be evaluated.
• No published case description of an affected carrier of POLD1 p.(Val1069Ile) was found in the collected sources; such a report, if it exists, would be the alternative route to a PP4 assessment.

Sources

Reasoning

BS3 (Richards et al. 2015, PMID:25741868) requires well-established functional studies showing no damaging effect on protein function. The same systematic search that failed to find PS3-supporting data also found no variant-level assay of POLD1 p.(Val1069Ile) in any consulted source: OncoKB reports no variant-specific reviewed functional evidence for POLD1 V1069I, the three ClinVar submissions for VCV000537093 cite no functional assay, COSMIC has no record of POLD1 c.3205G>A, Cancer Hotspots lists neither the variant nor a significant hotspot at V1069, the one retrieved full text (PMID:25394175) does not mention POLD1 or the variant, and PubMed/PMC searches for 'V1069I', 'Val1069Ile' and 'c.3205G>A' returned zero records (accessed 2026-09-18). Computational predictions are explicitly not a substitute for BS3: this is a missense variant, so only REVEL could support PP3/BP4 on the appropriate predictor path, and neither REVEL nor the unavailable SpliceAI output is functional evidence for BS3 - the SpliceAI query timed out with max_delta_score null, and a missing score must not be read as evidence of no splice impact. No functional assay result, damaging or neutral, exists for this variant, so BS3 is not_assessed; importantly, the absence of functional data must not be converted into BS3 support (or into PS3 rejection) since the criterion is defined by a positive functional demonstration of non-damaging effect.

Rule

generic_acmg_2015:BS3

Evidence considered

• Criterion definition: BS3 requires well-established functional studies showing no damaging effect on protein function; adjudicated under generic ACMG/AMP 2015 (Richards et al., PMID:25741868) because no POLD1 CSPEC/VCEP specification was available for this case.
• OncoKB variant page for POLD1 V1069I: classification 'Unknown Oncogenic Effect', not_reviewed=true, empty citations - no functional study (damaging or neutral) reported for this variant.
• ClinVar VCV000537093: three single-submitter records (Labcorp/Invitae Uncertain significance, GeneDx Uncertain significance, Ambry Genetics Likely benign). Even the Likely-benign submission cites no functional assay, so it provides no BS3-supporting experimental evidence; the one validated PMID trail (28492532, Sherloc) reports no variant-level data.
• COSMIC: no record for POLD1 c.3205G>A. Cancer Hotspots: no hotspot row for POLD1 V1069 and no exact variant. Neither source reports functional or experimentally characterised behaviour for this variant.
• PMID:25394175 (Hampel et al. 2015, ACMG/NSGC referral-indications practice guideline): searched in full text for c.3205G>A, p.(V1069I), V1069I and NM_002691.4 with no hit; contains no POLD1 functional data of any kind.
• Independent literature search (NCBI E-utilities, PubMed and PMC, accessed 2026-09-18): 'POLD1 V1069I' = 0, 'POLD1 "c.3205G>A"' = 0, '"Val1069Ile"' = 0 records. No publication reports an assay result for this variant, so neither a damaging-effect nor a non-damaging-effect demonstration is available.
• Predictor data reviewed and deliberately excluded as functional evidence: REVEL score 0.197 (missense variant, so REVEL is the relevant PP3/BP4 path and it is not a functional assay), BayesDel -0.367908 (no generic ACMG-strength calibration established, treated as not_available), and SpliceAI with max_delta_score null (query timeout; missing data that must not be interpreted as absent splice impact).

Still missing

• No functional assay demonstrating normal or near-normal POLD1 function for NM_002691.4:c.3205G>A p.(Val1069Ile) exists in the consulted sources, so BS3 has no positive basis; a variant-specific proofreading/polymerase-activity or cellular mutagenesis assay performed against a validated benchmark set would be required.
• No systematic functional dataset covering POLD1 residue 1069 was identified that could support a range-based BS3 argument.
• No variant-level functional annotation is available from OncoKB, ClinVar submitters, COSMIC or Cancer Hotspots to substitute for a primary assay.
• In-silico output cannot close this gap for a missense variant: there is no SpliceAI score (query timed out, max_delta_score null) and the missense path is REVEL-based, which is not a functional assay.

Sources

Reasoning

BS4 (Richards et al. 2015, PMID:25741868) requires an observed lack of segregation in affected members of a family - affected relatives who carry the variant without co-segregating, or affected relatives demonstrated not to carry it. No family data of any kind exists in this case: screening of the review manifest, case summary, source registry, evidence, literature pass, full-text map and paper extracts for pedigree, affected-relative, proband, parental and segregation content returned nothing, and direct reads of prefetch.json, vcep_materials.json, clinvar_submission_audit.json and gene_context.json added none. Absence of family data is not itself BS4 evidence: non-segregation must actually be observed, and nothing in this record demonstrates it. Neither the ClinVar record (VCV000537093 - three single-submitter germline entries, two Uncertain significance and one Likely benign, with no family data and no segregation statement) nor the single retrieved full text (PMID:25394175, an ACMG/NSGC referral guideline that does not mention POLD1 or this variant) provides a tested affected relative, and the gene-context papers screened contain no occurrence of c.3205G>A or p.(V1069I). BS4 therefore cannot be evaluated on this record, and generic ACMG/AMP 2015 applies because no ClinGen CSPEC or POLD1 gene-specific specification exists.

Rule

ACMG/AMP 2015 BS4 (PMID:25741868): lack of segregation in affected members of a family

Evidence considered

• BS4 definition applied from the generic ACMG/AMP 2015 guideline (PMID:25741868); no ClinGen CSPEC or POLD1 gene-specific specification governs this case (case_summary compact_evidence.cspec.found = false; vcep_materials.json 'no cspec url available'; final_classification_framework.found = false, source = generic_acmg_fallback).
• Search of all case evidence files for non-segregation evidence ('segregation', 'not segregate', 'pedigree', 'affected relative', 'proband', 'family history') returned no matches; prefetch.json, vcep_materials.json, clinvar_submission_audit.json and gene_context.json read directly also contain none.
• ClinVar (source_registry key 'clinvar'), VCV000537093: three single-submitter germline records (SCV000767598 Labcorp/Invitae VUS; SCV001995559 GeneDx VUS; SCV006051457 Ambry Genetics Likely benign), no expert panel, conditions listed as 'Colorectal cancer, susceptibility to, 10', 'Hereditary cancer-predisposing syndrome' and 'not provided'; no submitter reports tested relatives or any segregation or non-segregation observation.
• Paper-extract review (paper_extracts.json, source_registry key 'PMID:25394175'): only that practice guideline was fetched as full text; it is a general cancer-referral framework, variant_mentioned = false, and provides no segregation or non-segregation observations.
• Gene-context publications (PMIDs 41673532, 42286372, 36291559, 37030525, 38749661) were screened for 'c.3205G>A', '3205G' and 'V1069I' with no occurrence, so no lack-of-segregation observation for this variant is available from the retrieved literature.
• Absence of family data was not converted into BS4 support: BS4 requires demonstrated non-segregation, which is not present in this record.

Still missing

• No pedigree or relative genotypes, so non-segregation can neither be shown nor excluded.
• No affected family members reported in any ClinVar submission or in the retrieved literature for this variant.

Sources

Reasoning

Generic ACMG/AMP 2015 BP5 (Richards et al. 2015, PMID:25741868) requires that the variant be found in a case with an alternate molecular basis for the disease, for example a co-occurring pathogenic variant in a different gene that explains the phenotype. This case contains no proband-level genotype data, no second-gene or panel testing results, and no statement of an alternative molecular cause for any affected individual; the ClinVar records for VCV000537093 are variant-level submissions that report no accompanying alternate pathogenic finding (their only extracted criterion signals are PM2 and PP3 from uncertain-significance submissions, and the Ambry Genetics Likely benign submission reports no criteria at all). No case report describing an affected carrier of p.(Val1069Ile) was retrieved, so no published instance of this variant co-occurring with an alternative molecular explanation could be examined either. Because the criterion's prerequisite - a case in which this variant was observed together with another molecular explanation for disease - has neither been established nor excluded, BP5 is recorded as not_assessed rather than not_met. No numeric threshold (and therefore no comparison operator or inequality) is involved in BP5.

Rule

generic_acmg_2015: variant found in a case with an alternate molecular basis for disease (Richards et al. 2015, PMID:25741868)

Evidence considered

• ClinVar VCV000537093 submissions report no co-occurring alternate pathogenic finding and no case-level molecular details beyond classification and review status (SCV000767598 Uncertain significance, SCV001995559 Uncertain significance, SCV006051457 Likely benign).
• Extracted submission criterion signals are PM2 and PP3 only; no second-gene or alternate-basis evidence appears in the ClinVar submission audit for this variant.
• No proband-level data exist in the case record: review_manifest.json, case_summary.json, evidence.json, and prefetch.json contain no patient-level testing results, and no retrieved paper mentions this variant.

Still missing

• No proband-level genotype/panel data (or published case) is available to determine whether an alternate molecular basis for disease was present in an individual carrying this variant.
• The ClinVar submissions are variant-level only and do not report co-occurring pathogenic findings, so they cannot close this gap.