POLD1 c.187G>A, p.(Glu63Lys)

Reasoning

Scope first: NM_002691.4:c.187G>A in POLD1 is a missense substitution (NP_002682.2:p.(Glu63Lys)), so BP4 is evaluated from the missense protein-impact path (REVEL) only; the SpliceAI splice path does not apply to this variant class. No POLD1 ClinGen CSPEC/VCEP or gene-specific PP3/BP4 rule set exists for this case, so the generic ClinGen SVI REVEL calibration applies. The local REVEL v1.3 lookup gives 0.018. Against the published bands (Pejaver et al. 2022, PMID:36413997): strong BP4 requires <= 0.016, moderate <= 0.183, supporting <= 0.29. A score of 0.018 does not reach the strong band but sits below both the moderate and supporting cutoffs, so BP4 is applied at moderate strength. The score is only marginally above the strong band boundary (0.016 vs 0.018), which does not change the band assignment under the published cutoffs but is noted for transparency. SpliceAI max delta 0.008 is numerically <= 0.1, yet it was deliberately not used as BP4 support here: BP4 for a missense variant is taken only from REVEL, so the clean splice delta neither strengthens nor creates the criterion. BayesDel (-0.510638) is available but has no generic ACMG-strength calibration and was not used for BP4.

Rule

BP4 (moderate): missense/REVEL path; REVEL 0.018 <= 0.183 and > 0.016 (Pejaver et al. 2022, PMID:36413997)

Evidence considered

• Variant consequence is missense (NP_002682.2:p.(Glu63Lys)), so BP4 is taken from the REVEL missense path only; the SpliceAI score cannot support or negate BP4 for a missense variant.
• Local REVEL v1.3 lookup (revel source registry entry) returned 0.018 for chr19:50399038 G>A, matching this variant's GRCh38 position.
• Applied thresholds (ClinGen SVI REVEL calibration, Pejaver et al. 2022, Am J Hum Genet, PMID:36413997): BP4 strong <= 0.016, moderate <= 0.183, supporting <= 0.29. REVEL 0.018 is <= 0.183 but > 0.016, so BP4 is met at moderate strength.
• SpliceAI max delta 0.008 (<= 0.1) was not used as BP4 support because the splice path is not the applicable path for a missense variant; only the REVEL score feeds BP4 here.
• BayesDel -0.510638 is available but has no established generic ACMG-strength calibration, so it was treated as not_available for BP4.
• No POLD1-specific CSPEC/VCEP or gene-specific PP3/BP4 lookup exists (no cspec url; final_classification_framework found = false), so generic calibrated thresholds govern.

Sources

Reasoning

PS1 requires that the same amino-acid change (p.Glu63Lys) has been established as pathogenic when produced by a different nucleotide change. ClinVar holds exactly one record carrying the p.Glu63Lys protein change, NM_002691.4(POLD1):c.187G>A, variation ID 407981 - i.e. this variant itself - and it has no pathogenic or likely pathogenic submission (5 submissions: 3 Likely benign, 2 Uncertain significance; conflicting classifications, no expert-panel review). A ClinVar search for POLD1 plus 'Glu63Lys' returns only that single record, so no alternative nucleotide change yielding p.Glu63Lys has been reported at all. Structurally this is expected: codon 63 is GAR (Glu) and Lys (AAA/AAG) is reachable from it by a single-nucleotide substitution only at c.187, so the PS1 comparator set is effectively empty rather than merely negative. PS1 is not met.

Rule

PS1: same amino-acid change caused by a different nucleotide change, previously established as pathogenic

Evidence considered

• ClinVar variation 407981 (NM_002691.4(POLD1):c.187G>A, p.Glu63Lys) aggregates 5 single-submitter classifications: 3 Likely benign (GeneDx, Ambry Genetics, CeGaT) and 2 Uncertain significance (Quest Diagnostics, Labcorp/Invitae); 0 expert-panel submissions; no pathogenic or likely pathogenic submission.
• ClinVar search 'POLD1[gene] AND Glu63Lys' returned exactly 1 record (variation 407981), i.e. no alternative nucleotide change producing p.Glu63Lys has been submitted.
• Codon-level reasoning: residue 63 is encoded by GAR (Glu); p.Glu63Lys requires AAR (Lys), achievable from GAR only by the single-nucleotide substitution c.187G>A, the variant under assessment.

Still missing

• No same-amino-acid-change comparator exists to evaluate: the single ClinVar p.Glu63Lys record is this variant itself, classified Likely benign/Uncertain significance without expert-panel review.

Sources

Reasoning

PS3 requires a well-established functional assay (in vivo or in vitro) demonstrating a damaging effect on POLD1 protein function (ACMG/AMP 2015, PMID:25741868). NM_002691.4:c.187G>A is a missense substitution, NP_002682.2:p.(Glu63Lys), and no functional assay of any kind has been performed or reported for it. The variant-level OncoKB record returned no variant-specific functional evidence ('Unknown Oncogenic Effect', not_reviewed = true, empty biological effect, no associated PMIDs), so no assay result exists to calibrate for strength (supporting/moderate/strong). Neither retrieved full-text paper mentions POLD1 or the variant, and neither reports a functional assay: PMID:26467025 is a methodology report on a rules-based scoring system in 338 non-POLD1 genes, and PMID:25394175 is a cancer-predisposition referral guideline. No somatic or functional record for this change was found in COSMIC. The in-silico results available for this variant (REVEL 0.018, BayesDel -0.510638, SpliceAI max delta 0.008) are computational predictions belonging to the PP3/BP4 evidence type and are not functional assay evidence, so they cannot support PS3 at any strength. Residue Glu63 lies N-terminal to and outside the POLD1 exonuclease/proofreading domain (reported as residues 268-471), so even the published proofreading-domain assays used for other POLD1 variants would not cover this residue. PS3 is therefore not met for lack of any qualifying assay data, not because function was tested and found normal.

Evidence considered

• PS3 (ACMG/AMP 2015, PMID:25741868) requires well-established functional studies showing a damaging effect; no such study exists for POLD1 c.187G>A p.(Glu63Lys).
• OncoKB variant-level page for POLD1 E63K reported no variant-specific reviewed functional evidence: classification 'Unknown Oncogenic Effect', is_variant_specific = false, not_reviewed = true, empty biological_effect and no citations_pmids.
• The two retrieved full texts (PMID:26467025, a rules-based scoring methodology study in 338 genes; PMID:25394175, an ACMG/NSGC cancer-predisposition referral guideline) do not mention POLD1 or c.187G>A/p.(Glu63Lys), and neither reports a functional assay.
• A COSMIC search for POLD1 c.187G>A returned no record ('no result'), providing no somatic or functional evidence for this change.
• In-silico outputs available for this variant (REVEL 0.018, BayesDel -0.510638, SpliceAI max delta 0.008) are computational predictor scores that inform PP3/BP4, not PS3/BS3 functional evidence.
• Residue Glu63 is N-terminal to the POLD1 exonuclease/proofreading domain (described as amino acids 268-471 in the germline POLD1/P0LE literature retrieved for gene-level context, PMID:41673532), so published proofreading-domain functional assays would not have tested this residue.
• No ClinGen VCEP/CSPEC specification or local gene-specific framework is available for POLD1 (cspec found = false; framework_mode = generic_acmg), so generic ACMG/AMP 2015 criteria definitions were applied.

Still missing

• No functional assay (polymerase fidelity/proofreading, protein stability, or cell-based readout) has been performed or reported for POLD1 p.Glu63Lys; PS3 cannot be upgraded without a validated assay with appropriate controls for this variant.
• No assay result exists at this residue, which lies outside the POLD1 exonuclease/proofreading domain, so calibration of a hypothetical assay would additionally require establishing biological relevance.

Sources

Reasoning

PS4 requires statistically significant enrichment of the variant in affected individuals versus controls (or a well-powered case series). No POLD1-specific CSPEC/VCEP framework exists, so the generic ACMG/AMP definition applies. No affected-versus-control counts, odds ratio, confidence interval, or p-value are available for NM_002691.4:c.187G>A from any source: the ClinVar record (variation ID 407981) contains five single-submitter laboratory classifications with no case data, the variant has no reported somatic associations (COSMIC), and neither full-text paper reviewed (the ACMG/NSGC cancer-referral practice guideline; the Karbassi standardized scoring-system paper) mentions POLD1 or this variant. Conversely, gnomAD v2.1 non-cancer data show the allele at 6.746e-05 (13/192,696 alleles, 0 homozygotes), i.e. it is observed in population reference individuals rather than enriched in cases. With no enrichment statistic reported anywhere, there is no numeric value to compare against any threshold, and the criterion's required direction (excess in cases) is not supported. PS4 is therefore not met; this is absence of case-control evidence, not a demonstrated depletion.

Rule

Generic ACMG/AMP PS4: the prevalence of the variant in affected individuals is significantly increased compared with the prevalence in controls; no applicable POLD1-specific VCEP/CSPEC rule was retrieved. Enrichment statistics strengthen as the value rises (met requires value >= threshold), but no value was reported in this case to compare.

Evidence considered

• ClinVar variation ID 407981 (NM_002691.4(POLD1):c.187G>A, p.Glu63Lys) holds five germline submissions from clinical laboratories (three Likely benign, two Uncertain significance) with no affected-case counts, odds ratio, or enrichment statistic.
• gnomAD v2.1 non-cancer exomes report the variant at total AF 6.746e-05 (13/192,696 alleles; 0 homozygotes), showing it occurs in reference population samples rather than being concentrated in cases.
• COSMIC reports no previous somatic cancer occurrence for this variant, so there is no tumour-case enrichment signal either.
• PMID:25394175 (ACMG/NSGC cancer-predisposition referral practice guideline) does not mention POLD1 or c.187G>A; it provides only general referral context and cannot support PS4.
• PMID:26467025 (Karbassi standardized variant scoring system) does not mention POLD1 or c.187G>A; it is a methodology paper on evidence weighting and contains no case data for this variant.
• ClinVar-cited PMIDs attached to the submissions (26467025, 28492532) are scoring-system/methodology references rather than primary case reports of this variant.
• No numeric enrichment threshold from a governing specification applies because no POLD1-specific VCEP/CSPEC framework was found and no odds ratio was reported to compare.

Still missing

• No case-control cohort, odds ratio, confidence interval, or p-value specific to NM_002691.4:c.187G>A was available.
• No proband-level case series with this variant was available to assess enrichment.

Sources

Reasoning

PM1 requires the variant to lie in a mutational hotspot and/or a critical, well-established functional domain without benign variation. No governing specification exists for POLD1: no ClinGen CSPEC/VCEP framework was retrieved for this gene and the gene's domain_tables entry is empty, so there is no approved critical-domain list to check the residue against and the call rests on the gene's annotated domains and on hotspot data. POLD1's functionally critical regions are the 3'-5' exonuclease (proofreading) domain, residues 243-477 (Pfam PF03104; UniProt P28340), the DNA polymerase family B domain, residues 541-973 (Pfam PF00136), the N-terminal domain PF24055 at residues 143-220, and the nuclear localization signal at residues 4-19, with residues 1-34 disordered. p.Glu63Lys alters residue 63, which lies outside every one of these regions, in the N-terminal segment between the nuclear localization signal and the first folded domain. This is consistent with the residue-level databases: Cancer Hotspots returns no hotspot row for POLD1 codon 63 (query 'POLD1 E63'), the variant is absent from COSMIC, and the ClinVar codon-63 variants are confined to Uncertain significance/Likely benign classifications. PM1 is not met.

Rule

PM1: variant in a mutational hotspot or in a critical, well-established functional domain without benign variation

Evidence considered

• OncoKB POLD1 curation localizes the disease-relevant alterations to the exonuclease/proofreading domain: '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 'Alterations within the exonuclease domains of POLD1 and POLE... result in the accumulation of single nucleotide variants and an ultra-mutated phenotype'.
• Cancer Hotspots lookup for 'POLD1 E63' returned no hotspot row ('This variant does not lie in a statistically significant hotspot').
• COSMIC search for POLD1 c.187G>A returned no result, i.e. the variant is not a recurrent somatic change.
• Pfam/InterPro annotation retrieved for POLD1 (UniProt P28340) during this assessment: PF03104 3'-5' exonuclease domain residues 243-477; PF00136 DNA polymerase family B residues 541-973; PF24055 N-terminal domain residues 143-220; UniProt features nuclear localization signal 4-19 and disordered region 1-34. Residue 63 falls outside all of these.

Still missing

• No ClinGen CSPEC/VCEP or local domain table exists for POLD1 (domain_tables is empty), so no VCEP-approved critical-domain list could be checked; domain boundaries used here come from Pfam/InterPro annotation of UniProt P28340.
• Hotspot evidence is limited to Cancer Hotspots (somatic-focused) and COSMIC; no germline-specific POLD1 hotspot dataset was available.

Sources

Reasoning

PM2 is applied at supporting strength for a variant at or below an allele frequency of 1e-4 (ClinGen SVI 2020 recommendation to downgrade PM2 to supporting; Richards et al. 2015, PMID:25741868). The aggregate frequencies are below that cutoff: gnomAD v4.1 total AF 3.152e-05 (49/1,554,692), gnomAD v2.1 total AF 6.746e-05 (13/192,696), v2.1 non-cancer exomes 7.055e-05 (11/155,924) and v3.1 non-cancer genomes 8.785e-05 (13/147,980). However, PM2 is evaluated against the population-maximum (ancestry-specific) frequency rather than the pooled total, so that a variant is not declared rare merely because it is uncommon in the numerically dominant reference group. On that basis the variant exceeds the cutoff: the Admixed American ancestry frequency is 4.296e-04 (22/51,216) in gnomAD v4.1 and 4.262e-04 (11/25,808) in v2.1, and the population-maximum filtering allele frequency is 2.901e-04 (v4.1) / 2.173e-04 (v2.1) / 3.068e-04 (v3.1 non-cancer genomes). These values exceed 1e-4 by roughly three- to four-fold, and the variant is observed on 22 independent alleles in the v4.1 Admixed American cohort alone. The variant is therefore present at above-threshold frequency in a specific ancestry group and is not absent or ultra-rare in the general population, so PM2 is not met. This is flagged for human review because the aggregate frequency (3.15e-05) falls below 1e-4 while the ancestry-specific population-maximum frequency exceeds it; the call is made on the ancestry-specific value, consistent with standard practice of using the population-maximum filtering frequency, and would invert only if an aggregate-frequency interpretation were adopted. The gnomAD-Canada v1.0 lookup returned HTTP 500, so no Canadian HostSeq genome frequencies were available; that failure does not change the conclusion because above-threshold frequency is already documented within gnomAD itself.

Rule

PM2: population-maximum allele frequency <= 0.0001, supporting strength (PMID:25741868)

Evidence considered

• gnomAD v4.1 all-comers: total AF 3.152e-05 (49/1,554,692 alleles); Admixed American AF 4.296e-04 (22/51,216); grpmax FAF 2.901e-04; homozygotes 0.
• gnomAD v2.1 all-comers: total AF 6.746e-05 (13/192,696 alleles); Admixed American AF 4.262e-04 (11/25,808); grpmax FAF 2.173e-04; homozygotes 0.
• gnomAD v2.1 non-cancer exomes (AF 7.055e-05, 11/155,924) and gnomAD v3.1 non-cancer genomes (AF 8.785e-05, 13/147,980, grpmax FAF 3.068e-04) were reviewed as the non-cancer population sources; both remain above the 1e-4 cutoff on a population-maximum basis.
• gnomAD-Canada v1.0 query for chr19:50399038 G>A returned HTTP 500 (Internal Server Error) and no frequency data; the report of the variant as 'not available' in that cohort reflects the failed request, not a confirmed absence, and was therefore not used as evidence of rarity.
• Threshold applied: PM2 supporting at population frequency <=1e-4 (ClinGen SVI recommendation to apply PM2 at supporting strength; Richards et al. 2015, PMID:25741868). No POLD1 CSPEC/VCEP or local gene-specific framework exists (framework_mode=generic_acmg), so the generic threshold applies.
• Ancestry consideration: the highest observed frequency is consistently in the Admixed American group (gnomAD v2.1 and v4.1), which is above the cutoff and is the basis for not meeting PM2.

Still missing

• gnomAD-Canada v1.0 returned HTTP 500 (server error), so its frequencies could not be assessed; this is a failed lookup rather than a confirmed absence and adds no supporting evidence for rarity.
• No POLD1-specific PM2 frequency threshold exists (no CSPEC/VCEP, no local gene framework); the generic 1e-4 supporting-strength cutoff was applied.
• The PM2 determination depends on using the ancestry-specific population-maximum frequency rather than the pooled aggregate frequency; a reviewer preferring the aggregate value would reach the opposite result.

Sources

Reasoning

PM3 (Richards et al. 2015, ACMG/AMP, PMID:25741868) is met only when the variant is detected in trans with a pathogenic variant in a recessive disorder; both elements are absent here. First, no second POLD1 allele is reported anywhere in this case: ClinVar VCV000407981 (POLD1 NM_002691.4:c.187G>A, p.Glu63Lys) carries five germline submissions (3 Likely benign, 2 Uncertain significance; 'criteria provided, single submitter' for each, 1 star, no expert-panel submission) and none of them reports a co-occurring POLD1 variant, zygosity, or phase. Second, the recessive precondition is not satisfied for this gene: the germline disease context for POLD1 is dominantly inherited, with the OncoKB POLD1 gene record stating that germline heterozygous loss-of-function variants in the exonuclease domain cause polyposis and predispose to colorectal, endometrial and possibly brain cancer, and no recessive POLD1 disease is documented in any source assembled for this case (the pvs1_gene_context 'inheritance' field is empty and its only caution is that no official CSPEC/VCEP or local framework exists). No VCEP or CSPEC specification governs POLD1 in this case (vcep_materials.json: 'no cspec url available', empty summaries and files; final_classification_framework found=false, source 'generic_acmg_fallback', precedence_rank 3), so the generic ACMG/AMP 2015 criterion applies as written and is not met. Population data are concordant with the absence of a biallelic genotype: gnomAD v4.1 reports 49/1,554,692 alleles (AF 3.15175e-05) and gnomAD v2.1 reports 13/192,696 alleles (AF 6.74638e-05), with zero homozygotes in either dataset, so no carrier of two POLD1 alleles is observed. The two full texts retrieved for this case were checked for trans/cis or co-occurrence evidence and contain neither POLD1 nor c.187G>A (PMID:26467025, a generic variant-scoring methodology paper; PMID:25394175, an ACMG/NSGC cancer-referral practice guideline), so they supply no allelic evidence either. In the absence of any pathogenic variant in trans with c.187G>A—and with the trans construct itself not applicable to this dominantly inherited disease context—PM3 is not met.

Rule

ACMG/AMP 2015 PM3 (Richards et al. 2015, PMID:25741868): detected in trans with a pathogenic variant, recessive disorders; no POLD1 VCEP/CSPEC specification available (generic fallback, framework_complete=false) — trans partner absent and dominant gene-disease context

Evidence considered

• ClinVar VCV000407981 / ClinVarID 407981 (NM_002691.4(POLD1):c.187G>A, p.Glu63Lys): 5 germline submissions (Quest Diagnostics Uncertain significance 2017; Ambry Genetics Likely benign 2023; Labcorp/Invitae Uncertain significance 2026; GeneDx Likely benign 2019; CeGaT Tuebingen Likely benign 2023), API classification 'Conflicting classifications of pathogenicity', review status 'criteria provided, conflicting classifications', 1 star, expert_panels empty. Submitted conditions include 'Colorectal cancer, susceptibility to, 10' and 'Hereditary cancer-predisposing syndrome'. No submission reports a second POLD1 variant, zygosity, or cis/trans configuration.
• No governing gene-specific framework for POLD1: vcep_materials.json states 'no cspec url available' with empty summaries/files_found/user_files and gene_vcep_dir null; final_classification_framework found=false, source 'generic_acmg_fallback', precedence_rank 3, scoring_model 'ACMG/AMP 2015'; case_summary.json records framework_mode 'generic_acmg' and cspec.found=false. Generic ACMG/AMP 2015 definitions therefore apply, in which PM3 requires a pathogenic variant in trans in a recessive disorder and BP2 requires observation in cis (any inheritance pattern) or in trans (fully penetrant dominant) with a pathogenic variant.
• POLD1 germline disease is dominantly inherited: the OncoKB POLD1 gene record states that germline heterozygous loss-of-function mutations in the exonuclease domain cause polyposis and predispose individuals to colorectal, endometrial and possibly brain cancers (p125 catalytic subunit; exonuclease-domain alterations accumulate single-nucleotide variants with an ultra-mutated phenotype). No autosomal recessive POLD1 disease is documented in any source assembled for this case.
• Population allele data show heterozygous carriers only, with zero homozygotes: gnomAD v4.1 AF 3.15175e-05 (49/1,554,692 alleles; highest in Admixed American AF 0.000429553, 22/51,216; grpmax FAF 0.00029005; homozygotes 0) and gnomAD v2.1 AF 6.74638e-05 (13/192,696 alleles; Admixed American AF 0.000426224, 11/25,808; grpmax FAF 0.00021734; homozygotes 0). No biallelic carrier is reported in either dataset, so population data provide no support for a recessive biallelic mechanism.
• Both full texts retrieved for this case were checked for this variant and for allelic/co-occurrence content and contain neither: paper_extracts.json records variant_mentioned=false for PMID:26467025 (Karbassi et al., a rules-based variant-scoring methodology applied to 11,894 variants in 338 genes, which weights co-occurrence and segregation generically but presents no POLD1 or c.187G>A data) and for PMID:25394175 (ACMG/NSGC cancer-predisposition referral practice guideline; POLD1 is not mentioned). Neither provides phase, cis/trans, or co-occurring-variant evidence.
• The only POLD1 allele-phase observation in the case literature concerns a different POLD1 variant and cannot be transferred to c.187G>A: a report of two teenage siblings with multiple adenomas and colorectal cancer carrying a maternally inherited heterozygous PMS2 exon 12 deletion together with a paternally inherited POLD1 variant NP_002682.2:p.Asp316Asn (digenic MMR-plus-polymerase-proofreading deficiency). That variant is Asp316Asn, not Glu63Lys, and the trans relationship is between POLD1 and PMS2 rather than between two POLD1 alleles, so it is not evidence for or against PM3 for this variant.
• Case-wide check for sample-level allelic data found none: case_summary.json 'compact_evidence' contains only extracted/normalization fields, ClinVar, gnomAD, SpliceAI (max delta 0.008), REVEL (0.018), BayesDel (-0.510638), HCI prior (not found), OncoKB, COSMIC, hotspots, a 3-PMID literature packet and gene/PVS1/PM5 scaffolds; no phenotype, zygosity, proband, family, segregation or phase sections exist, and pvs1_gene_context repeats an empty 'inheritance' list.

Still missing

• No proband genotype, zygosity, trio, or phasing data are present anywhere in this case, so whether c.187G>A sits in trans with any other POLD1 allele could not be directly observed; only variant-level resources (normalization, ClinVar, gnomAD, predictors, gene context) were available.
• No CSPEC/VCEP specification or local gene-specific framework exists for POLD1 (cspec.found=false; vcep_materials.json reports no cspec url and no gene files; final_classification_framework found=false), so PM3 was evaluated against generic ACMG/AMP 2015 semantics.
• No recessive POLD1 disease association could be confirmed in the assembled sources and the pvs1_gene_context 'inheritance' field is empty; a gene-disease inheritance confirmation from a curated source (e.g. ClinGen gene-disease validity, OMIM) would harden the dominance premise.

Sources

Reasoning

PM5 requires a different missense change at the same codon (Glu63) to have been established as pathogenic. ClinVar contains three alternative codon-63 missense changes: c.187G>C p.(Glu63Gln) (variation 4141016, Uncertain significance), c.188A>C p.(Glu63Ala) (variation 1525500, Uncertain significance) and c.189G>T p.(Glu63Asp) (variation 584986, 4 submissions - 3 Uncertain significance and 1 Likely benign, conflicting classifications). None is classified pathogenic or likely pathogenic, and none carries expert-panel review, so no codon-63 comparator meets the criterion's requirement. The case's own PM5 comparator harvest independently returned no candidates (eligible_for_classic_pm5_search = false, total_candidates_seen = 0). PM5 is not met.

Rule

PM5: different missense change at the same codon established as pathogenic

Evidence considered

• ClinVar codon-63 missense variants (search 'POLD1[gene] AND E63'): c.187G>C p.(Glu63Gln) = Uncertain significance (single submitter); c.188A>C p.(Glu63Ala) = Uncertain significance (single submitter); c.189G>T p.(Glu63Asp) = conflicting classifications from 4 submitters (3 Uncertain significance, 1 Likely benign, no pathogenic/likely pathogenic).
• PM5 comparator harvest for this case: eligible_for_classic_pm5_search = false, pm5_mode = 'unknown', total_candidates_seen = 0, same_residue_candidates = 0.

Still missing

• No pathogenic/likely pathogenic same-codon comparator exists; the strongest same-residue evidence available is a single Likely benign submission for p.Glu63Asp.

Sources

Reasoning

PP2 has two arms: missense variation must be a common disease mechanism in the gene, and the gene must have a low rate of benign missense variation. POLD1 clearly satisfies the first arm - germline missense changes in the exonuclease/proofreading domain are the established cause of polymerase proofreading-associated polyposis, and OncoKB characterizes POLD1 disease as proofreading-domain loss of function. The second arm is not supported: gnomAD gene-level constraint for POLD1 (ENSG00000062822) gives mis_z = 2.75 and observed/expected missense ratio 0.84 with a 90% interval of 0.80-0.88 (gnomAD; mis_z as defined by Lek et al. 2016, PMID:27535533), i.e. near the neutral expectation rather than the depletion expected of a gene with a low rate of benign missense variation. With no gene-specific PP2 specification available for POLD1 (no ClinGen CSPEC/VCEP, no local gene framework), the second arm cannot be established and PP2 is not met. Because arm one is satisfied and the missense-constraint metric is only moderately below the usual significance range, this call is flagged for human review.

Rule

PP2: missense variant in a gene with a low rate of benign missense variation and in which missense variants are a common mechanism of disease

Evidence considered

• gnomAD gene-level constraint for POLD1 (ENSG00000062822): mis_z = 2.75, o/e missense = 0.838 (90% interval 0.802-0.877), pLI ~ 0, o/e LoF = 0.609; the registry's gnomAD entries index this variant's records in the same gnomAD datasets.
• mis_z is the missense constraint metric of Lek et al. 2016 (Nature, PMID:27535533); a value of 2.75 is not in the range used to define significant missense depletion.
• OncoKB POLD1 gene-level curation: missense alterations in the exonuclease/proofreading domain are the established germline disease mechanism, so missense is a common mechanism of disease in this gene.

Still missing

• Arm one of PP2 is satisfied (exonuclease-domain missense is an established POLD1 disease mechanism) but arm two (low rate of benign missense variation) rests only on gene-level gnomAD constraint; no POLD1-specific PP2 specification or benign-missense-rate analysis is available.

Sources

Reasoning

Scope first: NM_002691.4:c.187G>A is a coding missense substitution in POLD1 predicted to yield NP_002682.2:p.(Glu63Lys) (consequence_class = missense, no canonical splice-consensus involvement), so PP3 must be taken from the missense protein-impact path, i.e. REVEL, and not from any splice prediction. No POLD1 ClinGen CSPEC/VCEP or gene-specific PP3/BP4 lookup table exists for this case (cspec found = false; final_classification_framework found = false, source = generic_acmg_fallback), so the generic calibrated thresholds are the governing specification. The local REVEL v1.3 lookup returns 0.018 for chr19:50399038 G>A (same GRCh38 position as this variant). That is far below the ClinGen SVI REVEL supporting PP3 cutoff of >= 0.644 (Pejaver et al. 2022, PMID:36413997), so no computational evidence of a damaging missense effect exists and PP3 is not met. The SpliceAI result in the case data (max delta 0.008) was deliberately not used: for a missense variant the splice path is not applicable to PP3, so a clean splice delta can never be converted into PP3 support or its absence. BayesDel (-0.510638) is present but has no established generic ACMG-strength calibration, so it is treated as not_available for PP3 rather than being assigned a strength.

Rule

PP3 not met: missense/REVEL path; REVEL 0.018 < 0.644 supporting cutoff (Pejaver et al. 2022, PMID:36413997)

Evidence considered

• Variant consequence is missense (NP_002682.2:p.(Glu63Lys)), so PP3/BP4 are drawn from the REVEL missense path only; the SpliceAI splice path is not applicable to this variant class.
• Local REVEL v1.3 lookup (revel source registry entry, file revel-v1.3_chrom_19_049589733_053770083.csv.zip) returned a score of 0.018 for chr19:50399038 G>A.
• Applied threshold: PP3 supporting >= 0.644 (ClinGen SVI REVEL calibration, Pejaver et al. 2022, Am J Hum Genet, PMID:36413997). REVEL 0.018 is below the supporting cutoff, so PP3 is not met.
• SpliceAI max delta 0.008 (SpliceAI lookup, PMID:30661751 cutoffs) was not used as PP3 support or as evidence against PP3, because PP3 for a missense variant is taken only from REVEL.
• BayesDel -0.510638 is available but has no generic ACMG-strength calibration, so it is not_available for PP3 and was not assigned a strength.
• No POLD1-specific CSPEC/VCEP or gene-specific PP3 lookup exists for this case (no cspec url; final_classification_framework found = false), so the generic calibrated thresholds govern.

Sources

Reasoning

PP5 is applied here only to an exact-variant ClinVar expert-panel classification of Pathogenic or Likely pathogenic. The exact variant, NM_002691.4(POLD1):c.187G>A (p.Glu63Lys), is present in ClinVar as variation ID 407981 (VCV000407981), but its expert-panel field is empty and it has zero expert-panel submissions. All five germline submissions are from individual clinical laboratories (Quest Diagnostics, Ambry Genetics, Labcorp/Invitae, GeneDx, CeGaT) with review status "criteria provided, single submitter", and the record-level review status is "criteria provided, conflicting classifications" (1 star), with three Likely benign and two Uncertain significance submissions. The record does not carry a Pathogenic or Likely pathogenic assertion of any kind, expert-panel or otherwise. ClinVar-cited PMIDs (26467025 Karbassi scoring system; 28492532 Sherloc refinement) are methodology papers, not expert-panel variant assertions. The PP5 trigger is therefore absent and ordinary laboratory submissions were not used to satisfy it.

Rule

PP5 applies only to an exact-variant ClinVar expert-panel (or equivalent expert-panel) classification of Pathogenic or Likely pathogenic; ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions do not qualify.

Evidence considered

• ClinVar variation ID 407981 matches NM_002691.4(POLD1):c.187G>A (p.Glu63Lys) exactly (VCV000407981, last updated 2026-06-20, 5 submissions from 5 submitters).
• The record's expert-panel field is empty and expert-panel submissions = 0; review status is "criteria provided, conflicting classifications" (1 star).
• Submitting laboratories are Quest Diagnostics, Ambry Genetics, Labcorp/Invitae, GeneDx and CeGaT, each with review status "criteria provided, single submitter"; classification breakdown is three Likely benign and two Uncertain significance.
• No Pathogenic or Likely pathogenic assertion exists in this record, so the exact-variant expert-panel trigger required for PP5 is absent; no numeric threshold comparison was applicable.
• PMID:26467025 and PMID:28492532, cited by two submissions, are variant-classification methodology papers and do not constitute an expert-panel assertion for this variant.

Sources

Reasoning

BA1 requires a stand-alone benign allele frequency of >=0.05 (Richards et al. 2015, PMID:25741868). NM_002691.4:c.187G>A (POLD1 p.Glu63Lys; GRCh38 chr19:50399038 G>A) is present in gnomAD v2.1 at a total allele frequency of 6.746e-05 (13/192,696 alleles, 0 homozygotes) and in gnomAD v4.1 at 3.152e-05 (49/1,554,692 alleles, 0 homozygotes). The highest frequency in any ancestry group is 4.296e-04 in Admixed American individuals in v4.1 (22/51,216), matching the v2.1 Admixed American figure of 4.262e-04 (11/25,808); the population-maximum filtering allele frequency (grpmax FAF) is 2.901e-04 in v4.1 and 2.173e-04 in v2.1. The non-cancer subsets are concordant (v2.1 non-cancer exomes AF 7.055e-05, 11/155,924; v3.1 non-cancer genomes AF 8.785e-05, 13/147,980). The single highest subpopulation value in any dataset is 1KG:MXL at 0.00794 in gnomAD v3.1 non-cancer, which rests on 1 allele in 126 individuals and is a small-sample estimate that is itself still below 0.01, let alone 0.05. No queried population approaches the 0.05 stand-alone threshold: the top ancestry frequency is roughly 170-fold below it and the aggregate frequency roughly 1,500-fold below it. BA1 is not met. gnomAD-Canada v1.0 could not be queried (HTTP 500), but the variant is already detected at low frequency across gnomAD v2.1 and v4.1, so the failure does not plausibly hide a >=5% frequency.

Rule

BA1: stand-alone benign allele frequency >= 0.05 (PMID:25741868)

Evidence considered

• gnomAD v2.1 all-comers: total AF 6.746e-05 (13/192,696 alleles), homozygotes 0; highest ancestry group Admixed American AF 4.262e-04 (11/25,808), homozygotes 0; grpmax FAF 2.173e-04.
• gnomAD v4.1 all-comers: total AF 3.152e-05 (49/1,554,692 alleles), homozygotes 0; highest ancestry group Admixed American AF 4.296e-04 (22/51,216), homozygotes 0; grpmax FAF 2.901e-04.
• gnomAD v2.1 non-cancer exomes: AF 7.055e-05 (11/155,924 alleles), homozygotes 0. gnomAD v3.1 non-cancer genomes: AF 8.785e-05 (13/147,980 alleles), homozygotes 0; highest subpopulation estimate 1KG:MXL 0.00794 (1/126 alleles), which is below the BS1/BA1 thresholds and is a single-allele small-sample estimate.
• Threshold applied: BA1 stand-alone benign frequency >=0.05 (Richards et al. 2015, PMID:25741868). No ClinGen CSPEC/VCEP and no local gene-specific framework exists for POLD1 (framework_mode=generic_acmg), so the generic ACMG/AMP 2015 threshold applies.
• No population source reaches or approaches the BA1 threshold, including the ancestry-specific maximum and the non-cancer subsets.

Sources

Reasoning

BS1 requires a strong benign allele frequency of >=0.01 (generic ClinGen-SVI default; per-gene derivation preferred, Whiffin et al. 2017, PMID:28518168). NM_002691.4:c.187G>A (POLD1 p.Glu63Lys) has a gnomAD v2.1 total allele frequency of 6.746e-05 (13/192,696, 0 homozygotes) and a gnomAD v4.1 total frequency of 3.152e-05 (49/1,554,692, 0 homozygotes). The highest ancestry-group frequency is 4.296e-04 in Admixed American individuals (v4.1, 22/51,216) and 4.262e-04 (v2.1, 11/25,808); grpmax FAF is 2.901e-04 (v4.1) and 2.173e-04 (v2.1). The non-cancer subsets give the same picture (v2.1 non-cancer exomes AF 7.055e-05; v3.1 non-cancer genomes AF 8.785e-05, grpmax 3.068e-04). The single highest subpopulation estimate observed anywhere is 1KG:MXL 0.00794 (1/126 alleles) in gnomAD v3.1 non-cancer, which is itself below the 0.01 threshold and rests on one allele in 126 individuals, so neither the ancestry maximum nor that small-sample subpopulation estimate reaches BS1. The variant frequency is more than 20-fold below the 0.01 threshold in the most common ancestry group and roughly 300-fold below it at the aggregate level. BS1 is not met.

Rule

BS1: strong benign allele frequency >= 0.01 (PMID:28518168)

Evidence considered

• gnomAD v2.1 all-comers: total AF 6.746e-05 (13/192,696 alleles); Admixed American AF 4.262e-04 (11/25,808); grpmax FAF 2.173e-04; homozygotes 0.
• gnomAD v4.1 all-comers: total AF 3.152e-05 (49/1,554,692 alleles); Admixed American AF 4.296e-04 (22/51,216); grpmax FAF 2.901e-04; homozygotes 0.
• gnomAD v2.1 non-cancer exomes AF 7.055e-05 (11/155,924) and gnomAD v3.1 non-cancer genomes AF 8.785e-05 (13/147,980, grpmax FAF 3.068e-04) were reviewed because non-cancer subsets can be the preferred population source for cancer-predisposition genes; both are below the BS1 threshold.
• Threshold applied: BS1 strong benign frequency >=0.01 (generic ClinGen-SVI default; per-gene derivation preferred per Whiffin et al. 2017, PMID:28518168). No POLD1 CSPEC/VCEP or local gene framework exists, so the generic threshold applies.
• The 1KG:MXL subpopulation value of 0.00794 (1/126 alleles) in gnomAD v3.1 non-cancer was considered and rejected as a threshold-crossing observation because it is a single-allele estimate in a very small sample and is itself below 0.01.

Still missing

• No POLD1-specific BS1 frequency threshold has been derived for this gene (Whiffin et al. 2017 method); the generic 0.01 SVI default was applied in its place.

Sources

Reasoning

BS2 (Richards et al. 2015, PMID:25741868) is met by observation of the variant in a healthy individual in whom full penetrance is expected, classically a homozygote for a recessive condition or an adult for a fully penetrant, early-onset dominant condition. Two independent lines show BS2 is not satisfied here. First, this is not a recessive disease context: germline POLD1 variants in the exonuclease domain cause polymerase proofreading-associated polyposis, an adult-onset, incompletely penetrant autosomal dominant colorectal/endometrial cancer predisposition, so observation in a healthy adult would not constitute BS2 evidence. Second, the empirical homozygote count is zero in every dataset queried: gnomAD v2.1 (0 homozygotes for 13 variant alleles), gnomAD v4.1 (0 homozygotes for 49 variant alleles), gnomAD v2.1 non-cancer exomes (0) and gnomAD v3.1 non-cancer genomes (0). No homozygote, and no documented healthy-individual observation, supports BS2. Note that the absence of homozygotes in gnomAD is not treated as evidence for pathogenicity here (that would be a different criterion); for BS2 the relevant point is simply that the criterion's required observation does not exist.

Rule

BS2: observed in a healthy individual in whom full penetrance is expected (PMID:25741868)

Evidence considered

• gnomAD v2.1 all-comers homozygote count 0 (13 variant alleles); gnomAD v3.1/v4.1 homozygote count 0 (49 variant alleles in v4.1).
• gnomAD v2.1 non-cancer exomes homozygote count 0 (11/155,924 alleles); gnomAD v3.1 non-cancer genomes homozygote count 0 (13/147,980 alleles).
• POLD1 germline disease context: heterozygous loss-of-function variants in the exonuclease domain cause polymerase proofreading-associated polyposis with colorectal, endometrial and possibly brain cancer predisposition, an adult-onset and incompletely penetrant dominant condition (OncoKB gene-level curation; literature-derived gene context).
• Threshold/applicability statement: BS2 requires observation in a healthy individual with full penetrance expected (Richards et al. 2015, PMID:25741868); the disease context here is adult-onset and incompletely penetrant, and no homozygotes are reported in any dataset.

Still missing

• No homozygote or healthy-adult carrier observation exists in the queried population datasets, so BS2 cannot be satisfied; gnomAD-Canada v1.0 was unavailable (HTTP 500) but is not a required source for BS2.

Sources

Reasoning

BS3 requires a well-established functional assay showing no damaging effect on protein function or splicing (ACMG/AMP 2015, PMID:25741868). For POLD1 NM_002691.4:c.187G>A p.(Glu63Lys) no functional assay of any kind was performed or reported, so there is no assay result demonstrating preserved function. The variant-level OncoKB record returned no variant-specific functional evidence (not_reviewed = true, no biological effect stated, no PMIDs), and no functional record was returned by COSMIC. Neither retrieved full text mentions the variant or the gene in a functional-assay context: PMID:26467025 is a scoring-methodology study and PMID:25394175 a referral guideline. This criterion is not met for absence of functional data rather than for presence of data showing normal function; benign computational predictions (REVEL 0.018, BayesDel -0.510638) and the absence of predicted splice impact (SpliceAI max delta 0.008) belong to BP4/BP7/in-silico evidence types and cannot be converted into BS3. ClinVar submissions for this variant were used only to identify the PMID trail and not to infer any functional criterion, and no submission reported a functional assay.

Evidence considered

• BS3 (ACMG/AMP 2015, PMID:25741868) requires well-established functional studies showing no damaging effect; no functional study was reported for this variant.
• OncoKB variant-level page for POLD1 E63K returned no variant-specific functional or assay evidence (not_reviewed = true, is_variant_specific = false, no PMIDs), so it neither supports nor refutes a benign functional effect.
• COSMIC search for POLD1 c.187G>A returned no record, giving no functional or somatic evidence in either direction.
• PMID:26467025 and PMID:25394175 were screened for this variant and gene; neither reports a functional assay, and neither mentions c.187G>A/p.(Glu63Lys).
• REVEL 0.018 and BayesDel -0.510638 (both benign-direction) and SpliceAI max delta 0.008 were noted but are computational, not functional-assay, evidence and therefore do not contribute to BS3.
• No POLD1-specific specification exists (cspec found = false; framework_mode = generic_acmg), so the generic ACMG/AMP 2015 BS3 definition was applied.

Still missing

• No functional assay demonstrating normal POLD1 function (or normal splicing) exists for c.187G>A p.(Glu63Lys); BS3 cannot be met without one.
• Benign in-silico prediction scores for this variant are real evidence but are evaluated under the in-silico (BP4) criterion, not BS3.

Sources

Reasoning

BP1 applies only to a missense variant in a gene for which primarily truncating variants are known to cause disease, i.e. a gene where missense change is not an established mechanism. POLD1 is the opposite case: the established germline disease mechanism is loss of proofreading activity from exonuclease-domain missense substitutions (OncoKB: 'Germline heterozygous loss-of-function mutations in the exonuclease domain of POLD1 cause polyposis...'), and the gene shows no evidence of being haploinsufficiency-driven - gnomAD gives pLI ~ 0, o/e LoF 0.61 (LOEUF 0.73) and mis_z 2.75, inconsistent with truncating variants being the primary disease mechanism. The case's own PVS1 gene-context artifact records loss of function as a supported POLD1 mechanism, but that loss of function is realized through proofreading-domain missense, not through truncation. Because missense variation is an established disease mechanism in POLD1, BP1's premise is not satisfied and the criterion is not met.

Rule

BP1: missense variant in a gene for which primarily truncating variants are known to cause disease

Evidence considered

• OncoKB POLD1 gene-level curation attributes germline disease to proofreading (exonuclease) domain loss of function and states that somatic POLD1 variants are extremely rare, i.e. missense is the established mechanism.
• gnomAD gene-level constraint for POLD1 (ENSG00000062822): pLI ~ 5.4e-15 (effectively 0), o/e LoF = 0.609 (90% upper bound 0.734), mis_z = 2.75 - no evidence that truncating variants drive disease through haploinsufficiency.
• PVS1 gene-context artifact for this case records lof_mechanism_supported = true (loss of function as a POLD1 germline mechanism); this concerns loss of proofreading function, which in POLD1 is conferred by exonuclease-domain missense changes rather than by truncating variants.

Sources

Reasoning

BP2 (Richards et al. 2015, ACMG/AMP, PMID:25741868) is met when the variant is observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder, or observed in cis with a pathogenic variant in any inheritance pattern; each arm requires a co-occurring pathogenic allele whose phase relative to NM_002691.4:c.187G>A has been determined. No such observation exists in any source assembled for this case. ClinVar VCV000407981 lists five germline submissions for POLD1 c.187G>A (3 Likely benign, 2 Uncertain significance; 'criteria provided, single submitter', 1 star, no expert panel) and reports no second POLD1 variant, no cis/trans statement, and no haplotype information; its conditions ('Colorectal cancer, susceptibility to, 10', 'Hereditary cancer-predisposing syndrome') describe dominant cancer-predisposition phenotypes rather than an allelic configuration. The two retrieved full texts were searched for this variant and for allelic/co-occurrence data and contain neither POLD1 nor c.187G>A (PMID:26467025 is a generic variant-scoring methodology; PMID:25394175 is an ACMG/NSGC referral guideline), so neither can supply a cis/trans observation. Population data cannot substitute: gnomAD v4.1 (49/1,554,692 alleles) and gnomAD v2.1 (13/192,696 alleles) report only heterozygous carriers with zero homozygotes and give no phase-resolved haplotypes pairing c.187G>A with a pathogenic POLD1 allele. The single POLD1 phase observation in the case literature involves a different allele (NP_002682.2:p.Asp316Asn inherited paternally in trans with a maternally inherited PMS2 exon 12 deletion in two teenage siblings), so it is not a cis/trans observation for c.187G>A. No POLD1 VCEP/CSPEC specification governs this case (vcep_materials.json: no cspec url, empty files; final_classification_framework found=false, generic_acmg_fallback, precedence_rank 3), so BP2 is applied in its generic ACMG/AMP 2015 form. Because no co-occurring pathogenic variant or phase information is available for this variant, the criterion's requisite observation is absent and BP2 is not met.

Rule

ACMG/AMP 2015 BP2 (Richards et al. 2015, PMID:25741868): observed in trans with a pathogenic variant in a fully penetrant dominant gene/disorder, or in cis with a pathogenic variant in any inheritance pattern; no POLD1 VCEP/CSPEC specification available (generic fallback) — no co-occurring pathogenic allele or phase observation available

Evidence considered

• ClinVar VCV000407981 / ClinVarID 407981 (NM_002691.4(POLD1):c.187G>A, p.Glu63Lys): five germline submissions (3 Likely benign, 2 Uncertain significance), 1 star, review status 'criteria provided, conflicting classifications', no expert-panel submission, conditions 'Colorectal cancer, susceptibility to, 10' / 'Hereditary cancer-predisposing syndrome' / 'not provided' / 'not specified'. No co-occurring POLD1 variant and no cis/trans (phase) statement is present in any submission, so neither the cis arm nor the trans arm of BP2 has an observable trigger.
• POLD1 germline disease is dominantly inherited (OncoKB POLD1 record: germline heterozygous loss-of-function variants in the exonuclease domain cause polyposis with colorectal/endometrial/brain cancer predisposition), which is the disease context in which a trans observation with a pathogenic allele could become BP2-relevant; no such trans observation exists for c.187G>A.
• No governing gene-specific framework for POLD1: vcep_materials.json reports 'no cspec url available' with empty summaries/files_found and gene_vcep_dir null; final_classification_framework found=false, source 'generic_acmg_fallback', precedence_rank 3; case_summary.json records framework_mode 'generic_acmg' and cspec.found=false. BP2 was therefore applied in its generic ACMG/AMP 2015 form (cis with a pathogenic variant in any inheritance pattern, or trans in a fully penetrant dominant disorder).
• Both retrieved full texts were searched for the exact variant and for allelic/co-occurrence content and contain neither: paper_extracts.json records variant_mentioned=false for PMID:26467025 (rules-based variant-scoring methodology over 11,894 variants in 338 genes; no POLD1 and no c.187G>A data) and for PMID:25394175 (ACMG/NSGC cancer-predisposition referral guideline; POLD1 not mentioned). Neither paper reports a phased or co-occurring pathogenic allele for this variant.
• Population data cannot supply the missing observation: gnomAD v4.1 reports 49/1,554,692 alleles (AF 3.15175e-05, homozygotes 0) and gnomAD v2.1 reports 13/192,696 alleles (AF 6.74638e-05, homozygotes 0); only heterozygous carriers appear and no phase-resolved haplotypes pairing c.187G>A with a pathogenic POLD1 allele are available. gnomAD-Canada v1.0 returned no record (HTTP 500), so it contributes nothing.
• The only POLD1 phase observation in the case literature concerns a different allele: two teenage siblings with multiple adenomas and colorectal cancer carried a maternally inherited heterozygous PMS2 exon 12 deletion together with a paternally inherited POLD1 NP_002682.2:p.Asp316Asn, i.e. a digenic MMR/polymerase-proofreading configuration involving a variant other than c.187G>A; it is therefore not a cis/trans observation for this variant.
• Case-wide check for sample-level allelic data found none: case_summary.json 'compact_evidence' holds only variant-level resources (normalization, ClinVar, gnomAD, SpliceAI max delta 0.008, REVEL 0.018, BayesDel -0.510638, HCI prior not found, OncoKB, COSMIC, hotspots, 3-PMID literature packet, PVS1/PM5 scaffolds); no proband, family, zygosity, segregation, or phase sections exist and pvs1_gene_context 'inheritance' is empty.

Still missing

• No proband genotype, zygosity, trio, or phasing data are present in this case, so cis/trans relationships with any co-occurring POLD1 allele could not be observed directly; BP2 can only be re-evaluated if a phased genotype (e.g. long-read or trio phasing) becomes available.
• No CSPEC/VCEP specification or local gene-specific framework exists for POLD1, so BP2 was evaluated against generic ACMG/AMP 2015 semantics; a gene-specific specification could define additional conditions under which BP2 applies.

Sources

Reasoning

BP6 is applied here only to an exact-variant ClinVar expert-panel classification of Benign or Likely benign. The exact variant NM_002691.4(POLD1):c.187G>A (p.Glu63Lys) is present in ClinVar as variation ID 407981 (VCV000407981), but the record carries zero expert-panel submissions and an empty expert-panel field; the record-level review status is "criteria provided, conflicting classifications" (1 star). Its Likely benign label reflects an aggregate of three single-submitter clinical laboratories (Ambry Genetics, GeneDx, CeGaT) against two Uncertain significance submissions (Quest Diagnostics, Labcorp/Invitae). Per the governing rule, this aggregate non-expert label must not trigger BP6, even though it leans benign. The exact-variant expert-panel Benign/Likely benign assertion required for BP6 is therefore absent.

Rule

BP6 applies only to an exact-variant ClinVar expert-panel Benign or Likely benign classification; ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions do not qualify.

Evidence considered

• ClinVar variation ID 407981 matches NM_002691.4(POLD1):c.187G>A (p.Glu63Lys) exactly (VCV000407981, 5 submissions, 1-star review status "criteria provided, conflicting classifications").
• Expert-panel submissions = 0 and the expert-panel field is empty, so there is no expert-panel Benign or Likely benign assertion for this variant.
• The Likely benign signal is an aggregate of three single-submitter clinical laboratories (Ambry Genetics, GeneDx, CeGaT) conflicting with two Uncertain significance submissions (Quest Diagnostics, Labcorp/Invitae); aggregate laboratory labels do not qualify for BP6.
• No numeric threshold comparison was applicable to BP6 in this case.

Sources

Reasoning

PS2 requires that NM_002691.4:c.187G>A (p.Glu63Lys) was observed to arise de novo in an affected individual, that both parental samples were tested and shown by identity testing to be the biological parents, that the family history is consistent with de novo inheritance, and that the phenotype matches the POLD1 gene-disease association with reasonable specificity (Richards et al. 2015, PMID:25741868). No proband-level genotype, inheritance status or phenotype is documented anywhere in this case, and no parental testing exists: the ClinVar record (variation 407981, VCV000407981) comprises five single-submitter germline submissions (three Likely benign, two Uncertain significance, no expert panel), none of which asserts a de novo occurrence, trio testing, or parental genotypes, and neither full text retrieved for this case mentions POLD1 c.187G>A. The criterion therefore cannot be reached at any level of confidence: with no de novo observation at all, PS2 is not assessable rather than refuted. No ClinGen CSPEC or VCEP specification exists for POLD1, so the generic ACMG/AMP 2015 definitions govern.

Evidence considered

• Governing framework: no ClinGen CSPEC and no gene-specific VCEP were retrieved for POLD1 (vcep_materials.json records 'no cspec url available'; final_classification_framework.json returns found=false with source 'generic_acmg_fallback'), so PS2 was evaluated under the generic ACMG/AMP 2015 definitions (Richards et al. 2015, PMID:25741868) referenced by generic_acmg_combination_rules.
• ClinVar variation 407981 for NM_002691.4(POLD1):c.187G>A (p.Glu63Lys) holds five germline, single-submitter submissions - Quest Diagnostics and Labcorp/Invitae as Uncertain significance, GeneDx, Ambry Genetics and CeGaT Tuebingen as Likely benign - with no expert-panel submission and conflicting overall classification. No submission reports a de novo event, a proband phenotype, parental genotypes or parentage confirmation; the only criterion signal extracted from any submission text is PM2 (population frequency), so the ClinVar record contributes no PS2 evidence.
• No proband, trio, parental genotype or de novo observation for c.187G>A was located in any source collected for this case. The two full texts retrieved for this case - PMID:26467025 (a general 7-point variant-scoring methodology study across 338 genes) and PMID:25394175 (an ACMG/NSGC hereditary-cancer referral practice guideline) - neither name POLD1 nor this variant, and neither carries case-level segregation or de novo data.
• A final classification recorded elsewhere (e.g. 'Likely benign' from three clinical laboratories in ClinVar) is not evidence of a de novo event and was not used to infer PS2.

Still missing

• No proband-level genotype, indication for testing, phenotype or inheritance status is documented for the individual in whom POLD1 c.187G>A was observed.
• No parental testing data: neither parental genotypes nor maternity/paternity confirmation (identity testing) have been reported for this variant.
• No de novo occurrence of POLD1 c.187G>A is reported in the ClinVar record (variation 407981, five germline submissions), in PMID:25394175, in PMID:26467025, or in any other source collected for this case.

Sources

Reasoning

PM6 requires that NM_002691.4:c.187G>A (p.Glu63Lys) is assumed to have arisen de novo on clinical grounds, without confirmation that both parental samples are the biological parents, while still requiring a family history consistent with de novo inheritance and reasonable phenotype specificity (Richards et al. 2015, PMID:25741868). Even this weaker, unconfirmed premise cannot be established here: no source in this case reports a de novo occurrence of the variant, nor any parental genotypes (confirmed or assumed). The ClinVar record (variation 407981) contains five single-submitter germline submissions with no de novo assertion and no family data, and neither retrieved full text mentions the variant. With no de novo event documented at any level of confidence, PM6 is not assessable rather than refuted. Generic ACMG/AMP 2015 definitions apply because no POLD1-specific specification exists.

Evidence considered

• Governing framework: no ClinGen CSPEC or gene-specific VCEP was available for POLD1, so PM6 was evaluated under generic ACMG/AMP 2015 definitions (Richards et al. 2015, PMID:25741868), in which PM6 covers a de novo occurrence assumed on clinical grounds where parentage has not been confirmed by identity testing.
• ClinVar variation 407981 for c.187G>A: five germline single-submitter submissions (three Likely benign, two Uncertain significance; no expert panel). None reports de novo inheritance, trio testing or parental samples; the only extracted criterion signal is PM2, so no PM6 premise is present in the ClinVar record.
• Neither retrieved full text (PMID:26467025, PMID:25394175) mentions POLD1 c.187G>A or provides family/de novo data; no other source collected for this case reports the variant in a patient, so there is no clinical basis on which a de novo event could even be assumed.
• The variant is observed in population databases (gnomAD v2.1 total AF 6.75e-05, 13/192,696 alleles; gnomAD v4.1 AF 3.15e-05, 49/1,554,692 alleles, no homozygotes), which is population-frequency context only and provides no de novo information.

Still missing

• No de novo occurrence of POLD1 c.187G>A has been reported in any source, so even the unconfirmed de novo premise required by PM6 cannot be established.
• Parental genotypes are unavailable, so neither confirmed parentage (PS2) nor presumed de novo status (PM6) can be evaluated.
• No clinical phenotype or indication for testing is documented for any carrier of this variant, which is required to judge whether the presentation matches the POLD1 (polymerase proofreading-associated polyposis / colorectal cancer susceptibility) disease association.

Sources

Reasoning

PP1 requires co-segregation of NM_002691.4:c.187G>A (p.Glu63Lys) with disease in multiple affected members of a family, ideally with the number of informative meioses stated (Richards et al. 2015, PMID:25741868). No pedigree, no genotype for any affected or unaffected relative, and no meioses are reported for this variant in any source collected for this case. The ClinVar record (variation 407981) consists of five single-submitter germline submissions that cite no segregation data and include no family structure, and neither full text retrieved for this case names the variant, let alone presents a family study. Because there is no family data at all, co-segregation cannot be scored at any strength, and PP1 is not assessable rather than refuted. Generic ACMG/AMP 2015 definitions apply, as no POLD1-specific specification exists.

Evidence considered

• Governing framework: no ClinGen CSPEC or gene-specific VCEP was available for POLD1, so PP1 was evaluated under generic ACMG/AMP 2015 definitions (Richards et al. 2015, PMID:25741868), where PP1 is supporting pathogenic evidence for co-segregation with disease in multiple affected family members.
• ClinVar variation 407981 for c.187G>A: five germline single-submitter submissions (Quest Diagnostics and Labcorp/Invitae as Uncertain significance; GeneDx, Ambry Genetics and CeGaT Tuebingen as Likely benign), none expert-panel. No submission reports a pedigree, co-segregation or a meiosis count, and the only extracted criterion signal is PM2, so the ClinVar record contributes no segregation evidence.
• Neither retrieved full text (PMID:26467025 - retrospective multi-gene variant-scoring methodology; PMID:25394175 - ACMG/NSGC hereditary-cancer referral guideline) mentions POLD1 c.187G>A, and neither provides family data for it. No numeric segregation threshold could be applied because no segregation data exist for this variant.

Still missing

• No family pedigree and no number of informative meioses are documented for POLD1 c.187G>A.
• No genotype data for affected or unaffected relatives are available, so co-segregation cannot be scored at any strength (supporting or otherwise).
• No publication or database record collected for this case reports this variant within a family or provides segregation data.

Sources

Reasoning

PP4 requires that the patient's phenotype or family history be highly specific for a disease with a single genetic etiology. No POLD1-specific CSPEC/VCEP framework exists, so the generic ACMG/AMP definition applies. This case is a variant-level classification: no individual clinical phenotype, HPO terms, diagnosis, age at onset, tumour spectrum, or family history is present in any available case material. The only disease-level information available is the submitter-assigned ClinVar condition list ("Colorectal cancer, susceptibility to, 10"; "Hereditary cancer-predisposing syndrome"; "not provided"; "not specified"), which reflects laboratory panel/ordering context rather than a documented proband phenotype in whom this variant was found. General POLD1 polymerase-proofreading-associated polyposis and hereditary cancer context is gene-level and cannot substitute for phenotype specificity in the individual case. PP4 is therefore not assessed.

Rule

Generic ACMG/AMP PP4: the patient's phenotype or family history is highly specific for a disease with a single genetic etiology; no applicable POLD1-specific VCEP/CSPEC rule was retrieved.

Evidence considered

• No individual clinical phenotype, HPO annotation, diagnosis, or family history accompanies NM_002691.4:c.187G>A in the available case materials.
• The ClinVar record lists submitter-level conditions ("Colorectal cancer, susceptibility to, 10", "Hereditary cancer-predisposing syndrome", "not provided", "not specified") aggregated across five laboratory submissions; these are order/panel context labels and do not document a specific proband phenotype.
• Gene-level POLD1 polymerase-proofreading-associated polyposis and hereditary cancer context is not phenotype-specific evidence for PP4 in this case.

Still missing

• No patient-level phenotype, disease diagnosis, or family history was provided for this case.
• No phenotype-specificity or likelihood assessment for a POLD1-associated presentation is available.

Sources

Reasoning

BS4 requires evidence that NM_002691.4:c.187G>A (p.Glu63Lys) does not segregate with disease in a family with multiple affected members - typically an affected relative who carries the phenotype but not the variant (Richards et al. 2015, PMID:25741868). No family study, pedigree or relative genotype data exist for this variant. The ClinVar record (variation 407981) reports five single-submitter germline submissions, none of which states a segregation or non-segregation observation, and no source collected for this case reports the variant in a family. Because the non-segregation premise cannot be tested, BS4 is not assessable rather than met or refuted. Notably, the three Likely benign ClinVar submissions are final classifications without accompanying segregation evidence, and a criterion may not be derived from another source's final classification, so they cannot be used to support BS4. Generic ACMG/AMP 2015 definitions apply, as no POLD1-specific specification exists.

Evidence considered

• Governing framework: no ClinGen CSPEC or gene-specific VCEP was available for POLD1, so BS4 was evaluated under generic ACMG/AMP 2015 definitions (Richards et al. 2015, PMID:25741868), where BS4 is strong benign evidence requiring lack of segregation in affected family members.
• ClinVar variation 407981 for c.187G>A: five germline single-submitter submissions with no expert-panel review and conflicting overall classification (three Likely benign, two Uncertain significance). None reports family structure, a pedigree, or the variant status of affected relatives, and none asserts a non-segregation observation, so no BS4 evidence is present in the ClinVar record.
• The three Likely benign submissions (GeneDx SCV000572306, Ambry Genetics SCV003745311, CeGaT Tuebingen SCV004183692) are classifications carrying no segregation evidence and no cited PMIDs; they were not used to infer BS4, because a criterion may not be derived from another source's final classification.
• Neither retrieved full text (PMID:26467025, PMID:25394175) mentions POLD1 c.187G>A or supplies family data, so no non-segregation evidence exists anywhere for this variant.

Still missing

• No family with multiple affected members is documented for POLD1 c.187G>A, which BS4 requires.
• No genotypes for affected relatives are available, so non-segregation (an affected, variant-negative individual) cannot be demonstrated.
• The ClinVar Likely benign submissions (GeneDx, Ambry Genetics, CeGaT Tuebingen) provide no stated segregation or non-segregation data and cannot substitute for family evidence, since a criterion may not be inferred from another source's final classification.

Sources

Reasoning

BP5 requires the variant to be observed in a case in which an alternate molecular basis for the disease was established (for example a pathogenic variant in another gene explaining the proband's phenotype). No POLD1-specific CSPEC/VCEP framework exists, so the generic ACMG/AMP definition applies, and no numeric likelihood-ratio-style threshold is defined in any governing source for this criterion in this case. The available case materials contain no proband, no disease assignment, no co-occurring variant results (POLD1 is on multi-gene hereditary cancer panels), and no statement of an alternative molecular diagnosis; the ClinVar record contains only aggregate laboratory classifications without case-level detail, and the two full-text papers reviewed contain neither this variant nor POLD1. BP5 is therefore not assessed for lack of the required case-level evidence.

Rule

Generic ACMG/AMP BP5: the variant is found in a case with an alternate molecular basis for the disease; no applicable POLD1-specific VCEP/CSPEC rule was retrieved and no numeric threshold for this criterion is defined in the governing source.

Evidence considered

• No proband-level clinical or molecular data (alternate pathogenic variant, second hit, or exclusion of other causes) are present in the available case materials.
• The ClinVar record holds only aggregate laboratory classifications (three Likely benign, two Uncertain significance) with no case-level detail on alternate molecular explanations.
• PMID:25394175 and PMID:26467025 contain neither POLD1 nor c.187G>A and provide no case information for BP5.

Still missing

• No patient-level case with a confirmed alternative molecular explanation (e.g. a pathogenic variant in another hereditary cancer gene) is available.
• No co-occurring variant or second-hit results, and no disease-exclusion information, are available for the proband.