POLE c.4039C>T, p.(Leu1347=)

Reasoning

PM2 is met at Supporting strength under the supplied generic supporting threshold of allele frequency <=0.0001 when the all-comers dataset is the default. gnomAD v4.1 (all-comers) reports an overall frequency of 1.85954e-06 (3/1613304 alleles) with a group-maximum filtering AF of 2.8e-07 and a highest subpopulation frequency of 1.09801e-05 (1/91074, South Asian); gnomAD v2.1 reports the variant as absent (0/249534). All of these values are well below 0.0001, so the variant qualifies as extremely rare/absent in population data. The local León-Castillo et al. 2020 custom POLE framework has no PM2 population-source or threshold specification, so the supplied generic PM2 threshold and the default all-comers gnomAD v2.1/v4.1 datasets were used; the non-cancer subsets are concordant (v2.1 non-cancer exomes 0/235300; v3.1 non-cancer absent), so no substitution changes the call.

Rule

Generic PM2 (Supporting): allele frequency <=0.0001; use all-comers gnomAD by default unless a governing VCEP specifies a different population source; no applicable POLE-specific PM2 rule.

Evidence considered

• gnomAD v4.1 (all-comers): total AF 1.85954e-06 (3/1613304 alleles); group-maximum filtering AF 2.8e-07; highest subpopulation AF 1.09801e-05 (1/91074, South Asian); zero homozygotes.
• gnomAD v2.1 (all-comers, exome): variant absent, 0/249534 alleles, AF 0.
• gnomAD v2.1 non-cancer exomes (0/235300) and gnomAD v3.1 non-cancer genomes (absent) are concordant and equally support PM2, but are not substituted because no governing POLE specification requires a non-cancer or exome-only source.
• The local POLE custom framework has no PM2 population-source or threshold specification, so the supplied generic PM2 threshold of <=0.0001 and the default all-comers gnomAD datasets were used.

Sources

Reasoning

Scope: BP7 applies to this variant, which is a synonymous (silent) substitution, NM_006231.4:c.4039C>T producing NP_006222.2:p.(Leu1347=) (consequence_class 'synonymous'). The variant is not a canonical +/-1 or 2 splice-consensus change (canonical_splice_consensus false) and lies within the coding sequence, so the relevant question is whether splicing is predicted to be affected. The SpliceAI Lookup result for this exact transcript/variant reports no site loss or gain: DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01, i.e. a maximum delta score of 0.014, with Pangolin SG 0.003 and SL -0.017 concordantly near zero, indicating neither loss of a native splice site nor creation of a cryptic one. The published SpliceAI no-impact benchmark of <=0.1 (Jaganathan et al. 2019, Cell, PMID:30661751) is far exceeded by 0.014. No POLE-specific BP7 specification exists in the governing custom framework (which customizes only PM1, PS4, PP3 and BP4), so the generic ACMG/AMP BP7 criterion applies, and its splicing-prediction clause is satisfied. The single SpliceAI prediction is used here and is not double-counted for PP3 or BP4.

Evidence considered

• The normalized consequence is synonymous: NM_006231.4:c.4039C>T produces NP_006222.2:p.(Leu1347=) (protein_1l p.(L1347=)), and the variant-level PVS1 scaffold records consequence_class 'synonymous' with canonical_splice_consensus false.
• SpliceAI Lookup result for NM_006231.4:c.4039C>T (12-133225625-G-A, hg37): DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01; maximum delta score 0.014; Pangolin SG 0.003, SL -0.017 - no native acceptor/donor site lost and no new splice site created.
• Published SpliceAI calibration used verbatim: BP4/no-significant-impact cutoff at max delta <=0.1 (Jaganathan et al. 2019, Cell, PMID:30661751); the observed 0.014 is well below it, applied here as the calibrated splice-impact basis for BP7.
• The governing framework for this case is the local custom POLE framework (Leon-Castillo et al. 2020, custom_internal), which customizes only PM1, PS4, PP3 and BP4 and explicitly falls back to generic ACMG/AMP outside those rules; no POLE-specific BP7 specification exists.
• This single SpliceAI prediction is allocated to BP7 and is not reused for PP3 or BP4.

Still missing

• Nucleotide-level conservation at chr12:132649039G>A (e.g. PhyloP/PhastCons/GERP) was not retrieved in this case, so the generic ACMG/AMP BP7 sub-condition that the nucleotide is not highly conserved could not be independently checked; the verdict rests on the splicing-prediction clause (SpliceAI max delta 0.014, Pangolin near zero).
• MaxEntScan/VarSeak were not retrieved; SpliceAI and Pangolin (two independent algorithms) both predict no splice impact, so the absence of further algorithms cannot create evidence either way.

Sources

Reasoning

PS3 requires well-established in vitro or in vivo functional studies supportive of a damaging effect on POLE. The variant is NM_006231.4:c.4039C>T, NP_006222.2:p.(Leu1347=) - a synonymous substitution that leaves the amino acid unchanged, so the only assay classes that could establish a damaging effect are a splicing/RNA assay (minigene, RT-PCR, RNA-seq) or a protein-function assay. No such experimental study was identified for this variant in any source retrieved for this case: the case evidence bundle contains no functional-assay data source, and the only functional-adjacent result is the SpliceAI lookup, which is an in-silico predictor (max delta 0.014; DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01) and by definition cannot be counted as PS3 evidence. Both retrievable publications are framework/methodology papers with no variant-level data (PMID:25394175, ACMG/NSGC cancer-predisposition referral guideline, variant_mentioned=false; PMID:28492532, Sherloc refinement of ACMG-AMP criteria, variant_specific=false, no full text in this case). The governing gene-specific specification for this case - the local custom POLE framework derived from Leon-Castillo et al. 2020 - customizes only PM1, PS4, PP3 and BP4 and defines no PS3 rule, so the generic ACMG/AMP 2015 PS3 definition applies and is not satisfied. No strength calibration is reachable because there is no assay to calibrate. Because the criterion fails on absence of qualifying experimental evidence rather than on a negative assay result, PS3 is reported not_met rather than not_assessed; the underlying data gap is recorded in gaps_remaining.

Rule

PS3 (ACMG/AMP 2015, generic): well-established in vitro or in vivo functional studies supportive of a damaging effect on the gene or gene product. No gene-specific PS3 specification exists in the governing POLE package.

Evidence considered

• Governing framework check: the case runs in framework_mode 'custom_gene_framework'. The local POLE specification (final_classification_framework, version leon-castillo-2020-custom-framework-v1, labelled 'Leon-Castillo et al. 2020 custom POLE framework', official_clingen=false) declares customized_criteria = [PM1, PS4, PP3, BP4] only and defines no PS3 rule, so generic ACMG/AMP 2015 PS3 was applied for this criterion.
• Variant consequence: NM_006231.4:c.4039C>T -> NP_006222.2:p.(Leu1347=), i.e. synonymous (Mutalyzer protein description p.(=), predicted protein identical to reference). Any damaging-effect functional readout would therefore have to come from a splicing/RNA assay or a protein-function assay; a missense-type in-vitro model does not exist for this allele.
• SpliceAI lookup (source_registry key 'spliceai', NM_006231.4:c.4039C>T): max delta score 0.014 (DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01; Pangolin SG 0.003, SL -0.017). SpliceAI is a computational predictor, not an experimental assay, so it provides no PS3 support regardless of score.
• Paper PMID:25394175 (Hampel et al., ACMG/NSGC practice guideline on referral indications for cancer predisposition assessment): paper_extracts record variant_mentioned=false, direct_quote empty, experimental_context empty, searched_for = [c.4039C>T, NP_006222.2:p.(L1347=), p.L1347=, L1347, c.4039C>T]. No functional assay, no POLE content.
• Paper PMID:28492532 (Nykamp et al., Sherloc: a comprehensive refinement of the ACMG-AMP variant classification criteria): methodology/guidelines review applied to >40,000 clinically observed variants; classified variant_specific=false, paper_type guidelines_review. No full text is available in this case and no variant-level functional data is reported.
• Other retrieved sources were checked for any functional assay on this variant and report none: ClinVar variation 699633 (Likely benign, 2 submitters; no functional-assay evidence statement), OncoKB POLE L1347= (Unknown Oncogenic Effect; is_variant_specific=false, no citations), CancerHotspots POLE L1347 (no hotspot row found), COSMIC POLE c.4039C>T (no result). None of these is an experimental functional study, and none was used as functional evidence here.

Still missing

• No experimental functional assay (RNA/minigene/RT-PCR splicing assay or POLE protein-function assay) has been performed or reported for POLE c.4039C>T; the case evidence bundle contains no functional-assay source type, and no functional study was retrievable from the literature pass.
• A targeted search of the functional/splicing literature for POLE c.4039C>T would be required before PS3 could be scored at any strength; the two PubMed records retrieved here (PMID:25394175, PMID:28492532) are ClinVar-derived framework papers with no variant-level content.
• Patient RNA (e.g. blood or tumour RNA) is not available in this case, so a splicing-outcome assay cannot be substituted by data already on hand.

Sources

Reasoning

Governing specification: the León-Castillo et al. 2020 custom POLE framework (local gene-specific framework, not an official ClinGen CSPEC/VCEP) customizes PS4 and caps it at Supporting. Its PS4_Supporting rule reads, verbatim: 'Apply PS4_Supporting only when the exact POLE missense variant is recurrent in both the COSMIC and TCGA endometrial carcinoma cohorts in Supplementary Table S1, has a combined EC count >=10, and belongs to the established/pathogenic set highlighted by León-Castillo et al. 2020. Under the current package this supports p.P286R, p.V411L, p.A456P, and p.S297F. Do not upgrade above Supporting.' The exact comparison operator is '>=' applied to the combined EC recurrence count. This variant is NM_006231.4:c.4039C>T, NP_006222.2:p.(Leu1347=) — a synonymous (silent) substitution, not a missense variant, and codon 1347 lies far outside the POLE exonuclease domain (approximately amino acids 268-471; the paper's hotspots are P286R, V411L, S297F, A456P, S459F). Searches of the full text and Supplementary Table S1 for every notation of this variant ('c.4039C>T', 'p.Leu1347=', 'p.L1347=', 'L1347', '4039', '1347') returned zero hits: the variant is absent from Table S1. Its combined EC recurrence count is therefore 0. Substituting into the governing inequality: 0 >= 10? No — the rule is not satisfied. No other PS4-eligible evidence exists: COSMIC reports no somatic record for this variant, cancerhotspots.org shows no significant hotspot, and no germline case-control enrichment study is available for POLE c.4039C>T. The framework itself notes the underlying evidence is somatic EC recurrence rather than germline case-control enrichment, which is why PS4 is capped at Supporting even for qualifying variants. With the variant non-missense, outside the exonuclease domain, and absent from the recurrence table, PS4 is not met at any strength.

Rule

León-Castillo et al. 2020 custom POLE framework — PS4_Supporting: exact POLE missense variant recurrent in both COSMIC and TCGA EC cohorts (Supplementary Table S1) with combined EC count >=10 and in the established pathogenic set (P286R, V411L, A456P, S297F); capped at Supporting.

Evidence considered

• Searched PATH-250-323-s002.xlsx (Supplementary Table S1, 'POLE mutations reported in EC in COSMIC or TCGA') for c.4039C>T, p.Leu1347=, p.L1347=, L1347, 1347 and 4039: zero matches. The table lists 102 recurrent POLE variants with exonuclease-domain hotspots at the top (p.P286R c.857C>G, total 181 EC; p.V411L c.1231G>T, total 114); p.Leu1347= is not present, so its combined COSMIC+TCGA EC recurrence count is 0.
• Searched PATH-250-323.pdf full text for the same variant notations: zero matches. The paper discusses only the five established hotspot missense substitutions (P286R, V411L, S297F, A456P, S459F) and their somatic/ultramutated phenotype; it contains no reference to c.4039C>T or p.Leu1347=.
• The variant is synonymous (p.Leu1347=), whereas the governing PS4_Supporting rule applies only to missense variants; it is therefore also outside the rule's scope.
• Codon 1347 lies well outside the POLE exonuclease domain (approx. aa 268-471) in which the paper's pathogenic variants reside.
• COSMIC evidence in this case reports 'This variant has not previously been reported in somatic cancers (COSMIC)', and cancerhotspots.org reports no statistically significant hotspot at this residue.
• Governing operator quoted verbatim from the framework: combined EC count >=10; substitution 0 >= 10 is false, so PS4 is not met.

Still missing

• PATH-250-323-s001.tif (declared by the gene index for PM1/PS4) could not be converted to text and was not visually reviewed; it is a supplementary figure, and no extractable variant-level recurrence or case-control data is expected from it.
• No germline case-control enrichment dataset (affected vs. control allele frequencies) exists for POLE c.4039C>T, so PS4 cannot be evaluated under a generic ACMG prevalence framework either.

Sources

Reasoning

The governing local POLE framework derived from Leon-Castillo et al. 2020 implements PM1 as an exact-substitution rule: PM1_Strong for the five established exonuclease-domain hotspot missense substitutions emphasised by the paper (p.P286R, p.V411L, p.S297F, p.A456P, p.S459F); PM1_Moderate for the exact non-hotspot exonuclease-domain missense substitutions the paper classifies as pathogenic (p.F367S, p.L424I, p.M295R, p.P436R, p.M444K, p.D368Y); and PM1_Supporting for exact recurrent exonuclease-domain missense substitutions placed in the POLE-score 3/uncertain tier (p.A465V, p.L424V, p.T278M, p.A428T). None of these enumerated substitutions is p.Leu1347=; no rule in the framework applies to a synonymous change, and the framework explicitly warns that the Moderate tier applies only to those exact substitutions 'not all substitutions at the same codon'. Independently of the exact-substitution lists, residue 1347 lies well distal to the N-terminal exonuclease domain that contains all five hotspot residues (286-459); the paper states that >95% of POLE mutations outside the exonuclease domain are classified as non-pathogenic by POLE-score, and that only 1/23 (4%) tumours with a non-exonuclease-domain mutation showed the ultramutated genomic alterations. CancerHotspots (queried POLE L1347) reported no statistically significant hotspot at this residue. The case's vcep_materials.json domain_tables key is empty ({}), so no authoritative, VCEP-approved critical-domain list exists for POLE to check the residue against; the PM1 assessment therefore rests on the framework's enumerated substitutions, the exonuclease-domain localisation, and the hotspot databases, all of which are negative. p.Leu1347= also does not appear in the VCEP supplementary tables: grep for c.4039C>T / p.Leu1347= / p.L1347= / 1347 across PATH-250-323.pdf.txt and PATH-250-323-s002.xlsx.txt returned no entry. PM1 is therefore not met.

Evidence considered

• The governing framework's PM1 rules were read in full from the 'final_classification_framework' payload: PM1_Strong is limited to p.P286R, p.V411L, p.S297F, p.A456P, p.S459F; PM1_Moderate to p.F367S, p.L424I, p.M295R, p.P436R, p.M444K, p.D368Y ('use this only for these exact substitutions, not all substitutions at the same codon'); PM1_Supporting to p.A465V, p.L424V, p.T278M, p.A428T. p.Leu1347= matches none of them.
• Leon-Castillo et al. 2020 (PATH-250-323.pdf), Results: 'The 59 exonuclease domain mutations comprised 21 unique variants, the five most common of which (P286R, 21 cases; V411L, 13 cases; S297F, 3 cases; A456P, 2 cases; and S459F, 2 cases) were classified as pathogenic based on previous reports [7,8,26] and designated as hotspot POLE mutations for the purpose of this study (Table 1).' All five hotspot residues (286-459) fall inside the N-terminal exonuclease domain; residue 1347 is far distal to that region.
• Leon-Castillo et al. 2020 (PATH-250-323.pdf), Discussion: 'nearly all (>95%) POLE mutations outside the exonuclease domain are classified as non-pathogenic by POLE-score', and Abstract: 'Only 1/23 (4%) tumours with non-EDM showed these genomic alterations, indicating that a large majority of mutations outside the exonuclease domain are not pathogenic.'
• Supplementary Table S1 (PATH-250-323-s002.xlsx) lists 103 recurrent POLE variants with COSMIC/TCGA counts; the rows flanking the queried region are p.G1343D (c.4028G>A), p.Q1335* (c.4003C>T), p.S1380L (c.4139C>T) and p.A1375V (c.4124C>T). No row for c.4039C>T / p.Leu1347= exists in the table.
• CancerHotspots (queried 'POLE L1347') returned no result: 'This variant does not lie in a statistically significant hotspot.'
• vcep_materials.json 'domain_tables' is an empty object, so no authoritative list of approved POLE critical functional domains was available for the residue check.

Still missing

• vcep_materials.json domain_tables is empty ({}): no VCEP-approved critical-functional-domain list exists for POLE, so the residue was checked against the Leon-Castillo et al. 2020 framework rules and the exonuclease-domain hotspot set (residues 286-459) rather than an authoritative domain table.
• PATH-250-323-s001.tif (declared for PM1 and PS4 by the gene VCEP index) could not be converted to searchable text and was not visually reviewed; it is a supplementary figure and is not expected to carry variant-level PM1 evidence.
• No entry for this variant exists in any searchable file of the POLE package: grep for 'c.4039C>T', 'p.Leu1347=', 'p.L1347=', 'L1347' and '4039' over PATH-250-323.pdf.txt, PATH-250-323-s002.xlsx.txt, PATH-250-323-s003.xlsx.txt and PATH-250-323-s004.xlsx.txt returned zero matches, so the package provides no pre-assigned hotspot/domain code for p.Leu1347=.

Sources

Reasoning

Scope first: NM_006231.4:c.4039C>T is a synonymous change (NP_006222.2:p.(Leu1347=), consequence_class 'synonymous', genomic NC_000012.11:g.133225625G>A / chr12:132649039G>A hg38), so PP3 must be taken from the splice-prediction path (SpliceAI), not from REVEL; REVEL is not applicable to a synonymous variant and was in any case not found for this position. The applicable SpliceAI Lookup result for this variant reports DS_AG 0.014, DS_AL 0.008, DS_DG 0.001 and DS_DL 0.01, i.e. a maximum delta score of 0.014, with Pangolin SG 0.003 and SL -0.017. The published SpliceAI calibration supplied for this task places PP3 supporting at max delta >=0.2 (Jaganathan et al. 2019, Cell, PMID:30661751), and PP3 moderate at >=0.5. The observed maximum delta of 0.014 is far below the supporting cutoff, and no second splice predictor predicts a deleterious effect. The POLE gene-specific PP3 rule derived from Leon-Castillo et al. 2020 (Supplementary Table S2/S3: missense variant present in the table, REVEL class 'likely disease causing', <=1 benign in silico results) is not triggered, because this synonymous variant does not appear in either supplementary table. PP3 is therefore not met.

Evidence considered

• Case normalization: NM_006231.4:c.4039C>T = NP_006222.2:p.(Leu1347=), consequence_class 'synonymous', variant_bucket 'other', canonical_splice_consensus false; genomic NC_000012.11:g.133225625G>A (hg37) / chr12:132649039G>A (hg38).
• The Supplied path-selection rule assigns synonymous variants to the SpliceAI path only, so REVEL is not applicable and was not consulted for this criterion.
• SpliceAI Lookup result for NM_006231.4:c.4039C>T: DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01 (maximum delta score 0.014); Pangolin SG 0.003, SL -0.017; spliceai_available true.
• Published SpliceAI calibration used verbatim: PP3 supporting at max delta >=0.2 and PP3 moderate at >=0.5 (Jaganathan et al. 2019, Cell, PMID:30661751). Observed 0.014 is below both.
• Local custom POLE framework (Leon-Castillo et al. 2020) PP3 rule requires the exact missense variant to be listed in Supplementary Table S2 or S3 with a 'likely disease causing' REVEL class and <=1 benign in silico results; this variant is absent from both tables, so the rule does not trigger and the generic workflow applies.
• Search of the declared POLE VCEP materials (PATH-250-323.pdf, PATH-250-323-s003.xlsx, PATH-250-323-s004.xlsx) for 'c.4039C>T', 'p.Leu1347=', 'p.L1347=', '1347' and '4039' returned no entry for this variant; the main paper's in silico section concerns prediction of the effect of non-synonymous variants only (SIFT, PANTHER, SNAP2, PolyPhen-2, PROVEAN, REVEL).

Still missing

• The POLE gene-specific PP3 rule from Leon-Castillo et al. 2020 could not be applied: the exact variant c.4039C>T / p.(Leu1347=) appears in neither Supplementary Table S2 (PATH-250-323-s003.xlsx, in silico results for TCGA endometrial carcinoma variants) nor Supplementary Table S3 (PATH-250-323-s004.xlsx, in silico results for non-TCGA endometrial carcinoma cohorts); both tables list only non-synonymous/truncating entries. Searches for 'c.4039C>T', 'p.Leu1347=', 'p.L1347=', '1347' and '4039' across the whole POLE fulltext directory returned no hits.

Sources

Reasoning

PP5 (reputable source recently reports the variant as pathogenic) is restricted here to an exact-variant ClinVar expert-panel classification of Pathogenic or Likely pathogenic. The ClinVar record for this exact variant (VCV000699633, NM_006231.4(POLE):c.4039C>T, p.Leu1347=, exact transcript-HGVS match) is classified 'Likely benign', with a review status of 'criteria provided, single submitter' / ClinVar API 'criteria provided, multiple submitters, no conflicts' (2 stars). The ClinVar submission audit confirms 2 submissions (SCV002623430, Ambry Genetics; SCV001008760, Labcorp Genetics formerly Invitae), both originating from clinical testing laboratories, both with is_expert_panel = false, and expert_panel_submissions = 0 out of 2. There is no ClinGen Variant Curation Expert Panel assertion for this variant or this gene in the record. The direction of the only available classification is benign, not pathogenic, and it derives from ordinary laboratory submissions rather than an expert panel. PP5 therefore is not met: the required expert-panel Pathogenic/Likely pathogenic classification does not exist.

Rule

PP5 requires an exact-variant ClinVar expert-panel (ClinGen VCEP) classification of Pathogenic or Likely pathogenic; ordinary laboratory submissions and aggregate multi-submitter labels do not qualify.

Evidence considered

• ClinVar VCV000699633 (variation 699633): classification 'Likely benign'; review status 'criteria provided, single submitter' (ClinVar API: 'criteria provided, multiple submitters, no conflicts', 2 stars); expert_panels empty.
• clinvar_submission_audit.json: submissions_total 2 (Ambry Genetics SCV002623430; Labcorp Genetics/Invitae SCV001008760), expert_panel_submissions 0, both single-submitter clinical-testing submissions, is_expert_panel false for each.
• No ClinGen expert-panel assertion exists for POLE c.4039C>T, and the only available classification is benign, so the pathogenic direction required by PP5 cannot be satisfied.

Sources

Reasoning

BA1 is not met. Under the generic ACMG/AMP stand-alone benign threshold of allele frequency >=0.05, the highest reported frequency for this variant is 1.09801e-05 (1/91074 alleles) in the South Asian population of gnomAD v4.1, and the overall all-comers frequency is 1.85954e-06 (3/1613304 alleles) with a group-maximum filtering AF of 2.8e-07; it is entirely absent from gnomAD v2.1 (0/249534 alleles). Every value sits roughly four orders of magnitude below 0.05. The governing framework for this case is the local León-Castillo et al. 2020 custom POLE framework, which customizes only PM1, PS4, PP3 and BP4 and provides no population-specific BA1 rule, so the generic BA1 threshold applies. Substituting the non-cancer subsets does not change the result (gnomAD v2.1 non-cancer exomes 0/235300; gnomAD v3.1 non-cancer absent).

Rule

Generic ACMG/AMP BA1: allele frequency >=0.05; no applicable POLE-specific BA1 rule.

Evidence considered

• gnomAD v4.1 (all-comers): total 3/1613304 alleles, AF 1.85954e-06; group-maximum filtering AF 2.8e-07; highest subpopulation frequency 1.09801e-05 (1/91074) in South Asian individuals; zero homozygotes.
• gnomAD v2.1 (all-comers, exome): 0/249534 alleles, AF 0; zero homozygotes.
• gnomAD v2.1 non-cancer exomes: 0/235300 alleles, AF 0; gnomAD v3.1 non-cancer genomes: absent. These concordantly exclude a benign frequency signal.
• No applicable POLE-specific BA1 rule is provided by the local custom POLE framework; the generic ACMG/AMP BA1 threshold of 0.05 was therefore used.

Sources

Reasoning

BS1 is not met. Under the generic strong benign threshold of allele frequency >=0.01, the highest reported frequency is 1.09801e-05 (1/91074 alleles) in the South Asian population of gnomAD v4.1, the overall all-comers frequency is 1.85954e-06 (3/1613304), and the group-maximum filtering AF is 2.8e-07; the variant is absent from gnomAD v2.1 (0/249534). None of these approaches 0.01. The local León-Castillo et al. 2020 custom POLE framework does not specify BS1, so the supplied generic threshold applies, and the non-cancer subsets are concordant (v2.1 non-cancer exomes 0/235300; v3.1 non-cancer absent).

Rule

Generic BS1: allele frequency >=0.01; no applicable POLE-specific BS1 rule.

Evidence considered

• gnomAD v4.1 (all-comers): total AF 1.85954e-06 (3/1613304); group-maximum filtering AF 2.8e-07; highest subpopulation AF 1.09801e-05 (1/91074, South Asian).
• gnomAD v2.1 (all-comers, exome): 0/249534 alleles, AF 0.
• gnomAD v2.1 non-cancer exomes 0/235300 and gnomAD v3.1 non-cancer genomes absent are concordant with the all-comers result.
• No applicable POLE-specific BS1 frequency threshold is provided; the supplied generic BS1 threshold of 0.01 was used.

Sources

Reasoning

BS2 is not met. BS2 requires the variant to be observed in a healthy adult individual — homozygous for a recessive condition, heterozygous for a dominant condition with full penetrance expected at an early age, or hemizygous for an X-linked condition. No such observation exists here: the variant is seen on only 3 alleles in gnomAD v4.1 (3/1613304) and is absent from gnomAD v2.1 (0/249534), with zero homozygotes in every dataset, so there is no documented unaffected carrier or homozygote that could support BS2. The residual gap is that population databases do not provide per-individual phenotype or health status, but the near-total absence of the allele means BS2 cannot be satisfied regardless. The local POLE framework provides no BS2 rule.

Rule

BS2 requires observation in a healthy adult (homozygous for recessive, heterozygous for a fully penetrant dominant, or hemizygous for X-linked disease); no applicable POLE-specific BS2 rule is provided.

Evidence considered

• gnomAD v4.1 reports 0 homozygotes among 1613304 alleles and the variant on only 3 alleles (AF 1.85954e-06).
• gnomAD v2.1 reports 0 homozygotes and 0 carriers among 249534 alleles; gnomAD v2.1 non-cancer exomes likewise 0/235300.
• Population allele-frequency records do not document per-individual phenotype, age or health status, so no unaffected-adult observation is available to satisfy BS2; no POLE-specific BS2 rule is provided.

Still missing

• gnomAD records do not document individual clinical phenotype, age or health status, so BS2 cannot be positively satisfied from these data; the near-absence of the allele (3/1613304, 0 homozygotes) makes BS2 unmeetable in any case.

Sources

Reasoning

BS3 requires well-established in vitro or in vivo functional studies showing no damaging effect on the gene or gene product. No experimental functional study of any kind was identified for NM_006231.4:c.4039C>T, NP_006222.2:p.(Leu1347=): the case evidence bundle carries no functional-assay source, and the literature pass returned only two ClinVar-derived framework papers (PMID:25394175 referral guideline, PMID:28492532 Sherloc criteria refinement), neither reporting variant-level data. Two things that superficially look like BS3 support do not qualify. First, the SpliceAI result (max delta 0.014, below the <=0.1 BP4 cutoff) is an in-silico prediction, not a functional assay, and BS3 explicitly requires experimental functional evidence; it also belongs to the in-silico splicing pathway, not to this group. Second, the ClinVar Likely benign classification and the absence of an amino-acid change from a synonymous substitution are not functional datasets - converting a classification into BS3 would be inferring a criterion from another source's final classification, which is not permitted, and a synonymous codon change is not by itself a functional assay of splicing. The governing POLE specification defined for this case customizes only PM1, PS4, PP3 and BP4, so generic ACMG/AMP 2015 BS3 applies and is not satisfied. The absence of a negative assay result is a data gap, not evidence of no effect, so BS3 is reported not_met with the gap recorded rather than being credited at supporting strength.

Rule

BS3 (ACMG/AMP 2015, generic): well-established in vitro or in vivo functional studies show no damaging effect on protein function or splicing. No gene-specific BS3 specification exists in the governing POLE package.

Evidence considered

• Governing framework check: framework_mode 'custom_gene_framework'; the local POLE specification (final_classification_framework, leon-castillo-2020-custom-framework-v1) customizes only PM1, PS4, PP3, BP4 and contains no BS3 rule, so generic ACMG/AMP 2015 BS3 ('well-established functional studies show no damaging effect') was applied.
• Functional-adjacent data in this case is limited to the in-silico SpliceAI lookup: max delta score 0.014 (DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01). A computational prediction is not an assay and cannot satisfy BS3; it is also the wrong evidence pathway for BS3 in this group.
• No functional-assay entry exists anywhere in the retrieved evidence: ClinVar 699633 has no functional-assay evidence statement; OncoKB POLE L1347= is not variant-specific and lists no citations; CancerHotspots POLE L1347 returned no row; COSMIC POLE c.4039C>T returned no result. No source performed or cited an in vitro/in vivo functional study on c.4039C>T.
• Paper PMID:25394175 (Hampel et al., ACMG/NSGC referral guideline): variant_mentioned=false, no direct quote, no experimental context - is a referral-indication practice guideline, contains no functional assay data.
• Paper PMID:28492532 (Nykamp et al., Sherloc): methodological refinement of the ACMG-AMP criteria applied to >40,000 variants; variant_specific=false, no full text available in this case, no functional data on this variant.
• Deliberately not used as BS3 evidence: the ClinVar Likely benign classification for variation 699633 (2 submitters) and the synonymous nature of p.(Leu1347=). A classification is not a functional dataset, and a synonymous change is not an assay; BS3 was left unmet rather than credited at supporting strength from these.

Still missing

• No experimental functional study showing absence of a damaging effect (e.g. a normal splicing outcome on patient RNA or in a minigene system, or a normal proofreading/POLE function readout) exists for POLE c.4039C>T in this case; the case evidence bundle contains no functional-assay source type.
• BS3 could not be credited even at supporting strength without a qualifying assay with proper controls; no such assay was retrievable from the literature pass (PMID:25394175 and PMID:28492532 contain no variant-level functional data).
• The ClinVar Likely benign record for this variant carries no functional-assay evidence statement, so it cannot be leveraged indirectly to close this gap.

Sources

Reasoning

Scope first: BP4 for this variant can only come from the splice-prediction path, because NM_006231.4:c.4039C>T is synonymous (NP_006222.2:p.(Leu1347=)) and the missense/protein-predictor path (REVEL) is not applicable to it. The only applicable prediction available is the SpliceAI Lookup result, with a maximum delta score of 0.014 (DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01) plus Pangolin SG 0.003 / SL -0.017, which is numerically below the published BP4 supporting threshold of <=0.1 (Jaganathan et al. 2019, Cell, PMID:30661751). However, that single SpliceAI prediction is the calibrated basis for BP7, the synonymous-specific criterion, and this group's rules forbid double-counting one prediction across PP3/BP4/BP7; the prediction is therefore allocated to BP7 and is not reused for BP4. With no independent computational evidence left for BP4 (REVEL is out of scope for a synonymous variant, and the POLE Supplementary Table S2/S3 species-specific BP4 rule does not trigger because the variant is absent from both tables), BP4 is not met on this evidence.

Evidence considered

• Case normalization: NM_006231.4:c.4039C>T = NP_006222.2:p.(Leu1347=), consequence_class 'synonymous'; a synonymous variant takes BP4 evidence from the SpliceAI path only.
• SpliceAI Lookup result: maximum delta score 0.014 (DS_AG 0.014, DS_AL 0.008, DS_DG 0.001, DS_DL 0.01); Pangolin SG 0.003, SL -0.017.
• Published SpliceAI calibration used verbatim: BP4 supporting at max delta <=0.1 (Jaganathan et al. 2019, Cell, PMID:30661751). The observed 0.014 would satisfy that cutoff, but the same single prediction is the basis of BP7 and is not double-counted here.
• Local custom POLE framework (Leon-Castillo et al. 2020) BP4 rule requires the exact missense variant to be listed in Supplementary Table S2 or S3 with a 'Likely benign'/'Likely-benign' REVEL class and >=4 benign in silico results; this variant is absent from both tables, so the custom rule does not trigger and the generic workflow applies.
• The declared POLE VCEP materials (PATH-250-323.pdf, PATH-250-323-s003.xlsx, PATH-250-323-s004.xlsx) contain no pre-assigned BP4 code for this variant; the in silico tables cover non-synonymous variants only.

Still missing

• No independent computational prediction is available for BP4 after allocating the SpliceAI result to BP7; REVEL/BayesDel are not applicable to a synonymous variant.
• The POLE gene-specific BP4 rule from Leon-Castillo et al. 2020 (Supplementary Table S2/S3: missense variant present in the table, REVEL class 'Likely benign'/'Likely-benign', >=4 benign in silico results) cannot be applied: the exact variant c.4039C>T / p.(Leu1347=) has no entry in PATH-250-323-s003.xlsx or PATH-250-323-s004.xlsx, and neither table contains any synonymous variant. Searches for 'c.4039C>T', 'p.Leu1347=', 'p.L1347=', '1347' and '4039' across the POLE fulltext directory returned no hits.

Sources

Reasoning

Under the governing instruction for this group, BP6 may be applied only for an exact-variant ClinVar expert-panel classification of Benign or Likely benign; aggregate ClinVar labels and ordinary laboratory submissions must never trigger it. The exact-variant ClinVar record VCV000699633 (NM_006231.4(POLE):c.4039C>T, p.Leu1347=) is classified 'Likely benign', but the clinvar_submission_audit.json shows 2 submissions, both from clinical-testing laboratories (Ambry Genetics; Labcorp Genetics formerly Invitae), both single-submitter, with is_expert_panel = false for each and expert_panel_submissions = 0. There is no ClinGen Variant Curation Expert Panel assertion for this variant. The benign-direction ClinVar label therefore does not meet the BP6 requirement, which demands an expert-panel provenance. BP6 is not met: the classification is real but its source is not an expert panel, so it cannot be used.

Rule

BP6 requires an exact-variant ClinVar expert-panel (ClinGen VCEP) classification of Benign or Likely benign; ordinary laboratory submissions and aggregate multi-submitter labels must never trigger BP6.

Evidence considered

• ClinVar VCV000699633: 'Likely benign', review status 'criteria provided, single submitter' (API 'criteria provided, multiple submitters, no conflicts', 2 stars); no expert-panel assertion.
• clinvar_submission_audit.json: expert_panel_submissions 0 of 2; both submissions (Ambry Genetics; Labcorp Genetics/Invitae) are single-submitter clinical-testing submissions with is_expert_panel false, i.e. ordinary laboratory submissions.
• Because no expert-panel benign classification exists for this exact variant, the BP6 expert-panel-source requirement is not satisfied.

Sources

Reasoning

Generic ACMG/AMP 2015 PS2 requires a de novo variant (confirmed maternity and paternity) in a patient with the disease and no family history (PMID:25741868). The governing POLE framework (Leon-Castillo et al. 2020 custom POLE framework) customizes only PM1, PS4, PP3 and BP4, so PS2 falls back to generic ACMG/AMP, and no gene-specific PS2 specification is available. The case is a variant-level assessment (NM_006231.4:c.4039C>T, NP_006222.2:p.(Leu1347=)); the collected evidence comprises only ClinVar (VCV000699633, Likely benign, 2 submitters), gnomAD v2.1/v4.1 frequency, SpliceAI (max delta 0.014), OncoKB, COSMIC and Cancer Hotspots lookups. There is no proband, no trio or parental testing, no confirmation of parentage, no phenotype record and no family history anywhere in the case evidence, and a systematic search of the case evidence for de novo / parental / proband / meiosis / pedigree terms returns no records. Neither reviewed publication mentions the variant or supplies de novo data (PMID:25394175 is a cancer-referral practice guideline with no POLE content; PMID:28492532 is a methods-level refinement of ACMG-AMP criteria). PS2 therefore cannot be assessed from the available data; the criterion is neither met nor refuted.

Evidence considered

• Generic ACMG/AMP 2015 criterion PS2 definition used because the governing POLE framework does not customize PS2: de novo (both maternity and paternity confirmed) in a patient with the disease and no family history (PMID:25741868).
• The local POLE framework in use (Leon-Castillo et al. 2020 custom POLE framework) lists customized criteria PM1, PS4, PP3 and BP4 only, with explicit instruction to fall back to generic ACMG/AMP outside that scope.
• Case evidence is limited to variant-level sources: ClinVar VCV000699633 (Likely benign, 2 germline submitters, no conflicts, 2 stars), gnomAD v2.1 (0/249534 alleles) and v4.1, SpliceAI max delta 0.014, OncoKB (Unknown Oncogenic Effect, non-variant-specific), COSMIC (no record) and Cancer Hotspots (no significant hotspot).
• No case record contains a proband, parental genotypes, parentage confirmation, affected-relatives data or a phenotype, so no de novo observation can be documented for PS2.

Still missing

• No proband-level clinical record (phenotype, age at onset, disease status) is available for this variant.
• No trio or other parental testing data confirming maternity and paternity of c.4039C>T are available.
• No family-history documentation exists to establish whether the variant is absent from unaffected relatives or arises de novo.
• PS2 could be resolved only with a proband plus confirmed parental genotypes (or clinical de novo testing report).

Sources

Reasoning

PM3 requires a pathogenic (or likely pathogenic) variant confirmed in trans with the variant under assessment in a recessive gene/disorder. NM_006231.4:c.4039C>T is a synonymous change (NP_006222.2:p.(Leu1347=)) and the case evidence contains no allelic data of any kind: no second POLE variant, no parental or family genotypes, no affected-proband genotype, and no phase determination (no long-read or read-backed phasing). ClinVar variation 699633 holds only this variant, reported Likely benign by 2 clinical laboratories (Ambry Genetics SCV002623430; Labcorp/Invitae SCV001008760), germline origin, criteria provided single-submitter, with no condition labels and no criteria signals or phase information in either submission record. Neither paper retrieved for this case reports the variant or any trans configuration: PMID:25394175 is an ACMG/NSGC referral-indications practice guideline that does not mention POLE, and PMID:28492532 (Sherloc) is a criteria-refinement methodology paper with no variant-level data. The governing framework for this case is the local POLE framework derived from Leon-Castillo et al. 2020, which specifies custom rules only for PM1, PS4, PP3 and BP4 and therefore provides no PM3 specification; generic ACMG/AMP 2015 applies. Germline POLE cancer predisposition is heterozygous/dominant, and the only recessive POLE phenotype (FILS) requires biallelic variants, which cannot be evaluated without genotype or phenotype data. The criterion therefore cannot be assessed in either direction.

Evidence considered

• ClinVar variation 699633: NM_006231.4(POLE):c.4039C>T (p.Leu1347=) classified Likely benign by 2 clinical laboratories (Ambry Genetics SCV002623430 and Labcorp/Invitae SCV001008760), germline origin, review status 'criteria provided, single submitter'; no expert-panel submission and no condition labels or criteria signals in either record.
• Both ClinVar submissions describe the same single variant with no second POLE variant and no family or parental genotypes, so the cis/trans configuration of c.4039C>T is undetermined from ClinVar.
• PMID:25394175 was checked: it is an ACMG/NSGC practice guideline on referral indications for cancer predisposition assessment, does not mention POLE or this variant, and reports no allelic data.
• PMID:28492532 was checked: it is the Sherloc criteria-refinement paper (iterative ACMG/AMP rework applied across >40,000 clinically observed variants), a methodological reference with no variant-level or allelic data.
• Framework check: the local POLE framework derived from Leon-Castillo et al. 2020 (J Pathol 250:323-335) customizes only PM1, PS4, PP3 and BP4; it contains no PM3 rule, so the generic ACMG/AMP 2015 definition of PM3 (detected in trans with a pathogenic variant for a recessive disorder) governs.
• Gene-level context: germline heterozygous loss-of-function POLE variants cause dominant polyposis/colorectal cancer predisposition, while a recessive POLE phenotype (facial dysmorphism, immunodeficiency, livedo, short stature) requires biallelic variants; neither can be assessed here without genotype or phenotype data.

Still missing

• No phase/trans data: no second (likely) pathogenic POLE variant co-occurring with c.4039C>T, no proband genotype, and no parental or family testing to establish an in-trans configuration.
• No case indication or phenotype recorded, so it is not possible to determine whether a recessive (biallelic/FILS-type) POLE scenario is even in scope for this case.
• No ClinGen CSPEC/VCEP specification exists for POLE; the governing local custom framework does not specify PM3.

Sources

Reasoning

Generic ACMG/AMP 2015 PM6 requires an assumed de novo occurrence in a patient, without confirmation of maternity and paternity (PMID:25741868). The governing POLE framework customizes only PM1, PS4, PP3 and BP4 and provides no PM6 specification, so generic ACMG/AMP applies. Even at this lower evidentiary bar, PM6 still requires an observed de novo event in a proband with the relevant phenotype. The case provides none: there is no proband record, no parental testing of any kind, no phenotype, and no indication of the variant's parental origin. A systematic search of the case evidence for de novo / parental / proband / pedigree / family terms returns no records, and neither reviewed publication mentions the variant (PMID:25394175, PMID:28492532). PM6 therefore cannot be assessed.

Evidence considered

• Generic ACMG/AMP 2015 PM6 definition used because the governing POLE framework does not customize PM6: assumed de novo, without confirmation of parentage (PMID:25741868).
• The POLE framework in use customizes only PM1, PS4, PP3 and BP4 and directs generic ACMG/AMP fallback for all other criteria.
• No case evidence documents a proband, parental genotypes, or an asserted de novo occurrence for NM_006231.4:c.4039C>T / NP_006222.2:p.(Leu1347=).

Still missing

• No proband with an associated phenotype is documented for this variant.
• No parent-of-origin information (even unconfirmed parental genotypes or clinical assertion of de novo status) is available.
• PM6 could be resolved only with a proband and at least an unconfirmed but reported de novo observation, ideally with parental testing.

Sources

Reasoning

Generic ACMG/AMP 2015 PP1 requires co-segregation of the variant with the disease in multiple affected family members (segregation analysis with documented meioses) (PMID:25741868). The governing POLE framework customizes only PM1, PS4, PP3 and BP4 and provides no PP1 or segregation specification, so generic ACMG/AMP applies. The case contains no pedigree, no family members, no affected relative genotypes, and no meiosis count; the collected evidence is limited to variant-level databases (ClinVar VCV000699633, gnomAD v2.1/v4.1, SpliceAI max delta 0.014, OncoKB, COSMIC, Cancer Hotspots). A systematic search of the case evidence for segregation / pedigree / family / meiosis terms returns no records, and the only full-text paper examined (PMID:25394175) is a cancer-referral practice guideline that does not mention POLE or the variant. PP1 cannot be assessed because there is no family to segregate within; this is an absence of segregation data, not evidence against segregation.

Evidence considered

• Generic ACMG/AMP 2015 PP1 definition used because the governing POLE framework does not customize PP1: co-segregation with disease in multiple affected family members (PMID:25741868).
• The POLE framework in use specifies only PM1, PS4, PP3 and BP4, with generic ACMG/AMP fallback otherwise; it contains no segregation rule.
• No case record contains a pedigree, family member genotypes, or a meiosis count for NM_006231.4:c.4039C>T.

Still missing

• No pedigree or family structure is documented for this variant.
• No genotypes from affected or unaffected relatives are available, so no co-segregation can be counted.
• PP1 requires a multi-generation family with genotypes for affected members; none exists in this case.

Sources

Reasoning

PP4 requires that the patient's phenotype or family history be highly specific for a disease with a single genetic etiology. This case is a variant-level record (NM_006231.4:c.4039C>T) with no proband phenotype, no HPO terms, no disease indication and no family history recorded anywhere in the case evidence (review_manifest.json, case_summary.json, evidence.json and the full criteria evidence bundle were searched for phenotype/HPO/indication/proband fields; none were found). The only disease labels attached to the variant are ClinVar condition strings — 'Hereditary cancer-predisposing syndrome' and 'not provided' — which are generic cancer-predisposition labels, not a highly specific phenotype, and are attached to a benign-classified record. The variant itself is synonymous (p.Leu1347=), which would in any case argue against protein-level phenotypic specificity. Because the necessary clinical input (proband phenotype/family history) is absent, PP4 cannot be assessed and is recorded as not_assessed rather than not_met. Generic ACMG/AMP has been used in place of a VCEP specification, since the local León-Castillo framework explicitly does not customize phenotype criteria.

Evidence considered

• Searched the consolidated case-evidence bundle and case_summary.json for phenotype, HPO, indication, proband and family-history fields: none present.
• ClinVar (VCV000699633) lists conditions 'Hereditary cancer-predisposing syndrome' and 'not provided' — generic cancer-predisposition labels, not a specific phenotype.
• The local León-Castillo 2020 POLE framework explicitly states it does not customize phenotype criteria, so generic ACMG/AMP applies.

Still missing

• No patient phenotype, HPO terms, disease indication or family history is present in the case evidence bundle; PP4 cannot be evaluated without this clinical input.
• The ClinVar record carries only generic condition labels ('Hereditary cancer-predisposing syndrome', 'not provided'), which are not a highly specific single-gene phenotype.

Reasoning

Generic ACMG/AMP 2015 BS4 requires lack of segregation in affected members of a family (i.e. a tested family in which the variant is present in unaffected individuals or absent from affected ones) (PMID:25741868). The governing POLE framework customizes only PM1, PS4, PP3 and BP4 and provides no BS4 or non-segregation specification, so generic ACMG/AMP applies. BS4 is by definition false in the absence of family data: a non-segregation observation requires a family with genotyped affected and unaffected members. This case has no pedigree, no relatives, and no family genotypes; its evidence is variant-level only (ClinVar VCV000699633 as Likely benign from 2 submitters, gnomAD v2.1/v4.1, SpliceAI max delta 0.014, OncoKB, COSMIC, Cancer Hotspots). The ClinVar Likely benign classification itself is not non-segregation evidence and was not used to infer BS4. A systematic search of the case evidence for segregation / non-segregation / family terms returns no records. BS4 cannot be assessed.

Evidence considered

• Generic ACMG/AMP 2015 BS4 definition used because the governing POLE framework does not customize BS4: lack of segregation in affected members of a family (PMID:25741868).
• The POLE framework in use specifies only PM1, PS4, PP3 and BP4, with generic ACMG/AMP fallback otherwise; it contains no non-segregation rule.
• ClinVar VCV000699633 reports this variant as Likely benign from 2 germline submitters (Labcorp Genetics/Invitae SCV001008760, Ambry Genetics SCV002623430) with no expert-panel review; that classification is not a segregation observation and provided no non-segregation evidence for BS4.
• No case record contains a family, pedigree, or genotypes from affected/unaffected relatives for NM_006231.4:c.4039C>T.

Still missing

• No family with both affected and unaffected genotyped members is available to test non-segregation.
• No pedigree data exist that could show the variant in unaffected relatives or absent from affected relatives.
• BS4 could be resolved only with a family in which the variant fails to track with disease.

Sources

Reasoning

BP2 requires a demonstrated allelic relationship with a (likely) pathogenic variant: either 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. Neither condition can be evaluated here because no pathogenic POLE variant co-occurring with c.4039C>T was identified and no phase data exist (no parental/family genotypes, no read-backed or long-read phasing, no allele-specific testing). The variant itself is synonymous, NP_006222.2:p.(Leu1347=), so cis-configuration with a pathogenic variant in the same allele would be the only plausible route to BP2 support, and that configuration is unevidenced. ClinVar variation 699633 contains only this variant, reported Likely benign by 2 clinical laboratories (Ambry Genetics; Labcorp/Invitae) with germline origin, no condition labels, and no criteria signals; neither cited paper (PMID:25394175, PMID:28492532) reports the variant or any cis/trans configuration. Because the governing local POLE framework (Leon-Castillo et al. 2020) specifies no BP2 rule, generic ACMG/AMP 2015 applies. Without any phase observation the criterion can be neither met nor excluded.

Evidence considered

• ClinVar variation 699633: NM_006231.4(POLE):c.4039C>T (p.Leu1347=) reported Likely benign by 2 clinical laboratories (Ambry Genetics SCV002623430 and Labcorp/Invitae SCV001008760), germline origin, review status 'criteria provided, single submitter'; the record contains no phased genotypes, no co-occurring pathogenic variant, and no condition labels.
• The variant is synonymous, p.(Leu1347=), determined by normalization (NP_006222.2); it is not itself a pathogenic variant that another variant could be placed in cis or in trans against.
• Neither retrieved paper provides allelic or phasing information: PMID:25394175 is an ACMG/NSGC referral-indications practice guideline that never mentions POLE, and PMID:28492532 (Sherloc) is a criteria-refinement methodology paper with no variant-level data.
• Framework check: the local POLE framework derived from Leon-Castillo et al. 2020 customizes PM1, PS4, PP3 and BP4 only and gives no BP2 specification, so the generic ACMG/AMP 2015 BP2 definition governs.
• germline POLE predisposition is dominant (heterozygous loss-of-function), which would in principle make BP2 applicable if a pathogenic variant were shown on the opposite allele or in cis, but no such variant or phase information exists in this case.

Still missing

• No phase data: no determination of whether c.4039C>T sits in cis or in trans, since no parental/family genotypes, allele-specific testing, or long-read phasing are available.
• No co-occurring (likely) pathogenic POLE variant identified in the case, so BP2's requirement for a pathogenic variant in cis or in trans cannot be evaluated in either direction.
• No ClinGen CSPEC/VCEP specification exists for POLE; the governing local custom framework does not specify BP2.

Sources

Reasoning

BP5 requires that the variant be found in a case with an alternate molecular basis for the disease (for example a second, established pathogenic variant explaining the phenotype). Adjudicating it requires case-level genotype information — the affected individual's full variant profile or a reported co-occurring/alternate pathogenic finding. This case is a variant-only record: the case evidence bundle, case_summary.json, evidence.json and prefetch.json contain no proband genotype, no co-occurring variant list, no phenotype and no family data (searched for proband, co-occurring, alternate-molecular-basis and second-variant fields; none present). The variant is also absent from COSMIC and from the León-Castillo recurrence table, and there is no published case report linking c.4039C>T to an individual with a competing molecular diagnosis. With no case to evaluate, BP5 cannot be supported or excluded on the available data and is recorded as not_assessed rather than not_met.

Evidence considered

• Searched the case-evidence bundle, case_summary.json and prefetch.json for proband/co-occurring/second-variant/alternate-molecular-basis fields: none present.
• The variant is absent from COSMIC and from Supplementary Table S1, and no case report linking c.4039C>T to an individual with an alternate molecular basis was identified.

Still missing

• No proband genotype, co-occurring variant list or clinical narrative is present in the case evidence, so BP5's precondition (an affected case with an alternate molecular basis) cannot be evaluated.
• No published case report was identified that places c.4039C>T in an individual with a competing molecular diagnosis.