KEAP1 c.340G>A, p.(Gly114Arg)

Reasoning

PM2 (supporting) requires absence or extremely low frequency in population databases; the governing generic ACMG non-VCEP threshold is AF <0.1%. gnomAD v4.1 total AF = 6.19519e-07 (0.00006%, 1/1,614,156 alleles, 0 homozygotes) and highest subpopulation AF = 8.47427e-07 (0.00008%, NFE, 1/1,180,042 alleles), both far below 0.1%; the variant is absent from gnomAD v2.1 and gnomAD-Canada v1.0. PM2 is met at supporting strength. Caveat noted: the variant is not strictly absent (single NFE exome singleton; grpmax FAF null), and classic ACMG 2015 wording is 'absent in controls (or extremely low frequency if recessive)', but the governing pipeline rule and SVI 2020 guidance support PM2 for variants at this extremely low frequency, and the observed frequency is ~4 orders of magnitude below the 0.1% threshold.

Rule

ACMG/AMP 2015 (PMID:25741868) PM2; non-VCEP AF threshold <0.1% (supporting)

Evidence considered

• gnomad_v4
• gnomad_v2
• gnomad_canada
• clinvar
• pvs1_gene_context

Still missing

• gnomAD v4.1 grpmax FAF is null for this singleton allele; re-verify frequency on future gnomAD releases.

Sources

Reasoning

PM1 (Richards et al. 2015, PMID:25741868) requires the variant to lie in a mutational hot spot and/or a critical, well-established functional domain without benign variation. The mutational-hotspot arm is affirmatively negative for this residue: the Cancer Hotspots query 'KEAP1 G114' returned no_result (evidence.json results.hotspots: 'This variant does not lie in a statistically significant hotspot.'), and OncoKB provides no variant-specific evidence for G114R (classification 'Unknown Oncogenic Effect'). No VCEP/gene-specific PM1 rule exists for KEAP1, and gene_context_KEAP1.json contains no curated critical-domain map that would permit a generic-domain PM1 call (Gly114 sits near the C-terminus of the BTB/POZ dimerization/Cul3-binding region, but no domain pathogenicity rule is present in the case evidence to support this formally). Under the generic ACMG/AMP 2015 fallback, with no hotspot designation and no domain rule, PM1 is not met on the available evidence.

Rule

ACMG/AMP 2015 (PMID:25741868): PM1 - Located in a mutational hot spot and/or critical and well-established functional domain (e.g., active site of an enzyme) without benign variation

Evidence considered

• Cancer Hotspots query 'KEAP1 G114': no_result - G114 not in a statistically significant hotspot (evidence.json results.hotspots; no source_registry key exists for this source)
• OncoKB: no variant-specific reviewed functional evidence; classification 'Unknown Oncogenic Effect' (evidence.json results.oncokb)
• gene_context_KEAP1.json: gene summary only, no curated domain/hotspot map

Still missing

• No VCEP/gene-specific PM1 rule for KEAP1 and no curated critical-domain map in the case evidence
• Cancer Hotspots coverage is somatic-cancer-focused; germline disease-relevant domain annotation for KEAP1 (familial multinodular goiter) was not available

Sources

Reasoning

PP3 requires multiple lines of computational evidence supporting a deleterious effect on the gene or gene product (Richards et al. 2015, PMID:25741868). For NM_012289.4:c.340G>A (KEAP1 p.Gly114Arg), the missense in-silico path does not supply a calibrated deleterious prediction: REVEL score 0.600 is below the PP3-supporting threshold (>=0.644) of the project-verified ClinGen SVI calibration, and BayesDel noAF 0.171 is directionally pathogenic but comes from a predictor whose threshold table is explicitly flagged as an unverified placeholder in this pipeline (REVEL is the only verified in-silico predictor used in production). The splice path adds no PP3 contribution: SpliceAI max delta is 0.01 (DS_AG 0.01, DS_AL 0.01, DS_DG 0.0, DS_DL 0.0), far below any splice-altering threshold (>=0.2). The HCI prior was unavailable (gene_not_supported). With no VCEP/gene-specific PP3/BP4 lookup for KEAP1 and only one directionally suggestive, unverified prediction, the multiple-line requirement for PP3 is not met.

Rule

generic_acmg_2015_pp3: multiple concordant lines of calibrated computational evidence for a deleterious effect (missense path: REVEL >=0.644 PP3-supporting per project-verified ClinGen SVI calibration, PMID 36413997; splice path: SpliceAI max delta >=0.2); none met - REVEL 0.600 neutral, BayesDel 0.171 single unverified-tool prediction, SpliceAI 0.01 no impact

Evidence considered

• REVEL score 0.600 - below PP3-supporting threshold (>=0.644) of the verified ClinGen SVI calibration; no PP3 contribution
• BayesDel noAF score 0.171 - directionally pathogenic but single-tool prediction with unverified placeholder thresholds; insufficient alone for PP3
• SpliceAI max delta 0.01 (DS_AG 0.01, DS_AL 0.01, DS_DG 0.0, DS_DL 0.0; variant_consequence sequence_variant) - no predicted splice impact; no PP3 contribution
• HCI prior not found (gene_not_supported); no prior probability evidence
• No CSPEC/VCEP or local gene-specific PP3/BP4 framework exists for KEAP1 (gene_vcep_dir null, cspec found=false); generic ACMG/AMP 2015 fallback applied

Still missing

• No validated second missense predictor threshold table is available (BayesDel thresholds flagged unverified in ACMG_V4/svc4/insilico.py); REVEL is the only project-verified missense predictor, and it is neutral for this variant.
• HCI prior probability not available for KEAP1 (gene_not_supported), so no Bayesian prior context could be weighed.

Sources

Reasoning

PP5 requires an exact-variant ClinVar expert-panel (3-star) classification of Pathogenic/Likely pathogenic. ClinVar search for KEAP1 c.340G>A returned no result (evidence.json clinvar: search_status 'no_result', found false; case_summary: 'This variant is absent from ClinVar.', variation_ids []). With no ClinVar record at all, there is no expert-panel classification to apply, and per governing policy ordinary laboratory submissions or aggregate labels would never trigger PP5 even if present. Criterion definitively not met.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), PP5: reputable source recently reports variant as pathogenic but evidence is not available for independent evaluation. Governing case policy narrows PP5 to an exact-variant ClinVar expert-panel classification (Pathogenic/Likely pathogenic); ordinary laboratory submissions and non-expert assertions never trigger PP5.

Evidence considered

• ClinVar absence of the exact variant (no expert-panel Pathogenic/Likely pathogenic assertion exists)

Sources

Reasoning

BA1 requires an allele frequency >5% in a large population database (pipeline non-VCEP proxy: >1%). Observed frequencies are orders of magnitude below this: gnomAD v4.1 total AF = 6.19519e-07 (0.00006%, 1/1,614,156 alleles, 0 homozygotes) with the highest subpopulation AF = 8.47427e-07 (0.00008%, NFE 1/1,180,042 alleles); the variant is absent from gnomAD v2.1 and gnomAD-Canada v1.0. The highest observed population frequency is ~4-5 orders of magnitude below the 1% non-VCEP BA1 proxy, so BA1 is definitively not met. Disease-prevalence context is consistent: KEAP1 germline disease (familial nontoxic multinodular goiter, autosomal dominant) is very rare, with germline KEAP1 mutations reported in only 2 families prior to PMID:39373520, so a benign polymorphism-frequency explanation is implausible.

Rule

ACMG/AMP 2015 (PMID:25741868) BA1; non-VCEP AF threshold proxy >1% (classic BA1 >5%)

Evidence considered

• gnomad_v4
• gnomad_v2
• gnomad_canada
• pvs1_gene_context

Sources

Reasoning

BS1 requires an allele frequency greater than expected for the disorder (pipeline non-VCEP threshold: >0.3%). The highest observed AF is 8.47427e-07 (0.00008%) in gnomAD v4.1 NFE (single allele), with total AF 6.19519e-07 (0.00006%); the variant is absent from gnomAD v2.1 and gnomAD-Canada v1.0. The observed frequency is ~4 orders of magnitude below the 0.3% threshold. Disease penetrance/prevalence context (rare autosomal dominant familial multinodular goiter; germline KEAP1 mutations reported in only a handful of families, PMID:39373520) cannot plausibly accommodate a 0.3% population allele frequency. BS1 not met.

Rule

ACMG/AMP 2015 (PMID:25741868) BS1; non-VCEP AF threshold >0.3%

Evidence considered

• gnomad_v4
• gnomad_v2
• gnomad_canada
• pvs1_gene_context

Still missing

• No quantitative disease prevalence/penetrance estimate for KEAP1 germline phenotypes is available; qualitative rarity support only (PMID:39373520). This does not affect the outcome, as observed AF is far below threshold.

Sources

Reasoning

BS2 requires observation of the variant in healthy adult individuals (heterozygous state for a dominant disorder such as KEAP1 germline disease) with full penetrance expected. Available data show only a single heterozygous exome carrier in gnomAD v4.1 (NFE, 1/1,614,156 alleles, 0 homozygotes) and absence from gnomAD v2.1 and gnomAD-Canada v1.0; no curated unaffected-adult cohort observation is available. A single rare heterozygote in a general-population database is not an observation in a phenotyped healthy adult and is far below any frequency required to support BS2. BS2 not met.

Rule

ACMG/AMP 2015 (PMID:25741868) BS2 (observation in unaffected adult, heterozygous for dominant disorder)

Evidence considered

• gnomad_v4
• gnomad_v2
• gnomad_canada

Still missing

• No curated unaffected-adult cohort (e.g., ESP/1000 Genomes) query is available for this case; gnomAD data alone show a single rare heterozygote and 0 homozygotes, which is insufficient for BS2.

Sources

Reasoning

BP1 (Richards et al. 2015, PMID:25741868) requires that primarily truncating variants are known to cause the disease, making a missense variant unlikely to be pathogenic. This premise fails for KEAP1 germline disease: pvs1_gene_context.json (lof_mechanism_supported=true) embeds PMID:39373520, which reports 5 probands with pathogenic heterozygous KEAP1 mutations including three established pathogenic missense variants (p.L136P, p.R415C, p.R483H) causing familial multinodular goiter, alongside two truncating variants (p.Q86*, p.V411fs). The documented coexistence of pathogenic missense and truncating alleles demonstrates that missense is a bona fide disease mechanism in KEAP1 germline disease, so a missense variant cannot be discounted as likely benign on mechanism alone. BP1 is therefore not met.

Rule

ACMG/AMP 2015 (PMID:25741868): BP1 - Missense variant in a gene for which primarily truncating variants are known to cause disease

Evidence considered

• pvs1_gene_context.json: PMID:39373520 - pathogenic heterozygous KEAP1 mutations in familial multinodular goiter include missense p.L136P, p.R415C, p.R483H (missense is an established disease mechanism)
• ClinVar: absent for this variant (no record supporting any variant class)

Sources

Reasoning

BP2 (Richards et al. 2015, PMID:25741868, via generic_acmg_combination_rules) requires an actual observation of the variant either (a) in trans with a pathogenic variant for a fully dominant disease/disorder, or (b) in cis with a pathogenic variant in any inheritance pattern. No such observation has been reported for NM_012289.4:c.340G>A (p.Gly114Arg) in any source: the variant is absent from ClinVar (clinvar: no_result), absent from the literature pass (0 PMIDs), and is not among the KEAP1 germline mutations reported in the familial multinodular goiter case series (PMID:39373520 reports p.Q86*, p.L136P, p.V411fs, p.R415C, p.R483H). The only non-clinical observation is a single heterozygous carrier in gnomAD v4.1 (AF=6.19519e-07, 1/1,614,156 alleles, 0 homozygotes), which is population-level and unphased with respect to any pathogenic variant. The trans/cis observation required by BP2 therefore does not exist in the evidence base, so BP2 is not met.

Rule

generic_acmg_2015: BP2 requires observation of the variant in trans with a pathogenic variant (fully dominant disease) or in cis with a pathogenic variant (any inheritance pattern); no such observation is reported for this variant

Evidence considered

• Generic ACMG/AMP 2015 rules (generic_acmg_combination_rules, source PMID:25741868): BP2 requires observation in trans with a pathogenic variant (dominant disease) or in cis with a pathogenic variant (any pattern).
• clinvar: no ClinVar record for KEAP1 c.340G>A (no_result); no reported co-occurrence with any pathogenic variant.
• gnomad_v4: single heterozygous allele (1/1,614,156; 0 homozygotes), population observation only, no phase information, no pathogenic variant co-occurrence.
• pvs1_gene_context / PMID:39373520: 5 KEAP1 germline probands carry other mutations; c.340G>A is not reported among them, so no trans/cis observation of the queried variant exists.
• literature_pass.json: 0 PMIDs identified, criteria_index empty.

Still missing

• No proband observations of c.340G>A reported anywhere (ClinVar absent, literature_pass 0 PMIDs); no segregation or phase (trans/cis) data relative to any pathogenic KEAP1 variant.

Sources

Reasoning

BP4 requires multiple lines of computational evidence suggesting no impact on the gene or gene product (Richards et al. 2015, PMID:25741868). For NM_012289.4:c.340G>A (KEAP1 p.Gly114Arg), only the splice path contributes a no-impact line: SpliceAI max delta 0.01 predicts no significant splice impact. The missense path does not support BP4: REVEL 0.600 lies in the neutral zone (above the BP4-supporting threshold of <=0.183 and below the PP3-supporting threshold), and BayesDel noAF 0.171 is directionally pathogenic, actively contradicting a benign prediction. Only one of the three computational lines suggests no impact, so the multiple-line requirement of BP4 is not satisfied; applying BP4 on the SpliceAI line alone would be asymmetric with the PP3 determination.

Rule

generic_acmg_2015_bp4: multiple lines of computational evidence suggesting no impact (missense path: REVEL <=0.183 BP4-supporting per project-verified ClinGen SVI calibration, PMID 36413997; splice path: SpliceAI max delta <0.1); only SpliceAI (0.01) meets a no-impact threshold, while REVEL 0.600 is neutral and BayesDel 0.171 is pathogenic-direction - not met

Evidence considered

• SpliceAI max delta 0.01 (DS_AG 0.01, DS_AL 0.01, DS_DG 0.0, DS_DL 0.0) - single line suggesting no splice impact
• REVEL score 0.600 - neutral zone; does not meet BP4-supporting threshold (<=0.183) of the verified ClinGen SVI calibration
• BayesDel noAF score 0.171 - pathogenic-direction prediction; contradicts a benign no-impact prediction
• No CSPEC/VCEP or local gene-specific PP3/BP4 framework exists for KEAP1 (gene_vcep_dir null, cspec found=false); generic ACMG/AMP 2015 fallback applied

Still missing

• Only a single computational line (SpliceAI) predicts no impact; no benign-direction missense prediction is available (REVEL neutral, BayesDel pathogenic-direction).

Sources

Reasoning

BP6 requires an exact-variant ClinVar expert-panel (3-star) classification of Benign/Likely benign. ClinVar search for KEAP1 c.340G>A returned no result (evidence.json clinvar: search_status 'no_result', found false; case_summary: 'This variant is absent from ClinVar.'). With no ClinVar record at all, there is no expert-panel classification to apply, and per governing policy ordinary laboratory submissions or aggregate labels would never trigger BP6 even if present. Criterion definitively not met.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), BP6: reputable source recently reports variant as benign but evidence is not available for independent evaluation. Governing case policy narrows BP6 to an exact-variant ClinVar expert-panel classification (Benign/Likely benign); ordinary laboratory submissions and non-expert assertions never trigger BP6.

Evidence considered

• ClinVar absence of the exact variant (no expert-panel Benign/Likely benign assertion exists)

Sources

Reasoning

PS1 (Richards et al. 2015, PMID:25741868) requires a previously established pathogenic variant resulting in the same amino-acid change (p.Gly114Arg) regardless of nucleotide change. No such comparator was identified: ClinVar has no record for NM_012289.4:c.340G>A (evidence.json results.clinvar: 'This variant is absent from ClinVar.'), literature_pass.json contains 0 PMIDs, and the pm5_candidates artifact (pm5_mode 'unknown') explicitly skipped automatic same-residue comparator harvesting, so p.Gly114Arg was never enumerated against established-pathogenic records. Because the comparator evidence is absent rather than affirmatively refuted, PS1 is returned not_assessed pending a manual same-residue comparator search (e.g., ClinVar/LOVD queries for p.Gly114Arg). Governing framework: no ClinGen CSPEC/VCEP for KEAP1 exists (cspec found=false, vcep_materials empty), so the generic ACMG/AMP 2015 fallback (generic_acmg_combination_rules) applies.

Rule

ACMG/AMP 2015 (PMID:25741868): PS1 - Same amino acid change as a previously established pathogenic variant regardless of nucleotide change

Evidence considered

• ClinVar: variant absent, no submissions (evidence.json results.clinvar)
• pm5_candidates.json: found=false, pm5_mode unknown, no comparator candidates harvested
• literature_pass.json: 0 PMIDs available for this case

Still missing

• No established pathogenic variant resulting in the same amino-acid change (p.Gly114Arg) identified in ClinVar or literature; ClinVar has no record for c.340G>A
• pm5_candidates.json found=false with pm5_mode 'unknown' - same-residue comparator harvest was skipped, so an established-pathogenic p.Gly114Arg record may have been missed

Sources

Reasoning

PS2 requires a confirmed de novo occurrence of the variant in the proband (with maternity and paternity confirmed) in a phenotype-consistent case. No proband genotype data, no parental testing results, and no family-history information are present anywhere in the case bundle (case_summary, evidence.json, literature_pass.json, prefetch.json). The variant is absent from ClinVar (no submissions), the literature pass returned 0 PMIDs, and fulltext_map.json is empty, so no published de novo case could be identified. Evidence is insufficient to evaluate PS2; it is therefore not assessed rather than met or not met.

Rule

PS2: De novo variant (confirmed maternity and paternity) in a patient with the disease and no family history - strong evidence of pathogenicity (Richards et al. 2015, PMID:25741868; generic ACMG/AMP 2015 framework used because no KEAP1 CSPEC/VCEP exists).

Still missing

• Proband genotype data for NM_012289.4:c.340G>A
• Parental testing results (maternity and paternity confirmation) for the proband
• Phenotype consistency documentation for the proband relative to the gene/disease phenotype
• Any published de novo case reports of this exact variant (none found via ClinVar or literature pass)

Sources

Reasoning

PS3 (Richards et al. 2015, PMID:25741868; generic ACMG/AMP 2015 fallback applies because no ClinGen CSPEC/VCEP specification exists for KEAP1 - cspec found=false, vcep_materials.json framework_complete=false, precedence_rank=3) requires well-established in vitro or in vivo functional studies supportive of a damaging effect on the gene or gene product. No functional evidence exists in this case's evidence base: literature_pass.json retrieved 0 PMIDs; the publications/ directory and fulltext_map.json are empty; ClinVar and COSMIC have no entry for NM_012289.4:c.340G>A (p.(Gly114Arg)); and the OncoKB KEAP1/G114R record reports no variant-specific functional evidence (is_variant_specific=false, not_reviewed=true, classification 'Unknown Oncogenic Effect', biological_effect empty, no citations). The pvs1_articles retrieved for the PVS1 group do not mention this variant (no Gly114/G114/c.340G>A hits) and are gene-context only. In silico predictors (REVEL 0.6, BayesDel 0.171, SpliceAI max delta 0.01) are not functional studies and cannot satisfy PS3. Because the evidence base contains no functional assay data for p.(Gly114Arg), PS3 cannot be assessed; confidence 'high' refers to confidence in the completeness of this determination, not to any functional claim.

Evidence considered

• OncoKB KEAP1/G114R entry: no variant-specific functional evidence (is_variant_specific=false, not_reviewed=true, biological_effect empty); pipeline evidence_sentence: 'OncoKB did not identify variant-specific reviewed functional evidence for this variant; gene-level curated context is available for reviewer follow-up.'
• literature_pass.json: total_pmids=0, criteria_index empty
• fulltext_map.json: {} (no full texts); publications/ directory: no files
• evidence.json COSMIC: no_result for 'KEAP1 c.340G>A'
• evidence.json ClinVar: variant absent
• pvs1_articles (PMID_42279597, PMID_42232598, PMID_39373520, PMID_37838086, PMID_30717682): no mention of Gly114/G114/c.340G>A - gene-context only, not variant-specific functional evidence

Still missing

• No well-established functional assay evidence for p.(Gly114Arg) was retrieved (e.g., Keap1-Nrf2 reporter/degradation assays, NRF2 pathway activation assays, complementation/rescue experiments, or in vivo models). A targeted literature search for functional studies of KEAP1 p.Gly114Arg would be required to assess PS3.

Sources

Reasoning

No case-control enrichment data exists for this exact variant. literature_pass.json reports total_pmids=0 with an empty criteria_index, so no affected-vs-control cohort study has been identified. The variant is also unreported in ClinVar (no variation record) and COSMIC (no somatic record). gnomAD v4.1 population frequency (total AF 6.19519e-07, 1/1614156 alleles) is not case-control data and is evaluated under the population-frequency group, not PS4. Without any published or proband-derived case-control comparison for c.340G>A, PS4 cannot be assessed.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), PS4: prevalence of the variant in affected individuals is significantly increased compared with prevalence in controls (case-control studies; strength adjustable to moderate/supporting for weaker study designs). No KEAP1 ClinGen CSPEC/VCEP exists, so generic fallback applies (framework_mode: generic_acmg).

Still missing

• No case-control cohort studies reporting c.340G>A in affected individuals vs controls (literature_pass total_pmids=0)
• Variant unreported in ClinVar and COSMIC; no published observations of this variant in any affected cohort

Reasoning

PM5 (Richards et al. 2015, PMID:25741868) requires a novel missense change at an amino-acid residue where a different missense change has been determined to be pathogenic. The pm5_candidates artifact was unable to confirm classic same-residue PM5 semantics safely (pm5_mode 'unknown') and skipped automatic candidate harvesting, returning found=false with zero candidates; ClinVar has no record for this variant and literature_pass.json contains no PMIDs. No alternate pathogenic missense at residue Gly114 (e.g., p.Gly114Glu, p.Gly114Val) was identified by any source in the bundle. Because the same-residue comparator search was not performed, absence cannot be asserted as established; PM5 is returned not_assessed pending a manual comparator search for pathogenic missense changes at Gly114.

Rule

ACMG/AMP 2015 (PMID:25741868): PM5 - Novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before

Evidence considered

• pm5_candidates.json: found=false, pm5_mode unknown, 'Unable to confirm classic same-residue PM5 semantics safely; skipping automatic candidate harvesting.'
• ClinVar: variant absent, no submissions (evidence.json results.clinvar)
• literature_pass.json: 0 PMIDs available for this case

Still missing

• No different pathogenic missense change at residue Gly114 identified; pm5_candidates.json found=false (pm5_mode 'unknown', harvest skipped)
• ClinVar absent for c.340G>A and 0 PMIDs available, so alternate-change pathogenicity at Gly114 could not be evaluated

Sources

Reasoning

PM6 applies to an assumed de novo occurrence when parental relationships are NOT confirmed. No de novo occurrence evidence of any kind exists for this variant: no proband/parental trio data in the case bundle, variant absent from ClinVar, literature pass found 0 PMIDs, and fulltext_map.json is empty. With no proband or parental data, PM6 cannot be evaluated and is not assessed.

Rule

PM6: Assumed de novo occurrence, without confirmation of parental relationships - moderate evidence of pathogenicity (Richards et al. 2015, PMID:25741868; generic ACMG/AMP 2015 framework used because no KEAP1 CSPEC/VCEP exists).

Still missing

• Any evidence of a de novo occurrence (proband tested, parents tested or inferred)
• Statement of whether parental relationships were or were not confirmed (needed to choose PS2 vs PM6)
• Published reports of the variant arising de novo (none found via ClinVar or literature pass)

Sources

Reasoning

PP1 requires co-segregation of the variant with disease in multiple affected family members (number of informative meioses determines strength). The case bundle contains no pedigree, no family member genotypes, no segregation analysis, and no affected relatives described in any source (case_summary, evidence.json, literature_pass.json). ClinVar has no submissions for this variant and the literature pass returned 0 PMIDs, so no published segregation data could be identified. PP1 cannot be evaluated and is not assessed.

Rule

PP1: Co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease - supporting (may be upgraded with more meioses per gene-specific guidance) (Richards et al. 2015, PMID:25741868; generic ACMG/AMP 2015 framework used because no KEAP1 CSPEC/VCEP exists).

Still missing

• Pedigree/family structure for the proband
• Genotype status of the variant in affected and unaffected family members
• Count of informative meioses showing co-segregation (needed to set PP1 strength, e.g., supporting vs strong)
• Published segregation data for this exact variant (none found via ClinVar or literature pass)

Sources

Reasoning

PP2 (Richards et al. 2015, PMID:25741868) requires BOTH a low rate of benign missense variation in the gene AND missense variants being a common mechanism of disease. The second prong is partially supported: pvs1_gene_context.json embeds PMID:39373520 (familial multinodular goiter), which reports 5 probands with pathogenic heterozygous KEAP1 mutations including missense variants p.L136P, p.R415C, p.R483H alongside truncating p.Q86* and p.V411fs, and OncoKB notes KEAP1 is recurrently mutated in lung cancer (gene-level, predominantly inactivating). However, the first prong requires gene-level missense constraint metrics (e.g., gnomAD missense Z-score / observed-expected ratio) showing a low rate of benign missense variation, and no such data are present anywhere in the case evidence (gnomAD entries contain per-variant frequencies only). Without constraint data and without a KEAP1 VCEP rule, PP2 cannot be met and is returned not_assessed pending gene-level constraint metrics.

Rule

ACMG/AMP 2015 (PMID:25741868): PP2 - Missense variant in a gene that has a low rate of benign missense variation and in which missense variants are a common mechanism of disease

Evidence considered

• pvs1_gene_context.json: germline LoF mechanism supported; PMID:39373520 lists established pathogenic germline KEAP1 missense variants (p.L136P, p.R415C, p.R483H) and truncating variants (p.Q86*, p.V411fs)
• OncoKB: KEAP1 recurrently mutated in lung cancer, tumor suppressor context (gene-level, no variant-specific review)
• gnomAD v2/v4 entries contain only per-variant allele frequencies; no gene-level constraint metrics in case evidence

Still missing

• Missing gene-level missense constraint metrics (e.g., gnomAD missense Z-score or observed/expected ratio) required for the 'low rate of benign missense variation' prong
• No KEAP1 VCEP/gene-specific PP2 rule available

Sources

Reasoning

No proband phenotype or family-history information is present anywhere in the case bundle, and no publication reporting this exact variant was retrieved (literature_pass total_pmids=0). Gene-level germline context exists (KEAP1 loss-of-function linked to familial multinodular goiter in PMID 39373520), but that study reports only p.Q86*, p.L136P, p.V411fs, p.R415C and p.R483H and does not mention c.340G>A/p.Gly114Arg; per citation policy, gene-level disease discussion cannot support PP4, which requires a phenotype/family history highly specific for a disease with a single genetic etiology in a variant-carrying individual. Cannot assess without case-level phenotype data.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), PP4: phenotype or family history is highly specific for a disease with a single genetic etiology. No KEAP1 ClinGen CSPEC/VCEP exists, so generic fallback applies.

Still missing

• No proband phenotype or family history supplied for this case
• No variant-specific phenotype publication for c.340G>A (KEAP1 familial multinodular goiter literature does not mention this variant)

Reasoning

BS3 (Richards et al. 2015, PMID:25741868; generic ACMG/AMP 2015 fallback, no KEAP1 VCEP/CSPEC) requires well-established functional studies showing no damaging effect on protein function or splicing. The case contains no such studies: literature_pass.json retrieved 0 PMIDs; publications/ and fulltext_map.json are empty; the variant is absent from ClinVar and COSMIC; and OncoKB provides only gene-level KEAP1 tumor-suppressor context with no variant-specific functional data. SpliceAI predicts no significant splice impact (max delta score = 0.01; DS_AG=0.01, DS_AL=0.01, DS_DG=0.0, DS_DL=0.0) and REVEL (0.6) / BayesDel (0.171) scores are available, but these are in silico predictions that belong to PP3/BP4-type computational evidence, not well-established functional studies, and cannot satisfy BS3. Because no functional study demonstrating absence of a damaging effect for p.(Gly114Arg) was retrieved, BS3 cannot be assessed; confidence 'high' refers to confidence in this determination of evidence absence, not to a benign functional claim.

Evidence considered

• literature_pass.json: total_pmids=0; fulltext_map.json: {}; publications/ directory: no files
• OncoKB KEAP1/G114R entry: gene-level context only, no variant-specific functional evidence (is_variant_specific=false, not_reviewed=true)
• SpliceAI: max delta score = 0.01 (DS_AG=0.01, DS_AL=0.01, DS_DG=0.0, DS_DL=0.0) - in silico only, not a well-established functional study
• REVEL score 0.6 and BayesDel score 0.171058 - in silico predictors, not functional evidence
• evidence.json COSMIC: no_result; evidence.json ClinVar: variant absent

Still missing

• No functional studies showing no damaging effect on protein function or splicing for p.(Gly114Arg) were retrieved. BS3 would require e.g., Keap1-Nrf2 functional assays demonstrating wild-type-like activity, or validated splicing assays; only in silico predictions (SpliceAI max delta 0.01, REVEL 0.6, BayesDel 0.171) are available, which are insufficient for BS3.

Sources

Reasoning

BS4 (benign) requires non-segregation: the variant present in an affected proband but absent in one or more affected family members. No family members were tested or reported in any source in the case bundle, ClinVar has no submissions for this variant, and the literature pass returned 0 PMIDs. There is therefore no non-segregation evidence and no basis to apply BS4; it is not assessed.

Rule

BS4: Non-segregation of the variant with disease in affected family members - strong evidence of benign effect (Richards et al. 2015, PMID:25741868; generic ACMG/AMP 2015 framework used because no KEAP1 CSPEC/VCEP exists).

Still missing

• Genotype status of the variant in affected family members (non-segregation would require an affected relative without the variant)
• Any family-based testing data (none present in case bundle)
• Published reports of non-segregation for this exact variant (none found via ClinVar or literature pass)

Sources

Reasoning

BP5 requires the variant to be found in a case with an alternate molecular basis for disease. No proband or case-level data is available in this bundle (no phenotype, no segregation, no alternative diagnosis reported), and no publication reporting this variant was retrieved. There is therefore no information on whether c.340G>A coexists with an alternate molecular explanation. Cannot assess.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), BP5: variant found in a case with an alternate molecular basis for disease. No KEAP1 ClinGen CSPEC/VCEP exists, so generic fallback applies.

Still missing

• No proband/case-level data to determine whether an alternate molecular basis for the phenotype is present