MSH3 c.2041C>T, p.(Pro681Ser)

Reasoning

BP1 (Richards et al. 2015, PMID:25741868) applies when a missense variant occurs in a gene for which primarily truncating variants are known to cause disease. The germline disease associated with MSH3 is autosomal recessive familial adenomatous polyposis 4 (FAP4), mediated by biallelic loss of function (pvs1_gene_context: lof_mechanism_supported=true; gene_context_MSH3.json: autosomal recessive, minimal evidence for Lynch-syndrome role). The MSH3 mutation spectrum is frameshift/truncation-predominant: recurrent somatic frameshift mutations in the (A)8 tract and biallelic frameshifts in tumors (PMID:11470537), and no pathogenic missense variant is established in MSH3 - all five MSH3 missense variants in the Singh et al. 2020 cohort, including p.(Pro681Ser), were classified class 3 (VUS) (PMID:32634176). The queried variant is a missense change in this truncation-driven gene, so BP1 is met at supporting strength. Flagged for human review because no VCEP/CSPEC specification for MSH3 exists and the pathogenic MSH3 variant spectrum is small.

Rule

BP1: missense variant in a gene for which primarily truncating variants are known to cause disease (Richards et al. 2015, PMID:25741868)

Evidence considered

• pvs1_gene_context.json: MSH3 germline disease mechanism is loss-of-function (biallelic FAP4); lof_mechanism_supported=true
• gene_context_MSH3.json: MSH3 acts as a tumor suppressor; inherited mutations associated with autosomal recessive FAP, minimal evidence for Lynch-syndrome role
• PMID:11470537 (Ohmiya et al. 2001): hMSH3 frameshift mutations in the (A)8 tract are the recurrent secondary-mutator alterations; biallelic frameshift mutations found in tumors; only one germline missense (p.Pro681Ser) detected and not established pathogenic
• PMID:32634176 (Singh et al. 2020): five MSH3 missense variants (Tyr465Cys, Thr282Ile, Arg411His, Ser466Gly, Pro681Ser) all classified class 3 (VUS); no pathogenic MSH3 missense established

Still missing

• No VCEP/CSPEC specification governs MSH3; BP1 application is based on generic ACMG/AMP 2015 plus literature mechanism evidence and should be confirmed by human review

Sources

Reasoning

PS1 (Richards et al. 2015, PMID:25741868) requires the same amino-acid change as a previously established pathogenic variant. The exact same change, NM_002439.5:c.2041C>T p.(Pro681Ser), was reported as a germline variant in colon tumor 440 by Ohmiya et al. 2001 (PMID:11470537, section 3.2) and in an endometrial cancer cohort by Singh et al. 2020 (PMID:32634176, Table 2 and Results text), but neither publication establishes pathogenicity: Ohmiya only notes the variant and speculates generally that some missense variants may be pathogenic, and Singh classifies it class 3 (VUS). ClinVar variation 656200 is conflicting (7 uncertain-significance and 6 likely-benign clinical submissions, no expert panel). No source establishes p.(Pro681Ser) as pathogenic, so the PS1 prerequisite (established pathogenic same-change) is not met.

Rule

PS1: same amino acid change as a previously established pathogenic variant regardless of nucleotide change (Richards et al. 2015, PMID:25741868)

Evidence considered

• ClinVar variation 656200: p.(Pro681Ser) has 7 uncertain-significance and 6 likely-benign clinical submissions (conflicting, single-submitter stars, no expert panel) - no established pathogenic classification
• PMID:11470537 (Ohmiya et al. 2001): germline Pro681Ser in colon tumor 440 reported without a pathogenicity determination
• PMID:32634176 (Singh et al. 2020): c.2041C>T p.(Pro681Ser) classified class 3 (VUS) in an endometrial cancer cohort

Sources

Reasoning

No case-control or cohort-enrichment study exists for MSH3 c.2041C>T (p.Pro681Ser). The only variant-specific literature observation (PMID:32634176) is a case-only targeted-sequencing study of 199 unselected sporadic endometrial cancer patients with no control arm, no statistical test, and a single carrier (sample 051330, allele fraction 0.48) whom the authors classified as class 3 (VUS); a crude 1/199 case carrier frequency does not exceed the population frequency of this variant (gnomAD v2.1 AF 0.07998%, 226/282,556 alleles; gnomAD v4.1 AF 0.17438%, 2,814/1,613,708 alleles, including 7 NFE homozygotes), so no enrichment is demonstrated. Other ClinVar-linked candidate papers were checked: PMID:11470537 (germline/somatic hMSH3 mutations in GI cancers) does not mention this variant in the available full text, and PMID:28944238 full text is unavailable (empty retrieval). PS4 requires statistically significant increased prevalence in affected individuals versus controls, which is absent.

Rule

PS4 (ACMG/AMP 2015, PMID 25741868): 'The prevalence of the variant in affected individuals is significantly increased compared with controls.'

Evidence considered

• PMID:32634176 - Case-only NGS study of 22 MMR-pathway genes in 199 unselected sporadic endometrial cancer patients (no control arm); Table 2 lists c.2041C>T p.(Pro681Ser) in a single patient (sample 051330), authors' class 3 (VUS); no enrichment analysis possible.
• gnomad_v2 - Variant present at total AF 0.07998% (226/282,556 alleles, 0 homozygotes; NFE 0.15745%), i.e. a relatively common polymorphism, contradicting a rare disease-allele enrichment hypothesis.
• gnomad_v4 - Variant present at total AF 0.17438% (2,814/1,613,708 alleles) with 7 NFE homozygotes; population frequency at or above the crude 1/199 case carrier frequency.
• PMID:11470537 - Full text searched for the exact variant (c.2041C>T / p.Pro681Ser); variant not mentioned; no case data for this variant (cannot support PS4).
• PMID:28944238 - Full text empty/unavailable in this case; variant mention could not be verified (cannot support PS4).

Still missing

• No case-control study or statistically powered cohort comparison for c.2041C>T exists; a formal case-control analysis would be needed to ever meet PS4.
• PMID:28944238 (36 CRC susceptibility genes, ClinVar PMID trail) has empty full-text retrieval in this case, so any variant mention there could not be verified.

Sources

Reasoning

PM1 (Richards et al. 2015, PMID:25741868) requires location in a mutational hotspot and/or critical functional domain WITHOUT benign variation. (1) Hotspot: the cancerhotspots.org query 'MSH3 P681' returned no statistically significant hotspot (evidence.json hotspots). (2) Domain: residue 681 is conserved in the mouse MSH3 homologue (PMID:11470537) and falls within the C-terminal MutS ATPase region immediately N-terminal to the Walker A P-loop (GPNMGGKSSY visible in the normalized protein sequence, prefetch.json). (3) However, the 'without benign variation' condition fails decisively: the variant residue itself is polymorphic in the general population - gnomAD v2.1 AF 0.07998% (226/282,556 alleles, 0 homozygotes) and gnomAD v4.1 AF 0.17438% (2,814/1,613,708 alleles, 7 homozygotes, grpmax FAF 0.22155%). The presence of 7 homozygotes in gnomAD v4.1 is strong evidence of benign variation at this exact residue, so PM1 is not met.

Rule

PM1: located in a mutational hotspot and/or critical functional domain without benign variation (Richards et al. 2015, PMID:25741868)

Evidence considered

• cancerhotspots.org query 'MSH3 P681': no statistically significant hotspot (evidence.json hotspots result)
• gnomAD v2.1: AF 0.07998% (226/282,556 alleles, 0 homozygotes) at the variant residue itself
• gnomAD v4.1: AF 0.17438% (2,814/1,613,708 alleles, 7 homozygotes; grpmax FAF 0.22155%) at the variant residue itself
• PMID:11470537 (Ohmiya et al. 2001): residue 681 is conserved in the mouse MSH3 homologue
• Normalized protein sequence (prefetch.json): position 681 lies N-terminal to the Walker A P-loop (GPNMGGKSSY) of the MSH3 MutS ATPase region

Sources

Reasoning

PM2 (moderate pathogenic) requires the variant to be absent from, or at extremely low frequency in, a large race-matched general population/control cohort (>1000 individuals). This variant is neither absent nor rare: gnomAD v4.1 total AF is 0.174% (2814/1,613,708 alleles) with 7 homozygotes, and the highest population AF is 0.229% (NFE); grpmax FAF is 0.22155%. gnomAD v2.1 shows total AF 0.07998% but NFE AF 0.157% and grpmax FAF 0.13962%. The lab's non-VCEP PM2 operating threshold is < 0.1%, applied to the highest observed population frequency: both gnomAD v4.1 (0.229%) and v2.1 (0.157% NFE / 0.140% grpmax FAF) exceed 0.1%. Threshold artifact considered: the gnomAD v2.1 overall AF (0.07998%) falls below 0.1% only because the allele is diluted across ancestries; using the overall AF rather than the population-maximum AF would be a threshold artifact, and the 7 v4.1 homozygotes plus NFE AF of 0.157-0.229% make clear the variant is a common NFE allele, far too frequent for a pathogenic allele of an ultra-rare recessive disorder. PM2 not met.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), PM2: absent in large race-matched general population/control cohort (>1000 individuals), or below expected carrier frequency if recessive; lab non-VCEP gnomAD operating threshold < 0.1% (highest population frequency).

Evidence considered

• gnomAD v4.1: total AF 0.00174381 (2814/1,613,708), NFE AF 0.00228783 (2699/1,179,722), grpmax FAF 0.00221552, 7 homozygotes
• gnomAD v2.1: total AF 0.000799841 (226/282,556), NFE AF 0.00157447 (203/128,932), grpmax FAF 0.00139623
• gnomAD-Canada v1.0 (registry key gnomad_canada): variant absent (search_status 'absent'), dataset not usable to establish rarity given presence at 0.17-0.23% in gnomAD v4.1
• ACMG/AMP 2015 (PMID 25741868): PM2 requires absence or below expected carrier frequency in a large race-matched control cohort (>1000 individuals)
• Lab non-VCEP gnomAD operating threshold (user framework): PM2 < 0.1% on highest population frequency; v4.1 NFE 0.229% and v2.1 NFE 0.157% both exceed it
• PMID:32634176 (Singh 2020): reports c.2041C>T at 0.0787% in an affected-cohort study with no case enrichment, consistent with a common population allele rather than a rare pathogenic allele

Still missing

• Race-matched control cohorts beyond gnomAD (e.g., study-specific controls) were not separately available; gnomAD v2.1/v4.1 population-maximum frequencies already exceed the PM2 threshold, so this gap does not change the outcome.

Sources

Reasoning

PP2 (Richards et al. 2015, PMID:25741868) requires a missense variant in a gene with a LOW rate of benign missense variation AND in which missense variants are a COMMON mechanism of disease. Both prongs fail for MSH3. (1) The gene is missense-tolerant: this very variant is present at gnomAD v4.1 AF 0.17438% with 7 homozygotes, and additional MSH3 missense variants (Tyr465Cys, Thr282Ile, Arg411His, Ser466Gly) are common enough to be seen in a small endometrial cancer cohort (PMID:32634176). (2) Missense is not the established disease mechanism: MSH3 germline disease (autosomal recessive FAP4) is loss-of-function mediated (pvs1_gene_context: lof_mechanism_supported=true, minimal evidence for a Lynch-syndrome role per gene_context_MSH3.json), the recurrent tumor mutations are frameshifts in the (A)8 tract (PMID:11470537), and all five MSH3 missense variants in the Singh cohort were class 3 (VUS) with none pathogenic. PP2 is therefore not met.

Rule

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 (Richards et al. 2015, PMID:25741868)

Evidence considered

• gnomAD v4.1: this missense variant is present at AF 0.17438% (2,814/1,613,708 alleles) with 7 homozygotes, indicating low missense constraint at this residue
• pvs1_gene_context.json: MSH3 germline disease mechanism is loss-of-function (biallelic, FAP4); lof_mechanism_supported=true
• gene_context_MSH3.json: inherited MSH3 mutations associated with autosomal recessive FAP with minimal evidence for a Lynch-syndrome role
• PMID:11470537 (Ohmiya et al. 2001): hMSH3 tumor mutations are predominantly frameshifts in the (A)8 tract (secondary mutator mutations)
• PMID:32634176 (Singh et al. 2020): all five MSH3 missense variants in the cohort (including p.Pro681Ser) were class 3 (VUS), none pathogenic

Still missing

• No quantitative missense-constraint metric for MSH3 (e.g., gnomAD missense Z-score or similar) was retrieved in the case bundle; the assessment relies on observed allele frequencies and literature mechanism evidence

Sources

Reasoning

PP3 requires multiple lines of computational evidence supporting a deleterious effect on the gene or gene product. For NM_002439.5:c.2041C>T (p.Pro681Ser), all available computational evidence was evaluated against published calibration thresholds: (1) SpliceAI max delta score 0.03 (DS_AG 0.00, DS_AL 0.03, DS_DG 0.00, DS_DL 0.00; variant 5-80063896-C-T, exon 14 missense, distances to affected splice sites 43-131 nt) is far below the 0.2/0.5 delta thresholds for predicted splice alteration (Jaganathan et al., Cell 2019; PMID 30661751), i.e., no splice impact predicted; (2) REVEL score 0.496 is below the ClinGen SVI PP3 supporting threshold of >= 0.773 (Pejaver et al., Am J Hum Genet 2022; PMID 36413997) and falls in the uninformative range (0.016 < score < 0.773); (3) BayesDel score -0.131816 is treated as insufficiently calibrated because no verified published threshold is available to this pipeline, so it cannot be counted; (4) no aGVGD or other missense predictor scores are present in the case bundle. No single line of computational evidence supports a deleterious effect and the multiple-line requirement is not met, so PP3 is not_met. The determination rests on evaluated scores versus named published thresholds, not on any other source's classification.

Rule

Generic ACMG/AMP 2015 PP3 (PMID 25741868): multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.). Applied with ClinGen SVI missense-predictor calibration for REVEL (PP3 supporting >= 0.773; Pejaver et al. 2022, PMID 36413997) and SpliceAI delta thresholds (>= 0.2 possibly, >= 0.5 likely splice-altering; Jaganathan et al. 2019, PMID 30661751).

Evidence considered

• REVEL score 0.496 (local REVEL v1.3 lookup, variant 5-80768077-C-T) is below the ClinGen SVI PP3 supporting threshold REVEL >= 0.773 (Pejaver V et al., Am J Hum Genet. 2022;109(12):2163-2177; PMID 36413997) - does not support a deleterious missense effect.
• SpliceAI max delta score 0.03 (DS_AG 0.00, DS_AL 0.03, DS_DG 0.00, DS_DL 0.00) is below the delta >= 0.2 (possibly splice-altering) and >= 0.5 (likely splice-altering) thresholds from Jaganathan K et al., Cell. 2019;176(3):535-548.e24; PMID 30661751 - no splice impact predicted.
• BayesDel score -0.131816 (local BayesDel noAF lookup, variant 5-80063896-C-T) is treated as insufficiently calibrated for PP3/BP4 per pipeline policy because no verified published threshold can be named - not counted as evidence.
• No aGVGD scores or gene-specific PP3/BP4 lookup table were present in vcep_materials.json (cspec not found, gene_vcep_dir null); generic ACMG/AMP 2015 PP3 definition applied (PMID 25741868).

Still missing

• No additional calibrated missense predictors (e.g., aGVGD, PrimateAI) were available in the case bundle; PP3 is not met by the available REVEL and SpliceAI scores regardless.

Sources

Reasoning

Under the governing instruction, PP5 is applied only for an exact-variant ClinVar expert-panel classification of Pathogenic/Likely pathogenic. ClinVar VCV000656200 (NM_002439.5(MSH3):c.2041C>T, p.Pro681Ser) has zero expert-panel submissions (expert_panels: {}; review status 'criteria provided, conflicting classifications', 1 star; 14 submissions all from clinical laboratories: 7 Uncertain significance, 6 Likely benign, 1 no-assertion). No expert-panel Pathogenic/Likely pathogenic classification exists, and the aggregate ClinVar label is 'Conflicting classifications of pathogenicity' (leaning benign). Ordinary laboratory submissions and aggregate labels must not trigger PP5. Criterion not met.

Rule

PP5 (ACMG/AMP 2015, PMID 25741868): 'Reputable source recently reports variant as pathogenic, but the evidence is not available to the laboratory to perform an independent evaluation.' Governing constraint applied: only an exact-variant ClinVar expert-panel (>=3-star) Pathogenic/Likely pathogenic classification may trigger PP5; laboratory submissions and aggregate labels are excluded.

Evidence considered

• clinvar - VCV000656200: 0 expert-panel submissions of 14 total (7 VUS, 6 Likely benign clinical-lab submissions, 1 no-assertion); review_stars=1; aggregate 'Conflicting classifications of pathogenicity' with no Pathogenic/Likely pathogenic expert-panel assertion.

Sources

Reasoning

No VCEP/CSPEC exists for MSH3 (vcep_materials.json: generic_acmg_fallback, precedence_rank 3), so generic ACMG/AMP 2015 (PMID 25741868) governs. BA1 requires an allele frequency >5% in a reference population dataset (stand-alone benign). The highest observed frequency for NM_002439.5:c.2041C>T (p.Pro681Ser) is 0.229% in gnomAD v4.1 European (non-Finnish) (AF 0.00228783, 2699/1,179,722 alleles); grpmax FAF is 0.22155% (0.00221552); gnomAD v2.1 NFE AF is 0.157% and total AF 0.07998%. Even against the lab's stricter non-VCEP gnomAD operating threshold (BA1 > 1%), the highest AF (0.229%) is far below 1%, and it is more than an order of magnitude below the ACMG 2015 5% stand-alone threshold. BA1 not met.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), BA1: allele frequency > 5% in ExAC/ESP/1000 Genomes or equivalent reference population dataset; lab operating threshold for gnomAD non-VCEP is 1%.

Evidence considered

• gnomAD v4.1 (registry key gnomad_v4): total AF 0.00174381 (2814/1,613,708 alleles), NFE AF 0.00228783 (2699/1,179,722), grpmax FAF 0.00221552
• gnomAD v2.1 (registry key gnomad_v2): total AF 0.000799841 (226/282,556 alleles), NFE AF 0.00157447 (203/128,932), grpmax FAF 0.00139623
• ACMG/AMP 2015 (PMID 25741868): BA1 stand-alone benign requires allele frequency over 5% in control populations
• Lab non-VCEP gnomAD operating threshold (user framework): BA1 > 1%; highest observed AF 0.229% is below this

Still missing

• No separate ESP/1000 Genomes allele-frequency check was performed; gnomAD v2.1/v4.1 (which subsume these populations at far larger scale) both place the variant well below every BA1 threshold, so this gap does not change the outcome.

Sources

Reasoning

BS1 requires an allele frequency greater than expected for the disorder. The highest observed AF is 0.229% (gnomAD v4.1 NFE; grpmax FAF 0.22155%; gnomAD v2.1 NFE 0.157%), which is below the lab's non-VCEP gnomAD BS1 operating threshold of 0.3%. Disease-prevalence reasoning was considered: MSH3-associated polyposis is an ultra-rare, adult-onset, incompletely penetrant autosomal-recessive condition; a fully-penetrant-recessive expected allele frequency for prevalence ~1:100,000-1:1,000,000 would be roughly 0.1-0.3%, straddling the observed 0.23% frequency, so BS1 cannot be robustly established from population frequency alone under either a fixed-threshold or prevalence-based reading. The 0.229% frequency is near the 0.3% operating threshold and the prevalence/penetrance assumptions materially affect the call, so the criterion is flagged for human review, but under the operative generic-ACMG calibration BS1 is not met.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), BS1: allele frequency greater than expected for the disorder in a control population; lab non-VCEP gnomAD operating threshold 0.3%.

Evidence considered

• gnomAD v4.1: NFE AF 0.00228783 (0.229%), grpmax FAF 0.00221552 (0.222%)
• gnomAD v2.1: NFE AF 0.00157447 (0.157%), grpmax FAF 0.00139623 (0.140%)
• ACMG/AMP 2015 (PMID 25741868): BS1 is strong benign evidence when control-population AF is greater than expected for the disorder
• Lab non-VCEP gnomAD operating threshold (user framework): BS1 > 0.3%; highest AF 0.229% below threshold
• PMID:32634176 (Singh 2020): reports c.2041C>T at 0.0787% population frequency in an endometrial-cancer MMR sequencing study with no control-cohort enrichment, corroborating the gnomAD population frequency considered for BS1

Still missing

• Penetrance and prevalence of biallelic MSH3-associated polyposis were not formally quantified in this case; a prevalence-based BS1 calibration could shift the threshold assessment and warrants human review.

Sources

Reasoning

BS2 (strong benign) under ACMG/AMP 2015 requires the variant to be observed in a well-documented healthy adult - for a recessive disorder, in the homozygous state - with full penetrance expected at an early age. gnomAD v4.1 reports 7 homozygotes (all in European non-Finnish; 6 exome, 1 genome) for this variant, and gnomAD v2.1 reports 0. The 7 population homozygotes are material evidence that the homozygous state is tolerated in the general population, consistent with reduced penetrance of biallelic MSH3-associated polyposis. However, the generic rule is not satisfied: (1) MSH3-associated polyposis is adult-onset and not fully penetrant at an early age, the explicit ACMG 2015 qualifier for BS2; and (2) gnomAD subjects are not phenotypically documented as healthy for polyposis, and the ACMG 2015 text itself warns that population databases 'cannot be assumed to include only healthy individuals.' BS2 (strong) is therefore not met under the generic framework; a reviewer could consider whether the homozygote observation warrants a downgraded/supporting benign application under a VCEP-style framework, which is outside generic ACMG scope.

Rule

Generic ACMG/AMP 2015 (PMID 25741868), BS2: observed in a healthy adult individual for a recessive (homozygous) disorder, with full penetrance expected at an early age.

Evidence considered

• gnomAD v4.1: total_hom 7 (7 in NFE; exome_hom 6, genome_hom 1), total AF 0.00174381
• gnomAD v2.1: total_hom 0, total AF 0.000799841
• ACMG/AMP 2015 (PMID 25741868): BS2 requires observation in a well-documented healthy adult with full penetrance expected at an early age (recessive: homozygous); population databases cannot be assumed to contain only healthy individuals
• ACMG/AMP 2015 (PMID 25741868): biallelic MSH3-associated polyposis is adult-onset with incomplete penetrance, so early-age full-penetrance qualifier for BS2 is not satisfied

Still missing

• Phenotype and age data for the 7 gnomAD v4.1 homozygotes are not available; without documented healthy-adult status and known early-age full penetrance, BS2 cannot be upgraded from the generic-rule reading.
• No published report of this variant in phenotypically documented healthy homozygous adults was found in the retrieved full texts.

Sources

Reasoning

BP4 requires multiple lines of computational evidence suggesting no impact on the gene or gene product. Available evidence: REVEL score 0.496 is ABOVE the ClinGen SVI BP4 supporting threshold of <= 0.016 (Pejaver et al. 2022; PMID 36413997) and lies in the uninformative range, so REVEL does not support a benign effect; BayesDel -0.131816 is insufficiently calibrated (no verified published threshold available to this pipeline) and cannot be counted; SpliceAI max delta 0.03 indicates no predicted splice impact (Jaganathan et al. 2019; PMID 30661751), which constitutes a single line of benign-direction computational evidence. Because the SVI-calibrated missense predictors provide no benign signal and only one line (SpliceAI) suggests no impact, BP4's multiple-line requirement is not satisfied. BP4 is not_met; the primary gap is the absence of additional calibrated missense predictors (or a verified BayesDel threshold) rather than evidence of a deleterious effect.

Rule

Generic ACMG/AMP 2015 BP4 (PMID 25741868): multiple lines of computational evidence suggest no impact on the gene or gene product (conservation, evolutionary, splicing impact, etc.). Applied with ClinGen SVI missense-predictor calibration for REVEL (BP4 supporting <= 0.016; Pejaver et al. 2022, PMID 36413997) and SpliceAI delta thresholds (Jaganathan et al. 2019, PMID 30661751).

Evidence considered

• REVEL score 0.496 (local REVEL v1.3 lookup, variant 5-80768077-C-T) is above the ClinGen SVI BP4 supporting threshold REVEL <= 0.016 (Pejaver V et al., Am J Hum Genet. 2022;109(12):2163-2177; PMID 36413997) - does not support a benign missense effect.
• SpliceAI max delta score 0.03 (DS_AG 0.00, DS_AL 0.03, DS_DG 0.00, DS_DL 0.00) indicates no predicted splice impact (thresholds per Jaganathan K et al., Cell. 2019;176(3):535-548.e24; PMID 30661751) - a single line of benign-direction evidence, insufficient alone for BP4.
• BayesDel score -0.131816 (local BayesDel noAF lookup, variant 5-80063896-C-T) is treated as insufficiently calibrated for PP3/BP4 per pipeline policy because no verified published threshold can be named - not counted as evidence.
• Generic ACMG/AMP 2015 BP4 definition applied (PMID 25741868): no MSH3 CSPEC/VCEP or gene-specific PP3/BP4 lookup table exists in vcep_materials.json.

Still missing

• Only one line of computational evidence (SpliceAI: no predicted splice impact) points toward no impact; REVEL 0.496 is uninformative (BP4 supporting threshold <= 0.016 per PMID 36413997 not reached) and BayesDel is uncalibrated. A verified published BayesDel threshold or an additional calibrated missense predictor would be required for BP4 to be met.

Sources

Reasoning

Under the governing instruction, BP6 is applied only for an exact-variant ClinVar expert-panel classification of Benign/Likely benign. ClinVar VCV000656200 has zero expert-panel submissions (expert_panels: {}; review_stars=1). The 6 'Likely benign' submissions (Quest Diagnostics, Ambry Genetics, Labcorp/Invitae, CeGaT, Dasa, PreventionGenetics) are ordinary clinical-laboratory assertions, which by explicit governing rule must never trigger BP6, and the aggregate label is 'Conflicting classifications of pathogenicity'. Criterion not met.

Rule

BP6 (ACMG/AMP 2015, PMID 25741868): 'Reputable source recently reports variant as benign, but the evidence is not available to the laboratory to perform an independent evaluation.' Governing constraint applied: only an exact-variant ClinVar expert-panel (>=3-star) Benign/Likely benign classification may trigger BP6; laboratory submissions and aggregate labels are excluded.

Evidence considered

• clinvar - VCV000656200: 0 expert-panel submissions; 6 Likely benign classifications all from ordinary clinical laboratories (Quest, Ambry, Labcorp/Invitae, CeGaT, Dasa, PreventionGenetics); aggregate 'Conflicting classifications of pathogenicity' (1 star).

Sources

Reasoning

PS2 (ACMG/AMP 2015) requires a de novo occurrence of the variant in a proband with documented paternity and maternity confirmation. No proband-parent trio or parental testing data exists for MSH3 c.2041C>T (p.Pro681Ser) in any source in this case: none of the 14 ClinVar submissions (VCV000656200) report a de novo origin or parental testing, and the two publications that report the variant (PMID 11470537, germline carrier in colon tumor 440; PMID 32634176, endometrial cancer patient 060161) provide no parental or de novo data. Full-text checks of PMID 28944238 and PMID 20981092 likewise found no de novo occurrence for this variant (the 'de novo' content of PMID 20981092 concerns the 1000 Genomes trio pilot mutation-rate analysis, not this variant). With no de novo evidence at all, PS2 cannot be assessed.

Rule

PS2 (ACMG/AMP 2015, PMID 25741868): de novo variant in a proband with confirmed paternity and maternity (strong)

Evidence considered

• ClinVar VCV000656200 submission audit: 14 submissions (7 VUS, 6 likely benign, conflicting 1-star), none reporting de novo origin or parental testing (prefetch.json submission_audit)
• Full-text review of PMID 11470537 and PMID 32634176, the two papers reporting this variant: no de novo occurrence or parental testing data present
• Full-text check of PMID 28944238 and PMID 20981092: no variant-specific de novo evidence in available text

Still missing

• No proband-parent trio or parental testing data; no reported de novo occurrence of c.2041C>T with confirmed paternity/maternity in any ClinVar submission or publication

Sources

Reasoning

No functional assay evidence for MSH3 c.2041C>T (p.Pro681Ser) exists in any reviewed source. OncoKB reports no variant-specific reviewed functional evidence (classification 'Unknown Oncogenic Effect', not reviewed). ClinVar submitter SCV002525954 (St. Jude) explicitly states 'to our knowledge functional studies have not been performed'. The only paper in the bundle that mentions the variant, PMID:32634176, reports it as an in silico-only class 3 (VUS) case observation in an endometrial cancer cohort with no functional assay. The functional-theme papers PMID:11470537 (hMSH3/hMSH6 mutation screen in MMP gastrointestinal cancers) and PMID:28944238 (CRC susceptibility panel) do not contain this variant in their full texts. No ClinGen CSPEC or VCEP applies to MSH3, so generic ACMG/AMP 2015 (PMID 25741868) governs: PS3 requires well-established functional studies showing a deleterious effect. With no such study available, PS3 cannot be applied and is recorded as not_assessed.

Rule

PS3: Well-established functional studies show a deleterious effect (Richards et al. 2015, ACMG/AMP, PMID 25741868)

Evidence considered

• clinvar: VCV000656200 has 14 submissions (7 VUS, 6 Likely benign, conflicting); no submission cites a functional assay; SCV002525954 (St. Jude) states 'to our knowledge functional studies have not been performed'
• oncokb: 'OncoKB did not identify variant-specific reviewed functional evidence for this variant'; classification 'Unknown Oncogenic Effect', not reviewed, no citations
• PMID:32634176: reports c.2041C>T p.(Pro681Ser) (rs115198722) as class 3 (VUS) in an endometrial cancer cohort; classification based on in silico features only; no functional assay performed
• PMID:11470537: hMSH3/hMSH6 mutation study of MMP gastrointestinal cancers; exact variant c.2041C>T p.(Pro681Ser) absent from full text; no functional assay of this variant
• PMID:28944238: targeted sequencing of 36 CRC susceptibility genes; exact variant absent from full text (MSH3 variants reported are p.Gln74fs, p.Gly896*, p.Leu911Trp, p.Asn524Thr, p.Arg669Trp, p.Ser490Tyr, p.His827Arg); no functional assays performed

Still missing

• No published functional assay of MSH3 p.(Pro681Ser) (e.g., MMR repair/activity assay, protein stability or expression assay) was identified in the evidence bundle, OncoKB, ClinVar-cited literature, or the available full texts
• ClinVar submissions with Likely benign calls (Quest, Ambry, PreventionGenetics, Labcorp, CeGaT, Dasa) cite no functional studies; whether any laboratory relied on unpublished or in-house functional data cannot be verified from this bundle

Sources

Reasoning

PM3 (ACMG/AMP 2015) applies to recessive disorders and requires the variant to be detected in trans with a pathogenic/likely pathogenic variant in an affected proband. No ClinGen CSPEC/VCEP specification exists for MSH3, so the generic ACMG/AMP 2015 rule (PMID:25741868) governs. MSH3-associated familial adenomatous polyposis 4 (FAP4) is an autosomal recessive (biallelic) condition, so PM3 is applicable in principle. However, no source in this case provides phase or second-allele data: (1) PMID:11470537 reports the exact variant as a germline missense Pro->Ser at codon 681 in colon cancer case 440, but it is labeled a germline sequence variant (G*) with no zygosity, no second MSH3 alteration, and no trans confirmation; (2) PMID:32634176 reports the exact variant c.2041C>T p.(Pro681Ser) (rs115198722) as a Class 3 VUS in a single endometrial cancer patient (sample 060161) with no phase or second-variant data; (3) PMID:28944238 does not report this variant at all, and its MSH3 biallelic/compound-heterozygote cases (Table 5) involve Gly896*/Leu911Trp and Asn524Thr/Arg669Trp, not Pro681Ser; (4) ClinVar record VCV000656200 (7 VUS + 6 likely benign submissions, no expert panel) contains no phase information and per instructions no criterion was inferred from final classifications; (5) gnomAD v4.1 shows 7 homozygotes but population homozygosity is not trans-phase evidence. Evidence is insufficient to assess PM3; the variant has not been shown to co-occur in trans with a pathogenic MSH3 variant in any affected proband.

Rule

PM3: For recessive disorders, detected in trans with a pathogenic variant (ACMG/AMP 2015, PMID:25741868). MSH3-FAP4 is autosomal recessive; criterion applicable in principle but unassessable without phase data.

Evidence considered

• PMID:11470537 (Ohmiya 2001, full text): exact variant reported as germline missense Pro->Ser at codon 681 in colon cancer case 440, conserved in mouse MSH3, absent in 190 other individuals; labeled germline sequence variant (G*); no phase or second-allele data — does not satisfy PM3
• PMID:32634176 (Singh 2020, full text): exact variant c.2041C>T p.(Pro681Ser), rs115198722 listed as Class 3 VUS (REVEL 0.48, gnomAD 0.0787%) in endometrial cancer patient sample 060161; no zygosity, no second MSH3 variant, no trans observation — does not satisfy PM3
• PMID:28944238 (DeRycke 2017, full text): variant not reported; MSH3 biallelic/compound-heterozygote carriers in Table 5 carry other variants (Gly896*/Leu911Trp; Asn524Thr/Arg669Trp) — no trans observation involving this variant
• PMID:20981092 (1000 Genomes): variant not mentioned; population resource only, no phase data
• PMID:25741868 (ACMG/AMP 2015): governing rule text for PM3 ('for recessive disorders, detected in trans with a pathogenic variant')
• clinvar (VCV000656200): 14 submissions (7 VUS, 6 LB, conflicting), no expert panel, no phase/co-occurrence data; classifications not used to infer criteria
• gnomad_v4 (v4.1): 7 homozygotes in 1,613,708 alleles; population homozygosity without disease context is not trans-phase evidence for PM3

Still missing

• No phase (trans vs cis) determination available for any observation of c.2041C>T/p.Pro681Ser
• No affected-proband genotype data demonstrating a second (pathogenic/likely pathogenic) MSH3 allele in trans
• Zygosity of the variant in the reported carriers (PMID:11470537 case 440; PMID:32634176 sample 060161) is not stated
• No segregation or parental testing data for the variant

Sources

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 previously determined pathogenic. No such same-residue comparator variant could be identified: the automated PM5 candidate harvest was not performed for this case (pm5_candidates.json: pm5_mode='unknown', eligible_for_classic_pm5_search=false, 0 same-residue candidates collected), and none of the reviewed publications (PMID:11470537, PMID:32634176) or ClinVar variation 656200 submissions report any alternate missense change at codon 681, pathogenic or otherwise. Evidence is therefore insufficient to confirm or definitively exclude PM5; a curated inventory of other variants at codon 681 (ClinVar/LOVD/UMD) is required before this criterion can be resolved.

Rule

PM5: novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before (Richards et al. 2015, PMID:25741868)

Evidence considered

• pm5_candidates.json: pm5_mode 'unknown', 0 same-residue candidate variants collected (harvest skipped per pipeline recommendation)
• ClinVar variation 656200: submissions list no alternate (different) missense variant at codon 681
• PMID:32634176 (Singh et al. 2020): no variant other than c.2041C>T reported at codon 681 in the endometrial cancer cohort

Still missing

• No systematic ClinVar/LOVD query for other missense changes at codon 681 was performed (pm5_candidates.json pm5_mode='unknown'; candidate harvest skipped) - a curated same-residue variant inventory is needed to definitively resolve PM5

Sources

Reasoning

PM6 (ACMG/AMP 2015) applies to assumed de novo occurrence without confirmation of paternity and maternity (moderate). No source reports any de novo or assumed-de-novo occurrence of MSH3 c.2041C>T (p.Pro681Ser): the ClinVar record (VCV000656200) contains no de novo assertion or parental-testing note in any submission, and the variant-bearing publications (PMID 11470537; PMID 32634176) report only singleton germline carriers without parent data. Evidence is insufficient to evaluate PM6.

Rule

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

Evidence considered

• ClinVar VCV000656200 submission audit: no de novo or parental-testing evidence signals in any of the 14 submissions (prefetch.json submission_audit)
• Full-text review of PMID 11470537 and PMID 32634176: variant reported as germline in single cancer cases only; no parent data

Still missing

• No proband-parent trio or parental testing data; no assumed-de-novo observation (even without parentage confirmation) reported for c.2041C>T

Sources

Reasoning

PP1 (ACMG/AMP 2015) requires co-segregation of the variant with disease in multiple affected family members in a gene definitively known to cause the disease. No segregation analysis exists for MSH3 c.2041C>T (p.Pro681Ser) in this case: PMID 11470537 reports a single germline carrier (colon tumor 440) with no affected-family-member testing; PMID 32634176 reports the variant in endometrial cancer patient 060161 whose recorded family history of cancer is negative; and the ClinVar submissions report no co-segregation data, with one submitter (St. Jude, SCV002525954) noting the variant 'has not been reported in individuals with msh3-related attenuated familial adenomatous polyposis'. No affected relatives were genotyped in any source, so co-segregation cannot be established or refuted.

Rule

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

Evidence considered

• PMID 11470537 (Section 3.2): germline missense variant (Pro to Ser at codon 681) in colon tumor 440; no segregation data; not detected in 190 other individuals
• PMID 32634176 (Table 2, sample 060161): c.2041C>T p.(Pro681Ser) in an endometrial cancer patient with no other cancers and no family history of cancer recorded
• ClinVar VCV000656200: no submission reports co-segregation; SCV002525954 (St. Jude) states the variant has not been reported in individuals with MSH3-related AFAP (absence of reported affected carriers, not a segregation analysis)
• PMID 28944238: main text contains no segregation data for this variant (variant-level data in supplementary tables not included in this case's full-text snapshot)

Still missing

• No segregation analysis of affected family members; no genotyping of affected relatives for c.2041C>T in any publication or ClinVar submission

Sources

Reasoning

PP4 requires a phenotype or family history highly specific for a disease with a single genetic etiology. The case bundle contains no proband phenotype or family-history data; the only phenotype-adjacent information is ClinVar's aggregated condition labels across 14 ordinary laboratory submissions (Endometrial carcinoma, Familial adenomatous polyposis 4, Hereditary cancer-predisposing syndrome, MSH3-related disorder, not provided, not specified), which are submission-level labels rather than a documented proband phenotype. The one variant-specific published case (PMID:32634176, sample 051330) had endometrial cancer plus multiple additional primary cancers (breast, spine, lung, kidney, liver) and no family history of cancer, which is not the MSH3-associated phenotype (MSH3-related attenuated adenomatous polyposis / FAP4, autosomal recessive) and was classified class 3 by the authors. MSH3-related disease is not a single-etiology phenotype that endometrial cancer alone would identify. Evidence is insufficient to assess PP4.

Rule

PP4 (ACMG/AMP 2015, PMID 25741868): 'Phenotype or family history highly specific for a disease with a single genetic etiology.'

Evidence considered

• PMID:32634176 - Only variant-specific case: sample 051330, endometrial cancer with multiple other primary cancers and no family history of cancer; phenotype not specific for MSH3-related disorder; authors classified the variant class 3 (VUS).
• clinvar - VCV000656200 condition labels (Endometrial carcinoma, FAP4, Hereditary cancer-predisposing syndrome, MSH3-related disorder, not provided/not specified) are aggregated across 14 lab submissions, mostly 'not provided'; do not constitute a proband phenotype of record.

Still missing

• No proband phenotype or family-history documentation exists in the case bundle (needed to assess phenotype specificity for MSH3-related disorder/FAP4).
• No evidence that the proband (or any carrier) exhibits the MSH3-specific phenotype (attenuated adenomatous polyposis / MSH3-related disorder).

Sources

Reasoning

BS3 requires well-established functional studies showing no damaging effect on protein function or splicing (generic ACMG/AMP 2015, PMID 25741868; no ClinGen CSPEC/VCEP applies to MSH3). No functional study of this variant exists in any reviewed source: OncoKB has no variant-specific functional evidence, ClinVar submissions cite no functional assays (SCV002525954 explicitly states none performed), and the only paper mentioning the variant (PMID:32634176) is a case observation without an assay. Population frequency (gnomAD v2.1 AF 0.08%, 226 alleles, 0 homozygotes) and ClinVar Likely benign submissions are observational/population evidence that belongs to other criteria (BA1/BS1/BP4), not BS3, which specifically requires functional assay evidence. SpliceAI (max delta 0.03) is in silico and cannot substitute for BS3. BS3 is therefore not assessable and is recorded as not_assessed.

Rule

BS3: Well-established functional studies show no damaging effect on protein function or splicing (Richards et al. 2015, ACMG/AMP, PMID 25741868)

Evidence considered

• clinvar: no submission cites functional studies demonstrating normal protein function for VCV000656200; SCV002525954 (St. Jude) states 'to our knowledge functional studies have not been performed'
• oncokb: no variant-specific reviewed functional evidence for MSH3 P681S; 'Unknown Oncogenic Effect', not reviewed
• PMID:32634176: class 3 (VUS) case observation only; contains no assay demonstrating normal function of p.(Pro681Ser)
• PMID:11470537: hMSH3/hMSH6 mutation study; exact variant absent from full text; no functional data for this variant
• PMID:28944238: CRC susceptibility panel; exact variant absent from full text; no functional assays of MSH3 variants

Still missing

• No well-established functional study demonstrating normal MSH3 protein function for p.(Pro681Ser) is available in the reviewed evidence
• No RNA/splicing assay evidence for c.2041C>T; SpliceAI max delta score 0.03 is an in silico prediction and does not qualify as BS3 functional evidence
• Whether Likely benign ClinVar submitters (e.g., Quest SCV002774182, Labcorp SCV000952874) used unpublished in-house functional data to support benign assertions cannot be verified from this bundle

Sources

Reasoning

BS4 (ACMG/AMP 2015) requires lack of segregation in affected family members (affected relatives shown not to carry the variant). No segregation-based absence data exists for MSH3 c.2041C>T (p.Pro681Ser): no affected family member of any carrier has been genotyped in the available literature or ClinVar submissions. The St. Jude ClinVar note that the variant has not been reported in individuals with MSH3-related attenuated familial adenomatous polyposis is an absence-of-literature-reports statement, not a segregation analysis, and is further contradicted at the case level by the variant being present in reported cancer patients (PMID 11470537; PMID 32634176). BS4 cannot be assessed without family segregation data.

Rule

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

Evidence considered

• ClinVar VCV000656200: no segregation or non-segregation observations in any submission; SCV002525954 note documents absence of reports of the variant in MSH3-related AFAP individuals, which is not a segregation analysis
• PMID 11470537: variant present in a cancer patient (germline carrier, colon tumor 440), contradicting absence-in-affected at the case level; no relative testing
• PMID 32634176: variant present in endometrial cancer patient 060161; carrier's family history of cancer recorded as negative but no relatives genotyped

Still missing

• No family segregation data (no affected relatives tested for the variant); no documented non-segregation observations
• Population-frequency context (gnomAD v4.1 AF 0.174%, 7 homozygotes) is relevant to benign population-frequency criteria handled by other groups, not to segregation-based BS4

Sources

Reasoning

BP2 (ACMG/AMP 2015) requires either (a) the variant observed in trans with a pathogenic variant for a fully dominant disease/mechanism, or (b) the variant observed in cis with a pathogenic variant in any inheritance pattern. No ClinGen CSPEC/VCEP specification exists for MSH3, so the generic ACMG/AMP 2015 rule (PMID:25741868) governs. The canonical MSH3 Mendelian phenotype (familial adenomatous polyposis 4) is autosomal recessive, not fully dominant, so the dominant-disease arm is not applicable; the cis arm is inheritance-independent but requires phase data that is entirely absent. No source reports the variant co-occurring with any pathogenic variant in cis or in trans: PMID:11470537 (germline Pro681Ser in case 440, no other MSH3 alteration reported), PMID:32634176 (single Class 3 VUS observation, sample 060161), PMID:28944238 (variant absent; MSH3 compound-heterozygote/homozygous carriers listed involve other variants), ClinVar (no co-occurrence/phase data). Evidence is insufficient to assess BP2.

Rule

BP2: Observed in trans with a pathogenic variant for a fully dominant disease/mechanism; or observed in cis with a pathogenic variant in any inheritance pattern (ACMG/AMP 2015, PMID:25741868). Dominant arm not applicable to autosomal recessive MSH3-FAP4; cis arm unassessable without phase data.

Evidence considered

• PMID:11470537 (full text): single germline observation of the exact variant (case 440) with no cis/trans co-occurrence data
• PMID:32634176 (full text): single Class 3 VUS observation (sample 060161) with no co-occurrence or phase data
• PMID:28944238 (full text): variant not reported; no cis/trans co-occurrence involving this variant (Table 5 checked)
• PMID:25741868 (ACMG/AMP 2015): governing rule text for BP2 ('observed in trans with a pathogenic variant for a fully dominant disease/mechanism; or observed in cis with a pathogenic variant in any inheritance pattern')
• clinvar (VCV000656200): no co-occurrence or phase data in any of the 14 submissions; conflicting VUS/LB classifications not used to infer criteria

Still missing

• No cis/trans co-occurrence data for this variant with any pathogenic variant
• No proband-level phased genotype data (parental or segregation testing)
• No documentation of the variant on the same allele (cis) as another variant in any carrier
• Zygosity of reported carriers is not stated in PMID:11470537 or PMID:32634176

Sources

Reasoning

BP5 requires the variant to be found in a case carrying a pathogenic variant in another gene that fully explains the phenotype (alternate molecular diagnosis). No proband genotype or phenotype data is present in the case bundle, and no co-occurring pathogenic variant in another gene is documented anywhere in the available evidence. In the single variant-specific case report (PMID:32634176, sample 051330) no alternate molecular diagnosis is reported, but that patient is not established as the proband of this case. Evidence is insufficient to assess BP5.

Rule

BP5 (ACMG/AMP 2015, PMID 25741868): 'Variant found in a case with a pathogenic variant in another gene that fully explains the phenotype.'

Evidence considered

• clinvar - No proband-level genotype information in VCV000656200 submissions; condition labels are aggregated and cannot support an alternate-molecular-diagnosis determination.
• PMID:32634176 - Sample 051330 (the one published c.2041C>T carrier) has no reported pathogenic variant in another gene explaining phenotype, but this is a single research cohort case, not this case's proband record.

Still missing

• No proband multi-gene genotype data (panel/exome) to evaluate presence of a pathogenic variant in another gene fully explaining the phenotype.
• Proband identity and phenotype of record are not documented in this case bundle.

Sources