MSH6 c.3674C>T, p.(Thr1225Met)

Reasoning

NM_000179.3:c.3674C>T (p.Thr1225Met) is a missense variant. The governing InSiGHT MSH6 VCEP specification (v2.0) directs missense PP3/BP4 calls to the local HCI-prior lookup table rather than to generic missense predictors (REVEL/BayesDel) or SpliceAI. The HCI-prior lookup for this exact coding substitution (c.3674C>T) returns a prior probability of pathogenicity of 0.7336, which falls in the VCEP-defined PP3_Supporting band (>0.68 and <=0.81). SpliceAI predicts no splice impact (max delta = 0.00), so the splice-impact PP3 path (non-canonical splice nucleotide, delta >=0.2) does not apply to this exonic missense substitution.

Rule

InSiGHT MSH6 VCEP v2.0 PP3: HCI prior probability >0.68 & <=0.81 = Supporting (via local HCI-PRIORS-MSH6.txt lookup table)

Evidence considered

• VCEP index guidance (HCI-PRIORS-MSH6.txt) specifies: for MSH6 missense substitutions, look up the exact coding change and apply HCI prior probability thresholds >0.81=PP3_Moderate, >0.68 and <=0.81=PP3_Supporting, <0.11=BP4_Supporting.
• cspec MSH6 v2.0 PP3 rule: 'Missense variant with HCI prior probability for pathogenicity >0.68 & <=0.81 ... OR Predicted splice defect for non-canonical splicing nucleotides using SpliceAI with delta score >= 0.2 as per Walker et al 2023.'
• Local HCI-prior lookup (LOVD_MSH6_priors file) for c.3674C>T / p.T1225M returned probability=0.7336, mapp_score=29.5, custom_pp2_score=0.594 -- probability falls in the >0.68 & <=0.81 PP3_Supporting band.
• REVEL (0.819) and BayesDel (0.356297) scores are present in prefetch data but are NOT used here because the governing MSH6 VCEP specification designates the HCI-prior tool, not REVEL/BayesDel, as the calibrated missense in-silico predictor for this gene; BayesDel additionally has no verified published threshold available to this pipeline regardless of gene.

Sources

Reasoning

One endometrial tumor with the exact germline variant had 3/5 microsatellite markers unstable, meeting the MSH6 VCEP PP4_Supporting requirement for one CRC/endometrial MSI-H tumor assessed with a standard 5–10 marker panel.

Rule

ClinGen InSiGHT MSH6 v2.0 PP4_Supporting: 1 CRC/endometrial MSI-H tumor using a standard panel of 5–10 markers or tumor genome and/or loss of MMR protein expression consistent with variant location.

Evidence considered

• The MSH6 VCEP PP4_Supporting rule requires one CRC/endometrial MSI-H tumor tested using a standard panel of 5–10 markers, or loss of MMR protein expression consistent with variant location.
• In an unselected endometrial cancer cohort, the exact-variant case 1038 was recorded as ‘1038 61 White N N 3/5 MSH6 c.3674C>T p.Thr1225Met UV ND’; the study performed tumor microsatellite instability testing, and 3/5 markers were unstable.

Sources

Reasoning

PMID:22102614 directly tested MSH6 T1225M (c.3674C>T) in a validated cell-free in vitro mismatch-repair complementation assay (variant MSH6 heterodimerized with wild-type MSH2, added to MSH2/MSH6-deficient LoVo nuclear extract, G.T mismatch-repair substrate) and reported the variant as repair proficient alongside all other tested MSH6 VUS in that series. This is variant-specific proficient function in a protein-based lab assay, matching the InSiGHT MSH6 VCEP BS3 Supporting rule ('Variant-specific proficient function in protein and mRNA-based lab assays as per MMR functional assay flowchart'). No calibrated functional-odds-for-pathogenicity value was available in the bundle to formally place this assay on the flowchart's odds scale, so strength is capped at Supporting (the categorical proficient-function tier) rather than Strong.

Rule

InSiGHT MMR VCEP cspec v2.0 BS3: Calibrated functional assays with odds for pathogenicity <=0.05 (Strong) OR variant-specific proficient function in protein/mRNA-based lab assays per MMR functional assay flowchart (Supporting); >0.05-0.48 (Supporting).

Evidence considered

• PMID:22102614: 'Surprisingly, all tested MSH6 VUS were repair proficient' (Table 1; Results, 'MMR Activities of MSH2 and MSH6 VUS'; Discussion) - MSH6 T1225M is explicitly listed among the 20 tested variants with positive IHC and repair-proficient result in the cell-free complementation assay.
• InSiGHT MSH6 VCEP cspec v2.0 BS3 rule: Supporting strength applies for 'Variant-specific proficient function in protein and mRNA-based lab assays as per MMR functional assay flowchart'; Strong strength requires calibrated functional odds for pathogenicity <=0.05, which was not available for this assay/variant in the bundle.

Still missing

• The VCEP 'Functional-assay-flowchart.pdf' and 'Functional-assay-SVI-documentation-MMR.xlsx' contents (which formally classify which assays/publications qualify and at what calibration tier) were not extracted as readable text in this bundle, so the specific placement of the PMID:22102614 assay on the flowchart could not be independently confirmed beyond the qualitative 'repair proficient' finding.
• No independent mRNA-based (splicing/expression) assay data for this variant was found to corroborate the protein-based result.

Sources

Reasoning

The exact variant was reported in a parent and child, each with an MSI-negative colorectal tumor at age 49, providing two independent CRC tumors with a phenotype inconsistent with MSH6-associated mismatch-repair deficiency and meeting MSH6 VCEP BP5_Supporting.

Rule

ClinGen InSiGHT MSH6 v2.0 BP5_Supporting: 2 or 3 CRC/endometrial tumors with MSS and/or no loss of MMR protein expression and/or LS-spectrum tumors with MMR-protein loss inconsistent with the altered gene.

Evidence considered

• The MSH6 VCEP BP5_Supporting rule requires 2 or 3 CRC/endometrial tumors with MSS and/or no loss of MMR protein expression, or an LS-spectrum tumor with MMR-protein loss inconsistent with the altered gene.
• The exact-variant report states: ‘The missense MSH6 mutation (c.3674C>T), not found in control subjects, was found in family 487, in which both parent and child had MSI-negative colon tumors at age 49 years.’ These are two independent colorectal tumors; MSI-negative is inconsistent with the MMR-deficiency tumor phenotype required for PP4.

Sources

Reasoning

The only variant-specific functional assay in the evidence bundle (PMID:22102614, a cell-free in vitro mismatch-repair complementation assay using in vitro-expressed variant MSH6 heterodimerized with wild-type MSH2 added to MSH2/MSH6-deficient LoVo nuclear extract, tested against a G.T mismatch substrate) explicitly tested MSH6 T1225M (c.3674C>T) among 20 MSH6 VUS and found it repair proficient, i.e. showing no damaging effect on MMR function. This is the opposite of the damaging effect required for PS3. No calibrated odds-for-pathogenicity value for this specific assay/variant, and no monoallelic-expression loss-of-expression data, were available in the bundle to otherwise satisfy the MSH6 VCEP PS3 rule (which requires calibrated functional odds >2.08, MMR function defect per the flowchart, or complete loss of expression). PS3 is therefore not met.

Rule

InSiGHT MMR VCEP cspec v2.0 PS3: Calibrated functional assays with functional odds for pathogenicity >2.08 (Supporting) / >4.3-18.7 (Moderate, or MMR function defect per flowchart, or monoallelic loss of expression <10% WT) / >18.7 (Strong).

Evidence considered

• PMID:22102614 (cell-free in vitro MMR complementation assay) explicitly lists MSH6 T1225M (c.3674C>T) among 20 tested MSH6 variants and reports it as repair proficient: 'Surprisingly, all tested MSH6 VUS were repair proficient' (Table 1; Results, 'MMR Activities of MSH2 and MSH6 VUS'; Discussion).
• InSiGHT MSH6 VCEP cspec v2.0 PS3 rule requires calibrated functional odds for pathogenicity >2.08 (Supporting), >4.3 (Moderate/MMR function defect per flowchart), or >18.7 (Strong) for a damaging-effect functional criterion; a repair-proficient result does not meet any tier of this rule.

Still missing

• No calibrated functional-odds-for-pathogenicity value specific to this assay/variant was available in the bundle (the VCEP 'Functional-assay-SVI-documentation-MMR.xlsx' calibration table content was not extracted as text) to formally cross-check strength thresholds.
• No mRNA/monoallelic expression assay data for this variant was found in the bundle.

Sources

Reasoning

The applicable ClinGen InSiGHT MSH6 VCEP PM2 rule assigns Supporting strength only when the allele frequency is less than 0.00002 in gnomAD v4. The variant is present in gnomAD v4.1 at total AF 0.0000576232 (93/1,613,932 alleles), with Grpmax FAF 0.00005137 and no homozygotes. These frequencies exceed the VCEP PM2 threshold, so PM2 is not met. The absence from controls reported in an older exact-variant study does not replace the VCEP-specified gnomAD v4 requirement.

Rule

Absent/extremely rare allele frequency <0.00002 (<1 in 50,000 alleles) in gnomAD v4 dataset

Evidence considered

• The MSH6 VCEP PM2 rule is absent/extremely rare allele frequency <0.00002 (<1 in 50,000 alleles) in the gnomAD v4 dataset.
• gnomAD v4.1 reports 93/1,613,932 alleles, total AF 5.762324558903349e-05, Grpmax FAF 5.1369999999999984e-05, and zero homozygotes; the South Asian population has the highest broad population AF at 9.882290934645116e-05.
• PMID:17312306 reports the exact variant as not found in the study's control subjects, but the applicable VCEP rule specifically requires gnomAD v4 AF <0.00002, which is not satisfied.
• The gnomAD v4 frequency is above the VCEP PM2 cutoff, so PM2 is not met.

Sources

Reasoning

The applicable ClinGen InSiGHT MSH6 VCEP BA1 rule requires a gnomAD v4 Grpmax filtering allele frequency of at least 0.0022 (0.22%) and exclusion as a founder pathogenic variant. gnomAD v4.1 reports a Grpmax FAF of 0.00005137 (0.005137%), with total AF 0.0000576232; therefore the frequency threshold is not met. No founder exception is needed to reach this conclusion.

Rule

GnomAD v4 Grpmax filtering allele frequency >= 0.0022 (0.22%) and variant is excluded as founder pathogenic variant.

Evidence considered

• The MSH6 VCEP BA1 rule is gnomAD v4 Grpmax filtering allele frequency >= 0.0022 (0.22%) and exclusion as a founder pathogenic variant.
• gnomAD v4.1 reports 93/1,613,932 alleles, total AF 5.762324558903349e-05, no homozygotes, and Grpmax FAF 5.1369999999999984e-05; the highest broad population AF is South Asian 9.882290934645116e-05.
• The observed gnomAD v4 frequency is substantially below the VCEP BA1 threshold, so BA1 is not met.

Sources

Reasoning

The applicable ClinGen InSiGHT MSH6 VCEP BS1 rule requires a gnomAD v4 Grpmax filtering allele frequency of at least 0.00022 and less than 0.0022 (0.022-0.22%), with the variant excluded as a founder pathogenic variant. The gnomAD v4.1 Grpmax FAF is 0.00005137 (0.005137%), below the lower threshold; BS1 is not met.

Rule

GnomAD v4 Grpmax filtering allele frequency >= 0.00022 and < 0.0022 (0.022-0.22%) and variant is excluded as founder pathogenic variant.

Evidence considered

• The MSH6 VCEP BS1 rule is gnomAD v4 Grpmax filtering allele frequency >= 0.00022 and < 0.0022 (0.022-0.22%) and exclusion as a founder pathogenic variant.
• gnomAD v4.1 reports Grpmax FAF 5.1369999999999984e-05 (0.005137%), below 0.00022; the highest broad population AF is 9.882290934645116e-05 in South Asian samples.
• Because the VCEP lower BS1 frequency threshold is not reached, BS1 is not met.

Sources

Reasoning

The MSH6 VCEP BP4 missense rule requires HCI prior probability <0.11 for Supporting; the observed HCI prior for c.3674C>T is 0.7336, well above this threshold, so the missense BP4 path is not met. The alternate VCEP BP4 path (SpliceAI predicts no splicing impact, delta <=0.1) applies only to intronic and synonymous variants per the cspec rule text; this variant is an exonic missense substitution, so that path is not applicable here, and in any case the SpliceAI signal (max delta 0.00) would not be usable to support BP4 for a missense change under this VCEP's own rule scope.

Rule

InSiGHT MSH6 VCEP v2.0 BP4: HCI prior probability <0.11 required for Supporting (missense); not met (0.7336)

Evidence considered

• Local HCI-prior lookup for c.3674C>T returned probability=0.7336, which is far above the VCEP BP4_Supporting threshold of <0.11, so BP4 is not met via the missense path.
• cspec MSH6 v2.0 BP4 rule text restricts the SpliceAI-based no-splice-impact path to intronic and synonymous variants specifically; this variant is missense/exonic so that alternate path does not apply.
• SpliceAI evidence (max delta score = 0.00, per prefetch) is noted for completeness but is outside this VCEP's defined BP4 missense scope and is not used to support BP4 here.

Sources

Reasoning

MSH6 VCEP (InSiGHT ClinGen CSPEC v2.0) defines PS1 as: a predicted missense substitution encoding the same amino acid change (p.(Thr1225Met)) via a different nucleotide change previously classified Pathogenic by this VCEP (Strong), or as Likely Pathogenic with normal RNA result plus PM2_Supporting met (Moderate). The automated comparator search (pm5_candidates.json, which also covers same-position lookups) returned zero same-residue candidates and logged an HTTP 429 rate-limit error during ClinVar querying, so it did not reliably enumerate whether an alternate p.(Thr1225Met)-encoding nucleotide change exists with a VCEP Pathogenic/Likely Pathogenic classification. No comparator variant encoding Thr1225Met was located in the reviewed sources.

Rule

MSH6 VCEP CSPEC v2.0 PS1: same amino acid change via different nucleotide substitution, previously classified Pathogenic (Strong) or Likely Pathogenic with normal RNA + PM2_Supporting (Moderate) by this VCEP.

Evidence considered

• pm5_candidates.json reports total_candidates_seen=11, same_residue_candidates=0, with warning 'PM5 candidate harvesting failed softly' and error 'HTTP Error 429: Too Many Requests', indicating an incomplete rather than a confirmed-negative search.

Still missing

• Automated same-amino-acid-change comparator search failed (HTTP 429) and did not complete reliably; a manual ClinVar/VCEP curated-variant list search for other nucleotide changes producing p.(Thr1225Met) is needed before PS1 can be adjudicated.

Sources

Reasoning

The applicable InSiGHT MSH6 specification requires documented de novo status in a proband and assigns points according to parental confirmation and tumor evidence. The available case evidence does not document a proband with confirmed maternity and paternity, an assumed de novo event, or a qualifying de novo tumor context for this variant.

Rule

InSiGHT MSH6 PS2: 2 de novo points per proband with both maternity and paternity confirmed and an MMR-deficient Lynch-spectrum tumor; 1 point with both parents confirmed but no tumor data; 0.5 points for an assumed de novo variant with maternity and/or paternity unconfirmed. Strength: supporting at 0.5 points, moderate at 1–1.5, strong at 2–3.5, very strong at ≥4.

Evidence considered

• The MSH6 InSiGHT specification defines PS2 as applicable only for a proband with a de novo variant; confirmed maternity and paternity with an MMR-deficient Lynch-spectrum tumor earns 2 points, confirmed maternity and paternity without tumor data earns 1 point, and an assumed de novo variant with unconfirmed maternity and/or paternity earns 0.5 points.
• The consolidated case evidence contains no documented parental testing, de novo determination, or qualifying proband tumor evidence for NM_000179.3:c.3674C>T (p.Thr1225Met).

Still missing

• Documented proband-level parental testing and phase/de novo status are unavailable.
• No case-specific tumor MSI/IHC/methylation data are available for applying the MSH6 PS2 point rules.

Sources

Reasoning

The applicable ClinGen InSiGHT MSH6 VCEP framework requires co-occurrence of the variant with a known pathogenic or likely pathogenic sequence variant in the same gene in a patient with clinical features consistent with CMMRD. It assigns 1.0 point for a confirmed pathogenic/likely pathogenic variant in trans and 0.5 points when phase is unknown. The reviewed case and variant-specific publications report MSH6 c.3674C>T/p.Thr1225Met in affected or cohort-observed individuals, but do not document a second pathogenic/likely pathogenic MSH6 variant, its zygosity, phase, or CMMRD-consistent clinical features. Therefore no PM3 points can be assigned from the available evidence.

Rule

Co-occurrence with a known pathogenic/likely pathogenic sequence variant in the same gene in a patient with clinical features consistent with CMMRD: pathogenic/likely pathogenic in trans = 1.0 point; phase unknown = 0.5 points; sum points across cases. Strength: supporting at 0.5 points, moderate at 1-1.5 points, strong at 2-3.5 points, very strong at >=4 points.

Evidence considered

• The InSiGHT MSH6 VCEP specifies PM3 for co-occurrence with a known pathogenic/likely pathogenic variant in the same gene in a patient with clinical features consistent with CMMRD; confirmed in-trans phase is 1.0 point and unknown phase is 0.5 points.
• PMID:17312306 reports c.3674C>T in family 487, where a parent and child had MSI-negative colon tumors at age 49, but states that segregation analysis was not possible and considered the variant of unclear biologic relevance; no second pathogenic/likely pathogenic MSH6 variant or phase is reported.
• PMID:16885385 lists c.3674C>T/p.Thr1225Met in a 61-year-old endometrial-cancer case with 3/5 unstable markers as an unclassified variant; no second MSH6 variant or phase information is provided.
• PMID:18566915 lists c.3674C>T/p.Thr1225Met in three Danish Lynch-syndrome families among MSH6 alterations of unknown clinical significance; no variant-specific segregation, second pathogenic/likely pathogenic allele, or phase is reported.

Still missing

• A documented pathogenic or likely pathogenic variant in MSH6 in the same individual is needed.
• Parental or offspring testing, or other validated phasing evidence, is needed to establish whether the variants are in trans.
• Clinical evidence consistent with CMMRD must be documented for the VCEP PM3 rule.

Sources

Reasoning

MSH6 VCEP CSPEC v2.0 PM5 requires a different missense change at the same residue (Thr1225) previously classified Pathogenic (Moderate) or Likely Pathogenic (Supporting) at the protein level by this VCEP, and additionally requires PP3 to be met/supporting for the current variant before PM5 can be applied. The automated same-residue comparator search (pm5_candidates.json) found 0 same-residue candidates among 11 candidates seen, but the search logged an HTTP 429 rate-limit error and was flagged as having 'failed softly,' so absence of a comparator cannot be confirmed. No alternate Thr1225 missense change with a VCEP classification was identified in the reviewed sources (paper_extracts.json papers discuss only p.Thr1225Met itself, not other residue-1225 substitutions).

Rule

MSH6 VCEP CSPEC v2.0 PM5: different missense change at same residue classified Pathogenic (Moderate) or Likely Pathogenic (Supporting) by this VCEP at the protein level, contingent on PP3 being met for the current variant.

Evidence considered

• pm5_candidates.json: same_residue_candidates=0 of 11 total_candidates_seen, with explicit warning that harvesting 'failed softly' due to 'HTTP Error 429: Too Many Requests', and recommendation.level='no_useful_candidates_found' instructing not to apply PM5 from this artifact alone.

Still missing

• Automated same-residue comparator search was rate-limited (HTTP 429) and did not reliably confirm absence of other pathogenic/likely pathogenic missense changes at residue 1225; needs manual ClinVar/VCEP curated-variant-list confirmation.
• Even if a comparator were found, this group cannot independently confirm PP3 status (a VCEP prerequisite for applying PM5) since PP3 is adjudicated in a different criteria group.

Sources

Reasoning

The MSH6-specific PP1 rule requires co-segregation in pedigrees quantified by a combined Bayes likelihood ratio: supporting >2.08–4.3, moderate >4.3–18.7, or strong >18.7 in at least two families for the strong tier. The exact-variant literature reports affected relatives in one family but explicitly states that segregation analysis was not possible; no pedigree-level Bayes likelihood ratio is available.

Rule

InSiGHT MSH6 PP1: co-segregation quantified by combined Bayes likelihood ratio; supporting >2.08–4.3, moderate >4.3–18.7, strong >18.7 in ≥2 families.

Evidence considered

• The InSiGHT MSH6 PP1 rule requires an overall Bayes factor from segregation analysis; supporting requires a combined Bayes likelihood ratio >2.08 and ≤4.3, moderate requires >4.3 and ≤18.7, and strong requires >18.7 in at least two families.
• PMID:17312306 reports c.3674C>T in family 487, where a parent and child both had colon tumors at age 49 years, but explicitly states that segregation analysis was not possible and the variant had unclear biologic relevance.
• PMID:16885385 lists the exact variant in one endometrial-cancer case but provides no parental testing, pedigree segregation, or meiosis analysis.
• PMID:18566915 reports the exact variant in three Danish families as an alteration of unknown clinical significance but provides no variant-specific segregation result or Bayes likelihood ratio.

Still missing

• A complete pedigree with genotype and phenotype status for informative relatives is unavailable.
• No combined Bayes likelihood ratio or COOL analysis is available.
• The reported family observation is insufficient to establish co-segregation because segregation analysis was not performed or was not possible.

Sources

Reasoning

The applicable MSH6 VCEP BS2 rule requires confirmed in-trans co-occurrence with a known pathogenic variant in the same gene in a patient meeting the specified age and clinical conditions, including no evidence of CMMRD, with phase confirmed through parents or offspring. The case bundle does not provide this co-occurrence, phase, or CMMRD-exclusion evidence.

Rule

Co-occurrence in trans with a known pathogenic sequence variant in the same gene in a patient with colorectal cancer after age 45 (or other LS cancer above the median age of onset for that cancer in LS), and who has no previous or current evidence of clinical manifestations of CMMRD. Confirmation of phase requires testing of parents or offspring.

Evidence considered

• The MSH6 VCEP BS2 rule requires co-occurrence in trans with a known pathogenic sequence variant in the same gene in a patient with colorectal cancer after age 45 or another Lynch-syndrome cancer above the median age of onset, no prior or current evidence of CMMRD, and phase confirmation through parents or offspring.
• The exact-variant literature extraction for PMID:17312306 describes family 487 with the variant in a parent and child with MSI-negative colon tumors at age 49, but states that segregation analysis was not possible; it does not establish the required in-trans co-occurrence or phase.
• No qualifying BS2 observation is available in the case evidence bundle.

Still missing

• Need a documented MSH6 pathogenic-variant co-occurrence in trans with c.3674C>T in an appropriately aged patient.
• Need parental or offspring testing confirming phase.
• Need documentation that the patient meets the VCEP age/clinical requirements and has no clinical manifestations of CMMRD.

Sources

Reasoning

No affected individuals documented to lack the variant, no informative non-segregating pedigree, and no Bayes likelihood ratio for lack of co-segregation are available. The report that segregation analysis was not possible is not itself evidence of non-segregation.

Rule

InSiGHT MSH6 BS4: lack of co-segregation with combined Bayes likelihood ratio <0.05 is strong; >0.05 and ≤0.48 is supporting.

Evidence considered

• The InSiGHT MSH6 BS4 rule requires lack of co-segregation quantified by a combined Bayes likelihood ratio <0.05 for strong or >0.05 and ≤0.48 for supporting evidence.
• PMID:17312306 states that segregation analysis was not possible; it does not report an affected non-carrier or a quantified lack of co-segregation.

Still missing

• No genotype-negative affected relatives or other informative non-segregation observations are documented.
• No combined Bayes likelihood ratio for lack of co-segregation is available.

Sources