MSH2 c.1571G>A, p.(Arg524His)

Reasoning

Governing specification is the ClinGen InSiGHT MSH2 Specification v2.0, whose PM2 rule is: 'Absent/extremely rare allele frequency < 0.00002 (< 1 in 50,000 alleles) in gnomAD v4 dataset', applied at Supporting strength (the VCEP lists the rule at Supporting strength, and its own pilot classifications consistently apply PM2_supporting). The VCEP names gnomAD v4 as the source, so the all-comers gnomAD v4.1 entry governs: total AF 1.36322e-05 (22/1,613,824 alleles) and joint grpmax filtering AF 1.03e-05, both below 0.00002 — corresponding to about 1 in 73,000 alleles overall. Exome-only (1.36828e-05) and genome-only (1.31465e-05) components are both below the threshold as well. Concordant checks: gnomAD v2.1 AF 1.06093e-05 (grpmax FAF 9.58e-06), gnomAD v2.1 non-cancer exomes 8.44431e-06, gnomAD v3.1 non-cancer genomes 1.35237e-05 and gnomAD-Canada v1.0 1.08566e-04 as well as every gnomAD v4.1 subpopulation (max European Finnish 3.12402e-05) remain under 0.00002 except the small gnomAD-Canada cohort. Homozygotes are zero in every dataset. The variant therefore qualifies as absent/extremely rare at the VCEP gnomAD v4 threshold, and PM2 is met at Supporting strength.

Rule

VCEP InSiGHT MSH2 v2.0 PM2 (Supporting): Absent/extremely rare allele frequency < 0.00002 (< 1 in 50,000 alleles) in gnomAD v4 dataset.

Evidence considered

• Governing framework: ClinGen InSiGHT MSH2 Specification v2.0; PM2 rule is absent/extremely rare allele frequency < 0.00002 (< 1 in 50,000 alleles) in the gnomAD v4 dataset, listed at Supporting strength (the general ACMG PM2 threshold of <= 0.0001 was not used because the VCEP specification takes precedence).
• gnomAD v4.1 (default all-comers source per the VCEP designation; GRCh38 chr2-47466718-G-A): total AF 1.36322e-05 (22/1,613,824 alleles), exome AF 1.36828e-05 (20/1,461,692), genome AF 1.31465e-05 (2/152,132), joint grpmax filtering AF 1.03e-05, 0 homozygotes.
• gnomAD v4.1 subpopulation detail: highest frequency is European (Finnish) 3.12402e-05 (2/64,020 alleles, 0 homozygotes); Admixed American 1.66678e-05 (1/59,996); European non-Finnish, African, East Asian, South Asian, Middle Eastern, Ashkenazi Jewish, Amish and Remaining individuals all 0.
• gnomAD v2.1: total AF 1.06093e-05 (3/282,770 alleles, 0 homozygotes), grpmax FAF 9.58e-06, highest subpopulation Admixed American 5.64366e-05.
• Non-cancer subsets reviewed as instructed: gnomAD v2.1 non-cancer exomes AF 8.44431e-06 (2/236,846 alleles) and gnomAD v3.1 non-cancer genomes AF 1.35237e-05 (2/147,888 alleles, grpmax FAF 4.88e-06, 0 homozygotes) — both remain below 0.00002, so the VCEP-specified all-comers gnomAD v4 result is not challenged by the non-cancer-only sources.
• gnomAD-Canada v1.0 (HostSeq genomes, 10,487 genomes): total AF 1.08566e-04 (2/18,422 alleles), European non-Finnish 1.70357e-04 (2/11,740 alleles), 0 homozygotes; this 2-allele estimate exceeds the VCEP PM2 threshold in a small non-gnomAD cohort but does not affect the VCEP rule, which is defined on the gnomAD v4 dataset.
• No homozygotes for this variant were reported in any dataset, which is the population observation relevant to a dominant high-penetrance mismatch-repair gene.
• The VCEP pilot table (VCEP-pilot-variants---MMR.xlsx) was searched for this variant and has no row for it, so no VCEP precedent PM2 assignment exists; the pilot table does show that the VCEP routinely applies PM2_supporting (e.g. for MSH2 c.1807G>A, c.1862G>T, c.929T>G) consistent with the Supporting-strength rule text.
• No fetched full text (PMID:11598466, PMID:15604628, PMID:22167527, PMID:23760103, PMID:25394175, PMID:25741868, PMID:27363726) reports this variant or adds variant-specific population-frequency data; none was cited for PM2.

Sources

Reasoning

The governing rule is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification to the ACMG/AMP guidelines for MSH2 v2.0: PM5_Moderate requires 'a different missense change at an amino acid residue where a different missense change was classified by this VCEP as Pathogenic on the protein level and not due to aberrant splicing', PM5 is to be downgraded to PM5_Supporting when the other variant is Likely Pathogenic due to a missense alteration, and either way the rule may only be used if PP3 is supporting for the missense change under assessment. Precondition, PP3: satisfied - the variant's VCEP-declared HCI prior probability for pathogenicity is 0.7525 (row c.1571G>A / p.R524H, DBID MSH2_03461 in HCI-PRIORS-MSH2.txt), which the same specification places in the PP3_Supporting band (>0.68 and <=0.81) rather than PP3_Moderate (>0.81). Comparator, same residue: Arg524 is a residue at which a different missense change has been classified pathologically by the expert panel - NM_000251.3:c.1571G>C (p.Arg524Pro) is classified 'Likely pathogenic for Lynch syndrome', review status 'Reviewed by expert panel', based on the June 2019 expert-panel submission by the International Society for Gastrointestinal Hereditary Tumours (InSiGHT, Guidelines v2.4) (ClinVar VCV000001759, SCV000107214.3); InSiGHT is the expert panel whose MSH2 v2.0 specification governs this case, and its MSH2 functional-domain document in this case's VCEP file set is itself 'adapted from InSiGHT criteria v2.4'. The InSiGHT submission's recorded basis is 'Abrogated function & 2 MSI-H tumours', i.e. a protein-level (functional assay plus tumour) determination, and c.1571G>C alters the second base of codon 524 rather than a splice junction, so the comparator is not a splice-defect classification and the 'not due to aberrant splicing' condition is met. Because the comparator is Likely Pathogenic rather than Pathogenic, the specification directs the Supporting weight (PM5_Supporting) rather than PM5_Moderate, and it is PM5_Supporting that is applied here. No residue-524 variant upgrades this to Moderate: in ClinVar, c.1570C>T (p.Arg524Cys, VCV000421654) and c.1570C>G (p.Arg524Gly, VCV001775501) are 'Conflicting classifications of pathogenicity' from single submitters, c.1571G>T (p.Arg524Leu, VCV000090701) is a Variant of uncertain significance, c.1570del (p.Arg524fs) is a Pathogenic truncating variant that cannot serve as a missense comparator, and c.1572T>A (p.Arg524=) is synonymous; no residue-524 variant appears in the InSiGHT VCEP pilot variant table. The ClinVar record for c.1571G>A itself (VCV000182564) has no expert-panel submission, so this PM5 call rests on the residue-level comparator, which is exactly what the criterion is designed for. PM5 is met at Supporting strength.

Rule

InSiGHT MMR VCEP MSH2 v2.0, PM5 (Supporting): 'Missense change at an amino acid residue where a different missense change was classified as Likely Pathogenic on the protein level and not due to aberrant splicing. Only use PM5_Supporting if PP3 is supporting for the missense change.' Applied value: same residue Arg524, p.Arg524Pro / c.1571G>C classified Likely pathogenic by expert panel (InSiGHT, ClinVar VCV000001759); PP3 supporting for c.1571G>A (HCI prior 0.7525).

Evidence considered

• Governing rule applied verbatim from the InSiGHT MMR VCEP MSH2 v2.0 specification for PM5 - Moderate: 'Missense change at an amino acid residue where a different missense change was classified by this VCEP as Pathogenic on the protein level and not due to aberrant splicing. Only use PM5 if PP3 is supporting for the missense change. Use PM5_Supporting if other variant is Likely Pathogenic due to a missense alteration.' Supporting: 'Missense change at an amino acid residue where a different missense change was classified as Likely Pathogenic on the protein level and not due to aberrant splicing. Only use PM5_Supporting if PP3 is supporting for the missense change.'
• PP3 precondition satisfied: HCI-PRIORS-MSH2.txt (the file the VCEP index declares for MSH2 missense prior probabilities) contains the exact row for this variant - '10' 'c.1571G>A' '-' 'p.R524H' '0.999' '14.41' '0.7525' '{PMID22949387:Thompson et al., 2013}' 'DNA' 'SEQ' 'MSH2_03461' - and the case evidence carries the same value as the HCI prior probability (0.7525), which falls in the VCEP's PP3_Supporting band (>0.68 and <=0.81).
• Comparator classification (direct ClinVar lookup, since the case's own same-residue harvest failed): ClinVar VCV000001759 = NM_000251.3(MSH2):c.1571G>C (p.Arg524Pro), 'Likely pathogenic for Lynch syndrome', review status 'Reviewed by expert panel'; the record page states the classification 'is based on the expert panel submission Jun 2019 by International Society for Gastrointestinal Hereditary Tumours (InSiGHT)', 'Guidelines v2.4', submitter accession SCV000107214.3, with the submitter comment 'Abrogated function & 2 MSI-H tumours'. Genomic position of c.1571G>C (chr2:47466718, GRCh38 / 47693857, GRCh37) is the same nucleotide as this case's c.1571G>A, i.e. the second base of codon 524, not a splice junction.
• Same-residue comparators checked for an upgrade to PM5_Moderate: ClinVar esummary for MSH2 records at codon 524 returned c.1570C>T (p.Arg524Cys, VCV000421654) conflicting classifications; c.1570C>G (p.Arg524Gly, VCV001775501) conflicting classifications; c.1571G>T (p.Arg524Leu, VCV000090701) Uncertain significance (multiple submitters, no conflicts); c.1570del (p.Arg524fs, VCV000422243) Pathogenic but truncating, not a missense comparator; c.1572T>A (p.Arg524=, VCV001153022) Likely benign and synonymous. None is a Pathogenic missense classification by the expert panel, so the Supporting weight applies.
• Searched the InSiGHT VCEP pilot variant table (VCEP-pilot-variants---MMR.xlsx, full text) for '524', 'R524', 'Arg524': no residue-524 variant is listed, so the VCEP's own pilot precedent does not supply a residue-524 classification and does not contradict the InSiGHT expert-panel record for p.Arg524Pro.
• The case's automated comparator artifact (pm5_candidates.json) is not usable: the ClinVar harvest for 'MSH2[GENE] AND R524[VARNAME]' failed softly with 'HTTP Error 429: Too Many Requests', returning 0 candidates and the recommendation 'no_useful_candidates_found'; the same-residue evidence was therefore obtained by direct ClinVar queries (E-utilities esearch 'MSH2[gene] AND 524[var_name]' and 'MSH2[gene] AND Arg524[All Fields]', esummary, and the VCV000001759 record page) rather than from that artifact.
• ClinVar aggregate for this variant itself (VCV000182564, variation 182564, 'Conflicting classifications of pathogenicity', 13-14 single submitters, 0 expert-panel submissions) was consulted only to confirm that p.Arg524His has no expert-panel classification of its own; no classification of this variant was inferred into PM5, which rests solely on the residue-level comparator.

Sources

Reasoning

Scope: NM_000251.3:c.1571G>A encodes a missense substitution, NP_000242.1:p.(Arg524His), so PP3 is in scope and the applicable computational path is the protein-impact path, not the splice path. The governing specification is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification to the ACMG/AMP guidelines for MSH2 v2.0 (CSPEC, captured in the case bundle), which for missense variants defines PP3 solely by the HCI prior probability for pathogenicity: PP3_Moderate if prior >0.81, PP3_Supporting if prior >0.68 and <=0.81 (hci-priors.hci.utah.edu/PRIORS). The gene-specific file declared by this VCEP index as governing PP3/BP4 for MSH2 missense substitutions, HCI-PRIORS-MSH2.txt, was searched for every notation of this variant (c.1571G>A, p.R524H, p.Arg524His); the exported LOVD lookup table contains an exact row: exon 10, DNA c.1571G>A, protein p.R524H, DBID MSH2_03461, reference PMID22949387 (Thompson et al., 2013), with the prior-probability column value 0.7525 (accompanying row values: Custom_PP2_score 0.999, MAPP_score 14.41). This prior of 0.7525 falls in the >0.68 and <=0.81 band, which the specification assigns PP3_Supporting. The pre-assigned VCEP code therefore governs and is applied directly at Supporting strength. SpliceAI (max delta 0.038) was not consulted for PP3 because a missense variant takes its PP3/BP4 evidence from the protein-impact path only, and no splice prediction may be used as support for a missense substitution. A REVEL score of 0.918 is present in the evidence, but the governing gene-specific specification supersedes the generic REVEL calibration for MSH2 missense variants; no second predictor was counted toward this criterion.

Rule

InSiGHT MMR VCEP MSH2 v2.0, PP3 (Pathogenic Supporting): 'Missense variant with HCI prior probability for pathogenicity >0.68 & <=0.81 as per https://hci-priors.hci.utah.edu/PRIORS'. Applied value: HCI prior 0.7525 for c.1571G>A (p.R524H, DBID MSH2_03461).

Evidence considered

• Searched the VCEP-declared file /Users/gs_agent/LYFESCI_HERMES_v9/vcep_db/MSH2/fulltext/HCI-PRIORS-MSH2.txt.txt for 'c.1571G>A', 'p.R524H', 'p.Arg524His', 'Arg524', 'R524' and '1571'; exactly one matching row was found (line 3461): "10" "c.1571G>A" "-" "p.R524H" "0.999" "14.41" "0.7525" "{PMID22949387:Thompson et al., 2013}" "DNA" "SEQ" "MSH2_03461".
• The file header defines the columns as Variant/Exon, Variant/DNA, Variant/RNA, Variant/Protein, Variant/Custom_PP2_score, Variant/MAPP_score, Variant/MAPP/PP2_Prior, Variant/Reference, Variant/Detection/Template, Variant/Detection/Technique, Variant/DBID; the prior-probability value for this variant is therefore 0.7525, matching the same value independently carried in the case evidence as the HCI prior probability for c.1571G>A.
• Governing rule applied verbatim from the InSiGHT MMR VCEP MSH2 v2.0 specification for PP3: 'Missense variant with HCI prior probability for pathogenicity >0.68 & <=0.81 as per https://hci-priors.hci.utah.edu/PRIORS' (Pathogenic Supporting); the Moderate rule requires >0.81.
• Variant consequence is missense: NM_000251.3:c.1571G>A / NP_000242.1:p.(Arg524His); exon 10 of MSH2; genomic NC_000011-alternate coordinates NC_000002.11:g.47693857G>A (GRCh37) / chr2:47466718 G>A (GRCh38).
• SpliceAI max delta 0.038 was available in the evidence but deliberately not used for this missense variant's PP3, per the single-path rule; a 'no predicted splice impact' result is not PP3 or BP4 evidence for an amino-acid substitution.
• REVEL 0.918 is present in the local predictor evidence but the InSiGHT MSH2 v2.0 specification defines PP3/BP4 for missense variants by the HCI prior, which takes precedence over the generic ClinGen SVI REVEL calibration; REVEL was not independently counted.

Sources

Reasoning

Scope: NM_000251.3:c.1571G>A is a missense substitution, NP_000242.1:p.(Arg524His), so BS3 is in scope and is judged on protein-level functional evidence. Governing framework: the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for MSH2 v2.0 (CSPEC), whose BS3 rule reads: Strong = 'Calibrated functional assays with functional odds for Pathogenicity <= 0.05'; Supporting = 'Calibrated functional assays with functional odds for Pathogenicity >0.05 & <=0.48 OR Variant-specific proficient function in protein and mRNA-based lab assays as per MMR functional assay flowchart'. The two gene-specific files declared by this VCEP index as governing PS3/BS3 were searched in full for c.1571G>A, p.Arg524His, p.R524H, R524 and 1571; neither contains a variant-level entry for this variant (Functional-assay-SVI-documentation-MMR.xlsx is an assay inventory and Functional-assay-flowchart.pdf is a generic decision aid), so no pre-assigned BS3 code exists. The governing documentation does, however, approve and calibrate the assay that carries variant-level data for this variant: the Jia/Scott massively parallel MSH2 loss-of-function screen (PMID 33357406; 36550560), described there as a 'CALIBRATED ASSAY' with approval 'y', quantitative deep-sequencing readout of each MSH2 allele before and after 6-thioguanine selection in clonal HAP1 and HEK293 MSH2-knockout cells, biological and technical replicates met, basic positive control (known defective) and negative control (wild type) met, 22 pathogenic/likely pathogenic and 26 benign/likely benign validation controls, normal readout 'LOF score <= 0', abnormal readout 'LOF score > 0.4', proposed strength 'PS3, BS3'. In that assay p.Arg524His scores LOF -0.72011689603893503 (supplement Table S5 'AllVariants'), which is a normal readout under both the VCEP cutoff (<=0) and the paper's own a priori cutoff of 0: the variant is not enriched by 6-thioguanine, i.e. it retains mismatch-repair function. That is variant-specific proficient function from an assay the VCEP itself approves and lists as calibrated, so the criterion is met. Strength: the assay table's 'Proposed strength' entry for this assay is the unmodified 'PS3, BS3', which under the specification corresponds to the default strengths (BS3 default = Benign Strong), and the paper's functional map is 96% concordant with existing functional data and expert interpretations over 94% of all possible MSH2 missense variants - hence Strong. The conclusion is robust to the strength reading: on the more conservative path this same result is 'variant-specific proficient function' per the MMR functional assay flowchart, which the specification scores as BS3_Supporting, so BS3 is met either way. No conflicting functional evidence exists for this variant (no other approved assay reports a readout for it), and the benign direction of this evidence is not shored up by any pathogenic functional readout.

Rule

InSiGHT MMR VCEP MSH2 v2.0, BS3 (default Benign Strong): Strong = 'Calibrated functional assays with functional odds for Pathogenicity <= 0.05'; Supporting = 'Calibrated functional assays with functional odds for Pathogenicity >0.05 & <=0.48 OR Variant-specific proficient function in protein and mRNA-based lab assays as per MMR functional assay flowchart'. Applied value: VCEP-approved calibrated MSH2 assay (Jia et al., PMID 33357406; 36550560) LOF score for p.Arg524His = -0.72, a normal readout (normal = LOF <= 0), i.e. variant-specific proficient MMR function; the governing assay table lists this assay's proposed strengths as unmodified 'PS3, BS3', so the default (Strong) weighting is applied.

Evidence considered

• Searched both governing files with grep -rn -- 'c.1571G>A', 'p.Arg524His', 'p.R524H', 'R524' and '1571' over /Users/gs_agent/LYFESCI_HERMES_v9/vcep_db/MSH2/fulltext: neither Functional-assay-SVI-documentation-MMR.xlsx.txt nor Functional-assay-flowchart.pdf.txt contains this variant. The assay documentation was used as the governing calibration source: it is the file the VCEP index declares for PS3/BS3 with the instruction to use it 'to identify applicable calibrated assays and classification thresholds'.
• Assay used, from the governing file (column 'Jia; Scott', PMID 33357406; 36550560): approved calibrated MSH2 assay based on chemical selection for mismatch repair dysfunction plus deep sequencing of surviving MSH2 variants; clonal HAP1 and HEK293 MSH2-knockout cells transduced with a single-amino-acid saturation library; quantitative readout; biological replicates met, technical replicates met (triplicate); positive control met (known defective), negative control met (wild type); 22 P/LP and 26 B/LB validation controls; normal readout LOF score <= 0; abnormal readout LOF score > 0.4; proposed strength 'PS3, BS3'.
• Variant-level result supporting BS3: supplement mmc2.xlsx (TableS5_AllVariants) of Jia et al. 2021, Am J Hum Genet 108:163-175 (PMID 33357406, PMC7820803), retrieved from Europe PMC supplementary files during this assessment. Row R524H (residue 524): LOF score -0.72011689603893503, i.e. below zero and below the >0.4 abnormal cutoff, so a normal (proficient-function) readout. Codon-level internal control: neighbouring substitutions at the same residue do score as loss of function in the same assay (R524P +3.33, R524D +3.02, R524Y +1.08, R524G +0.96, R524L +0.45), which shows the assay discriminates at this codon and that -0.72 is a genuine negative rather than a missing or uninformative value.
• Assay calibration and reproducibility context from the same paper: functional effect map covering 94% of the 17,746 possible MSH2 missense variants, 96% concordant with existing functional data and expert clinicians' interpretations, and an a priori deleterious cutoff of LOF score 0 ('For LoF scores, a score cutoff of 0 was selected a priori, taking as deleterious any variants enriched by 6-TG'). This underpins weighting the normal readout as benign evidence rather than as noise.
• Flowchart cross-check (the specification's second, supporting-strength BS3 arm): for an in-frame/missense variant where a calibrated functional assay has been conducted, the flowchart directs converting the functional likelihood ratio to ACMG/AMP evidence weight; absent a calibrated assay it requires two independent assays with concordant results, and its 'proficient function' endpoint (repair levels similar to wild type, >=75% relative expression, nuclear localisation with DNA-damage sensitivity and no mutator phenotype) corresponds to the specification's 'variant-specific proficient function' wording. The calibrated readout here is normal, so this path also lands on BS3, at supporting strength, and therefore does not change the code - only the weighting.
• No conflicting functional data: the other approved assays in the same governing table (Drost CIMRA, Rath, four Thompson assays, Morak, Bouvet) contribute thresholds and control counts only and report no result for c.1571G>A, and the seven publications fetched into this case (PMID:11598466, PMID:15604628, PMID:22167527, PMID:23760103, PMID:25394175, PMID:25741868, PMID:27363726) contain no functional assay data for this or any individual MSH2 variant.

Still missing

• Functional-assay-SVI-documentation-MMR.xlsx (one of the two files declared by this VCEP index as governing PS3/BS3): searched for c.1571G>A, p.Arg524His, p.R524H, R524 and 1571 and found no entry for this variant. The file lists approved assays, their thresholds, control performance and proposed strengths, but no per-variant results, so it cannot supply a pre-assigned BS3 code for R524H.
• Functional-assay-flowchart.pdf (the other declared PS3/BS3 file): no entry for this variant; the PDF page has no extractable text layer and the flowchart gives generic decision paths only, so no variant-specific BS3 determination could be read from it.
• The governing assay table states this assay's thresholds in LOF-score units rather than in functional odds, so the assignment of BS3 at Strong strength rests on the table's 'Proposed strength' row (unmodified 'PS3, BS3' = the specification's default strengths) rather than on a directly tabulated odds value for this readout; the underlying variant-level value was retrieved during this assessment from the supplement of PMID 33357406 (PMC7820803) because that paper was not fetched into the case.

Sources

Reasoning

Scope: NM_000251.3:c.1571G>A encodes a missense substitution, NP_000242.1:p.(Arg524His), so PS1 is in scope and the governing text is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification to the ACMG/AMP guidelines for MSH2 v2.0 (CSPEC doc 1577628936, captured in the case evidence). The applicable PS1_Strong rule is: 'A predicted missense substitution that encodes the same amino acid change with a different underlying nucleotide change previously established by this VCEP as Pathogenic (not a predicted or confirmed splice defect)'; a second Strong branch covers variants affecting the same non-canonical splice nucleotide as a confirmed Pathogenic splice variant. Codon 524 is CGT: the ClinVar record for this exact variant carries the alias 'p.R524H:CGT>CAT' (VCV000182564), and it is independently confirmed by ClinVar VCV001153022 (c.1572T>A, p.Arg524=, Likely benign), which shows that the third base of the codon is T and that changing it is synonymous rather than histidine-encoding. Enumerating every single-nucleotide substitution in CGT codon 524: c.1570C>A/G/T give p.Arg524Ser/Gly/Cys; c.1571G>C and c.1571G>T give p.Arg524Pro and p.Arg524Leu (all present as separate rows in the VCEP-declared HCI-PRIORS-MSH2.txt substitution table); c.1572T>A/C/G are synonymous Arg. Only the observed c.1571G>A (CGT>CAT) encodes histidine, so no different underlying nucleotide change producing p.Arg524His exists at this codon - the precondition for the same-amino-acid branch of PS1 is structurally unsatisfiable for this variant, not merely unobserved. Consistent with that, no second p.Arg524His entry exists anywhere in the VCEP-governing files: HCI-PRIORS-MSH2.txt contains exactly one row for c.1571G>A/p.R524H (DBID MSH2_03461) and no c.1572 row at all, and the InSiGHT VCEP pilot variant table (VCEP-pilot-variants---MMR.xlsx) lists no residue-524 variant. The ClinVar record for this variant has no expert-panel submission (0 expert-panel submissions among 13-14 submitters), so no VCEP-established Pathogenic classification of p.Arg524His exists either. The splice-nucleotide branch of PS1 does not apply: c.1571G>A changes the second base of codon 524, is not a non-canonical splice nucleotide, and the SpliceAI max delta for this variant is 0.038. The residue-level finding that does exist - a different amino-acid change at the same residue, NM_000251.3:c.1571G>C (p.Arg524Pro), classified Likely pathogenic by expert panel review - belongs to PM5, not PS1, because PS1 requires the identical amino-acid change and PM5 is the criterion for a different change at the same codon.

Rule

InSiGHT MMR VCEP MSH2 v2.0, PS1 (Pathogenic Strong): 'A predicted missense substitution that encodes the same amino acid change with a different underlying nucleotide change previously established by this VCEP as Pathogenic (not a predicted or confirmed splice defect).' Not satisfied: no different nucleotide change encodes p.Arg524His (codon 524 CGT; only c.1571G>A gives CAT).

Evidence considered

• Governing rule applied verbatim from the InSiGHT MMR VCEP MSH2 v2.0 specification for PS1 (Pathogenic Strong): 'A predicted missense substitution that encodes the same amino acid change with a different underlying nucleotide change previously established by this VCEP as Pathogenic (not a predicted or confirmed splice defect). OR Variants affecting the same non-canonical splice nucleotide as a confirmed Pathogenic splice variant with similar or worse splicing in silico prediction using SpliceAI.' The specification's PS1 Moderate branch likewise requires a 'different underlying nucleotide change' encoding the same amino acid change.
• Reference codon established as CGT for Arg524 from two independent ClinVar records: VCV000182564 (this variant, alias 'p.R524H:CGT>CAT') and VCV001153022 (c.1572T>A, p.Arg524=, Likely benign, criteria provided by multiple submitters), the latter showing that the codon's third base is T and that a change there remains arginine.
• Complete single-nucleotide substitution inventory for codon 524 cross-checked against the VCEP-declared file /Users/gs_agent/LYFESCI_HERMES_v9/vcep_db/MSH2/fulltext/HCI-PRIORS-MSH2.txt.txt, searched for 'c.1571G>A', 'p.R524H', 'p.Arg524His', 'R524' and 'c.1572' (no c.1572 rows exist): the table's only histidine-encoding row at this codon is the exact row for this variant (line 3461: '10' 'c.1571G>A' '-' 'p.R524H' '0.999' '14.41' '0.7525' '{PMID22949387:Thompson et al., 2013}' 'DNA' 'SEQ' 'MSH2_03461'); the neighbouring rows are c.1570C>A/G/T (p.R524S/G/C) and c.1571G>C/T (p.R524P/L).
• Searched the InSiGHT VCEP pilot variant table (VCEP-pilot-variants---MMR.xlsx, converted in full to text) for '524', 'R524' and 'Arg524': no residue-524 variant of any kind appears, so there is no VCEP-established Pathogenic (or Likely Pathogenic) p.Arg524His comparator on record.
• ClinVar aggregate for this variant (VCV000182564, variation 182564) shows conflicting classifications from single submitters only, with 0 expert-panel submissions, so p.Arg524His itself has no VCEP/expert-panel Pathogenic classification.
• Splice-nucleotide branch of PS1 not applicable: c.1571G>A is the second base of codon 524 (exon 10), not a non-canonical splice nucleotide; SpliceAI max delta for this variant is 0.038 in the case evidence.
• Literature extraction set reviewed for any report of a second nucleotide change encoding p.Arg524His (paper_extracts.json): none of the extracted papers names the variant, and no alternate-nucleotide p.Arg524His report was found.

Sources

Reasoning

Scope: NM_000251.3:c.1571G>A encodes a missense substitution, NP_000242.1:p.(Arg524His), so PS3 is in scope and is judged on protein-level functional evidence (the MMR function/repaired-capacity assays), not on splice predictions. Governing framework: the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification to the ACMG/AMP guidelines for MSH2 v2.0 (CSPEC), whose PS3 rule is anchored on calibrated functional assays: Strong = functional odds for pathogenicity >18.7; Moderate = >4.3 and <=18.7, or an MMR function defect following the MMR functional assay flowchart, or complete loss of monoallelic expression (<10% of wild-type in cDNA without puromycin); Supporting = calibrated odds >2.08 and <=4.3. The two gene-specific files this VCEP index declares as governing PS3/BS3 were searched in full for every notation of this variant (c.1571G>A, p.Arg524His, p.R524H, R524, 1571): Functional-assay-SVI-documentation-MMR.xlsx contains no variant-level row at all (it is an assay inventory: PMIDs, materials, readout type, replicate/control performance, validation-control counts, normal/abnormal thresholds, approval flag and proposed strength), and Functional-assay-flowchart.pdf contains no variant-level entry either (its page yields no extractable text layer and the flowchart is variant-agnostic). No pre-assigned PS3 code for this variant therefore exists in either governing file. The assay with variant-level data is one the governing documentation itself approves: the Jia/Scott massively parallel MSH2 loss-of-function screen (PMID 33357406; 36550560), listed as 'CALIBRATED ASSAY' with approval 'y', quantitative readout (deep sequencing of each MSH2 allele before/after 6-thioguanine selection in clonal HAP1 and HEK293 MSH2-knockout cells), 22 pathogenic/likely pathogenic and 26 benign/likely benign validation controls, basic positive and negative controls met, biological and technical replicates met, and thresholds LOF score >0.4 = abnormal, LOF score <=0 = normal. In that assay the p.Arg524His variant has LOF score -0.72011689603893503 (supplement Table S5 'AllVariants'), i.e. a normal, non-loss-of-function readout: the value sits below zero, and below the >0.4 abnormal cutoff, and the paper states its own a priori deleterious cutoff is a LOF score of 0 (deleterious = enriched by 6-TG). The readout therefore demonstrates retained MMR function for this variant rather than a damaging effect, and reaches none of the PS3 odds bands (not even the >2.08 supporting band, which requires an abnormal calibrated readout). No other VCEP-approved assay (Drost CIMRA, Rath, Thompson cell-free/complementation/expression/localisation, Morak, Bouvet) reports a readout for this variant in the case evidence, and none of the seven publications fetched into this case reports a functional assay result for c.1571G>A. PS3 is therefore not met.

Rule

InSiGHT MMR VCEP MSH2 v2.0, PS3 (default Pathogenic Strong): Strong = 'Calibrated functional assays with functional odds for pathogenicity > 18.7'; Moderate = 'Calibrated functional assays with functional odds for pathogenicity >4.3 and <= 18.7 OR MMR function defect following functional assay flowchart* OR Variants with monoallelic expression: complete loss of expression (<10% of wild-type in cDNA without puromycin)'; Supporting = 'Calibrated functional odds for pathogenicity >2.08 and <= 4.3'. Applied value: VCEP-approved calibrated MSH2 assay (Jia et al., PMID 33357406; 36550560) LOF score for p.Arg524His = -0.72, i.e. a normal readout (normal = LOF <= 0; abnormal = LOF > 0.4) - no PS3 band reached.

Evidence considered

• Searched both governing files with grep -rn -- 'c.1571G>A', 'p.Arg524His', 'p.R524H', 'R524' and '1571' over /Users/gs_agent/LYFESCI_HERMES_v9/vcep_db/MSH2/fulltext. Functional-assay-SVI-documentation-MMR.xlsx.txt and Functional-assay-flowchart.pdf.txt returned no hit for this variant. The only file in that directory containing the variant is HCI-PRIORS-MSH2.txt.txt (line 3461: "c.1571G>A" ... "p.R524H" ... prior 0.7525, PMID22949387), which the VCEP index declares for PP3/BP4 only and which is an in-silico prior model, not a functional assay; it was therefore not used here.
• Assay inventory read from the governing file (columns are per assay; the relevant column is 'Jia; Scott', PMID 33357406; 36550560): CALIBRATED ASSAY, 'Chemical selection for mismatch repair dysfunction and deep sequencing to identify the surviving MSH2 variants'; material 'Clonal HAP1 MSH2 knockout cells and Clonal HEK293 MSH2 knockout cells transduced with mutant library produced by single amino acid saturation mutagenesis'; readout quantitative (deep sequencing read counts of each MSH2 allele before and after 6-thioguanine selection); biological replicates met, technical replicates met (triplicate); basic positive control met (known defective), basic negative control met (wild type); validation controls 22 P/LP and 26 B/LB; threshold for normal readout 'LOF score <= 0'; threshold for abnormal readout 'LOF score > 0.4'; approved assay 'y'; proposed strength 'PS3, BS3'.
• Variant-level result: supplement mmc2.xlsx (TableS5_AllVariants) of Jia et al. 2021, Am J Hum Genet 108:163-175 (PMID 33357406, PMC7820803), retrieved from Europe PMC supplementary files during this assessment. Row for R524H (residue 524): LOF score -0.72011689603893503, FoldX 6.24, PON_MMR 0.432, PolyPhen2 1, MAPP_MMR 10.06, REVEL 0.918, spliceAI deltaMax 0, gnomAD allele count 5, clinical classification VUS. The same table's flanking residues at codon 524 behave very differently (R524P LOF +3.33, R524D +3.02, R524Y +1.08, R524L +0.45, R524G +0.96), so the assay resolves this codon rather than being uninformative here, and R524H is the most clearly non-loss-of-function value in that set.
• Assay-wide calibration context from the same paper: the functional effect map covers 94% of the 17,746 possible MSH2 missense variants and is 96% concordant with existing functional data and expert clinicians' interpretations; the main text states 'For LoF scores, a score cutoff of 0 was selected a priori, taking as deleterious any variants enriched by 6-TG.' The VCEP's own approved-assay row uses LOF <=0 as normal and LOF >0.4 as abnormal, so -0.72 is normal readout under both the paper's and the VCEP's cutoffs.
• Threshold comparison for PS3: the governing specification's weakest PS3 band requires a calibrated functional odds for pathogenicity >2.08, and every stronger band requires a larger odds value or a flowchart-defined MMR function defect. The only calibrated readout available for this variant is in the normal range (LOF -0.72 <= 0), so no PS3 band is reached. The flowchart's missense branch leads to PS3 only down its 'Deficient function' arms (repair levels similar to a deficient cell line or pathogenic controls, <=25% relative expression, or cytoplasmic/tolerant/high-mutator phenotype); none of those applies, because the calibrated assay result is a normal readout.
• Cross-check of the remaining approved assays against this variant: the governing file's other columns (Drost CIMRA <23% abnormal for MLH1/MSH2; Rath OddsPath >2.08; four Thompson assays <25% in two independent assays / cytoplasmic / allelic loss; Morak full-length transcript; Bouvet mean survival >45.54% for MSH2) supply thresholds and controls only, and no column reports a result for c.1571G>A. The seven publications fetched into this case (PMID:11598466, PMID:15604628, PMID:22167527, PMID:23760103, PMID:25394175, PMID:25741868, PMID:27363726) were checked through their extractions and none reports a functional assay result for this variant or for any individual MSH2 variant; they are practice guidelines, an in-silico ESR/SIFT study and a clinical history-weighting algorithm.

Still missing

• Functional-assay-SVI-documentation-MMR.xlsx (one of the two files declared by this VCEP index as governing PS3/BS3): searched for c.1571G>A, p.Arg524His, p.R524H, R524 and 1571 and found no entry for this variant. That is a statement about the file's scope (it tabulates approved assays, thresholds and proposed strengths, not per-variant results), not a statement that functional evidence is absent.
• Functional-assay-flowchart.pdf (the other declared PS3/BS3 file): no entry for this variant; the PDF page has no extractable text layer and the flowchart is a generic, variant-agnostic decision aid, so no pre-assigned PS3/BS3 code for R524H could be read from it.
• No wet-lab functional readout for p.Arg524His was present among the publications fetched into this case (none of the seven fetched papers reports an assay result for this variant), and none of the case evidence files (evidence.json, prefetch.json) carries an assay readout for this variant. The LOF score used here was retrieved during this assessment from the supplement of PMID 33357406 (Jia et al. 2021, PMC7820803), which ClinVar submissions for this variant cite but which was not fetched into the case; the value should be confirmed against the VCEP's own variant record if one is later added.

Sources

Reasoning

Governing specification: ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis VCEP specifications for MSH2 v2.0 (CSPEC doc 1577628936), where PM3 is Applicable and is scored by points rather than applied at default strength. The VCEP rule is: co-occurrence of this variant with a known pathogenic/likely pathogenic sequence variant in the same gene in a patient with clinical features consistent with constitutional mismatch repair deficiency (CMMRD) as per Aronson et al. 2022 (PMID:33622763), using the CMMRD scoring table; a pathogenic/likely pathogenic partner variant confirmed in trans scores 1.0 point and one of unknown phase scores 0.5 point, with the sum mapping to =0.5 points Supporting, 1-1.5 points Moderate, 2-3.5 points Strong, and >=4 points Very Strong. No element of that rule is satisfied by the evidence available for this variant. This is a variant-level curation (NM_000251.3:c.1571G>A, p.Arg524His) with no proband report: there is no second MSH2 variant, no parental or offspring testing to establish phase, no in-trans observation, and no CMMRD-consistent clinical or tumour documentation (no MSI/IHC or malignancy age data that could populate the CMMRD table). ClinVar holds 13 submissions (11 Uncertain significance, 2 Likely benign) with no expert-panel review and none that reports a phase-resolved trans co-occurrence for this variant. None of the seven full-text papers fetched for this case names c.1571G>A or p.Arg524His, so no literature report of biallelic/trans occurrence exists to score. Zero points accrue, which is below the 0.5-point Supporting threshold, so PM3 is not met. Generic ACMG PM3 would point the same way, since it likewise requires trans with a pathogenic variant with phase confirmed by testing of parents or offspring, and MSH2-associated Lynch syndrome is inherited autosomal dominant in the great majority of probands (MSH2 is recessive only in the CMMRD setting).

Rule

CSPEC (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis VCEP) MSH2 v2.0, PM3 - Applicable, points-based: co-occurrence with a known pathogenic/likely pathogenic sequence variant in the same gene in a patient with clinical features consistent with CMMRD (Aronson et al. 2022, CMMRD diagnosis table). Pathogenic/Likely Pathogenic in trans = 1.0 point; Pathogenic/Likely Pathogenic, phase unknown = 0.5 points. Sum across cases: =0.5 points = Supporting; 1-1.5 points = Moderate; 2-3.5 points = Strong; >=4 points = Very Strong.

Evidence considered

• CSPEC MSH2 v2.0 (source key 'cspec') lists PM3 as Applicable with instructions: co-occurrence with a known pathogenic/likely pathogenic sequence variant in the same gene in a patient with clinical features consistent with CMMRD as per Aronson et al. 2022, with the CMMRD diagnosis table used to establish that the patient's features meet CMMRD; pathogenic/likely pathogenic in trans = 1.0 point, pathogenic/likely pathogenic with phase unknown = 0.5 points, summing to Supporting (=0.5), Moderate (1-1.5), Strong (2-3.5) or Very Strong (>=4) points.
• The variant is NM_000251.3:c.1571G>A, NP_000242.1:p.(Arg524His), a missense substitution in MSH2 (NM_000251.3); the case holds no proband, no tumour data and no family data, so no allelic-phase observation can be scored.
• No second MSH2 variant and no phase determination exist anywhere in the case evidence: prefetch, evidence, ClinVar submission detail and the full-text literature pass were all checked for a trans/cis observation or a CMMRD-consistent patient, and none is present.
• ClinVar record 182564 (source key 'clinvar') carries 13 submissions (11 Uncertain significance, 2 Likely benign; no expert-panel submission) and conflicting classifications; no submission reports the variant in trans with a pathogenic MSH2 allele or provides phase information, and the record's final classifications were not used to infer this criterion.
• The CMMRD diagnostic scoring table supplied with the MSH2 VCEP materials (source key 'vcep_table_for_cmmrd_diagnosis') could not be applied because the case contains no malignancy, age, immunohistochemistry or family-history data for any patient.
• All seven full-text papers fetched for this case (PMID:11598466, 15604628, 22167527, 23760103, 25394175, 25741868, 27363726) were screened in their pre-computed extractions for c.1571G>A / p.Arg524His; none of them names the variant, so no paper reports it in trans, in cis, or with phase information.
• PMID:25394175 (source key 'PMID:25394175') states that CMMRD is a recessive condition caused by biallelic mutations in the mismatch repair genes (MLH1, MSH2, MSH6, PMS2) - this defines the only inheritance context in which a trans MSH2 configuration would be expected, and no such biallelic patient is present in this case.
• PMID:25741868 (source key 'PMID:25741868') provides the generic ACMG/AMP PM3 definition (detected in trans with a pathogenic variant for a recessive disorder, requiring testing of parents or offspring to determine phase); the CSPEC MSH2 v2.0 points-based PM3 takes precedence over that generic form, and neither is satisfied here.
• No in-silico, functional, population-frequency or tumour evidence bears on the allelic-phase question, so none was used for PM3; the REVEL 0.918, HCI prior 0.7525, SpliceAI max delta 0.038 and gnomAD v4.1 AF 1.36e-05 values in this case are irrelevant to PM3.

Still missing

• No proband-level allelic data: no second (likely) pathogenic MSH2 variant, no parental or offspring testing to establish phase, and no in-trans or in-cis observation for c.1571G>A.
• No CMMRD-consistent clinical documentation (malignancy type/age, MSI/IHC, family history) that could be scored against the Aronson et al. 2022 CMMRD table; a biallelic/CMMRD proband would need to be identified before PM3 could be scored.
• None of the seven full-text papers retrieved for this case reports c.1571G>A (p.Arg524His), so no published case provides trans or phase information.

Sources

Reasoning

Governing specification: InSiGHT MSH2 v2.0 cspec, PP4 (Applicable, default strength Pathogenic Supporting). Exact rule text for the Supporting strength: '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 the variant location. MSI-H tumor with inconsistent protein expression does not meet PP4.' Moderate requires 2 independent such tumors and Strong requires >=3 independent CRC/endometrial MSI-H tumors in >=2 families; for MSH2 the IHC loss must be MSH2 and/or MSH6 loss (MutSalpha), per the cspec instructions. Applying this criterion literally: the number of documented CRC/endometrial MSI-H tumors with MSI testing on a standard 5-10 marker panel (or tumor genome) and/or MMR-protein loss consistent with MSH2/MSH6 is 0, therefore 0 >= 1? no; 0 >= 2? no; 0 >= 3? no. PP4 is not met at any strength. No tumor MSI/IHC/methylation data for any proband are present anywhere in the case materials (evidence.json, case_summary.json, prefetch.json, evidence_brief.json), and none of the seven extracted papers reports tumor MSI/IHC status for this variant. The two VCEP files declared for PP4 were searched directly and neither provides a pre-assigned PP4 code for this variant: 'VCEP-pilot-variants---MMR.xlsx' contains no MSH2 c.1571G>A row (its only 1571 entry is the unrelated PMS2 c.1571dup, p.Gly525ArgfsTer17), and 'Table-for-CMMRD-diagnosis.pdf' is a CMMRD (biallelic) malignancy point-scoring table whose own text assigns those points to PM3 ('>=3 points = CMMRD features meets PM3 criteria'), not to PP4, and which contains no entry for this variant or for heterozygous Lynch-syndrome evaluation. The criterion is therefore not met on absent tumor phenotype evidence, with the data gap recorded below.

Rule

InSiGHT MSH2 v2.0 cspec PP4: Supporting = 1 CRC/Endometrial MSI-H tumor (standard 5-10 marker panel or tumor genome) and/or MMR-protein loss consistent with variant location; Moderate = 2 independent tumors; Strong = >=3 independent tumors in >=2 families.

Evidence considered

• cspec (InSiGHT MSH2 v2.0) PP4 rules: Supporting = 1 CRC/endometrial MSI-H tumor (standard 5-10 marker panel or tumor genome) and/or MMR-protein loss consistent with variant location; Moderate = 2 such independent tumors; Strong = >=3 independent tumors in >=2 families. MSI-H with inconsistent protein expression does not meet PP4. IHC consistency for MSH2 requires MSH2 and/or MSH6 loss.
• Documented tumor count for this variant's cases = 0 (no MSI, IHC, BRAF V600E or MLH1 methylation data anywhere in evidence.json, case_summary.json, prefetch.json or evidence_brief.json); explicit comparisons: 0 >= 1? no; 0 >= 2? no; 0 >= 3? no.
• vcep_vcep_pilot_variants_mmr (VCEP-pilot-variants---MMR.xlsx): grep for c.1571G>A, p.R524H, p.Arg524His and '1571' found no MSH2 entry for this variant; the file's MSH2 PP4 precedents (e.g. c.1807G>A PP4_strong, c.1862G>T PP4_strong, c.490G>T PP4_moderate) all carry independent tumor data that is absent here, so no precedent code can be transferred to this variant.
• vcep_table_for_cmmrd_diagnosis (Table-for-CMMRD-diagnosis.pdf): searched - no entry for c.1571G>A/p.R524H. Table scores malignancies/premalignancies and CMMRD features (>=3 points); its header states the >=3-point threshold determines PM3, and it applies to biallelic MMR variant carriers (CMMRD), not to a heterozygous MSH2 missense variant being assessed for Lynch syndrome.
• No proband tumor phenotype, family cancer history or indication for testing is recorded in the case bundle, so the presence of an MSI-H tumor cannot be inferred from the reason for testing.

Still missing

• No tumor MSI status (standard 5-10 marker panel or tumor genome) and no MMR-protein IHC result for any proband carrying c.1571G>A appears in the case materials; PP4 could only have been met by such tumor phenotype data and cannot be scored higher until it is provided.
• Searched VCEP-pilot-variants---MMR.xlsx: no entry for MSH2 c.1571G>A (p.R524H) - the file's only '1571' row is PMS2 NM_000535.7:c.1571dup. Searched Table-for-CMMRD-diagnosis.pdf: no entry for this variant; that table scores CMMRD (biallelic) features and its own text assigns >=3 points to PM3, not PP4.

Sources

Reasoning

PP5 requires an exact-variant classification from a reputable/expert source that the laboratory cannot independently evaluate; under this case's framework it may be applied only when the exact variant carries a ClinVar expert-panel (3-star, expert-panel reviewed) classification - Pathogenic/Likely pathogenic for PP5, Benign/Likely benign for BP6. Gate check for MSH2 c.1571G>A (ClinVar variation 182564): expert_panels is an empty object in both the ClinVar scrape (evidence.json) and the ClinVar API record, submission_audit_summary reports expert_panel_submissions = 0, all 13 submissions are 'criteria provided, single submitter' from clinical laboratories (11 Uncertain significance, 2 Likely benign), review_stars = 1, and the aggregate label is 'Conflicting classifications of pathogenicity'. The expert-panel condition is therefore 0 >= 1? no, and PP5 is not met. Independently, the governing InSiGHT MSH2 v2.0 cspec marks PP5 'Not Applicable for this VCEP' with the instruction 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee', so the gene-specific specification would also not permit PP5 here. The five papers triaged to PP5 (PMID:11598466, 15604628, 20301390, 22167527, 25394175) were checked and none classifies this exact variant; they are guidelines or gene-level documents (variant_mentioned = false in every extract), so none is a reputable source asserting a variant-level classification.

Rule

Exact-variant ClinVar expert-panel (3-star) classification required: Pathogenic/Likely pathogenic -> PP5 (supporting). No expert-panel classification present; governing InSiGHT MSH2 v2.0 cspec additionally lists PP5 as 'Not Applicable for this VCEP'.

Evidence considered

• clinvar (variation 182564, MSH2 c.1571G>A p.Arg524His): expert_panels = {} (empty), expert_panel_submissions = 0, 13 submissions all 'criteria provided, single submitter', 11 Uncertain significance / 2 Likely benign, 1 star, aggregate 'Conflicting classifications of pathogenicity'. Explicit comparison: expert-panel classifications available = 0; 0 >= 1 required? no.
• cspec (InSiGHT MSH2 v2.0): PP5 applicability 'Not Applicable for this VCEP'; instructionsToUse 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee.'
• PMID:11598466, PMID:15604628, PMID:20301390, PMID:22167527, PMID:25394175 (the five papers this case triaged as PP5 candidates) were checked for a variant-level reputable-source classification of c.1571G>A/p.R524H: none names the variant; all are practice guidelines, referral-criteria documents or GeneReviews-style gene-level reviews.
• Ordinary laboratory submissions in ClinVar (Color Health, Ambry, GeneDx, Invitae/Labcorp, Myriad, Fulgent, Baylor, CHEO, Mendelics, All of Us, etc.) and the aggregate ClinVar label were explicitly not used to trigger PP5.

Still missing

• No expert-panel (3-star) ClinVar classification and no other reputable-source variant-level classification of MSH2 c.1571G>A exists at the time of review, so PP5 cannot be triggered; this is an absence of a qualifying classification, not a search gap.

Sources

Reasoning

Governing specification is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel MSH2 Specification v2.0, whose BA1 rule is: 'GnomAD v4 Grpmax filtering allele frequency >= 0.001 (0.1%) and variant is excluded as founder pathogenic variant.' The VCEP names gnomAD v4 explicitly, so the all-comers gnomAD v4.1 entry is the governing source and no non-cancer/exome-only substitute applies. NM_000251.3:c.1571G>A (p.Arg524His) has gnomAD v4.1 total AF 1.36322e-05 (22/1,613,824 alleles) with a joint grpmax filtering AF of 1.03e-05 (exome grpmax 9.31e-06, genome grpmax 4.88e-06). That is roughly 97-fold below the >= 0.001 stand-alone threshold. The highest subpopulation frequency in v4.1 is European (Finnish) at 3.12402e-05 (2/64,020 alleles), still ~32-fold below threshold. gnomAD v2.1 (AF 1.06093e-05, grpmax FAF 9.58e-06), the non-cancer subsets (v2.1 non-cancer exomes 8.44431e-06; v3.1 non-cancer genomes 1.35237e-05, grpmax FAF 4.88e-06) and gnomAD-Canada v1.0 (AF 1.086e-04, grpmax FAF 1.70e-04 in NFE) all likewise fall below 0.001. Homozygote count is zero in every dataset. The second VCEP condition (variant excluded as a founder pathogenic variant) is moot because the frequency condition is not satisfied. BA1 is therefore not met.

Rule

VCEP InSiGHT MSH2 v2.0 BA1 (Stand Alone): GnomAD v4 Grpmax filtering allele frequency >= 0.001 (0.1%) and variant is excluded as founder pathogenic variant.

Evidence considered

• Governing framework: ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel Specifications for MSH2 Version 2.0; BA1 rule requires gnomAD v4 grpmax filtering allele frequency >= 0.001 (0.1%) and exclusion as a founder pathogenic variant. Because the specification names gnomAD v4, the all-comers gnomAD v4.1 entry was used as the default source rather than a non-cancer or exome-only subset.
• gnomAD v4.1 (GRCh38 chr2-47466718-G-A): total AF 1.36322e-05, 22/1,613,824 alleles, 0 homozygotes; exome AF 1.36828e-05 (20/1,461,692), genome AF 1.31465e-05 (2/152,132); joint grpmax filtering AF 1.03e-05; highest subpopulation European (Finnish) AF 3.12402e-05 (2/64,020 alleles).
• gnomAD v2.1 (GRCh37 2-47693857-G-A): total AF 1.06093e-05, 3/282,770 alleles, 0 homozygotes; grpmax FAF 9.58e-06; highest subpopulation Admixed American AF 5.64366e-05 (2/35,438 alleles).
• gnomAD v2.1 non-cancer exomes AF 8.44431e-06 (2/236,846 alleles) and gnomAD v3.1 non-cancer genomes AF 1.35237e-05 (2/147,888 alleles; grpmax FAF 4.88e-06, highest subpopulation European non-Finnish 3.08632e-05) were also reviewed and are likewise far below the BA1 stand-alone threshold.
• gnomAD-Canada v1.0 (HostSeq genomes, 10,487 genomes): total AF 1.08566e-04 (2/18,422 alleles), grpmax FAF 1.7036e-04 in European non-Finnish, 0 homozygotes; this non-gnomAD resource is still ~6-fold below 0.001 and rests on 2 alleles.
• Homozygote counts were zero in every population dataset reviewed, which is the relevant observation for a dominant, high-penetrance mismatch-repair gene: an excess of homozygotes/compound states in population data would be benignity evidence, and none is present.

Sources

Reasoning

Governing specification is the ClinGen InSiGHT MSH2 Specification v2.0, whose BS1 rule is: 'GnomAD v4 Grpmax filtering allele frequency >= 0.0001 and < 0.001 (0.01-0.1%) and variant is excluded as founder pathogenic variant.' The VCEP names gnomAD v4, so the all-comers gnomAD v4.1 grpmax filtering AF of 1.03e-05 governs; this is ~10-fold below the 0.0001 threshold. The gnomAD v2.1 grpmax FAF (9.58e-06) and both non-cancer subsets (v2.1 non-cancer exomes 8.44431e-06; v3.1 non-cancer genomes grpmax FAF 4.88e-06) are also below threshold, as are all gnomAD v4.1 subpopulation frequencies (maximum European Finnish 3.12402e-05). One discordant observation is recorded for human review rather than for the code: gnomAD-Canada v1.0, a non-gnomAD HostSeq resource, shows a European non-Finnish grpmax filtering AF of 1.7036e-04, which would exceed the BS1 threshold, but it rests on only 2 of 11,740 NFE alleles, is not the source designated by the VCEP, and its total AF (1.08566e-04) is also a 2-allele estimate. BS1 is not met under the governing specification.

Rule

VCEP InSiGHT MSH2 v2.0 BS1 (Strong): GnomAD v4 Grpmax filtering allele frequency >= 0.0001 and < 0.001 (0.01-0.1%) and variant is excluded as founder pathogenic variant.

Evidence considered

• Governing framework: ClinGen InSiGHT MSH2 Specification v2.0; BS1 (Strong) requires gnomAD v4 grpmax filtering allele frequency >= 0.0001 and < 0.001 with exclusion as a founder pathogenic variant. Generic SVI default (>= 0.01) was not used because the VCEP specification takes precedence.
• gnomAD v4.1 (default per the VCEP source designation): total AF 1.36322e-05 (22/1,613,824 alleles), joint grpmax filtering AF 1.03e-05, exome grpmax 9.31e-06, genome grpmax 4.88e-06, 0 homozygotes; highest subpopulation European (Finnish) AF 3.12402e-05.
• gnomAD v2.1: total AF 1.06093e-05, grpmax FAF 9.58e-06; highest subpopulation Admixed American AF 5.64366e-05 (2/35,438 alleles).
• gnomAD v2.1 non-cancer exomes AF 8.44431e-06 (2/236,846 alleles) and gnomAD v3.1 non-cancer genomes AF 1.35237e-05 (2/147,888 alleles) were reviewed as the non-cancer proxies named in the task; both sit far below the BS1 band.
• gnomAD-Canada v1.0 (HostSeq genomes): total AF 1.08566e-04 (2/18,422 alleles), European non-Finnish grpmax FAF 1.7036e-04 (2/11,740 alleles), faf95 3.009e-05, 0 homozygotes; flagged for human review because its point estimate crosses the BS1 threshold while the VCEP-designated gnomAD v4.1 value does not.

Still missing

• The VCEP BS1/BA1 rules carry a second requirement (variant excluded as a founder pathogenic variant) that no source in this case evaluates directly; this does not change the verdict because the frequency condition fails first, so no founder-status lookup was recorded as an open gap.
• gnomAD-Canada v1.0 European non-Finnish grpmax filtering AF 1.7036e-04 (2/11,740 alleles) exceeds the BS1 threshold while the VCEP-designated gnomAD v4.1 source does not; a human reviewer may want to confirm whether this non-gnomAD Canadian cohort should bear on BS1 for this variant.

Sources

Reasoning

Scope: the variant is a missense substitution (NP_000242.1:p.(Arg524His)), so BP4 is in scope, and the applicable path is the protein-impact path. The governing InSiGHT MMR VCEP MSH2 v2.0 specification defines BP4 (Benign Supporting) for missense variants as an HCI prior probability of pathogenicity <0.11; its only other branch applies SpliceAI <=0.1 to intronic and synonymous variants. The VCEP-declared HCI-PRIORS-MSH2.txt table was searched for this variant and contains an exact row for c.1571G>A (p.R524H, DBID MSH2_03461, exon 10) with prior probability 0.7525, far above the <0.11 threshold, so the missense branch of BP4 is not satisfied. The intronic/synonymous SpliceAI branch of the VCEP rule is out of scope for this variant and, consistent with the single-path rule for a missense change, the SpliceAI max delta of 0.038 cannot be substituted as BP4 support. The REVEL score of 0.918 likewise gives no benign support (it is far above the ClinGen SVI REVEL BP4 supporting threshold of 0.29), though the gene-specific HCI-prior rule governs in any case. No benign computational evidence is present.

Rule

InSiGHT MMR VCEP MSH2 v2.0, BP4 (Benign Supporting), missense branch: 'Missense variant with HCI-prior probability of pathogenicity <0.11 as per https://hci-priors.hci.utah.edu/PRIORS'. Applied value: HCI prior 0.7525 for c.1571G>A (p.R524H, DBID MSH2_03461) - threshold not met.

Evidence considered

• Governing rule applied verbatim from the InSiGHT MMR VCEP MSH2 v2.0 specification for BP4 (Benign Supporting): 'Missense variant with HCI-prior probability of pathogenicity <0.11 as per https://hci-priors.hci.utah.edu/PRIORS OR For intronic and synonymous variants: SpliceAI predicts no splicing impact with delta score <= 0.1 as per Walker et al 2023.'
• Exact entry located by searching HCI-PRIORS-MSH2.txt (the file the VCEP index declares as governing PP3/BP4 for MSH2 missense): c.1571G>A / p.R524H / MSH2_03461 with prior probability 0.7525, which is not <0.11.
• SpliceAI max delta 0.038 is present in the case evidence; the SpliceAI branch of the VCEP BP4 rule is restricted to intronic and synonymous variants and a splice prediction may not be used as BP4 support for a missense substitution, so it was not applied.
• REVEL 0.918 (local REVEL v1.3 lookup) is well above the generic ClinGen SVI REVEL BP4_Supporting threshold of <=0.29 and provides no benign support; the gene-specific HCI-prior rule takes precedence over generic REVEL calibration for MSH2 missense variants.

Sources

Reasoning

Governing specification: InSiGHT MSH2 v2.0 cspec, BP5 (Applicable, default strength Benign Supporting), which replaces the generic 'alternate molecular basis' wording with a tumor-based rule. Exact rule text for Supporting strength: '2 or 3 tumors: CRC/Endometrial tumors with MSS and/or no loss of MMR protein expression and/or LS spectrum tumors with loss of MMR protein(s) that is inconsistent with the gene demonstrating genetic variation. OR 1 BRAF V600E (Colon only)/MLH1 methylation (in LS spectrum tumor only) with MSI-H/MLH1 loss.' Strong requires >=4 such tumors or >=2 BRAF V600E/MLH1-methylation events. Applying this criterion literally: the number of documented tumors showing MSS, no MMR-protein loss, or MMR-protein loss inconsistent with the variant gene, is 0, so 0 >= 2? no (and 0 >= 4? no, 0 >= 1 BRAF V600E/MLH1-methylation event? no). BP5 is not met at any strength. As with PP4, no tumor MSI/IHC/methylation data for any proband are present anywhere in the case materials, and no extracted paper reports tumor characteristics for this variant. Unlike the generic ACMG formulation, a benign alternative molecular basis in the case (e.g. a second pathogenic variant in another gene) would not satisfy this gene-specific rule unless accompanied by the specified tumor mis-match findings.

Rule

InSiGHT MSH2 v2.0 cspec BP5: Supporting = 2 or 3 CRC/Endometrial tumors with MSS and/or no MMR-protein loss and/or LS-spectrum tumors with MMR-protein loss inconsistent with the variant gene, OR 1 BRAF V600E (colon only)/MLH1 methylation with MSI-H/MLH1 loss; Strong = >=4 tumors, OR >=2 BRAF V600E/MLH1-methylation events.

Evidence considered

• cspec (InSiGHT MSH2 v2.0) BP5 rules: Supporting = 2 or 3 CRC/endometrial tumors with MSS and/or no MMR-protein loss and/or LS-spectrum tumors with MMR-protein loss inconsistent with the variant gene, OR 1 BRAF V600E (colon)/MLH1 methylation with MSI-H/MLH1 loss; Strong = >=4 such tumors, OR >=2 BRAF V600E/MLH1-methylation events.
• Documented tumor count for this variant's cases = 0; explicit comparisons: 0 >= 2? no; 0 >= 4? no; qualifying BRAF V600E / MLH1 methylation events = 0; 0 >= 1? no.
• The case materials (evidence.json, case_summary.json, prefetch.json, evidence_brief.json) contain no tumor phenotype, IHC, MSI/MSS, BRAF or methylation data and no family cancer history for this proband, and no extracted full text reports tumor characteristics for c.1571G>A.
• The generic ACMG wording of BP5 ('variant found in a case with an alternate molecular basis for disease') was not used, because the governing VCEP replaces it with the tumor-based rule quoted above.

Still missing

• No tumor MSS/MSI result, MMR-protein IHC result, BRAF V600E status or MLH1 promoter methylation result for any proband carrying c.1571G>A is present in the case materials, so this VCEP's tumor-based BP5 rules (2-3 tumors = Supporting; >=4 tumors = Strong) cannot be scored; this is missing phenotype data rather than absent negative evidence.

Sources

Reasoning

BP6 may be applied only when the exact variant carries a ClinVar expert-panel (3-star, expert-panel reviewed) classification of Benign or Likely benign. Gate check for MSH2 c.1571G>A (ClinVar variation 182564): expert_panels is empty, expert_panel_submissions = 0, all 13 submissions are 'criteria provided, single submitter' (2 of them Likely benign: Myriad Genetics 2024-12-10 and Labcorp/Invitae 2026-01-26), review_stars = 1, and the aggregate label is 'Conflicting classifications of pathogenicity'. The expert-panel condition is 0 >= 1? no, and BP6 is not met. The two single-submitter Likely benign laboratory assertions, and the aggregate ClinVar label, were explicitly not used, since ordinary laboratory submissions must never trigger BP6. Independently, the governing InSiGHT MSH2 v2.0 cspec marks BP6 'Not Applicable for this VCEP' with the instruction 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee', so the gene-specific specification would also not permit BP6 here.

Rule

Exact-variant ClinVar expert-panel (3-star) classification required: Benign/Likely benign -> BP6 (supporting). No expert-panel classification present; governing InSiGHT MSH2 v2.0 cspec additionally lists BP6 as 'Not Applicable for this VCEP'.

Evidence considered

• clinvar (variation 182564, MSH2 c.1571G>A p.Arg524His): expert_panels = {} (empty), expert_panel_submissions = 0, 13 submissions all single-submitter with 1 star, aggregate 'Conflicting classifications of pathogenicity'; the two Likely benign records (Myriad Genetics, Labcorp/Invitae) are ordinary laboratory submissions. Explicit comparison: expert-panel Benign/Likely benign classifications available = 0; 0 >= 1 required? no.
• cspec (InSiGHT MSH2 v2.0): BP6 applicability 'Not Applicable for this VCEP'; instructionsToUse 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee.'
• Aggregate ClinVar labels and single-submitter laboratory classifications were explicitly excluded from triggering BP6, as required.

Still missing

• No expert-panel (3-star) ClinVar classification of this variant as Benign/Likely benign, and no other reputable-source benign classification of the exact variant, exists at the time of review; BP6 therefore cannot be triggered.

Sources

Reasoning

The governing specification (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel, MSH2 v2.0) replaces the generic PS2 yes/no rule with a de novo point system: 2 points per proband with a de novo variant and both maternity and paternity confirmed who has an MMR-deficient LS-spectrum tumor (MSI/IHC consistent with the affected gene, no MLH1 methylation), 1 point per confirmed de novo proband with an LS-spectrum tumor but no MSI/IHC/methylation data, and 0.5 points per proband with an assumed de novo variant (maternity and/or paternity unconfirmed) and an LS-spectrum tumor; strength tiers are >=4 points = Very Strong, 2-3.5 = Strong, 1-1.5 = Moderate, 0.5 = Supporting. This case is a variant-level classification of NM_000251.3:c.1571G>A (p.Arg524His): there is no proband, no parental genotype data (maternity/paternity confirmation), no tumor MSI/IHC/methylation data and no pedigree. No de novo occurrence of this variant was found in the ClinVar record (VCV000182564, 14 submissions; the submission-audit signals capture only PP3/PS3/PM2-type statements, never de novo) or in any full-text paper retrieved for this case. With zero de novo probands available no points can be assigned, so the criterion cannot be scored. It is recorded as not_assessed rather than not_met so that the absence of family studies is not reported as evidence that the variant is not de novo.

Rule

VCEP MSH2 v2.0 PS2 de novo point system: >=4 points = Very Strong; 2-3.5 points = Strong; 1-1.5 points = Moderate; 0.5 points = Supporting (2 points per confirmed de novo proband with MMR-deficient LS-spectrum tumor; 1 point per confirmed de novo proband with LS-spectrum tumor without MSI/IHC/methylation data; 0.5 points per assumed de novo proband).

Evidence considered

• InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for MSH2 v2.0 (cspec) defines PS2 as a de novo point system: 2 points for a confirmed de novo proband with an MMR-deficient LS-spectrum tumor, 1 point for a confirmed de novo proband with an LS-spectrum tumor and no MSI/IHC/methylation data, 0.5 points for an assumed de novo proband with maternity and/or paternity unconfirmed; tiers >=4 / 2-3.5 / 1-1.5 / 0.5 points = Very Strong / Strong / Moderate / Supporting.
• The ClinVar record for this variant (VCV000182564, 14 submissions, 0 expert-panel submissions) contains no de novo claim; the submission audit captured only PP3, PS3 and PM2-type criteria sentences from submitters.
• No paper retrieved for this case reports MSH2 c.1571G>A (p.Arg524His) at all, so none can report a de novo occurrence; PMID:27363726 (Morris et al. 2016) reclassifies MSH2 variants using personal/family cancer history of probands, not de novo status.
• Richards et al. 2015 (PMID:25741868) defines generic PS2 as de novo (maternity and paternity confirmed) in a patient with the disease and no family history.

Still missing

• No proband-level data: no occurrence of this variant with documented maternity and paternity confirmation was found in the ClinVar record or in the retrieved literature.
• Tumor MSI/IHC/methylation status of any carrier is unknown, so even the 2-point 'confirmed de novo plus MMR-deficient LS-spectrum tumor' pathway cannot be evaluated.
• Parental testing (maternity and paternity confirmation) was not performed or reported for this variant.
• ClinVar-cited PMIDs whose full text was not retrieved (28526081, 29659569, 30267214, 31857677, 33357406, 33471991, 28492532) were not screened for de novo observations.

Sources

Reasoning

The governing MSH2 v2.0 VCEP specification requires a pedigree-based co-segregation analysis for PP1, with strength set by the combined Bayes likelihood ratio across pedigrees (multiplied together): >18.7 in at least 2 families = Strong, >4.3 and <=18.7 = Moderate, >2.08 and <=4.3 = Supporting, using COOL v3 with the overall Bayes factor. This case is a variant-level classification with no pedigree, no genotype data on relatives and no reported co-segregation of MSH2 c.1571G>A with colorectal/endometrial cancer. No source consulted in this case reports segregation of this variant: the ClinVar record (14 submissions) contains no segregation claim, and none of the retrieved full-text papers names the variant at all. With no pedigree, no Bayes likelihood ratio can be computed, so PP1 is not assessable rather than affirmatively not met.

Rule

VCEP MSH2 v2.0 PP1: co-segregation with disease in pedigree(s) with combined Bayes Likelihood Ratio >18.7 in >=2 families = Strong; >4.3 and <=18.7 = Moderate; >2.08 and <=4.3 = Supporting.

Evidence considered

• InSiGHT MMR VCEP specification for MSH2 v2.0 (cspec) sets PP1 strength from the combined Bayes likelihood ratio: >18.7 in >=2 families = Strong, >4.3 and <=18.7 = Moderate, >2.08 and <=4.3 = Supporting, with pedigrees analysed in COOL v3 and multiplied together across families.
• ClinVar VCV000182564 (14 submissions, 0 expert panels) provides no co-segregation or family data; the submission audit detected only PP3/PS3/PM2-type statements.
• PMID:27363726 (Morris et al. 2016) is the only source in this case that links MSH2 variant interpretation to family history data; it shows that even history-based reclassification requires about 10-20 proband observations of the specific variant (>96% of MLH1/MSH2 calls with <=20 observations, >63% with <=10), which are not available here, so no segregation likelihood ratio can be derived.
• PMID:11598466 states that for a missense MMR change such as this one the interpretation turns on whether it segregates with disease in the family or has been previously reported as disease-associated, and that without segregation the result is uninformative; it reports no segregation data for this variant.
• None of the retrieved full-text papers names MSH2 c.1571G>A (p.Arg524His), so none can supply segregation evidence.

Still missing

• No pedigree or family genotype data for MSH2 c.1571G>A was available, so the combined Bayes likelihood ratio (COOL v3 overall Bayes factor) cannot be calculated.
• No report of this variant co-segregating with colorectal/endometrial cancer in multiple affected family members was found in ClinVar or in any retrieved full-text paper.
• PMID:20301390 (GeneReviews: Lynch Syndrome) was registered as a source but its full text was not retrieved (screenshot only), so it could not be screened for family-level or segregation statements.
• ClinVar-cited PMIDs without retrieved full text (28526081, 29659569, 30267214, 31857677, 33357406, 33471991, 28492532) were not screened for segregation data.

Sources

Reasoning

The governing ClinGen InSiGHT MSH2 Specification v2.0 BS2 rule is not a population-frequency rule but a patient-level co-occurrence 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 as per Aronson et al 2022 (Refer to Table for CMMRD diagnosis.pdf). Confirmation of phase requires testing of parents or offspring.' Evaluating it requires, for this variant, a phase-confirmed in-trans second pathogenic MSH2 allele plus CMMRD-feature scoring. No such information exists anywhere in the case evidence: the case summary, ClinVar submission audit and criterion-lead submissions for variation ID 182564 describe only germline variant classification (14 submissions; 11 laboratories Uncertain significance, 2 Likely benign, 1 Uncertain Significance; no expert-panel submission), with no zygosity, phase, parental testing or second-variant data; no reviewed full-text paper reports this variant at all. The criterion is therefore not assessed rather than not met, because absence of the required testing data is not evidence that the condition is false. The generic ACMG/AMP BS2 concept (observed in trans with a pathogenic variant for a fully penetrant dominant disorder) does not apply in its original form to this gene, and the VCEP explicitly redirects BS2 to the CMMRD co-occurrence assessment.

Rule

VCEP InSiGHT MSH2 v2.0 BS2 (Strong): 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 as per Aronson et al 2022; confirmation of phase requires testing of parents or offspring.

Evidence considered

• Governing framework: ClinGen InSiGHT MSH2 Specification v2.0; BS2 (Strong) requires in-trans co-occurrence with a known pathogenic sequence variant in the same gene in a patient with colorectal cancer after age 45 (or another LS-spectrum cancer above the median LS onset age) without CMMRD features, with phase confirmed by parental or offspring testing.
• All ClinVar evidence for variation ID 182564 (14 submissions; 11 clinical laboratories Uncertain significance, 2 Likely benign, 1 Uncertain Significance; 0 expert-panel submissions; 5 criterion-lead submissions) was reviewed and contains no zygosity, phase or co-occurring-variant statement.
• The declared VCEP pilot table (VCEP-pilot-variants---MMR.xlsx, declared for PVS1, PM2, PP4) was searched in full for c.1571G>A / p.Arg524His / p.R524H and contains no row for this variant; it does supply the precedents that the VCEP applies PM2 at supporting strength in pilot classifications and uses BS2 for CMMRD/co-occurrence reasoning rather than for population frequency.
• The VCEP CMMRD diagnosis scoring table (Table-for-CMMRD-diagnosis.pdf: >=3 points indicates CMMRD features) was reviewed as the scoring instrument the BS2 rule depends on; no proband malignancy/clinical data are available in this case to score it.
• No full-text paper fetched for this case reports the variant, so no literature-based in-trans observation could be identified either (PMID:11598466, PMID:15604628, PMID:22167527, PMID:23760103, PMID:25394175, PMID:25741868, PMID:27363726 all record variant_mentioned = false).

Still missing

• No patient-level phase/zygosity, parental testing or co-occurring second pathogenic MSH2 variant information is available for this case (checked case_summary.json, clinvar_submission_audit.json, source_registry.json and paper_extracts.json); BS2 cannot be evaluated without it.
• Searched the declared VCEP file VCEP-pilot-variants---MMR.xlsx in full (vcep_db/MSH2/fulltext/VCEP-pilot-variants---MMR.xlsx.txt) for c.1571G>A, p.Arg524His and p.R524H: no entry for this variant exists in that table. It contains MSH2 pilot rows at other positions (e.g. c.1277-16T>C, c.1748A>G, c.913G>A, c.1563T>C, c.2210+2T>C, c.1807G>A, c.1862G>T, c.490G>T, c.929T>G, c.1276G>A, c.2075G>T, c.1012G>A, c.2459-12A>G, c.2634+5G>T, c.211G>C) and only a PMS2 c.1571dup row, so no VCEP precedent classification or PM2_pilot code assignment exists for this MSH2 variant.
• CMMRD-feature scoring per the VCEP Table-for-CMMRD-diagnosis.pdf requires clinical information for the proband that is not present in this case.

Sources

Reasoning

The governing MSH2 v2.0 VCEP specification defines BS4 as demonstrated lack of co-segregation with disease in pedigree(s): combined Bayes likelihood ratio <0.05 = Strong and >0.05 and <=0.48 = Supporting, again computed from pedigrees in COOL v3 and multiplied across families. This case contains no pedigree and no genotype data on affected or unaffected relatives, and no source consulted reports non-segregation of MSH2 c.1571G>A: the ClinVar record (14 submissions) contains no segregation-based claim, and no retrieved full-text paper names the variant. The combined Bayes likelihood ratio therefore cannot be computed and BS4 is not assessable rather than affirmatively not met. The absence of reported non-segregation is not treated as evidence of co-segregation, and the two ClinVar 'Likely benign' submissions (Myriad, Labcorp/Invitae) are not used here as BS4 evidence because their underlying pedigree data are not exposed.

Rule

VCEP MSH2 v2.0 BS4: lack of co-segregation with disease in pedigree(s) with combined Bayes Likelihood Ratio <0.05 = Strong; >0.05 and <=0.48 = Supporting.

Evidence considered

• InSiGHT MMR VCEP specification for MSH2 v2.0 (cspec) defines BS4 as lack of co-segregation with disease in pedigree(s) with a combined Bayes Likelihood Ratio <0.05 (Strong) or >0.05 and <=0.48 (Supporting), analysed with COOL v3 using the overall Bayes factor.
• ClinVar VCV000182564 (14 submissions) contains no non-segregation or pedigree evidence for this variant; two submitters classified it Likely benign but published no criteria or family detail.
• PMID:27363726 (Morris et al. 2016) was consulted as the only family-history-based resource in this case; it provides no non-segregation data for this variant and indicates that family-history-based reclassification of an MSH2 variant needs roughly 10-20 proband observations of that variant.

Still missing

• No pedigree or family genotype data for MSH2 c.1571G>A was available, so the combined Bayes likelihood ratio used to assign BS4 (COOL v3 overall Bayes factor) cannot be calculated.
• No report of this variant failing to co-segregate with disease in an affected family was found in ClinVar or in any retrieved full-text paper.
• The two Likely benign ClinVar submissions (SCV005897225 Myriad Genetics, SCV000284114 Labcorp/Invitae) expose no criteria or pedigree detail, so their basis could not be audited for BS4-relevant non-segregation evidence.
• PMID:20301390 (GeneReviews: Lynch Syndrome) full text was not retrieved (screenshot only), so it could not be screened for non-segregation or family-level statements.

Sources