MSH2 c.1847C>G, p.(Pro616Arg)

Reasoning

Scope first: NM_000251.3:c.1847C>G is a missense substitution, NP_000242.1:p.(Pro616Arg) (consequence class 'missense', canonical_splice_consensus false, SpliceAI variant_consequence 'sequence_variant'), so PP3/BP4 are adjudicated on the missense protein-impact path rather than the SpliceAI path. The governing specification is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel MSH2 v2.0 (cspec), which defines PP3 for missense variants exclusively by the HCI prior probability of pathogenicity: >0.81 = PP3_Moderate; >0.68 and <=0.81 = PP3_Supporting. The on-record HCI-PRIORS-MSH2 prior table supplied with this VCEP contains an exact entry for this coding substitution - c.1847C>G / p.P616R, HCI prior (Custom_PP2_score) = 0.900, MAPP 41.48 - and 0.900 exceeds the 0.81 PP3_Moderate cutoff, so the governing table pre-assigns PP3_Moderate. This is applied directly as the VCEP's own pre-assigned code. The same missense path is corroborated by the local REVEL score of 0.84, which independently meets the ClinGen SVI PP3_moderate threshold of >=0.773 (Pejaver et al. 2022, PMID:36413997); REVEL is recorded as agreement, not as a second, separate piece of evidence, so the same prediction is not double-counted. SpliceAI (max delta 0.001) was not used and does not contribute: the variant is neither intronic, synonymous nor in a splice region, and a splice prediction says nothing about tolerance of an amino-acid substitution. BayesDel (0.462389) was not used because no ACMG-strength calibration exists for it and this VCEP specifies no BayesDel cutoff.

Rule

VCEP MSH2 v2.0 PP3 (Moderate): missense variant with HCI prior probability for pathogenicity >0.81 -> PP3_Moderate

Evidence considered

• cspec (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel Specification for MSH2 v2.0) PP3 rules: 'Missense variant with HCI prior probability for pathogenicity >0.81 as per https://hci-priors.hci.utah.edu/PRIORS' = PP3_Moderate; 'Missense variant with HCI prior probability for pathogenicity >0.68 & <=0.81' = PP3_Supporting, the latter rule also carrying an alternative SpliceAI branch (delta score >= 0.2 for non-canonical splicing nucleotides) that is not applicable to this missense variant.
• hci_prior / vcep_hci_priors_msh2 (on-record HCI-PRIORS-MSH2.txt lookup table, the gene-specific PP3/BP4 supporting material declared by this gene's VCEP index): exact row found by grep for c.1847C>G, p.P616R at line 4078 of vcep_db/MSH2/fulltext/HCI-PRIORS-MSH2.txt.txt - exon 12, c.1847C>G, protein p.P616R, Custom_PP2_score (HCI prior probability of pathogenicity) 0.900, MAPP_score 41.48, MAPP/PP2_Prior 0.9565, reference {PMID22949387:Thompson et al., 2013}, DBID MSH2_04077. The variant is therefore not absent from the governing table; the pre-assigned code is read off that exact row.
• The same row confirms the amino-acid substitution: the table is keyed by coding substitution and lists c.1847C>G (p.P616R) explicitly, matching the case's normalized HGVS NP_000242.1:p.(Pro616Arg).
• revel (local REVEL v1.3 lookup, score 0.84, found true): 0.84 >= 0.773 ClinGen SVI REVEL PP3_moderate threshold (Pejaver et al. 2022, Am J Hum Genet, PMID:36413997). Used only as corroboration of the HCI-prior call on the same missense path; the governing VCEP code takes precedence.
• SpliceAI for this variant (max delta 0.001; DS_AG 0.0, DS_AL 0.0, DS_DG 0.0, DS_DL 0.001) was consulted only to establish scope and was deliberately not applied: PP3/BP4 for a missense variant take evidence from the missense path alone, and a splice prediction is not evidence that an amino-acid substitution is tolerated.
• bayesdel (score 0.462389) recorded but not used: no generic ACMG-strength calibration exists for BayesDel and this VCEP specifies no BayesDel cutoff, so it is not_available for PP3/BP4.
• The HCI prior table row's own reference is PMID:22949387 (Thompson et al., 2013), the source from which the priors were derived; that paper is not among this case's retrieved sources, so the prior is taken from the on-record VCEP table itself.

Sources

Reasoning

Governing specification: ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel MSH2 v2.0 (cspec). Its BS1 rule is: gnomAD v4 Grpmax filtering allele frequency >= 0.0001 and < 0.001 (0.01-0.1 percent) and the variant is excluded as founder pathogenic variant, at Benign Strong strength. The observed gnomAD v4.1 joint Grpmax FAF is 0.00017995 (0.018 percent), which sits inside the 0.0001-0.001 window, and the gnomAD v4.1 exome Grpmax FAF (0.00022385) and total all-comers AF (3.22189e-05, 52/1,613,962 alleles) are consistent. The variant is not designated as a founder pathogenic allele in any source consulted; it is present across several ancestries rather than as a founder-restricted pathogenic allele, so the exclusion clause is satisfied. The generic SVI default of 0.01 would not be reached, but the VCEP specification governs and its calibrated MSH2 window (0.01-0.1 percent) is met. Note the variant is never observed homozygous in gnomAD (0 homozygotes in v2.1, v4.1, and both non-cancer subsets), which is consistent with a benignly classified rare MSH2 allele rather than a highly penetrant dominant one. BS1 is met at Strong strength.

Rule

InSiGHT MSH2 v2.0 BS1: gnomAD v4 Grpmax FAF >= 0.0001 and < 0.001 (0.01-0.1%) and variant excluded as founder pathogenic variant. Strength: Strong.

Evidence considered

• InSiGHT MSH2 v2.0 cspec BS1 rule (Benign Strong): 'GnomAD v4 Grpmax filtering allele frequency >= 0.0001 and < 0.001 (0.01-0.1%) and variant is excluded as founder pathogenic variant.'
• gnomAD v4.1 (governing dataset named by the VCEP): joint Grpmax FAF 0.00017995, exome Grpmax FAF 0.00022385, genome Grpmax FAF 0.00001171, total AF 3.22189e-05 (52/1,613,962 alleles). The Grpmax FAF is the VCEP-specified metric and lands in the required 0.0001-0.001 band.
• gnomAD v2.1 (older build, context): Grpmax FAF 0.00015644 and total AF 6.71725e-05, also within the same order of magnitude; homozygotes 0.
• Non-cancer subsets: GNOMAD_V2_1_NON_CANCER exome Grpmax FAF 0.00015784 and GNOMAD_V3_1_NON_CANCER Grpmax FAF 0.00001171. The v2.1 non-cancer exome figure also exceeds 0.0001, so the BS1 verdict is robust to the choice of non-cancer versus all-comers source; the v3.1 non-cancer genome-only subset is small (AN 147,888) and lower.
• VCEP-pilot-variants---MMR.xlsx was searched for c.1847C>G, p.Pro616Arg and p.P616R: no row for this variant is present, so no InSiGHT precedent code assignment is available for it. Other MSH2 pilot variants that received BS1 (e.g. c.1277-16T>C, c.1563T>C) show the panel does apply BS1 to rare MSH2 alleles with benign-leaning evidence.

Still missing

• No independent founder-pathogenicity assessment was available; the 'excluded as founder pathogenic variant' clause of the VCEP BS1 rule was treated as satisfied on the basis that no consulted source designates c.1847C>G as a founder allele and the variant is not ancestry-restricted.
• The InSiGHT MMR VCEP pilot variant table (VCEP-pilot-variants---MMR.xlsx) contains no entry for NM_000251.3:c.1847C>G / p.Pro616Arg / p.P616R, so there is no VCEP precedent classification for this exact variant; the file was searched for all three notations.

Sources

Reasoning

NM_000251.3:c.1847C>G is a single-nucleotide substitution whose predicted protein consequence is a single amino-acid replacement, NP_000242.1:p.(Pro616Arg) (LRG_218p1:p.(Pro616Arg)); it introduces no premature termination codon, no frameshift, and no protein-length change, and it does not alter the initiation codon or a canonical splice consensus (it is not an IVS+/-1 or IVS+/-2 variant). SpliceAI predicts no splice impact (max delta 0.001, far below the 0.2 supporting cut-off), so there is no in-silico or assay basis for a splicing-aberration route either. The governing InSiGHT Hereditary Colon Cancer/Polyposis Expert Panel specification for MSH2 (ClinGen CSPEC v2.0) is applicable at the gene level (MSH2 loss of function is an established disease mechanism, and the CSPEC lists PVS1 as Applicable), but every PVS1 tier it defines requires a null-type consequence: Very Strong for nonsense/frameshift with a PTC at or before codon 891, large single- or multi-exon genomic alterations, IVS+/-1/-2 variants that disrupt the reading frame with predicted NMD, or mRNA-assay-proven splice aberrations; Strong for presumed in-tandem exon duplications producing a pre-terminal frameshift, the last-exon-base G>non-G rule, or IVS+/-1/-2 frame-preserving events in a functionally critical region; Moderate for nonsense/frameshift PTCs between codons 892 and 934. None of these descriptions match a missense substitution, so no tier is met and PVS1 is not met at any strength. Strength is recorded as null rather than as a downgraded tier because no SVI/VCEP downgrade rule is engaged - the criterion does not apply to this consequence type at all; it is not a met criterion whose strength was reduced.

Rule

MSH2 CSPEC v2.0 PVS1 (very strong/strong/moderate tiers) - not met: none of the nonsense/frameshift, large genomic alteration, IVS+/-1/-2 splice, mRNA-assay splice-aberration, or duplication-frameshift descriptions match a missense substitution.

Evidence considered

• Variant consequence: NM_000251.3:c.1847C>G encodes NP_000242.1:p.(Pro616Arg) (also given as LRG_218p1:p.(Pro616Arg)); the variant-level assessment records consequence_class 'missense', canonical_splice_consensus false, variant_bucket 'other', and states that the variant 'does not fall into the default generic PVS1 null-variant buckets of nonsense, frameshift, or canonical +/-1,2 splice consensus variants'.
• Governing specification (MSH2 CSPEC v2.0, InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel): PVS1 applicability 'Applicable', defaultStrength 'Pathogenic Very Strong'. Very Strong rule: 'Nonsense/frameshift variant introducing Premature Termination Codon (PTC) <= codon 891 in MSH2 ... OR Large genomic alterations of single or multi-exon size. OR Variants at IVS+/-1 or IVS+/-2 where exon skipping or use of a cryptic splice site disrupts reading frame and is predicted to undergo NMD ... OR Variants where mRNA assays ... indicate that the variant allele results in a splicing aberration'. Strong rule: presumed in-tandem exon duplication frameshift, last-exon-base G>non-G where the intron does not begin GTRRGT, or IVS+/-1/-2 frame-preserving events in a functionally critical region. Moderate rule: 'Nonsense/frameshift variant introducing premature termination codon between codons 892 & 934 in MSH2'.
• Gene-level eligibility: MSH2 loss of function is an established disease mechanism for Lynch syndrome (MSH2 pairs with MSH6/MSH3 in MutS-alpha/MutS-beta; loss causes MSI-high tumours and CMMRD in biallelic carriers), so the gene gate for PVS1 is permissive; the criterion nevertheless fails at the variant level because a missense substitution produces no null allele.
• SpliceAI lookup for NM_000251.3:c.1847C>G: available, max delta 0.001 (all donor/acceptor gain-loss components null; Pangolin gain 0.001, loss -0.008) - no predicted splice impact, so the IVS+/-1/-2 and mRNA-splicing-aberration branches of the MSH2 PVS1 rules are not triggered.
• InSiGHT pilot variant table searched for c.1847C>G / p.Pro616Arg / p.P616R: no row for this variant; the MSH2 precedent rows show PVS1 being applied only to nonsense, canonical splice, frameshift, large genomic and exon-level nulls (e.g. c.2210+2T>C, c.1276G>A, c.211G>C, g.47408556G>C), consistent with PVS1 being reserved for null-type consequences.
• No source consulted assigns any loss-of-function/null consequence to this specific substitution; the missense-specific evidence in this case (HCI prior probability 0.9565, REVEL 0.84, deep mutational scanning of MSH2 missense variants in PMID:33357406) speaks to whether the substituted protein is functional, not to whether the allele is null, and therefore cannot support PVS1.

Still missing

• PVS1_DecisionTree_MMR.pdf was searched for every notation of this variant (c.1847C>G, p.Pro616Arg, p.P616R, 1847, 616) and has NO entry for it; moreover the converted full text of that file is effectively empty (it contains only a page header, '## page 1', 12 bytes), so no variant-level or rule-level content could be read from that file at all. The governing PVS1 rules were therefore taken from the MSH2 CSPEC v2.0 criteria block and from the PVS1 decision-tree content carried in the case evidence materials (nonsense/frameshift NMD codons, GT-AG splice-site branches, deletion/duplication branches, initiation-codon branch for MSH2).
• VCEP-pilot-variants---MMR.xlsx was searched for every notation of this variant (c.1847C>G, p.Pro616Arg, p.P616R) and contains NO entry for it, so no InSiGHT VCEP precedent classification exists for this variant; the pilot's MSH2 rows include adjacent residues (p.Asp603Asn, c.1807G>A; p.Arg621Leu, c.1862G>T) but not codon 616.

Sources

Reasoning

The governing MSH2 specification (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel, MSH2 v2.0) lists PS1 as Applicable and defines PS1_Strong as 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); PS1_Moderate requires the same with a previously established Likely Pathogenic missense variant with normal RNA result plus PM2_supporting. Codon 616 is CCT (Pro). The VCEP HCI/MSH2 prior table enumerates all six single-nucleotide substitutions that change the encoded residue at codon 616: c.1846C>A p.P616T, c.1846C>G p.P616A, c.1846C>T p.P616S, c.1847C>A p.P616H, c.1847C>G p.P616R and c.1847C>T p.P616L. Converting Pro (CCT) to Arg requires a C>G change at the middle base of the codon, i.e. position 1847; every other single-nucleotide change at codon 616 encodes a different amino acid. Therefore c.1847C>G is the ONLY nucleotide change that can produce p.Pro616Arg, so there is no second, alternate nucleotide change available to satisfy the PS1 comparator requirement, and p.Pro616Arg itself is not a VCEP-established Pathogenic variant (it appears nowhere in the InSiGHT MMR VCEP pilot classification table, and ClinVar record 127636 carries 0 expert-panel submissions and conflicting classifications: Uncertain significance, Likely benign and Benign). The alternate PS1 branch (variants affecting the same non-canonical splice nucleotide as a confirmed Pathogenic splice variant with similar or worse SpliceAI prediction) is not applicable: this is a coding missense substitution at a non-splice nucleotide with no confirmed pathogenic splice variant at the same nucleotide. PS1 is therefore not met.

Rule

Panel specification (MSH2 v2.0), PS1_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).'

Evidence considered

• Searched every VCEP file declared for this gene for all three notations of the variant (grep -rn for 'c.1847C>G', 'p.Pro616Arg' and 'p.P616R' across /Users/gs_agent/LYFESCI_HERMES_v9/vcep_db/MSH2/fulltext); the only occurrence in any declared VCEP file is HCI-PRIORS-MSH2.txt line 4078: exon 12, c.1847C>G, RNA '-', p.P616R, Custom_PP2_score 0.900, MAPP_score 41.48, MAPP/PP2_Prior 0.9565.
• The same VCEP HCI/MSH2 prior table, read as the complete enumeration of codon-616 single-nucleotide substitutions, gives c.1846C>A p.P616T, c.1846C>G p.P616A, c.1846C>T p.P616S, c.1847C>A p.P616H, c.1847C>G p.P616R, c.1847C>T p.P616L. Only the c.1847C>G entry encodes arginine, so codon 616 (CCT) admits exactly one nucleotide change encoding p.Pro616Arg.
• Governing specification text: PS1_Strong requires '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)'; PS1_Moderate requires the same relative to a previously established Likely Pathogenic missense variant with normal RNA result and PM2_supporting.
• The InSiGHT MMR VCEP pilot variant table (provisional VCEP classifications with applied codes for MSH2 and PMS2 pilot variants) contains no variant at residue 616 and no p.Pro616Arg entry.
• ClinVar record 127636 (NM_000251.3(MSH2):c.1847C>G, p.Pro616Arg) is reported as conflicting classifications of pathogenicity with 0 expert-panel submissions; the variant itself is therefore not an established pathogenic comparator under PS1.
• PS1_Moderate cannot be reached either: no residue-616 comparator is classified Likely Pathogenic by the VCEP, so the PM2_supporting-plus-comparator pathway is unavailable irrespective of PM2 status.

Still missing

• The only VCEP classification list available is the InSiGHT MMR VCEP pilot variant table (48 variants, none at residue 616); absence of a VCEP-established p.Pro616Arg or residue-616 Pathogenic/Likely Pathogenic record is established from that table plus ClinVar's zero expert-panel submissions for ClinVar variation 127636, not from a complete InSiGHT variant database export.

Sources

Reasoning

Governing specification: ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel MSH2 v2.0 (cspec). Its PM2 rule is: absent/extremely rare allele frequency < 0.00002 (< 1 in 50,000 alleles) in the gnomAD v4 dataset, at Supporting strength (the specification's PM2 rule set contains only the Supporting tier). The variant is not absent and is not below that cutoff in gnomAD v4.1: total AF 3.22189e-05 (52/1,613,962 alleles), exome AF 3.28357e-05, genome AF 2.62916e-05, and joint Grpmax FAF 0.00017995 - all above 0.00002, by 1.6-fold even for the lowest of these point estimates. The older gnomAD v2.1 all-comers total AF (6.71725e-05) and v2.1 non-cancer exome AF (5.48662e-05) are also above 0.00002, so the verdict does not depend on build or on cancer-cohort exclusion. The single closest figure is the GNOMAD_V3_1_NON_CANCER genome-only AF of 2.02856e-05 (3/147,888 alleles), which is still marginally above the 0.00002 threshold; because the MSH2 v2.0 specification explicitly names gnomAD v4 as the PM2 source, the v4.1 all-comers figures are governing here and PM2 is not met. This is consistent with the variant's non-absent, multi-ancestry, recurrent observation in population databases.

Rule

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

Evidence considered

• InSiGHT MSH2 v2.0 cspec PM2 rule (Supporting): 'Absent/extremely rare allele frequency <0.00002 (<1 in 50,000 alleles) in gnomAD v4 dataset.' The specification defines only the Supporting tier for PM2.
• gnomAD v4.1 (all-comers default, and the dataset named by the VCEP): total AF 3.22189e-05 with 52/1,613,962 alleles; exome AF 3.28357e-05 (48/1,461,822); genome AF 2.62916e-05 (4/152,140); joint Grpmax FAF 0.00017995; homozygotes 0. Every point estimate exceeds 0.00002.
• gnomAD v2.1 (context): total AF 6.71725e-05 (19/282,854 alleles), highest ancestry Admixed American AF 0.000310401 (11/35,438 alleles); also far above 0.00002.
• Non-cancer subsets checked because MSH2 is a cancer-predisposition gene: GNOMAD_V2_1_NON_CANCER exomes AF 5.48662e-05 (13/236,940 alleles) and GNOMAD_V3_1_NON_CANCER genomes AF 2.02856e-05 (3/147,888 alleles). The v3.1 non-cancer genomes-only figure is the lowest of all datasets and remains only marginally above 0.00002; either way the PM2 cutoff is not satisfied, so the verdict is stable regardless of which population source is preferred.
• gnomAD-Canada v1.0 returned absent for this variant; absence in a small national cohort does not offset the gnomAD v4.1 observations, and the VCEP PM2 rule is anchored to gnomAD v4.
• VCEP-pilot-variants---MMR.xlsx searched for c.1847C>G, p.Pro616Arg and p.P616R: no row found. The pilot set shows PM2_supporting is used broadly for rare MSH2/MSH6/PMS2 variants whose frequency is below the same order of magnitude, i.e. the code is applied only when the frequency criterion is genuinely satisfied.

Still missing

• The InSiGHT MMR VCEP pilot variant table (VCEP-pilot-variants---MMR.xlsx) contains no entry for NM_000251.3:c.1847C>G / p.Pro616Arg / p.P616R (searched in all three notations), so no VCEP precedent classification exists for this exact variant; the PM2 verdict rests on the specification's own rule text and the gnomAD v4.1 data.

Sources

Reasoning

Governing specification: ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel MSH2 specification v2.0, which lists PM3 as Applicable but replaces the generic ACMG definition with a point-scored rule. The VCEP rule requires co-occurrence of this variant (NM_000251.3:c.1847C>G, p.(Pro616Arg)) with a known pathogenic or likely pathogenic sequence variant in the same gene (MSH2) in a patient whose clinical features are consistent with CMMRD as per Aronson et al. 2022 (VCEP 'Table for CMMRD diagnosis'): 1.0 point when the other variant is confirmed in trans, 0.5 points when phase is unknown, and the summed points set the strength (Very Strong >=4, Strong 2-3.5, Moderate 1-1.5, Supporting =0.5). This is the recessive/biallelic reading of MSH2, whose disease associations in the VCEP ruleset are Lynch syndrome 1 (autosomal dominant) and mismatch repair cancer syndrome 1 / CMMRD (autosomal recessive). The case contains no proband-level data at all: it is a variant-level query with no second MSH2 variant, no phase determination by parental or offspring testing, and no CMMRD phenotypic scoring. Screening of the case evidence found no report of such co-occurrence anywhere: none of the 14 ClinVar submissions (no expert-panel submission; conflicting classifications dominated by Uncertain significance and Likely benign) describes the variant in trans with a second pathogenic MSH2 allele, and none of the fetched full texts (PMID:25645574, 25741868, 26467025, 28135145, 28944238, 31265121, 32849802, 33357406) names c.1847C>G or p.(Pro616Arg) or reports a biallelic/compound-heterozygous MSH2 case involving codon 616. Population data likewise contain no biallelic observation (gnomAD v4.1: 52/1,613,962 alleles, 0 homozygotes; gnomAD v2.1: 19/282,854 alleles, 0 homozygotes). The PM3 point score is therefore 0, below the 0.5-point Supporting threshold, so PM3 is not met.

Rule

PM3 (ClinGen InSiGHT MMR VCEP, MSH2 v2.0, points-based): co-occurrence with a known P/LP sequence variant in the same gene in a patient with CMMRD-consistent features (Aronson et al. 2022 table); P/LP in trans = 1.0 point, P/LP with phase unknown = 0.5 point; Very Strong >= 4, Strong 2-3.5, Moderate 1-1.5, Supporting = 0.5. Observed score: 0 points -> not met.

Evidence considered

• InSiGHT MMR VCEP MSH2 v2.0 PM3 instruction: '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 - Refer to "Table for CMMRD diagnosis.pdf"... Pathogenic/Likely Pathogenic in trans: 1.0 point; Pathogenic/Likely Pathogenic - phase unknown: 0.5 points. Sum all cases with the above evidence to determine the PM3 strength.'
• VCEP MSH2 v2.0 PM3 strength rules: Very Strong >= 4 points; Strong 2-3.5 points; Moderate 1-1.5 points; Supporting = 0.5 points (Stand Alone not applicable).
• MSH2 inheritance context from the VCEP ruleset: Lynch syndrome 1 (MONDO:0007356, autosomal dominant) and mismatch repair cancer syndrome 1 (MONDO:0010159, autosomal recessive); the recessive configuration (biallelic/compound heterozygous) is the scenario PM3 evaluates.
• Phase and co-occurrence search of this case: no second MSH2 variant, no in trans/phase determination, and no CMMRD phenotype scoring are recorded anywhere in the case evidence; the query is variant-level only.
• ClinVar variation 127636 (NM_000251.3(MSH2):c.1847C>G, p.Pro616Arg): 14 submissions with conflicting classifications (predominantly Uncertain significance and Likely benign, 2 Benign), no expert-panel submission, and no submitter report of the variant in trans with another pathogenic MSH2 allele.
• gnomAD v4.1: 52/1,613,962 alleles (AF 3.22189e-05), grpmax FAF 0.00017995, 0 homozygotes; gnomAD v2.1: 19/282,854 alleles (AF 6.71725e-05), 0 homozygotes - no biallelic individual observed in either population dataset.
• Full texts of PMID:25645574, PMID:25741868, PMID:26467025, PMID:28135145, PMID:28944238, PMID:31265121, PMID:32849802 and PMID:33357406 were screened for c.1847C>G / p.(Pro616Arg) and for compound-heterozygous or biallelic MSH2 observations: none of them names this variant or reports two MSH2 alleles in one patient at codon 616.
• Published ACMG/AMP 2015 PM3 definition requires phase to be established by testing parents or offspring, which was not performed for any case in this evidence set.

Still missing

• No proband-level genotype or phenotype data in this case: no co-occurring pathogenic/likely pathogenic MSH2 variant, no phase determination (parental or offspring testing), and no CMMRD phenotypic scoring against the VCEP CMMRD table (Aronson et al. 2022).
• If a CMMRD-context biallelic case is later identified, the PM3 point score must be recomputed (1.0 point if the second P/LP MSH2 variant is in trans; 0.5 point if phase is unknown).

Sources

Reasoning

The governing MSH2 specification (ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel, MSH2 v2.0) lists PM5 as Applicable with two levels: PM5_Moderate requires a 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_Supporting applies where the other variant is classified Likely Pathogenic on the protein level; both require that PP3 is supporting (or stronger) for the variant under assessment. Residue 616 has five alternative amino-acid substitutions, all documented in the VCEP HCI/MSH2 prior table: p.Pro616Thr (c.1846C>A), p.Pro616Ala (c.1846C>G), p.Pro616Ser (c.1846C>T), p.Pro616His (c.1847C>A) and p.Pro616Leu (c.1847C>T). Their ClinVar records are: p.Pro616Leu (ClinVar 1781260) Likely benign, p.Pro616His (1781259) Uncertain significance, p.Pro616Ala (850718) Uncertain significance, p.Pro616Thr (920493) conflicting classifications, p.Pro616Ser (142665) conflicting classifications. None is classified Pathogenic or Likely Pathogenic, and none carries an expert-panel (VCEP) classification; the InSiGHT MMR VCEP pilot classification table contains no residue-616 variant, and ClinVar lists zero expert-panel submissions for this gene/variant set. The PP3-supporting gate is satisfied for the variant under assessment (HCI prior probability 0.9565, MAPP score 41.48; >0.81 = PP3_Moderate under this VCEP's PP3 rule), but PP3 is only a necessary precondition for PM5 and no qualifying same-residue comparator exists, so PM5 is not met at either strength. The pipeline's automated comparator harvest for this case failed soft with HTTP 429 and returned zero same-residue candidates; that gap was closed by a direct ClinVar query for all MSH2 codon-616 substitutions, which returned the five records above.

Rule

Panel specification (MSH2 v2.0), 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.'

Evidence considered

• Governing specification, 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.' PM5_Supporting applies the same requirement to a Likely Pathogenic comparator.
• Comparator enumeration: the VCEP HCI/MSH2 prior table lists every single-nucleotide substitution at codon 616 (p.P616T, p.P616A, p.P616S, p.P616H, p.P616R, p.P616L), i.e. five amino-acid-changing substitutions other than the variant under assessment.
• ClinVar query for MSH2 codon-616 substitutions (NCBI E-utilities esearch/esummary, performed for this review) returned variation 1781260 c.1847C>T p.Pro616Leu = Likely benign; 1781259 c.1847C>A p.Pro616His = Uncertain significance; 850718 c.1846C>G p.Pro616Ala = Uncertain significance; 920493 c.1846C>A p.Pro616Thr = conflicting classifications of pathogenicity; 142665 c.1846C>T p.Pro616Ser = conflicting classifications of pathogenicity. None is Pathogenic or Likely Pathogenic and none has expert-panel review.
• The InSiGHT MMR VCEP pilot variant table contains no variant at residue 616, and the ClinVar record for this variant (127636) shows 0 expert-panel submissions, so no VCEP-classified same-residue comparator exists.
• pp3 precondition for the variant under assessment is met independently of this group: HCI prior probability 0.9565 and MAPP score 41.48 for c.1847C>G (VCEP PP3 rule: >0.81 = PP3_Moderate), so PM5 fails on the comparator requirement, not on the PP3 gate.
• pm5_candidates.json for this case: found = false, same_residue_candidates = 0, total_candidates_seen = 7, errors include 'HTTP Error 429: Too Many Requests'.
• PMID:25741868 is the origin of the PM5 criterion ('novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before'), applied here through the stricter MSH2 v2.0 VCEP wording.

Still missing

• The case's automated PM5 comparator harvest (pm5_candidates.json) failed with 'HTTP Error 429: Too Many Requests' and reported 0 same-residue candidates out of 7 seen; this gap was closed by querying ClinVar directly for MSH2 codon-616 substitutions, which returned the five known alternative substitutions (all Uncertain significance, Likely benign or conflicting).
• No InSiGHT VCEP expert-panel classification exists for any codon-616 variant in ClinVar and none appears in the VCEP pilot variant table, so PM5 rests on the absence of a Pathogenic/Likely Pathogenic comparator rather than on a documented VCEP rejection of a specific comparator.

Sources

Reasoning

The governing PP4 rules for MSH2 (ClinGen InSiGHT v2.0) are tumor-phenotype counting rules: PP4_Strong requires >=3 independent CRC/endometrial MSI-H tumors (standard 5-10 marker panel or tumor genome) and/or loss of MMR protein expression consistent with the variant location in >=2 families; PP4_Moderate requires 2 such independent tumors; PP4_Supporting requires 1 such CRC/endometrial MSI-H tumor, and an MSI-H tumor with inconsistent protein expression does not meet PP4 at all. Explicit comparison against the lowest (Supporting) tier: number of documented MSI-H tumors with concordant MMR protein loss = 0; 0 >= 1? no. The higher tiers fail a fortiori (0 >= 2? no; 0 >= 3? no). Evidence actually available is phenotype-weak: ClinVar submitter statements report the variant 'in individuals with colorectal cancer and/or suspected Lynch syndrome (Yurgelun et al.)' and the LabCorp submission records a frequency-based sentence, but no MSI status, no MMR immunohistochemistry, no tumor genome MSI result, and no family-level tumor counts are recorded for this variant in the case evidence. The CMMRD scoring table declared to govern PP4 was also applied as a fallback phenotype-specificity rule: CMMRD points observed = 0; 0 >= 3 points? no (and the table's own header maps >=3 points to the PM3/biallelic-CMMRD pathway after excluding NF1 and LFS, so it is not a monoallelic PP4 driver on its own). PP4 is therefore not met for this variant on the evidence collected.

Rule

ClinGen InSiGHT MSH2 v2.0 PP4: 1 CRC/Endometrial MSI-H tumor (standard 5-10 marker panel or tumor genome) and/or loss of MMR protein expression consistent with the variant location = PP4_Supporting; 2 tumors = PP4_Moderate; >=3 independent tumors in >=2 families = PP4_Strong; MSI-H with inconsistent protein expression does not meet PP4.

Evidence considered

• Governing PP4 rule text (InSiGHT MSH2 v2.0): Strong = >=3 independent CRC/Endometrial MSI-H tumors in >=2 families or tumor genome and/or loss of MMR protein expression consistent with the variant location; Moderate = 2 independent MSI-H tumors; Supporting = 1 CRC/Endometrial MSI-H tumor; 'MSI-H tumor with inconsistent protein expression does not meet PP4'.
• Explicit inequality, lowest tier first: documented qualifying MSI-H tumors = 0, so 0 >= 1? no; 0 >= 2? no; 0 >= 3? no. The comparison operator read from this criterion's own source text is '>=' (a minimum-count rule, strengthening upward).
• ClinVar variation 127636 submitter statements: PreventionGenetics reports the variant 'in individuals with colorectal cancer and/or suspected Lynch syndrome (Yurgelun et al.)'; LabCorp's statement is a population-frequency argument ('this frequency is not significantly higher than expected for a pathogenic variant in msh2 causing hereditary nonpolyposis colorectal cancer'), not a tumor MSI-H phenotype observation; no submission documents MSI-H status or MMR protein expression.
• Table-for-CMMRD-diagnosis.pdf (declared to govern PP4) is the C4CMMRD point-scoring table adapted from Aronson et al 2022 (PMID 33622763); explicit inequality: CMMRD points observed = 0, so 0 >= 3? no. Its header states '>=3 points = CMMRD features' and routes that score to PM3 after excluding NF1/LFS.
• VCEP-pilot-variants---MMR.xlsx searched for c.1847C>G / p.Pro616Arg / p.P616R / '1847' / '616': no row for this variant; PP4 is used elsewhere in that table for other variants, so its absence for this variant is a genuine no-entry finding, not an unsearched file.

Still missing

• No tumor-level data (MSI status, MMR immunohistochemistry, tumor-genome MSI) and no family tumor counts exist for the tested individual/family, so PP4 could not be evaluated at any strength; if a proband CRC/endometrial tumor proven MSI-H with MMR protein loss consistent with MSH2 exists, PP4_Supporting would become applicable.
• VCEP-pilot-variants---MMR.xlsx was searched in full for every notation of this variant (c.1847C>G, p.Pro616Arg, p.P616R, '1847', '616') and contains no entry for this variant; the file does list PP4/PP4_strong/PP4_moderate codes for other MSH2 pilot variants (for example c.1807G>A, c.1862G>T, c.2075G>T), so it was searched for a pre-assigned code and genuinely has none for c.1847C>G.
• Table-for-CMMRD-diagnosis.pdf was searched for every notation of this variant and contains no variant-specific entry (it is a generic CMMRD point-scoring table), so no pre-assigned PP4 verdict could be taken from it.
• No malignancy, age-at-diagnosis, pilomatricoma, consanguinity, immunoglobulin, or family-history features are recorded for the tested individual, so the CMMRD point score is 0 but is also not evaluable as a positive phenotype.

Sources

Reasoning

Governing specification is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for MSH2 v2.0 (cspec), which supersedes the generic ACMG/AMP population thresholds. Its BA1 rule is: gnomAD v4 Grpmax filtering allele frequency (FAF) >= 0.001 (0.1 percent) and the variant excluded as a founder pathogenic variant. The observed gnomAD v4.1 joint Grpmax FAF for NM_000251.3:c.1847C>G is 0.00017995 (0.018 percent; 52/1,613,962 alleles, exome Grpmax FAF 0.00022385, genome Grpmax FAF 0.00001171), which is roughly 5.6-fold below the 0.001 VCEP stand-alone threshold and 278-fold below the generic 0.05 stand-alone threshold. The highest ancestry-specific frequency is Middle Eastern at 0.000329 (2/6,084 alleles), also far below any stand-alone cutoff. BA1 is therefore not met.

Rule

InSiGHT MSH2 v2.0 BA1: gnomAD v4 Grpmax FAF >= 0.001 (0.1%) and variant excluded as founder pathogenic variant.

Evidence considered

• InSiGHT MSH2 v2.0 cspec BA1 rule (stand-alone): 'GnomAD v4 Grpmax filtering allele frequency >= 0.001 (0.1%) and variant is excluded as founder pathogenic variant.' This VCEP rule outranks the generic ACMG/AMP stand-alone threshold of 0.05.
• gnomAD v4.1 (all-comers default dataset): total AC 52 / AN 1,613,962, total AF 3.22189e-05; exome AC 48 / AN 1,461,822 (AF 3.28357e-05); genome AC 4 / AN 152,140 (AF 2.62916e-05); homozygotes 0 in every subset; joint Grpmax FAF 0.00017995 (exome Grpmax FAF 0.00022385; genome Grpmax FAF 0.00001171).
• gnomAD v2.1 (context only): total AF 6.71725e-05 (19/282,854 alleles), Grpmax FAF 0.00015644, highest ancestry Admixed American AF 0.000310401 (11/35,438 alleles), homozygotes 0. Neither dataset reaches 0.001.
• Non-cancer subset context: GNOMAD_V2_1_NON_CANCER exomes AF 5.48662e-05 (13/236,940 alleles, homozygotes 0) and GNOMAD_V3_1_NON_CANCER genomes AF 2.02856e-05 (3/147,888 alleles, homozygotes 0). Neither approaches the BA1 threshold, so substituting a non-cancer source does not change the verdict. The MSH2 v2.0 specification explicitly names gnomAD v4 for BA1, so the all-comers gnomAD v4.1 entry is the governing source.
• Founder-pathogenic exclusion: no source consulted designates c.1847C>G as a founder pathogenic allele; the variant is observed across multiple ancestries (Admixed American, Middle Eastern, European non-Finnish) rather than as an ancestry-restricted pathogenic founder. This clause is not load-bearing because the frequency does not reach the BA1 cutoff in any case.

Sources

Reasoning

Scope first: this is a missense substitution (NP_000242.1:p.(Pro616Arg)), so BP4 rests on the missense protein-impact path. The governing ClinGen InSiGHT MSH2 v2.0 specification defines BP4_Supporting for missense variants as an HCI prior probability of pathogenicity <0.11; the on-record HCI-PRIORS-MSH2 row for c.1847C>G (p.P616R) gives an HCI prior of 0.900, far above that threshold, so the VCEP's own BP4 rule is not satisfied. The alternative SpliceAI branch of the same VCEP BP4 rule is explicitly restricted to 'intronic and synonymous variants' (SpliceAI delta score <= 0.1) and is therefore unavailable for this exonic missense change; the generic framework likewise forbids substituting a clean splice prediction for a missense variant's BP4. On the missense path itself the corroborating local REVEL score is 0.84, which is not <=0.29 (ClinGen SVI BP4_supporting, Pejaver et al. 2022, PMID:36413997), so BP4 is not met on either reading of the missense evidence. PP3 and BP4 therefore do not both apply, as required - the variant is benign-direction in neither predictor.

Rule

VCEP MSH2 v2.0 BP4 (Supporting): missense variant with HCI prior probability of pathogenicity <0.11 -> BP4_Supporting; SpliceAI delta <= 0.1 limb applies only to intronic and synonymous variants

Evidence considered

• cspec (ClinGen InSiGHT MSH2 v2.0) BP4 rule: '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.' The SpliceAI limb is scoped by the rule text to intronic and synonymous variants only and cannot be invoked for this missense variant.
• hci_prior / vcep_hci_priors_msh2 (HCI-PRIORS-MSH2.txt, line 4078): c.1847C>G / p.P616R, HCI prior 0.900, MAPP 41.48, reference {PMID22949387:Thompson et al., 2013}. 0.900 is not <0.11, so the VCEP's own BP4 threshold is not met; the entry exists, so this is a genuine negative result rather than an absence of data.
• revel (local REVEL score 0.84): not <=0.29 (BP4_supporting) and not <=0.183 (moderate) or <=0.016 (strong) under the ClinGen SVI REVEL calibration (Pejaver et al. 2022, PMID:36413997), so the missense path provides no BP4 support either.
• SpliceAI for this variant (max delta 0.001, below the <=0.1 rule cutoff) was deliberately not used as BP4 support: the VCEP's SpliceAI limb is restricted to intronic and synonymous variants, and a missense variant's PP3/BP4 may not be taken from the splice path even when the splice delta looks benign.
• bayesdel (score 0.462389) recorded but not used - no established ACMG-strength calibration and no VCEP-specified BayesDel cutoff, so not_available for BP4.
• PMID:32849802 (mRNA splicing / CIMRA study) was consulted on whether splicing data should carry in-silico weight for predicted missense substitutions; it does not mention this variant and reports that splicing data contributed to classification for only a minority of predicted missense/in-frame alterations, supporting the decision not to route this missense variant's PP3/BP4 through the splice path.

Sources

Reasoning

The governing BP5 rules for MSH2 (ClinGen InSiGHT v2.0) are tumor-phenotype counting rules in the benign direction: BP5_Strong requires >=4 tumors that are CRC/endometrial MSS and/or show no loss of MMR protein expression and/or are LS-spectrum tumors with loss of MMR protein(s) inconsistent with the gene harbouring the variant, OR >=2 BRAF V600E (CRC only)/MLH1 methylation (in an LS-spectrum tumor only) with MSI-H/MLH1 loss; BP5_Supporting requires 2 or 3 such tumors, OR 1 BRAF V600E (colon only)/MLH1 methylation (in an LS-spectrum tumor only) with MSI-H/MLH1 loss. Explicit inequalities read from this criterion's own source text (both operators are minimum-count '>=' rules, strengthening upward): qualifying benign-pattern tumors documented = 0, so 0 >= 4? no (BP5_Strong not met) and 2 <= 0 <= 3? no (BP5_Supporting not met); qualifying BRAF V600E / MLH1 methylation events = 0, so 0 >= 2? no and 0 >= 1? no. No MSS result, no MMR immunohistochemistry, no BRAF V600E result, and no MLH1 methylation result is recorded for this variant or the tested individual. Benign laboratory classifications exist in ClinVar (2 Benign, 5 Likely benign, 2 likely benign) but a submitter's final classification is not BP5 evidence and was not converted into BP5; note that this criterion uses tumor-count thresholds, not a likelihood-ratio threshold, so no <= operator applies to it in this framework.

Rule

ClinGen InSiGHT MSH2 v2.0 BP5: 2 or 3 CRC/endometrial MSS or no-MMR-loss tumors (or 1 BRAF V600E/MLH1 methylation event) = BP5_Supporting; >=4 tumors (or >=2 BRAF V600E/MLH1 methylation events) = BP5_Strong.

Evidence considered

• Governing BP5 rule text (InSiGHT MSH2 v2.0): Strong = '>=4 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 >=2 BRAF V600E (CRC only)/MLH1 methylation (in LS spectrum tumor only) with MSI-H/MLH1 loss'; Supporting = '2 or 3 tumors ... OR 1 BRAF V600E (Colon only)/MLH1 methylation (in LS spectrum tumor only) with MSI-H/MLH1 loss'.
• Explicit inequalities using the operators as literally written in the BP5 text (all '>=' minimum-count rules): 0 >= 4? no; 0 >= 2? no; 2 <= 0 <= 3? no; 0 >= 1? no. No 'lenient' <= threshold exists in this criterion's governing text.
• ClinVar variation 127636: 2 Benign, 5 Likely benign and 2 likely benign laboratory submissions exist, but none documents an alternate molecular basis or an MSS/no-MMR-loss tumor phenotype, and submitter final classifications were not used as BP5 evidence.
• No tumor MSI, MMR immunohistochemistry, BRAF V600E or MLH1 methylation data are present anywhere in the case evidence bundle for this variant.

Still missing

• No tumor-level benign-pattern evidence (MSS status, MMR protein expression, BRAF V600E, MLH1 methylation) is recorded for the tested individual; if 2 or more MSS/no-MMR-loss tumors exist, BP5_Supporting would become applicable.
• Table-for-CMMRD-diagnosis.pdf and VCEP-pilot-variants---MMR.xlsx were searched for this variant in every notation and contain no BP5 entry for it, so no pre-assigned benign verdict could be taken from the declared VCEP files.

Sources

Reasoning

The governing framework is the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Expert Panel specification for MSH2 Version 2.0 (default strength Pathogenic Strong), which replaces the generic ACMG de novo criterion with a de novo point score. It awards 2 points per proband for a de novo variant with both maternity and paternity confirmed in a case with an MMR-deficient Lynch-spectrum tumour (MSI/IHC consistent with the affected gene and no MLH1 methylation); 1 point per proband for a de novo variant with both maternity and paternity confirmed in a case with a Lynch-spectrum tumour but no MSI/IHC/methylation data; and 0.5 points per proband for an assumed de novo variant with maternity and/or paternity unconfirmed in a case with a Lynch-spectrum tumour and no tumour data. Strength is then read from the total: >=4 points Very Strong, 2-3.5 points Strong, 1-1.5 points Moderate, 0.5 points Supporting. The available evidence documents no proband, no parental genotypes, no maternity or paternity confirmation, no de novo observation, no Lynch-spectrum tumour or tumour MSI/IHC/MLH1-methylation result for this variant, so no de novo points can be awarded and PS2 cannot be assigned. Treating an unpopulated point total as 'not met' would misrepresent an evidence gap as a negative finding, so the criterion is left not assessed with the gap recorded.

Rule

InSiGHT MSH2 v2.0 PS2: >=4 de novo points = Very Strong; 2-3.5 = Strong; 1-1.5 = Moderate; 0.5 = Supporting.

Evidence considered

• The InSiGHT MSH2 Version 2.0 specification requires 2, 1, or 0.5 de novo points per qualifying proband depending on parental confirmation and on whether an MMR-deficient Lynch-spectrum tumour is documented, and converts the total to Very Strong (>=4), Strong (2-3.5), Moderate (1-1.5) or Supporting (0.5).
• No proband, parental genotype, maternity/paternity confirmation, de novo observation, Lynch-spectrum tumour diagnosis, or tumour MSI/IHC/methylation result is documented for this variant.
• ClinVar holds conflicting classifications for this variant from 14 submitters (Uncertain significance and Likely benign/Benign), with no expert-panel submission, and none of the 21 literature records was triaged to a de novo evidence theme; no source supplies a de novo occurrence for c.1847C>G.

Still missing

• Document the proband and phenotype, including whether a Lynch-spectrum tumour (colorectal/endometrial/ovarian/small bowel/renal pelvis/ureter/gastric carcinoma, sebaceous skin tumour, or glioma) is present.
• Provide variant testing results for the proband and for both biological parents, including confirmation of maternity and paternity.
• Provide tumour MSI, MMR protein IHC, and MLH1 methylation results, since these determine whether a proband scores 2, 1, or 0.5 de novo points.
• State whether the de novo status is confirmed or only assumed, and calculate the resulting de novo point total.

Sources

Reasoning

The InSiGHT MSH2 VCEP specification (v2.0) assigns PS3 only from functional evidence: calibrated functional assays with functional odds for pathogenicity >18.7 (Strong), >4.3 and <=18.7 (Moderate, or MMR function defect following the MMR functional assay flowchart, or monoallelic expression with complete loss of the variant allele at <10% of wild-type in cDNA without puromycin), or >2.08 and <=4.3 (Supporting). The VCEP-approved SVI functional assay documentation lists the calibrated assays available for MSH2, including the Drost CIMRA cell-free MMR assay (PMID:30504929; 31965077) with abnormal readout <23% for MSH2, the Jia/Scott massively parallel MSH2 deep-mutational-scanning assay (PMID:33357406; 36550560) with abnormal readout LoF score >0.4 and normal LoF score <=0, and the Bouvet methylating-agent survival assay (PMID:30998989) with abnormal MSH2 readout mean survival >45.54%, together with validated positive and negative controls (e.g. 22 P/LP and 26 B/LB controls for the Jia assay). All three assay notations for this variant - c.1847C>G, p.Pro616Arg, p.P616R - were searched in Functional-assay-SVI-documentation-MMR.xlsx and Functional-assay-flowchart.pdf and in every fetched full text. No variant-specific assay result, readout value, or functional odds value for MSH2 c.1847C>G (p.Pro616Arg) exists in the case evidence: the assay documentation file contains only assay-level rows with no variant-level entries, and PMID:33357406 (the calibrated MSH2 deep-mutational-scanning assay, whose map covers 94.4% of possible single-codon substitutions) reports no score for codon 616 in its text, its variant-level scores residing only in the Figure 2 heatmap and supplementary tables. The MMR functional assay flowchart likewise cannot be applied, because it requires an actual calibrated functional assay result (or two concordant independent assays) for the variant. OncoKB reports no variant-specific reviewed functional evidence. The HCI-PRIORS MSH2 table does contain a variant-level row for c.1847C>G, but that table holds computational pathogenicity priors used for PP3/BP4, not a functional assay measurement, and it cannot support PS3. A met PS3 therefore cannot be established; because the decisive missing item is a variant-specific calibrated assay measurement that plausibly exists in the published deep-mutational-scanning dataset, the criterion is recorded as not_assessed rather than not_met.

Rule

Calibrated functional assays with functional odds for pathogenicity > 18.7 (Strong); >4.3 and <= 18.7, OR MMR function defect following functional assay flowchart, OR monoallelic expression with complete loss of expression (<10% of wild-type in cDNA without puromycin) of the variant allele (Moderate); >2.08 and <= 4.3 (Supporting).

Evidence considered

• MSH2 VCEP specification v2.0 PS3: Strong at calibrated functional odds for pathogenicity >18.7; Moderate at >4.3 and <=18.7, or MMR function defect following the functional assay flowchart, or monoallelic expression with complete loss of the variant allele (<10% of wild-type in cDNA without puromycin); Supporting at >2.08 and <=4.3.
• Functional-assay-SVI-documentation-MMR.xlsx (sheet 'MMR assays') documents only assay-level rows. For MSH2 the applicable approved calibrated assays are: Drost CIMRA (PMID:30504929; 31965077; abnormal <23% for MSH2, normal >=70%; 10 MLH1/MSH2 + 3 MSH6 P/LP and 10 + 4 B/LB validation controls; proposed PS3/PS3_Moderate/PS3_Supporting/BS3/BS3_Supporting); Jia/Scott massively parallel deep mutational scan (PMID:33357406; 36550560; abnormal LoF score >0.4, normal LoF score <=0; 22 P/LP and 26 B/LB controls; proposed PS3/BS3); Thompson 2013 (PMID:24362816; MMR activity complementation, protein expression and subcellular localization; <25% vs >75% in two independent assays; 23/7/18 P/LP and 14/5/10 B/LB controls; PS3_moderate/BS3_supporting); Morak (PMID:31332305; cDNA full-length transcript analysis with and without puromycin, allelic loss <=10%; 25 P/LP and 33 B/LB controls); and Bouvet (PMID:30998989; methylating-agent cell survival, abnormal MSH2 mean survival >45.54%, normal <45.54%; 10 MSH2 P/LP and 11 B/LB controls).
• Functional-assay-flowchart.pdf is an image-only PDF whose text layer is empty; its logic was read from the converted vcep_materials summary: for an in-frame indel or missense variant not at the first/last 3 bases of an exon and without predicted splice impact, PS3/BS3 may be taken either from a calibrated functional assay (quantitative multifactorial analysis or conversion of the functional likelihood ratio to ACMG/AMP weight) or, when no calibrated assay was performed, from two independent assays with concordant results (mammalian MMR activity, protein expression/stability, subcellular localization and cell-based MMR activity).
• All three notations of this variant (c.1847C>G, p.Pro616Arg, p.P616R) were searched in both declared VCEP functional files; neither contains a variant-level row for this variant.
• PMID:33357406 (Jia et al. 2021) is the VCEP-approved calibrated MSH2 functional assay; it reports an MSH2 deep-mutational-scanning map in MSH2-knockout HAP1 cells with 6-thioguanine selection covering 94.4% of possible single-codon substitutions, concordant with prior functional studies (95.8%) and with expert clinical classifications (sensitivity 95.3%, specificity 95.5%), but the variant c.1847C>G / p.Pro616Arg is not named anywhere in the text and no LoF score for codon 616 is given.
• PMID:32849802 tested mRNA splicing in 24 MMR variants and CIMRA repair activity in 12 predicted missense variants; the MSH2 missense variants tested are c.1275A>G, c.1344C>T and c.2154A>G, and c.1847C>G is not among them, so no variant-specific proficient or deficient functional result is obtainable from this source.
• OncoKB (MSH2 P616R page): 'OncoKB did not identify variant-specific reviewed functional evidence for this variant; gene-level curated context is available for reviewer follow-up.'
• The HCI-PRIORS MSH2 table (vcep_hci_priors_msh2) contains a variant-level row for c.1847C>G (p.P616R; MAPP score 41.48, custom PP2 score 0.900, HCI prior probability 0.9565, reference PMID:22949387), but this table supplies computational pathogenicity priors that the MSH2 VCEP governs under PP3/BP4, not a functional assay readout, and it is therefore not PS3 evidence.

Still missing

• Functional-assay-SVI-documentation-MMR.xlsx was searched for c.1847C>G, p.Pro616Arg and p.P616R; the file documents assays, thresholds and controls only, and contains no variant-level entry for this variant.
• Functional-assay-flowchart.pdf was searched for c.1847C>G, p.Pro616Arg and p.P616R; no variant-specific entry was found, and the flowchart requires an assay result that does not exist in this record.
• PMID:33357406 (calibrated MSH2 deep-mutational-scanning assay) reports no variant-specific LoF score for codon 616 in its full text; the variant-level scores are held only in the Figure 2 heatmap and supplementary tables, which were not retrieved, so the LoF score for p.(Pro616Arg) remains unknown.
• No variant-specific functional odds value (needed for the VCEP PS3 Strong/Moderate/Supporting thresholds of >18.7, >4.3-18.7 and >2.08-4.3) is available for this variant.
• No patient-derived mRNA assay (with and without NMD inhibition) exists for this variant, so the VCEP PS3 Moderate monoallelic-expression route (complete loss of the variant allele, <10% of wild-type) cannot be evaluated.

Sources

Reasoning

The governing InSiGHT MSH2 Version 2.0 specification (default strength Pathogenic Supporting) scores PP1 from a combined segregation Bayes likelihood ratio: >2.08 and <=4.3 is Supporting, >4.3 and <=18.7 is Moderate, and >18.7 in at least two families is Strong, with results from multiple pedigrees combined by multiplication and the analysis performed with a co-segregation tool such as COOL v3 using the overall Bayes factor. No pedigree, affected or unaffected relative genotype, phenotype assessment, informative meiosis, family count, or calculated segregation Bayes likelihood ratio is documented for this case, so no PP1 strength band can be placed. Without any segregation observation the criterion is treated as not assessable rather than as a failed co-segregation result.

Rule

InSiGHT MSH2 v2.0 PP1: combined Bayes likelihood ratio >2.08 and <=4.3 = Supporting; >4.3 and <=18.7 = Moderate; >18.7 in >=2 families = Strong.

Evidence considered

• The InSiGHT MSH2 Version 2.0 specification sets PP1 at a combined segregation Bayes likelihood ratio >2.08 and <=4.3 for Supporting, >4.3 and <=18.7 for Moderate, and >18.7 in at least two families for Strong, with multi-pedigree results combined by multiplication.
• No pedigree, segregation genotype, informative meiosis, phenotype segregation observation, family count, or Bayes likelihood ratio is documented for this variant.
• None of the 21 literature records retrieved for this variant was triaged to a segregation theme, and the seven full texts extracted for this case report no family segregation data for MSH2 c.1847C>G.

Still missing

• Provide a pedigree identifying affected and unaffected relatives with their cancer phenotypes and ages at diagnosis.
• Provide variant testing results for informative relatives and the number of informative meioses.
• Provide the family-level and combined segregation Bayes likelihood ratio, including the tool, allele-frequency and model parameters used.
• State the number of independently segregating families, since PP1 Strong requires a combined likelihood ratio >18.7 in >=2 families.

Sources

Reasoning

The governing MSH2 v2.0 specification (cspec) does not define BS2 as a healthy-homozygote or population-frequency criterion; it repurposes BS2 as a phase criterion: 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), with no previous or current evidence of CMMRD clinical manifestations per Aronson et al 2022, and where phase is confirmed by testing of parents or offspring. No such evidence exists in this case's evidence set: no second same-gene pathogenic variant, no parental/offspring phase testing, and no CMMRD scoring data were captured for the proband. In the population direction, the variant is never observed homozygous in gnomAD (0 homozygotes in v2.1, v4.1, and the two non-cancer subsets), so there is no healthy-homozygote observation to support the classic BS2 formulation either. BS2 is therefore not assessed rather than met or refuted.

Rule

InSiGHT MSH2 v2.0 BS2: co-occurrence in trans with a known pathogenic same-gene variant in a CRC patient after age 45 without CMMRD features; phase confirmed by parental/offspring testing. Strength: Strong.

Evidence considered

• InSiGHT MSH2 v2.0 cspec BS2 rule (Benign Strong) is a trans/phase criterion, not a frequency criterion: '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.'
• Case evidence set contains no second MSH2 (or other MMR) pathogenic variant in the proband, no parental or offspring phase testing, and no CMMRD clinical scoring; the ClinVar submissions listed in this case are all for c.1847C>G itself and none report trans co-occurrence.
• gnomAD homozygote counts are 0 in every dataset queried (v2.1 total, v4.1 total, GNOMAD_V2_1_NON_CANCER, GNOMAD_V3_1_NON_CANCER), so there is no healthy adult homozygote observation bearing on BS2.
• VCEP-pilot-variants---MMR.xlsx was searched for c.1847C>G, p.Pro616Arg and p.P616R with no match; the file shows BS2 applied to other pilot variants (MSH6 c.107C>T; MSH2 c.913G>A), indicating the criterion is used by InSiGHT in its co-occurrence form.

Still missing

• No phase-confirmed in-trans companion variant in MSH2 is documented for the proband; parental/offspring testing results and CMMRD features (Aronson et al 2022 scoring, Table-for-CMMRD-diagnosis.pdf) are absent from the case evidence set.
• gnomAD provides no homozygous carrier observation for c.1847C>G in any dataset (0 homozygotes across v2.1, v4.1, v2.1 non-cancer exomes, v3.1 non-cancer genomes), so a healthy-homozygote argument cannot be constructed from population data alone.
• The InSiGHT MMR VCEP pilot variant table (VCEP-pilot-variants---MMR.xlsx) contains no entry for c.1847C>G / p.Pro616Arg / p.P616R; however the table does contain pilot variants assigned BS2 (e.g. MSH6 c.107C>T, MSH2 c.913G>A), confirming BS2 is used by this panel in its trans-co-occurrence sense rather than as a population code.

Sources

Reasoning

The InSiGHT MSH2 VCEP specification (v2.0) assigns BS3 only from functional evidence: calibrated functional assays with functional odds for pathogenicity <=0.05 (Strong), or >0.05 and <=0.48 (Supporting), or - for synonymous and intronic variants - laboratory mRNA assays with NMD inhibition showing no aberration, or variant-specific proficient function in protein- and mRNA-based lab assays according to the MMR functional assay flowchart. All three notations of this variant (c.1847C>G, p.Pro616Arg, p.P616R) were searched in Functional-assay-SVI-documentation-MMR.xlsx, Functional-assay-flowchart.pdf and every fetched full text. No variant-specific functional measurement exists for MSH2 c.1847C>G (p.Pro616Arg): the SVI assay documentation contains only assay-level rows (Drost CIMRA and the Jia deep-mutational-scanning assay are both approved calibrated MSH2 assays, with normal readouts defined as >=70% repair for CIMRA and LoF score <=0 for the Jia assay), and no variant-level entry; PMID:33357406 does not name the variant or report a codon 616 LoF score in its text; PMID:32849802 did not assay c.1847C>G; and OncoKB reports no variant-specific reviewed functional evidence. The MMR functional assay flowchart branch for 'similar repair levels to wild-type or likely benign controls' therefore cannot be evaluated because no assay was performed on this variant. Benign functional evidence under BS3 cannot be established on assay-wide tolerance rates or on the variant being classified as benign/likely benign by other laboratories, since a criterion may never be inferred from another source's final classification. The criterion is recorded as not_assessed rather than not_met because the decisive missing item - a variant-specific calibrated assay measurement, which plausibly exists for codon 616 in the published deep-mutational-scanning dataset - is a data-availability gap rather than evidence of preserved function.

Rule

Calibrated functional assays with functional odds for Pathogenicity <= 0.05 (Strong) OR synonymous substitutions and intronic variants with no associated mRNA aberration (splicing or allelic imbalance) by laboratory assays conducted with nonsense-mediated decay inhibition; functional odds >0.05 and <=0.48 OR variant-specific proficient function in protein- and mRNA-based lab assays as per MMR functional assay flowchart (Supporting).

Evidence considered

• MSH2 VCEP specification v2.0 BS3: Strong at calibrated functional odds for pathogenicity <=0.05, or synonymous/intronic variants with no mRNA aberration (splicing or allelic imbalance) by laboratory assays using NMD inhibition; Supporting at functional odds >0.05 and <=0.48, or variant-specific proficient function in protein- and mRNA-based lab assays as per the MMR functional assay flowchart.
• Functional-assay-SVI-documentation-MMR.xlsx (sheet 'MMR assays') lists approved calibrated MMR assays that would support BS3, with their normal-function thresholds and validation controls for MSH2: Drost CIMRA (PMID:30504929; 31965077; normal repair >=70% for MLH1/MSH2; 10 MLH1/MSH2 + 3 MSH6 P/LP and 10 + 4 B/LB controls); Jia/Scott deep mutational scan (PMID:33357406; 36550560; normal LoF score <=0; 22 P/LP and 26 B/LB controls); Thompson 2013 (PMID:24362816; >75% in two independent assays; 23/7/18 P/LP and 14/5/10 B/LB controls); Morak (PMID:31332305; biallelic 50 +/- 10% expression; 25 P/LP and 33 B/LB controls); Bouvet (PMID:30998989; normal MSH2 mean survival <45.54%; 10 MSH2 P/LP and 11 B/LB controls). The file contains no variant-level rows, so none of these thresholds can be evaluated for this variant.
• All three notations of this variant (c.1847C>G, p.Pro616Arg, p.P616R) were searched in both declared VCEP functional files; no variant-level entry was found in either.
• Functional-assay-flowchart.pdf is an image-only PDF with an empty text layer; its logic was read from the converted vcep_materials summary, in which the missense/in-frame-indel branch reaches BS3 only through a calibrated functional assay, or through protein expression/stability (>=75% relative expression or similar to wild-type control), subcellular localization (nuclear) and cell-based MMR activity assays showing sensitive-to-damage, non-mutator behaviour, or through patient-derived RNA splicing assays.
• PMID:33357406 (Jia et al. 2021) reports the calibrated MSH2 deep-mutational-scanning assay but does not name c.1847C>G / p.Pro616Arg and gives no LoF score for codon 616; the reported assay-wide observation that 89.4% of scored MSH2 missense variants are functionally neutral, and that more than half of residues tolerate any substitution, is a cohort-level statistic and cannot be applied to this specific substitution.
• PMID:32849802 assayed mRNA splicing for 24 MMR variants and CIMRA repair activity for 12 predicted missense variants; MSH2 c.1847C>G is not among the variants tested, so it provides no variant-specific evidence of preserved or impaired function.
• OncoKB (MSH2 P616R page) reports no variant-specific reviewed functional evidence, so no curated benign functional assertion is available.
• The HCI-PRIORS MSH2 table (vcep_hci_priors_msh2) records a variant-level row for c.1847C>G (p.P616R; HCI prior probability 0.9565, MAPP score 41.48, custom PP2 score 0.900). That table supplies computational pathogenicity priors governed by the MSH2 VCEP under PP3/BP4; it is not a functional assay readout and cannot evidence preserved function for BS3.
• ClinVar submissions include submissions classified as Benign or Likely Benign by several clinical laboratories, but a classification stated by another submitter is not functional assay evidence and was not used for BS3.

Still missing

• Functional-assay-SVI-documentation-MMR.xlsx was searched for c.1847C>G, p.Pro616Arg and p.P616R; the file contains assay thresholds and control counts but no variant-level entry for this variant.
• Functional-assay-flowchart.pdf was searched for c.1847C>G, p.Pro616Arg and p.P616R; no variant-specific entry was found, so the flowchart's proficient-function branch cannot be applied.
• PMID:33357406 (calibrated MSH2 deep-mutational-scanning assay) reports no LoF score for codon 616 in its text; the score is held only in the Figure 2 heatmap and supplementary tables, which were not retrieved, so it is unknown whether p.(Pro616Arg) has a normal (LoF score <=0) or abnormal (>0.4) readout.
• No variant-specific functional odds value is available for this variant, so the VCEP BS3 thresholds (<=0.05 Strong; >0.05 and <=0.48 Supporting) cannot be applied.
• No variant-specific protein- or mRNA-based assay result exists in this record to support BS3_Supporting via the MMR functional assay flowchart.

Sources

Reasoning

The governing InSiGHT MSH2 Version 2.0 specification (default strength Benign Strong) scores BS4 from a combined segregation Bayes likelihood ratio in the non-segregation direction: <0.05 is Strong and >0.05 and <=0.48 is Supporting, again with multi-pedigree results combined by multiplication. The specification's generic caveat applies in full here: phenocopies are common for Lynch-spectrum cancers and a family may carry a second pathogenic MMR variant, both of which can mimic lack of co-segregation and must be excluded before BS4 is applied. No pedigree, affected-relative genotype, phenotype assessment, non-segregating observation, family count, or calculated Bayes likelihood ratio is documented for this case, so no BS4 strength band can be placed and the criterion is left not assessed rather than recorded as a failed non-segregation finding.

Rule

InSiGHT MSH2 v2.0 BS4: combined Bayes likelihood ratio <0.05 = Strong; >0.05 and <=0.48 = Supporting.

Evidence considered

• The InSiGHT MSH2 Version 2.0 specification sets BS4 Strong at a combined non-segregation Bayes likelihood ratio <0.05 and BS4 Supporting at >0.05 and <=0.48, and warns that phenocopies for common cancers and additional pathogenic variants in the family can produce apparent non-segregation.
• No pedigree, affected-relative variant testing, phenotype assessment, non-segregating observation, family count, or Bayes likelihood ratio is documented for this variant.
• Neither the case-level sources nor the extracted full texts report any family in which MSH2 c.1847C>G was tested in affected and unaffected relatives.

Still missing

• Provide a pedigree identifying affected relatives and their disease phenotypes.
• Provide variant testing results for affected and unaffected relatives with the number of informative meioses.
• Document possible phenocopies and any second pathogenic MMR variant in the family, which can confound apparent non-segregation.
• Provide the family-level and combined non-segregation Bayes likelihood ratio, including model assumptions.

Sources