MLH3 c.3455G>A, p.(Arg1152His)

Reasoning

PM2 (moderate in ACMG/AMP 2015; applied at supporting strength per the local framework convention) supports pathogenicity when a variant is absent from - or at extremely low frequency (below the expected carrier frequency if recessive) in - a large general-population or control cohort (>1000 individuals), race-matched to the patient (Richards et al., PMID:25741868). MLH3 c.3455G>A is present in gnomAD but at extremely low frequency: total AF = 0.0128% (206/1,612,008 alleles) in gnomAD v4.1 and 0.00636% (16/251,476 alleles) in gnomAD v2.1, 0 homozygotes in both; grpmax filtering allele frequency 0.0144% (v4.1); and the variant is absent from gnomAD-Canada v1.0 (HostSeq). All values are below the local non-VCEP gnomAD PM2 calibration of <0.1% and are also below the expected disease allele frequency (~0.17% from path_MMR carrier prevalence ~1/300, PMID:34043773), supporting the 'extremely low frequency' interpretation. The recessive arm of PM2 is applicable because MLH3 has an established recessive germline phenotype (biallelic loss-of-function variants underlie polyposis predisposition; gene-context artifact pvs1_gene_context, literature-only, lists PMID:30573798). Threshold-artifact check: the cross-population spread of the 206 v4.1 alleles (NFE 192, AFR 5, AMR 3, remaining 3, FIN 1, EAS 1, SAS 1) with zero homozygotes and concordance between v2.1 and v4.1 argues that this is a genuine rare variant, not an artifact, and the grpmax FAF (0.0144%) remains below the 0.1% calibration even after filtering-allele-frequency adjustment. PM2 is met at supporting strength.

Rule

PM2 (moderate in ACMG/AMP 2015, PMID:25741868; applied at supporting strength per local framework convention): Absent in controls (or at extremely low frequency if recessive) in Exome Sequencing Project, 1000 Genomes Project, or Exome Aggregation Consortium; local non-VCEP gnomAD calibration <0.1%. Met: AF 0.0128% (gnomAD v4.1) / 0.00636% (gnomAD v2.1), 0 homozygotes, absent from gnomAD-Canada, grpmax FAF 0.0144%.

Evidence considered

• gnomad_v4: gnomAD v4.1 total AF = 0.0001278 (0.01278%, 206/1,612,008 alleles), 0 homozygotes, grpmax FAF = 0.00014369 (0.01437%); below the local non-VCEP PM2 calibration of <0.1% (PMID:25741868 defines PM2 as absent or extremely low frequency, with extremely low frequency if recessive).
• gnomad_v2: gnomAD v2.1 total AF = 6.36e-05 (0.00636%, 16/251,476 alleles), 0 homozygotes; independently confirms extremely low frequency.
• gnomad_canada: variant absent from gnomAD-Canada v1.0 (HostSeq Canadian general-population genomes), consistent with extremely low population frequency.
• pvs1_gene_context: gene-context artifact (literature-only) supporting MLH3 loss-of-function germline disease mechanism and listing PMID:30573798 ('Biallelic germline nonsense variant of MLH3 underlies polyposis predisposition'), used to justify the recessive 'extremely low frequency if recessive' arm of PM2; cited only as gene-context artifact, not as a direct variant-level source.
• PMID:25741868: ACMG/AMP 2015 rule - PM2 when the variant is absent from (or below the expected carrier frequency if recessive) a large general population or control cohort (>1000 individuals) race-matched to the patient; the paper cautions absence/rarity alone is not sufficient or strong evidence of pathogenicity, supporting supporting-strength application.
• PMID:34043773: expected path_MMR carrier prevalence ~1 in 300 (allele frequency ~0.17%); the observed AF of 0.0128% is below this expected disease allele frequency, consistent with the PM2 'below expected carrier frequency' condition.

Still missing

• The variant is present (not absent) in gnomAD; PM2 is applied via the 'extremely low frequency' / local <0.1% calibration arm rather than the strict absence arm, so the supporting strength (rather than moderate) reflects the reduced certainty from a present-but-rare variant.
• No ClinGen CSPEC/VCEP for MLH3 exists to confirm the applicable disease model (dominant vs recessive) and the corresponding PM2 threshold; the recessive-arm justification rests on the gene-context artifact pvs1_gene_context (literature-only, citing PMID:30573798 for biallelic MLH3 polyposis predisposition), which is not a formal VCEP determination.
• No MLH3-specific expected carrier frequency has been published; the comparison uses the classical path_MMR carrier prevalence (~1/300, PMID:34043773) as an upper-bound anchor.

Sources

Reasoning

PS1 (ACMG/AMP 2015, PMID:25741868) requires the same amino-acid change as a previously established pathogenic variant regardless of nucleotide change. The case variant NM_001040108.2:c.3455G>A (p.Arg1152His) is the only ClinVar record for this amino-acid change in the case data (VCV000544272) and is classified Uncertain significance by 6 clinical laboratories (criteria provided, single submitter; no expert-panel and no pathogenic/likely-pathogenic submissions; 2-star, no conflicts). No alternate nucleotide change producing p.Arg1152His (e.g., c.3455G>C) was identified anywhere in the case bundle, and the comparator itself (p.Arg1152His) is not established pathogenic. Therefore the PS1 prerequisite - a previously established pathogenic variant producing the same amino acid - is not met. Generic ACMG/AMP 2015 applies because no MLH3 CSPEC/VCEP exists (cspec found=false; vcep_materials empty; final_classification_framework=generic_acmg_fallback).

Rule

PS1: Same amino acid change as a previously established pathogenic variant regardless of nucleotide change. Strength: Strong. Source: Richards et al. 2015 (PMID:25741868), generic ACMG/AMP (no MLH3 CSPEC/VCEP available).

Evidence considered

• clinvar (VCV000544272): p.Arg1152His classified Uncertain significance by 6 clinical laboratories; no pathogenic/likely-pathogenic submissions; no alternate-nucleotide same-amino-acid comparator present in case data
• PMID:25741868 (Richards et al. 2015): PS1 rule definition - same amino acid change as a previously established pathogenic variant regardless of nucleotide change (strong)

Still missing

• No case-bundle search of ClinVar/literature for alternate nucleotide changes at codon 1152 producing p.Arg1152His was available; absence of such a variant is inferred from the single VCV record matched by transcript HGVS.

Sources

Reasoning

PM1 (ACMG/AMP 2015, PMID:25741868) requires the variant to lie in a mutational hot spot and/or a critical, well-established functional domain that lacks benign variation. Hotspot prong: the CancerHotspots query 'MLH3 R1152' returned no result (evidence bundle, evidence.json hotspots entry: 'This variant does not lie in a statistically significant hotspot.'), so the variant is not in a statistically significant mutational hotspot. Domain prong: no functional-domain annotation (e.g., Pfam/InterPro domain boundaries) for MLH3 or residue 1152 is present in the case bundle, OncoKB provides no variant-specific functional evidence (MLH3 infrequently altered in cancer), and the hotspot/domain-themed full texts retrieved for this case (PMID:33451724, PMID:33516529) contain no MLH3 or residue-1152 content. With neither prong supported by available evidence, PM1 is not met. Generic ACMG/AMP 2015 applies (no MLH3 CSPEC/VCEP).

Rule

PM1: Located in a mutational hot spot and/or critical and well-established functional domain without benign variation. Strength: Moderate. Source: Richards et al. 2015 (PMID:25741868), generic ACMG/AMP (no MLH3 CSPEC/VCEP available).

Evidence considered

• CancerHotspots query 'MLH3 R1152' returned no statistically significant hotspot (case evidence bundle, evidence.json hotspots entry; not a source_registry key, cited from bundle content only)
• oncokb (MLH3 R1152H): no variant-specific reviewed functional evidence; classification 'Unknown Oncogenic Effect'; gene-level context only
• PMID:25741868 (Richards et al. 2015): PM1 rule definition - mutational hot spot and/or critical and well-established functional domain without benign variation (moderate)

Still missing

• No domain annotation source for MLH3/Arg1152 (e.g., Pfam/InterPro MutL_C domain boundaries) was available in the case bundle, so the 'critical and well-established functional domain' prong could not be affirmatively evaluated; a reviewer with domain annotation could revisit PM1.
• No MLH3 CSPEC/VCEP mutational-hotspot or domain specification exists (vcep_materials empty).

Sources

Reasoning

No calibrated computational line supports a deleterious effect. SpliceAI max delta score = 0.06 (DS_AG 0.00, DS_AL 0.06, DS_DG 0.00, DS_DL 0.06), far below the 0.2 splice-altering threshold, so the splice path provides no PP3 support for this exonic missense variant. REVEL score = 0.361, below the ClinGen SVI-calibrated PP3 threshold of 0.764 (Pejaver et al. 2022, PMID 36413997). BayesDel score = -0.112721 is excluded because no verified published calibration threshold is available to this pipeline (no named publication/table), so it cannot be used for PP3. aGVGD was not retrieved for this variant. With every applicable predictor below/outside its calibrated pathogenic range, PP3 is not met.

Rule

PP3 (Richards et al. 2015, PMID 25741868): Multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.). Requires at least one calibrated computational line in the pathogenic range; none present here.

Evidence considered

• SpliceAI (threshold from Jaganathan et al. 2019, PMID 30661751, delta >= 0.2 = predicted splice-altering): max delta score 0.06 (DS_AG 0.00, DS_AL 0.06, DS_DG 0.00, DS_DL 0.06) - below threshold, no predicted splice impact; no PP3 support from the splice path.
• REVEL (thresholds from Pejaver et al. 2022 ClinGen SVI calibration, PMID 36413997: PP3 >= 0.764, BP4 <= 0.016): score 0.361 - below the PP3 threshold, no PP3 support.
• BayesDel: score -0.112721 retrieved but treated as insufficiently calibrated for PP3/BP4 because no verified published threshold (named publication and table) is available to this pipeline.
• aGVGD: no score available for this variant (not retrieved); not assessed.
• Governing framework: generic ACMG/AMP 2015 (Richards et al., PMID 25741868) - no ClinGen CSPEC or VCEP specification exists for MLH3; PP3 definition applied from the generic framework.

Still missing

• aGVGD score was not retrieved for this variant; only its absence is noted, no score to evaluate.
• BayesDel could not be used for PP3 because no verified published threshold is available to this pipeline.

Sources

Reasoning

PP5 is applied in this pipeline only for an exact-variant ClinVar expert-panel (VCEP) classification of Pathogenic/Likely pathogenic. ClinVar VCV000544272 (MLH3 c.3455G>A) has zero expert panel submissions (expert_panels = {}; expert_panel_submissions = 0; review stars = 2, 'criteria provided, multiple submitters, no conflicts'). All 6 submissions are from ordinary clinical laboratories and all classify the variant as Uncertain significance. No exact-variant expert-panel pathogenic classification exists, so PP5 is not met. The flagged PP5 candidate papers (PMID:20301390, PMID:33516529, PMID:34043773) were reviewed and none mentions this variant; aggregate ClinVar VUS labels were not used as a substitute for an expert-panel assertion.

Rule

Richards et al. 2015 (PMID:25741868) PP5: 'Reputable source recently reports variant as pathogenic, but the evidence is not available to the laboratory to perform an independent evaluation'. Pipeline applies PP5 only for an exact-variant ClinVar expert-panel (VCEP) Pathogenic/Likely pathogenic classification.

Evidence considered

• clinvar — VCV000544272 expert_panels = {}; expert_panel_submissions = 0; review_stars = 2; all 6 submissions VUS from ordinary clinical laboratories (TU Dresden, PreventionGenetics, Fulgent, Rigshospitalet Copenhagen, Ambry, Labcorp).
• PMID:25741868 — PP5 rule definition; group instruction restricts PP5 to exact-variant ClinVar expert-panel classifications.

Sources

Reasoning

BA1 (stand-alone benign) requires an allele frequency >5% in a large general population database per ACMG/AMP 2015 (Richards et al., PMID:25741868: 'if over 5% then it is considered as stand-alone support (BA1)'); the local non-VCEP gnomAD calibration for this pipeline is even more permissive at >1%. The highest allele frequency observed for MLH3 c.3455G>A (p.Arg1152His) in any population database in this case is gnomAD v4.1 European (non-Finnish) AF = 0.0001629 (0.0163%, 192/1,178,648 alleles); total AFs are 0.0128% (206/1,612,008 alleles) in gnomAD v4.1 and 0.00636% (16/251,476 alleles) in gnomAD v2.1, with 0 homozygotes in both. The highest observed frequency is ~60-fold below the 1% local threshold and ~300-fold below the 5% ACMG stand-alone threshold. Threshold-artifact check: the gnomAD v4.1 grpmax filtering allele frequency (0.0001437; 0.0144%) is also far below any BA1 threshold, and the variant is spread across many populations (NFE 192, AFR 5, AMR 3, remaining 3, FIN 1, EAS 1, SAS 1 alleles) with no homozygous calls, arguing against a localized sequencing/curation artifact inflating frequency. BA1 is not met under either the generic ACMG/AMP 2015 standard or the local calibration.

Rule

BA1 (stand-alone benign): Allele frequency is >5% in Exome Sequencing Project, 1000 Genomes Project, or Exome Aggregation Consortium (Richards et al. 2015, PMID:25741868, Table 4); local non-VCEP gnomAD calibration >1%. Not met: highest observed AF 0.0163% (gnomAD v4.1 NFE).

Evidence considered

• gnomad_v4: gnomAD v4.1 total AF = 0.0001278 (0.01278%, 206/1,612,008 alleles, 0 homozygotes); highest subpopulation NFE AF = 0.0001629 (0.01629%, 192/1,178,648 alleles, 0 homozygotes); grpmax FAF = 0.00014369 (0.01437%). All values far below the BA1 threshold of 5% (PMID:25741868) and below the local non-VCEP calibration of 1%.
• gnomad_v2: gnomAD v2.1 total AF = 6.36e-05 (0.00636%, 16/251,476 alleles, 0 homozygotes); highest subpopulation AFR AF = 0.000123 (0.0123%, 2/16,256 alleles). Far below any BA1 threshold.
• PMID:25741868: ACMG/AMP 2015 rule - allele frequency >5% in a general population database is stand-alone benign evidence (BA1); frequency observed here (max 0.0163%) is ~300-fold below 5%.

Still missing

• None material: the observed allele frequency is 2-3 orders of magnitude below every BA1 threshold (5% ACMG/AMP 2015, PMID:25741868; 1% local non-VCEP calibration), and no population database in this case reports an AF above 0.02%.
• No MLH3-specific (VCEP) BA1 threshold exists because no ClinGen CSPEC/VCEP specification is available for MLH3; the determination therefore relies on generic thresholds, which are far more permissive than any MLH3-specific calibration would plausibly be.

Sources

Reasoning

BS1 (strong benign) requires an allele frequency greater than expected for the disorder (Richards et al., PMID:25741868). For the hereditary colorectal cancer / Lynch syndrome context, the expected carrier prevalence of pathogenic MMR-gene variants is approximately 1 in 300 (PMID:34043773), corresponding to an allele frequency of ~0.17%; MLH3 is not among the classical Lynch genes addressed in those guidelines (PMID:34043773 contains no MLH3 content), and no MLH3-specific expected-frequency calibration exists. The observed allele frequencies for c.3455G>A are 0.0128% (gnomAD v4.1 total; 206/1,612,008 alleles), 0.00636% (gnomAD v2.1 total; 16/251,476 alleles), and at most 0.0163% in any single subpopulation (NFE, gnomAD v4.1) - approximately 13-fold BELOW the classical Lynch expected allele frequency of ~0.17% and far below the local non-VCEP BS1 calibration of >0.3%. Because the observed frequency is lower than (not greater than) the expected disease allele frequency, BS1 is not met; the low frequency is instead consistent with a rare, potentially disease-relevant variant.

Rule

BS1 (strong benign): Allele frequency is greater than expected for the disorder (Richards et al. 2015, PMID:25741868, Table 4); local non-VCEP gnomAD calibration >0.3%. Not met: observed total AF 0.0128% (gnomAD v4.1) / 0.00636% (gnomAD v2.1) is below the expected disease allele frequency (~0.17%, from path_MMR carrier prevalence ~1/300, PMID:34043773).

Evidence considered

• gnomad_v4: gnomAD v4.1 total AF = 0.0001278 (0.01278%, 206/1,612,008 alleles, 0 homozygotes); max subpopulation AF = 0.0001629 (0.01629%, NFE). Observed AF is below the expected disease allele frequency (~0.17% from path_MMR carrier prevalence 1/300, PMID:34043773), so BS1 (AF greater than expected) is not met.
• gnomad_v2: gnomAD v2.1 total AF = 6.36e-05 (0.00636%, 16/251,476 alleles); consistent with the v4.1 observation and below any BS1 threshold.
• PMID:25741868: ACMG/AMP 2015 rule - BS1 applies when allele frequency is greater than expected for the disorder (Table 4); not satisfied here because the observed frequency is lower than the expected disease allele frequency.
• PMID:34043773: EHTG/ESCP Mallorca Lynch guidelines state 'The prevalence of path_MMR carriers has been estimated to be around one in 300', i.e., ~0.33% carriers (~0.17% allele frequency), used as the expected-frequency anchor for the BS1 comparison; the document contains no MLH3 content, so no MLH3-specific anchor is available from this source.

Still missing

• No MLH3-specific expected allele-frequency calibration exists (no ClinGen CSPEC/VCEP for MLH3); the 'expected frequency for the disorder' anchor is approximated from the classical Lynch MMR genes (path_MMR carrier prevalence ~1/300, PMID:34043773), which likely overestimates the MLH3-specific expected frequency.
• MLH3's dominant germline disease contribution is not firmly established (it is absent from the Mallorca/EHTG-ESCP Lynch guidelines), so the applicable disease model and expected-frequency anchor carry residual uncertainty; this does not affect the BS1 conclusion because observed AF is below every reasonable anchor.

Sources

Reasoning

BS2 (strong benign) requires observation in a well-documented healthy adult individual for a disorder with full penetrance expected at an early age - homozygous for a recessive disorder, heterozygous for a dominant disorder, or hemizygous for an X-linked disorder (Richards et al., PMID:25741868). For MLH3, 0 homozygous carriers are observed in both gnomAD v2.1 (16/251,476 alleles) and gnomAD v4.1 (206/1,612,008 alleles), so the recessive-model route (healthy adult homozygote) fails. For a dominant/reduced-penetrance cancer model, Lynch syndrome is not fully penetrant at an early age - penetrance varies by gene and gender from very high to immeasurable (PMID:34043773) - so even the observed low-frequency heterozygous carriers in a general-population database (largely unselected for phenotype) would not satisfy BS2's early-onset full-penetrance requirement, and no carrier-level phenotype data are available to establish healthy-adult status. There is no evidence of homozygotes or of overrepresentation of this variant in unaffected individuals exceeding disease prevalence (the observed AF of 0.0128% is below, not above, the ~0.17% expected disease allele frequency). BS2 is not met.

Rule

BS2 (strong benign): Observed in a healthy adult individual for a recessive (homozygous), autosomal dominant (heterozygous), or X-linked (hemizygous) disorder, with full penetrance expected at an early age (Richards et al. 2015, PMID:25741868, Table 4). Not met: 0 homozygotes in gnomAD v2.1 and v4.1; disease not fully penetrant at an early age; no healthy-adult carrier documentation.

Evidence considered

• gnomad_v4: gnomAD v4.1 - 206/1,612,008 alleles, homozygotes = 0; no homozygous carrier observed to support the BS2 recessive route.
• gnomad_v2: gnomAD v2.1 - 16/251,476 alleles, homozygotes = 0; no homozygous carrier observed.
• PMID:25741868: ACMG/AMP 2015 rule - BS2 requires a well-documented healthy adult with full penetrance at an early age (recessive: homozygous; dominant: heterozygous); neither condition is satisfied here (0 homozygotes; no healthy-adult phenotype documentation; disease not fully penetrant early in life).
• PMID:34043773: Lynch syndrome penetrance 'vary(ies) by gene and gender from very high to immeasurable', i.e., not fully penetrant at an early age, which precludes the BS2 healthy-adult criterion for this dominant-cancer disease context.

Still missing

• No individual-level phenotype/health status is available for the 206 gnomAD v4.1 carriers; BS2's 'well-documented healthy adult' requirement cannot be directly verified from the aggregate database records.
• No homozygous carriers exist in either gnomAD version, and none would be expected at this allele frequency; BS2 (homozygous-healthy route for the recessive MLH3 polyposis model) has no supporting observation.
• Lynch syndrome is incompletely penetrant with adult onset (PMID:34043773), so the 'full penetrance at an early age' precondition of BS2 is not met for this disease context.

Sources

Reasoning

BP4 requires multiple lines of computational evidence suggesting no impact on the gene or gene product. Only one line qualifies: SpliceAI max delta = 0.06 (below the 0.2 splice-altering threshold) indicates no predicted splice impact. REVEL = 0.361 is above the SVI-calibrated BP4 threshold of <= 0.016 (Pejaver et al. 2022, PMID 36413997), so it falls in the indeterminate zone and does NOT support BP4. BayesDel is excluded (no verified published threshold). aGVGD was not retrieved. A single splice-prediction line is insufficient to satisfy the 'multiple lines' requirement of BP4, so BP4 is not met.

Rule

BP4 (Richards et al. 2015, PMID 25741868): Multiple lines of computational evidence suggest no impact on the gene or gene product (conservation, evolutionary, splicing impact, etc.). Requires multiple qualifying lines; only one (SpliceAI) is present, so the criterion is not met.

Evidence considered

• SpliceAI (threshold from Jaganathan et al. 2019, PMID 30661751): max delta score 0.06 - below 0.2, no predicted splice impact; counts as one (splice-path) line toward BP4.
• REVEL (thresholds from Pejaver et al. 2022 ClinGen SVI calibration, PMID 36413997): score 0.361 - above the BP4 threshold of <= 0.016, indeterminate zone; does not support BP4.
• BayesDel: score -0.112721 retrieved but treated as insufficiently calibrated for PP3/BP4 because no verified published threshold (named publication and table) is available to this pipeline.
• aGVGD: no score available for this variant (not retrieved); not assessed.
• Governing framework: generic ACMG/AMP 2015 (Richards et al., PMID 25741868) - no ClinGen CSPEC or VCEP specification exists for MLH3; BP4 definition applied from the generic framework.

Still missing

• Only one computational line (SpliceAI, no splice impact) supports a benign/no-impact prediction; a second calibrated benign-range prediction (e.g., REVEL <= 0.016 or another verified predictor) would be required to meet the 'multiple lines' bar for BP4.
• BayesDel could not be evaluated for BP4 because no verified published threshold is available to this pipeline.
• aGVGD score not retrieved for this variant.

Sources

Reasoning

BP6 is applied in this pipeline only for an exact-variant ClinVar expert-panel (VCEP) classification of Benign/Likely benign. ClinVar VCV000544272 has zero expert panel submissions (expert_panels = {}; expert_panel_submissions = 0; review stars = 2). All 6 submissions are from ordinary clinical laboratories and all classify the variant as Uncertain significance; no Benign or Likely benign assertion exists for this variant from any submitter. No exact-variant expert-panel benign classification exists, so BP6 is not met.

Rule

Richards et al. 2015 (PMID:25741868) BP6: 'Reputable source recently reports variant as benign, but the evidence is not available to the laboratory to perform an independent evaluation'. Pipeline applies BP6 only for an exact-variant ClinVar expert-panel (VCEP) Benign/Likely benign classification.

Evidence considered

• clinvar — VCV000544272 expert_panels = {}; expert_panel_submissions = 0; review_stars = 2; all 6 submissions VUS, none Benign/Likely benign.
• PMID:25741868 — BP6 rule definition; group instruction restricts BP6 to exact-variant ClinVar expert-panel classifications.

Sources

Reasoning

PS2 (de novo, parental identity confirmed) cannot be assessed: the case bundle contains no proband phenotype record and no parental testing results. ClinVar VCV000544272 (6 submissions, all germline 'Uncertain significance', reviewed 2021-2026) contains no de novo assertion and no parental/family testing data for NM_001040108.2:c.3455G>A (p.Arg1152His). No full-text publication in this case mentions the variant or any de novo occurrence. Governing framework is generic ACMG/AMP 2015 (PMID 25741868) because no ClinGen CSPEC/VCEP specification exists for MLH3 (vcep_materials.json: cspec found=false, framework_mode=generic_acmg). Per that rule, PS2 requires the variant to have arisen de novo with both parents shown by identity testing to be biological parents and tested negative; none of that information is available here.

Rule

PS2 (strong): De novo variant - variant observed to have arisen de novo (parental samples test negative), strong support for pathogenicity only if both parental samples were shown through identity testing to be the biological parents of the patient (Richards et al. 2015, PMID 25741868).

Evidence considered

• ClinVar VCV000544272 reviewed: 6 submissions, all germline, all 'Uncertain significance'; no submission asserts de novo occurrence or reports parental testing for c.3455G>A
• Rule definition (ACMG/AMP 2015, PMID 25741868): PS2 applies when the variant is observed de novo AND both parental samples are shown by identity testing to be the biological parents and test negative
• Full-text publications in this case (PMID 25741868, PMID 34043773, PMID 33516529, PMID 33451724) screened: none mention c.3455G>A, p.Arg1152His, or any de novo occurrence

Still missing

• No proband phenotype or clinical indication data in the case bundle
• No parental testing results (neither parent tested, or results not captured in pipeline sources)
• No identity/paternity-maternity confirmation data
• No de novo assertion for this variant in ClinVar submissions or literature

Sources

Reasoning

PS3 requires well-established in vitro or in vivo functional studies demonstrating a damaging effect on the gene or gene product (Richards et al. 2015, PMID:25741868). No such functional study evidence exists for MLH3 c.3455G>A (p.Arg1152His) in any source in this case. The governing framework is generic ACMG/AMP 2015 (no MLH3 VCEP/CSPEC exists; vcep_materials.json reports no cspec URL and final_classification_framework.json reports generic_acmg_fallback). Sources consulted: OncoKB explicitly states it identified no variant-specific reviewed functional evidence for R1152H (Unknown Oncogenic Effect); COSMIC has no record of this variant in somatic cancers (absence of a record is not functional evidence); SpliceAI max delta score 0.06 and REVEL 0.361 are in-silico predictions, which per PMID:25741868 are only predictions and not recommended as a sole source of evidence and do not constitute functional studies; ClinVar submissions (6 laboratories, all Uncertain significance) contain no functional assay assertion. With no functional data demonstrating a deleterious effect, PS3 cannot be applied; status is not_assessed rather than not_met because the absence of evidence is an evidence gap, not evidence of normal function.

Rule

PS3: Well-established in vitro or in vivo functional studies supportive of a damaging effect on the gene or gene product (Richards et al. 2015, PMID:25741868, Table 3). Not applicable here because no qualifying functional study evidence was identified.

Evidence considered

• PMID:25741868: Governing ACMG/AMP 2015 definitions of PS3/BS3 and the requirement that functional studies be well-established and reflect the biological environment; also states in-silico predictions are only predictions and should not be the sole basis of a clinical assertion.
• oncokb: OncoKB MLH3 R1152H entry reports no variant-specific reviewed functional evidence ('Unknown Oncogenic Effect'); gene-level note that MLH3, a DNA mismatch repair protein, is infrequently altered in cancer.
• spliceai: SpliceAI Lookup for NM_001040108.2:c.3455G>A reports max delta score 0.06 (DS_AL 0.06, DS_DL 0.06, DS_AG 0, DS_DG 0); in-silico splice prediction only, not a functional study and therefore not usable for PS3/BS3.
• evidence.json COSMIC result: variant has not previously been reported in somatic cancers (COSMIC); absence from a somatic database does not constitute functional assay evidence.
• clinvar: ClinVar variation 544272 classified Uncertain significance by 6 laboratories; no submission asserts functional assay evidence.

Still missing

• No published or database functional assay data for MLH3 p.Arg1152His (e.g., in vitro MMR activity assay, MMR-deficient cell-line complementation/rescue assay, or validated variant-specific functional readout) was retrieved in the literature pass or any case source.
• A validated functional assay for this exact variant, with appropriate positive/negative controls and demonstrated biological relevance to MLH3 MMR function, would be required before PS3 could be applied; without it the criterion remains not assessed.

Sources

Reasoning

PS4 (Richards et al. 2015, PMID:25741868) requires prevalence of the variant in affected individuals significantly increased versus controls, with Note 2 permitting a lower strength when multiple affected observations exist without control data. The evidence bundle contains no case-control study, case series, or affected-proband allele counts for MLH3 c.3455G>A (p.Arg1152His). The only case-observation candidates flagged by literature triage (PMID:20301390, PMID:33516529, PMID:34043773) are gene-level reviews/guidelines; full-text review of the retrievable ones confirms they do not mention this variant. ClinVar VCV000544272 submissions are all Uncertain significance laboratory labels with no per-proband phenotype or allele-count data and cannot serve as case observations. With no case/control enrichment evidence of any kind, PS4 cannot be assessed.

Rule

Richards et al. 2015 (PMID:25741868) PS4: 'The prevalence of the variant in affected individuals is significantly increased compared with the prevalence in controls'; Note 2 permits downgrade to moderate/supporting when multiple affected observations exist without control data.

Evidence considered

• PMID:25741868 — PS4 rule definition and Note 2 (case-control prevalence; downgrade option for multiple affected observations without controls).
• clinvar — VCV000544272: 6 laboratory submissions, all Uncertain significance, 0 expert panel submissions; no case-control or affected-proband data reported.
• gnomad_v2 — variant present in the general population (total AF 6.36e-05, 16/251476 alleles; max subpopulation AFR 1.23e-04); no case/control enrichment comparison available.

Still missing

• No case-control study or case series for MLH3 c.3455G>A identified in the evidence bundle; no affected-proband allele counts or enrichment statistics exist.
• ClinVar VCV000544272 submissions provide no per-proband phenotype or allele-count data that could support PS4 at reduced strength (Note 2).
• PMID:20301390 (GeneReviews 'Lynch Syndrome') was flagged as a PP5/PS4 candidate but no abstract or full text was retrievable in this case bundle, so it could not be screened for variant mentions.

Sources

Reasoning

Governing framework is generic ACMG/AMP 2015 (PMID 25741868): no MLH3 ClinGen CSPEC/VCEP exists (cspec.found=false, gene_vcep_dir=null, precedence_rank=3, framework_mode=generic_acmg). PM3 (moderate) requires detection of the variant in trans with a pathogenic variant in a recessive disorder, upgradeable to strong with multiple independent trans observations. The recessive-disease prerequisite is plausibly satisfied for MLH3: the pvs1_gene_context artifact identifies a biallelic germline MLH3 LOF mechanism ('Biallelic germline nonsense variant of MLH3 underlies polyposis predisposition', PMID 30573798), so PM3 is conceptually applicable. However, the decisive evidence - a documented observation of c.3455G>A in trans with a pathogenic/likely pathogenic MLH3 variant in an affected proband, or parental/phase testing, or a compound-heterozygous case - is entirely absent from the evidence bundle. ClinVar VCV000544272 contains 6 submissions, all 'Uncertain significance', no expert-panel submissions (0), and the 3 criterion-lead submissions cite only the ACMG methodology paper (PMID 25741868); no variant-specific case report or segregation/phase data was retrieved. Per the framework's cis/trans guidance, and because the variant is present in the general population (gnomAD v4.1 AF 1.28e-04, 206/1,612,008 alleles), any future trans co-occurrence claim would additionally require a statistical approach to control for random co-occurrence. Evidence is insufficient to apply PM3; status is not_assessed pending phase data.

Rule

PM3: For recessive disorders, detected in trans with a pathogenic variant (moderate evidence of pathogenicity); may be upgraded to strong with multiple observations of the variant in trans with other pathogenic variants (ACMG/AMP 2015, PMID 25741868).

Evidence considered

• PMID:25741868 - governing generic ACMG/AMP 2015 framework; PM3 rule text 'For recessive disorders, detected in trans with a pathogenic variant' and cis/trans testing guidance (moderate, upgradeable to strong with multiple observations).
• pvs1_gene_context - gene-context artifact supporting a germline biallelic (recessive) MLH3 LOF mechanism via PMID 30573798 ('Biallelic germline nonsense variant of MLH3 underlies polyposis predisposition'), establishing that PM3 is conceptually applicable to MLH3.
• clinvar - VCV000544272: 6 submissions, all Uncertain significance, 0 expert-panel submissions, 3 criterion-lead submissions citing only PMID 25741868; no trans-phase observation or case-level detail documented.
• gnomad_v4 - variant present in general population (AF 0.000127791, 206/1612008 alleles, 0 homozygotes), triggering the framework caution that trans co-occurrence claims require statistical control.

Still missing

• No affected-proband or family observation documenting c.3455G>A in trans with a pathogenic/likely pathogenic MLH3 variant in any source in the evidence bundle.
• No parental phase testing, segregation, or compound-heterozygosity data available for this variant.
• ClinVar submissions (VCV000544272) provide no criterion details or phase information and cite only the ACMG methodology paper (PMID 25741868); no variant-specific case reports were retrieved.
• If a trans observation is ever obtained, the variant's presence in gnomAD (AF 1.28e-04) requires a statistical co-occurrence control before PM3 can be applied.

Sources

Reasoning

PM5 (ACMG/AMP 2015, PMID:25741868) requires a novel missense change at an amino-acid residue where a different missense change has been determined to be pathogenic. The automated same-residue comparator harvest was not performed: output/pm5_candidates.json reports pm5_mode='unknown', eligible_for_classic_pm5_search=false, total_candidates_seen=0, with the explicit note 'Unable to confirm classic same-residue PM5 semantics safely; skipping automatic candidate harvesting' and an instruction not to apply PM5 from that artifact alone. The ClinVar data in the case bundle (VCV000544272) documents no other missense change at Arg1152 with a pathogenic/likely-pathogenic classification. Because no same-residue comparator variant was collected or identified, PM5 cannot be evaluated on available evidence and is returned as not_assessed. A manual ClinVar/literature search for other missense changes at Arg1152 is required to resolve this criterion.

Rule

PM5: Novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before. Strength: Moderate. Source: Richards et al. 2015 (PMID:25741868), generic ACMG/AMP (no MLH3 CSPEC/VCEP available).

Evidence considered

• pm5_candidates: comparator harvest skipped (pm5_mode='unknown'; total_candidates_seen=0); artifact instructs that PM5 must not be applied from it alone
• clinvar (VCV000544272): only record for codon 1152 in case data; no other missense at Arg1152 with P/LP classification documented
• PMID:25741868 (Richards et al. 2015): PM5 rule definition - novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before (moderate)

Still missing

• No automatic same-residue comparator harvest was performed (pm5_candidates.json: pm5_mode='unknown'; 0 candidates); manual ClinVar/literature search for other missense changes at Arg1152 (e.g., p.Arg1152Cys, p.Arg1152Ser, p.Arg1152Leu) with pathogenic/likely-pathogenic classification is required.
• No expert-panel or gene-specific PM5 rule exists for MLH3 (no CSPEC/VCEP), so the governing threshold for an acceptable same-residue comparator is not defined for this gene.

Sources

Reasoning

PM6 (assumed de novo without confirmation of parental identity) cannot be assessed: no parental testing results, no de novo assertion, and no proband family-history data exist anywhere in the case bundle. ClinVar submissions for VCV000544272 carry no family or parental testing information. No publication in this case reports the variant or a de novo event. PM6 would require at least that the variant is assumed de novo with parental samples testing negative but without identity confirmation (ACMG/AMP 2015, PMID 25741868); the pipeline contains no evidence either way.

Rule

PM6 (moderate): Assumed de novo, but without confirmation of paternity and maternity (Richards et al. 2015, PMID 25741868); 'PM6 applies if identity is assumed but not confirmed.'

Evidence considered

• ClinVar VCV000544272 reviewed: no submission contains a de novo or parental-testing assertion for c.3455G>A
• Rule definition (ACMG/AMP 2015, PMID 25741868): PM6 applies if identity is assumed but not confirmed, with parental samples negative
• Full-text publications screened: no de novo occurrence of the variant reported

Still missing

• No parental testing results of any kind (confirmed or assumed de novo status unknown)
• No statement of family history consistent or inconsistent with de novo inheritance
• No de novo assertion in ClinVar submissions or literature for c.3455G>A

Sources

Reasoning

PP1 (cosegregation with disease in affected family members) cannot be assessed: the case bundle contains no pedigree, no affected relatives, no segregation study, and no family testing data for NM_001040108.2:c.3455G>A. The literature_pass.json triage surfaced no segregation-focused paper for this variant, and none of the four full-text publications available mentions the variant or any family segregation. Per generic ACMG/AMP 2015 (PMID 25741868; no MLH3 CSPEC/VCEP exists), PP1 requires demonstration of cosegregation in affected family members, which is entirely absent.

Rule

PP1 (supporting): Cosegregation with disease in affected family members in a gene definitively known to cause the disease (Richards et al. 2015, PMID 25741868).

Evidence considered

• ClinVar VCV000544272 reviewed: no submission reports segregation or family studies for c.3455G>A
• Rule definition (ACMG/AMP 2015, PMID 25741868): PP1 = cosegregation with disease in affected family members
• Literature pass and full-text screening: no segregation evidence for this variant found in the case sources

Still missing

• No pedigree or family structure data in the case bundle
• No segregation testing results (no affected relatives tested for the variant)
• No literature report of cosegregation of c.3455G>A with disease
• No proband phenotype/disease status documented against which segregation could be scored

Sources

Reasoning

PP2 (ACMG/AMP 2015, PMID:25741868) requires (a) a gene with a low rate of benign missense variation and (b) missense variants being a common mechanism of disease. Prong (a): no missense-constraint data for MLH3 is present in the case bundle (hci_prior found=false; custom_pp2_score=null; no gnomAD missense Z-score). Prong (b): the gene-level evidence available (pvs1_gene_context) supports loss of function as the germline disease mechanism for MLH3 (five supporting literature sources, including biallelic germline nonsense MLH3 variants underlying polyposis predisposition, PMID 30573798), with no evidence that missense variants are a common mechanism of MLH3 disease. Because the missense-constraint prong is unevaluable and the mechanism evidence available favors LoF rather than missense, PP2 is returned as not_assessed pending gene-level constraint data and a defined MLH3-specific mechanism rule.

Rule

PP2: Missense variant in a gene that has a low rate of benign missense variation and in which missense variants are a common mechanism of disease. Strength: Supporting. Source: Richards et al. 2015 (PMID:25741868), generic ACMG/AMP (no MLH3 CSPEC/VCEP available).

Evidence considered

• pvs1_gene_context: germline loss-of-function mechanism supported for MLH3 (supporting_sources PMIDs 29568967, 30573798, 31822864, 36833239, 37427096); no support for missense as a common disease mechanism
• Case bundle contains no missense-constraint metric for MLH3 (hci_prior found=false; custom_pp2_score=null)
• PMID:25741868 (Richards et al. 2015): PP2 rule definition - missense variant in a gene that has a low rate of benign missense variation and in which missense variants are a common mechanism of disease (supporting)

Still missing

• No gnomAD missense Z-score or equivalent missense-constraint/benign-ratio metric for MLH3 was available in the case bundle to evaluate the 'low rate of benign missense variation' prong.
• MLH3 disease mechanism (LoF vs missense) is incompletely established; available germline literature (pvs1_gene_context) supports LoF, and no gene-specific/CSPEC rule defines PP2 applicability for MLH3.

Sources

Reasoning

PP4 (Richards et al. 2015, PMID:25741868) requires the variant to be present in a case with a highly specific phenotype or family history. No proband phenotype or family-history information is available in the case bundle: ClinVar submissions list asserted conditions (colorectal cancer, endometrial carcinoma, colorectal cancer HNPCC type 7, MLH3-related disorder, not provided/not specified) but these are laboratory submission labels, not a confirmed proband phenotype, and may not be used to infer criterion support. In addition, the MLH3 germline phenotype (Lynch-like/MMR spectrum) is not an established highly specific phenotype, and no phenotype-specificity evidence for p.Arg1152His exists. PP4 cannot be assessed without proband phenotype data.

Rule

Richards et al. 2015 (PMID:25741868) PP4: 'Variant is present in a case with a highly specific phenotype or family history'.

Evidence considered

• PMID:25741868 — PP4 rule definition (variant present in a case with a highly specific phenotype or family history).
• clinvar — VCV000544272 submissions list conditions but contain no confirmed proband phenotype or family-history data usable for PP4.

Still missing

• No proband phenotype or family-history information is available in the case bundle.
• MLH3 germline phenotype (Lynch-like/MMR spectrum) is not established as highly specific; no phenotype-specificity evidence exists for p.Arg1152His.

Sources

Reasoning

BS3 requires well-established in vitro or in vivo functional studies showing no damaging effect on protein function or splicing (Richards et al. 2015, PMID:25741868). No functional study for MLH3 c.3455G>A (p.Arg1152His) was identified in any case source. SpliceAI (max delta 0.06) predicts no significant splice impact and REVEL (0.361) is a low-to-moderate in-silico score, but both are computational predictions, not well-established functional studies, and PMID:25741868 explicitly advises that in-silico predictions should not be used as the sole source of evidence for a clinical assertion. gnomAD v2.1 presence (AF 6.36e-05, 0 homozygotes) is population-frequency evidence relevant to BA1/BS1/BS2 and does not satisfy BS3. Absence of demonstrated harm is not equivalent to demonstrated normal function, so BS3 is not_assessed pending validated functional assay data.

Rule

BS3: Well-established in vitro or in vivo functional studies show no damaging effect on protein function or splicing (Richards et al. 2015, PMID:25741868, Table 4). Not applicable here because no qualifying functional study evidence was identified.

Evidence considered

• PMID:25741868: Governing ACMG/AMP 2015 definition of BS3 (well-established functional studies show no damaging effect) and guidance that in-silico predictions are only predictions and not recommended as the sole source of evidence.
• spliceai: SpliceAI Lookup for NM_001040108.2:c.3455G>A reports max delta score 0.06; consulted to check for a splicing functional consequence but is in-silico only and does not meet the BS3 experimental standard.
• oncokb: OncoKB MLH3 R1152H entry reports no variant-specific reviewed functional evidence; absence of curated functional data does not support BS3.
• evidence.json gnomAD v2.1 result: AF 6.36e-05 (16/251476 alleles), 0 homozygotes; population data noted as not constituting functional evidence for BS3.

Still missing

• No functional assay evidence showing normal (or preserved) MLH3 function for p.Arg1152His (e.g., in vitro MMR activity or complementation assay with validated controls) was retrieved in the literature pass or any case source.
• In-silico predictions (SpliceAI max delta 0.06; REVEL 0.361) are insufficient to meet the BS3 'well-established functional studies' standard and cannot substitute for experimental data.
• Population frequency data (gnomAD v2.1 AF 6.36e-05, grpmax FAF 6.01e-05, 0 homozygotes) do not support a benign functional conclusion and belong to the BA1/BS1/BS2 evidence domain, not BS3.

Sources

Reasoning

BS4 (lack of segregation in affected family members) cannot be assessed: absence of segregation evidence is not the same as demonstrated non-segregation. No family/pedigree or segregation data exist in the case bundle, so it cannot be established that affected relatives lack the variant (or that unaffected carriers exist). Note the variant is present in gnomAD at low frequency (v2.1 AF 0.00636%, grpmax FAF 0.0060%; v4.1 AF 0.0128%, grpmax FAF 0.0144%), but BS4 under ACMG/AMP 2015 (PMID 25741868) is defined by family segregation data, not population frequency (population-frequency arguments belong to BA1/BS1, assessed elsewhere).

Rule

BS4 (strong): Lack of segregation in affected members of a family (Richards et al. 2015, PMID 25741868).

Evidence considered

• ClinVar VCV000544272 reviewed: no submission reports segregation or non-segregation data for c.3455G>A
• Rule definition (ACMG/AMP 2015, PMID 25741868): BS4 = lack of segregation in affected family members
• gnomAD population frequencies noted only to record why they were NOT used here (BS4 requires family data, not allele frequency)

Still missing

• No affected family members tested, so non-segregation cannot be demonstrated
• No pedigree data to identify informative meioses
• No literature report of non-segregation of c.3455G>A with disease

Sources

Reasoning

BP1 (ACMG/AMP 2015, PMID:25741868) requires a gene for which primarily truncating variants are known to cause disease. The case bundle documents that loss-of-function/truncating MLH3 variants can be disease-causing (pvs1_gene_context: biallelic germline nonsense MLH3 variant underlies polyposis predisposition, PMID 30573798), which is directionally consistent with BP1. However, no quantitative or gene-rule evidence establishes that MLH3 disease is primarily caused by truncating variants: MLH3 is a MutL-homolog DNA mismatch-repair gene (gene_context summary) whose disease association is incompletely defined, no pathogenic missense variant burden is characterized in the bundle, and no CSPEC/VCEP rule exists for MLH3. Because the 'primarily truncating variants are known to cause disease' standard cannot be confirmed from available evidence, BP1 is returned as not_assessed rather than applied.

Rule

BP1: Missense variant in a gene for which primarily truncating variants are known to cause disease. Strength: Supporting (benign). Source: Richards et al. 2015 (PMID:25741868), generic ACMG/AMP (no MLH3 CSPEC/VCEP available).

Evidence considered

• pvs1_gene_context: LoF/truncating MLH3 variants documented as disease-causing (e.g., PMID 30573798 biallelic nonsense variant, polyposis predisposition), but primacy of truncating variants is not established
• PMID:25741868 (Richards et al. 2015): BP1 rule definition - missense variant in a gene for which primarily truncating variants are known to cause disease (supporting, benign direction)

Still missing

• No quantitative or gene-rule evidence (e.g., proportion of established pathogenic MLH3 variants that are truncating vs missense) to establish that primarily truncating variants cause MLH3 disease.
• No MLH3 CSPEC/VCEP or local gene framework defining BP1 applicability for MLH3.

Sources

Reasoning

Governing framework is generic ACMG/AMP 2015 (PMID 25741868). BP2 is supporting evidence for a benign role when (a) the variant is observed in trans with a pathogenic variant for a fully dominant disease/mechanism, or (b) the variant is observed in cis with a pathogenic variant in any inheritance pattern. Neither arm can be evaluated: (1) A fully dominant MLH3 disease/mechanism is not established - the documented germline mechanism for MLH3 is biallelic/recessive (pvs1_gene_context artifact citing PMID 30573798), and the monoallelic HNPCC7 association is not a validated fully dominant mechanism; (2) no cis-phase or trans-phase observation of c.3455G>A with any pathogenic variant exists anywhere in the evidence bundle - ClinVar VCV000544272 (6 submissions, all Uncertain significance, no assertion-criteria detail, no expert panel) contains no parental testing, family, or phase data, and no variant-specific publication was retrieved. Evidence is insufficient to apply either arm of BP2; status is not_assessed pending phase data and confirmation of the disease mechanism.

Rule

BP2: Observed in trans with a pathogenic variant for a fully dominant disease/mechanism, or observed in cis with a pathogenic variant in any inheritance pattern (supporting evidence for a benign role; ACMG/AMP 2015, PMID 25741868).

Evidence considered

• PMID:25741868 - governing generic ACMG/AMP 2015 framework; BP2 rule text: 'In the context of dominant disorders, the detection of a variant in trans with a pathogenic variant can be considered supporting evidence for a benign impact (BP2)' and 'finding the second variant in cis would be supporting evidence, though not definitive, for a benign role (BP2)'.
• clinvar - VCV000544272: 6 submissions, all Uncertain significance, 0 expert-panel submissions; no cis/trans phase information or family-level data documented for this variant.

Still missing

• No cis-phase or trans-phase observation of c.3455G>A with a pathogenic variant is documented in any evidence source.
• Fully dominant MLH3 disease/mechanism is not established; the established germline mechanism is biallelic (recessive) per the pvs1_gene_context artifact (PMID 30573798), so the dominant arm of BP2 lacks a validated premise.
• No parental testing, family segregation, or second-variant configuration data available for this variant.

Sources

Reasoning

BP5 (Richards et al. 2015, PMID:25741868) requires the variant to be found in a case with an alternate molecular basis for disease (e.g., a pathogenic variant in another gene that explains the phenotype). The case bundle contains no proband-level molecular data: no additional germline findings, no alternate pathogenic variant, and no phenotype to attribute. Without any proband genotype/phenotype information, the presence or absence of an alternate molecular cause cannot be assessed.

Rule

Richards et al. 2015 (PMID:25741868) BP5: 'Variant found in a case with an alternate molecular basis for disease'.

Evidence considered

• PMID:25741868 — BP5 rule definition (variant found in a case with an alternate molecular basis for disease).
• clinvar — VCV000544272 record contains no proband-level molecular findings or alternative-cause data.

Still missing

• No proband-level molecular data available to determine whether an alternate molecular basis for disease exists.

Sources