MLH3 c.3828-24dup, p.?

Reasoning

Allele frequency far exceeds the frequency expected for a pathogenic MLH3 allele. gnomAD v4.1 total AF = 1.297% (16359/1261538 alleles, 8 homozygotes) exceeds the lab's non-VCEP BA1 operating threshold of AF > 1%. gnomAD v2.1 total AF = 0.998% is marginally below 1%, but the max-subpopulation AF (African/African American 3.15% v2.1 / 2.60% v4.1) and grpmax FAF (3.70% v2.1 exome, 2.50% v4.1 joint) clearly exceed the threshold, and the larger, more current v4.1 dataset clears it on total AF. The presence of 8 homozygotes in v4.1 supports a genuine common polymorphism rather than a sequencing artifact.

Rule

BA1 (Richards et al. 2015): allele frequency in population databases is too high to be consistent with a pathogenic variant for the disorder. Local non-VCEP operating threshold: AF > 1% in gnomAD.

Evidence considered

• gnomAD v4.1 (source key gnomad_v4): total allele frequency for 14-75030725-T-TA = 0.0129675 (1.297%, 16359/1261538 alleles) with 8 homozygotes; exceeds the lab's non-VCEP BA1 operating threshold of AF > 1%.
• gnomAD v2.1 (source key gnomad_v2): total AF = 0.0099843 (0.998%, 1246/124796 alleles) with 2 homozygotes; African/African American subpopulation AF = 0.0315 (3.15%) and exome grpmax FAF = 0.03699, both well above 1%.
• BA1 rule per Richards et al. 2015 (PMID:25741868): allele frequency in population databases too high to be consistent with pathogenicity. No ClinGen CSPEC/VCEP exists for MLH3 (cspec.found=false, framework_mode=generic_acmg), so the generic rule applies with the lab's local non-VCEP threshold (BA1 > 1% in gnomAD).
• Threshold-artifact check: gnomAD-Canada flags the locus as low-complexity (lcr=true), and indel calls in homopolymer/polypyrimidine tracts can be inflated; however, concordance across three independent datasets and the presence of multiple homozygotes make the >1% conclusion robust even if some fraction were artifact.

Still missing

• gnomAD is a general-population cohort not phenotyped for MLH3-related Lynch syndrome/polyposis; disease prevalence/penetrance of MLH3 germline disease is not precisely quantified, but no plausible disease prevalence approaches a 1.3% allele frequency.

Sources

Reasoning

Allele frequency is greater than expected for MLH3-associated disease. gnomAD v4.1 total AF 1.297% and gnomAD v2.1 total AF 0.998% both exceed the lab's non-VCEP BS1 operating threshold of AF > 0.3%. Every dataset and subpopulation clears the threshold, including the most conservative measurement (gnomAD-Canada overall AF 0.168% is below 0.3% overall but its AFR subpopulation AF is 1.86%; the v2.1/v4.1 totals are the governing values).

Rule

BS1 (Richards et al. 2015): allele frequency is greater than expected for the disorder. Local non-VCEP operating threshold: AF > 0.3% in gnomAD.

Evidence considered

• gnomAD v4.1 (source key gnomad_v4): total AF = 0.0129675 (1.297%, 16359/1261538 alleles) and gnomAD v2.1 (source key gnomad_v2): total AF = 0.0099843 (0.998%, 1246/124796 alleles); both exceed the lab's non-VCEP BS1 operating threshold of AF > 0.3%.
• Highest population frequencies: African/African American AF = 3.15% (v2.1) and 2.60% (v4.1); gnomAD-Canada overall AF = 0.168% with AFR subpopulation AF = 1.86%.
• BS1 rule per Richards et al. 2015 (PMID:25741868): allele frequency greater than expected for the disorder. MLH3 germline disease (Lynch syndrome susceptibility, biallelic polyposis predisposition) is rare; a 1%+ allele frequency (carrier frequency ~2.6%) is far above any plausible disease prevalence.
• No ClinGen CSPEC/VCEP exists for MLH3; generic ACMG/AMP 2015 rules apply with the lab's local non-VCEP threshold (BS1 > 0.3% in gnomAD).

Sources

Reasoning

The governing generic framework (no CSPEC/VCEP exists for MLH3; generic_acmg_combination_rules fallback) pre-assigns BP4 (supporting) for intronic variants when SpliceAI max delta < 0.1, and this variant's max delta is 0.056 (DS_AG 0.021, DS_AL 0.056, DS_DG 0.0, DS_DL 0.003), clearly below the threshold. The score is also well below the 0.2 splice-alteration cutoff of the SpliceAI publication (Jaganathan et al. 2019, PMID 30661751), consistent with no predicted impact on the polypyrimidine tract / splice acceptor. BP4 met at supporting strength; the same prediction is not counted toward PP3 or BP7 (no double counting).

Rule

Generic framework: intronic/synonymous/non-canonical-splice-position variants use SpliceAI only - max delta < 0.1 -> BP4 (supporting); max delta > 0.2 -> PP3 (supporting); [0.1, 0.2] -> neither met (gray zone).

Evidence considered

• SpliceAI Lookup (source key 'spliceai') reports a max delta score of 0.056 for NM_001040108.2:c.3828-24dup (DS_AG 0.021, DS_AL 0.056, DS_DG 0.0, DS_DL 0.003), i.e. no significant predicted splice impact.
• Governing generic framework (generic_acmg_combination_rules; source basis ACMG/AMP 2015, PMID 25741868) pre-assigns BP4 (supporting) for intronic/synonymous/non-canonical-splice-position variants when SpliceAI max delta < 0.1; observed max delta 0.056 < 0.1, so BP4 is met.
• The score is also below the 0.2 (80% recall) recommended threshold for predicting splice-altering variants from the SpliceAI publication, Jaganathan et al. 2019, Cell 176(3):535-548 (PMID 30661751), supporting a no-impact interpretation.
• Per the generic framework, BayesDel is never used for PP3/BP4 (no verified published threshold available to this pipeline); REVEL was skipped by the pipeline because the variant is not an SNV. The splice path is the only applicable computational path for this intronic variant.

Still missing

• Only a single computational line (SpliceAI) was available for BP4; no second splicing predictor (e.g. MaxEntScan, NNSplice) or conservation-based line was available to satisfy a strict 'multiple lines' reading of ACMG/AMP 2015 BP4 (PMID 25741868). The score is far below threshold so the call is robust, but a second predictor would strengthen confidence.
• The <0.1 BP4 cutoff is documented in the lab's generic framework (generic_acmg_combination_rules), which attributes it to SVI-recommended SpliceAI thresholds 'per Walker et al. 2023' without listing a PMID; provenance of that specific cutoff should be verified before external reuse, though the conclusion is independently supported by the 0.2 cutoff of PMID 30661751.

Sources

Reasoning

NM_001040108.2:c.3828-24dup is an intronic single-nucleotide duplication located 24 bp upstream of exon 9 (intron 8, VEP consequence 'splice_polypyrimidine_tract_variant' per gnomAD-Canada; variant_validator exon annotation 8i). It is not a null variant: it does not fall into any default PVS1 bucket (nonsense, frameshift, canonical +-1/2 splice-site variant), and the pipeline's PVS1 variant scaffold classifies it as variant_bucket 'other' with consequence_class 'unknown' and canonical_splice_consensus false, so the generic SVI PVS1 framework is not triggered (apply_generic_pvs1_framework false). SpliceAI reports a max delta score of 0.056 (DS_AG 0.021, DS_AL 0.056, DS_DG 0.0, DS_DL 0.003), i.e. no significant predicted splice impact that would produce a null (truncating/degraded) transcript. The MLH3 PVS1 gene gate is 'eligible' (germline LoF mechanism supported in the literature, e.g. biallelic germline nonsense MLH3 variants underlying polyposis predisposition), but gene-level eligibility does not make this intronic variant a null variant. No RNA or protein functional evidence demonstrates exon skipping, intronic inclusion, or NMD-triggering out-of-frame consequence. PVS1 therefore not met.

Evidence considered

• pvs1_variant_assessment: variant NM_001040108.2:c.3828-24dup has predicted protein consequence NP_001035197.1:p.? (no protein-level consequence predicted), consequence_class 'unknown', canonical_splice_consensus false, variant_bucket 'other', and apply_generic_pvs1_framework false, with rationale that it is not a default PVS1 null-variant bucket.
• pvs1_gene_context: MLH3 germline LoF mechanism is supported (pvs1_gene_gate 'eligible') citing PMID:30573798 (biallelic germline nonsense MLH3 variant underlies polyposis predisposition) among other germline-context papers, so the gene gate does not block PVS1; the variant itself still fails the null-variant requirement.
• pvs1_generic_framework (ClinGen SVI PVS1 recommendations, PMC6185798): default PVS1 applies to nonsense, frameshift, and canonical +-1/2 splice-site variants; an intronic -24 duplication is outside these default buckets.
• spliceai: SpliceAI max delta score 0.056 for this variant (DS_AG 0.021, DS_AL 0.056, DS_DG 0.0, DS_DL 0.003), far below the 0.2 threshold used for splice-impact predictions, indicating no significant predicted disruption of splicing.
• gnomad_canada: VEP annotation for 14-75030725-T-TA reports most_severe_consequence 'splice_polypyrimidine_tract_variant' (variant_class insertion) with region_flag lcr true - i.e., an intronic polypyrimidine-tract insertion, not a canonical splice-site or coding null variant.
• PMID:25741868 (ACMG/AMP 2015 framework): PVS1 is reserved for null variants (nonsense, frameshift, canonical +-1/2 splice, initiation codon, single/multi-exon deletion); a splice-region variant requires functional confirmation of a null effect, which is absent here.

Still missing

• No RNA-level analysis (RT-PCR/minigene) demonstrating presence or absence of aberrant splicing for this intronic -24 duplication.
• No functional studies establishing whether any altered transcript is produced and whether it would undergo NMD.

Sources

Reasoning

PS3 requires well-established functional studies demonstrating a deleterious effect on gene/protein function. No functional study of c.3828-24dup was found in the evidence bundle: the literature pass retrieved a single paper (PMID:25741868, the ACMG/AMP 2015 guideline), which does not mention MLH3 or this variant (paper_extracts confirms variant_mentioned=false). No RNA-based (RT-PCR/cDNA, minigene), splicing, or protein-based (e.g., mismatch-repair activity) assay exists for this intronic duplication in any collected source. Under the generic ACMG/AMP framework (no MLH3 CSPEC/VCEP available), PS3 cannot be applied without such evidence, so the criterion is not met.

Evidence considered

• PMID:25741868 (ACMG/AMP 2015) defines PS3 as requiring well-established functional studies showing a deleterious effect and, for splice-site variants, states that confirmation of impact requires functional analysis by RNA or protein analysis.
• Literature pass and paper_extracts identified no functional study of c.3828-24dup; the only retrieved paper (PMID:25741868) does not mention the variant (variant_mentioned=false).

Still missing

• No RNA-based splicing analysis (patient RNA/cDNA or minigene assay) for MLH3 c.3828-24dup.
• No protein-level functional assay (e.g., MMR activity) for MLH3 c.3828-24dup.

Sources

Reasoning

No case-control or case-series enrichment study for this variant exists in any fetched source. The literature pass contains a single paper (PMID:25741868, ACMG/AMP standards) whose pre-computed extract confirms it does not mention MLH3 or c.3828-24dup. The variant is instead a common polymorphism (gnomAD v4.1 total AF 1.297%, 8 homozygotes; gnomAD v2.1 total AF 0.998%, 2 homozygotes), which is incompatible with significant enrichment in affected individuals for a rare germline disorder; the population group independently scores BA1/BS1 on this frequency. The only variant-specific observation beyond population data is a somatic cancer occurrence in COSMIC (COSV113426260, n=2), which is not germline case-control evidence and cannot support PS4.

Rule

PS4 (Richards et al. 2015): the prevalence of the variant in affected individuals is significantly increased compared with controls (case-control study).

Evidence considered

• No case-control or case-series study of NM_001040108.2:c.3828-24dup was identified: the literature pass (literature_pass.json) contains a single paper, PMID:25741868 (ACMG/AMP standards), and its pre-computed extract (paper_extracts.json) records variant_mentioned=false for c.3828-24dup; no cohort enrichment data for this variant exists in any fetched source.
• The variant is common in the general population: gnomAD v4.1 (source key gnomad_v4) total AF = 0.0129675 (1.297%, 16359/1261538 alleles, 8 homozygotes) and gnomAD v2.1 (source key gnomad_v2) total AF = 0.0099843 (0.998%, 1246/124796 alleles, 2 homozygotes). A ~1% allele frequency is incompatible with significant case enrichment for a rare Mendelian disorder, so PS4 is not met.
• COSMIC lists this exact allele as a somatic cancer observation (COSV113426260, n=2; COSMIC is not a source_registry key and is cited here in prose only). Somatic tumor occurrence is not germline case-control evidence and cannot support PS4.
• PS4 rule per Richards et al. 2015 (PMID:25741868): prevalence of the variant in affected individuals is significantly increased compared with controls; requires case-control data, which are absent here.

Still missing

• No case-control study of c.3828-24dup in MLH3-related disease was identified in the literature pass; a dedicated case-control or large case-series study would be required to ever support PS4, and the ~1% population allele frequency makes such enrichment implausible.

Sources

Reasoning

PM2 requires absence from controls or extremely low population frequency. Observed allele frequencies are more than an order of magnitude above the lab's non-VCEP PM2 operating threshold of AF < 0.1% in every dataset: gnomAD v4.1 1.297%, gnomAD v2.1 0.998%, gnomAD-Canada v1.0 0.168%. Even the lowest observed frequency (gnomAD-Canada overall) exceeds the threshold. PM2 is not met.

Rule

PM2 (Richards et al. 2015): absent from controls (or at extremely low frequency if recessive). Local non-VCEP operating threshold: AF < 0.1% in gnomAD.

Evidence considered

• PM2 rule per Richards et al. 2015 (PMID:25741868): absent from controls (or at extremely low frequency if recessive). Local non-VCEP operating threshold: AF < 0.1% in gnomAD.
• Observed AF: gnomAD v4.1 (source key gnomad_v4) = 0.0129675 (1.297%); gnomAD v2.1 (source key gnomad_v2) = 0.0099843 (0.998%); gnomAD-Canada v1.0 (source key gnomad_canada) = 0.001683 (0.168%) - all exceed 0.1%.
• The variant is a common polymorphism (rs775178280), so the 'absent or extremely rare' condition of PM2 is unambiguously not satisfied; no PM2 supporting evidence can be drawn from population frequency for this variant.

Sources

Reasoning

Variant is an intronic 1-bp duplication (c.3828-24dup) in intron 8 of MLH3, 24 bp upstream of the exon 9 splice acceptor (gnomAD-Canada VEP consequence: splice_polypyrimidine_tract_variant). The splice-impact path therefore applies and the missense in-silico path does not (not a missense SNV; REVEL/BayesDel were skipped as 'not snv or no coords'). SpliceAI Lookup reports max delta score 0.056 (DS_AG 0.021, DS_AL 0.056, DS_DG 0.0, DS_DL 0.003), which is below the governing generic framework's PP3 threshold (SpliceAI max delta > 0.2) and far below the 0.2 (80% recall) splice-alteration cutoff recommended in the SpliceAI publication (Jaganathan et al. 2019, PMID 30661751). No computational evidence of splice impact; PP3 not met.

Rule

Generic framework: intronic/synonymous/non-canonical-splice-position variants use SpliceAI only - max delta > 0.2 -> PP3 (supporting); max delta < 0.1 -> BP4 (supporting); [0.1, 0.2] -> neither met (gray zone). BayesDel never used (no verified published threshold).

Evidence considered

• SpliceAI Lookup (source key 'spliceai') reports for NM_001040108.2:c.3828-24dup a max delta score of 0.056 with DS_AG 0.021, DS_AL 0.056, DS_DG 0.0, DS_DL 0.003 (variant consequence 'sequence_variant').
• The governing generic framework (generic_acmg_combination_rules, source basis PMID 25741868) assigns PP3 (supporting) only when SpliceAI max delta > 0.2 for intronic/synonymous/non-canonical-splice-position variants; observed 0.056 is far below this, so PP3 is not met.
• The 0.2 threshold used to flag likely splice-altering variants is the recommended 80% recall cutoff from the SpliceAI publication, Jaganathan et al. 2019, Cell 176(3):535-548 (PMID 30661751); a max delta of 0.056 is below it, indicating no predicted splice impact.

Still missing

• Missense in-silico predictors (REVEL, BayesDel, aGVGD) are not applicable because the variant is an intronic duplication, not a missense substitution; only the splice-impact sub-path was evaluable.

Sources

Reasoning

Per group policy, PP5 applies only for an exact-variant ClinVar expert-panel classification of Pathogenic or Likely pathogenic. ClinVar VCV001697684 has exactly one submission (SCV002551390, Center for Genomic Medicine, Rigshospitalet): Likely benign, review status 'criteria provided, single submitter' (1 star), germline clinical testing. Submission audit shows expert_panel_submissions = 0 and is_expert_panel = false for the only submission. No expert-panel pathogenic classification exists, so PP5 is not met.

Rule

PP5 (Richards et al. 2015): reputable source recently reports variant as pathogenic, but the evidence is not available to the laboratory to perform an independent evaluation. Pipeline restriction: only an exact-variant ClinVar expert-panel Pathogenic/Likely pathogenic classification may trigger PP5.

Evidence considered

• ClinVar VCV001697684 (source key clinvar) contains exactly one submission (SCV002551390): Likely benign, review status 'criteria provided, single submitter' (1 star), germline clinical testing, condition 'not specified'.
• Submission audit (prefetch.json / clinvar_submission_audit): submissions_total = 1, expert_panel_submissions = 0, is_expert_panel = false for the single submission, trust_tier = single_submitter. The Likely benign label is from an ordinary clinical laboratory, not an expert panel.
• PP5 rule per Richards et al. 2015 (PMID:25741868) with pipeline restriction: PP5 is triggered only by an exact-variant ClinVar expert-panel Pathogenic/Likely pathogenic classification; no such classification exists for this variant, so PP5 is not met.

Sources

Reasoning

Homozygotes are observed (8 in gnomAD v4.1, 7 of them African/African American; 2 in gnomAD v2.1), but BS2 requires observation in a healthy adult for a disorder with full penetrance expected at an early age. MLH3-associated Lynch syndrome/polyposis is adult-onset and incompletely penetrant, and gnomAD is a general-population cohort without verified healthy-adult phenotype (and includes some affected individuals). The criterion's phenotype conditions are therefore not satisfied, and the homozygote observations are instead fully consistent with a common benign polymorphism (captured by BA1/BS1).

Rule

BS2 (Richards et al. 2015): 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.

Evidence considered

• gnomAD v4.1 (source key gnomad_v4) reports 8 homozygotes (7 African/African American); gnomAD v2.1 (source key gnomad_v2) reports 2 homozygotes.
• BS2 rule per Richards et al. 2015 (PMID:25741868): 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. These conditions are not met: gnomAD is not a verified healthy-adult cohort and MLH3-related disease is late-onset with incomplete penetrance.
• Homozygote counts in a ~1% allele-frequency polymorphism are expected under Hardy-Weinberg equilibrium in a general cohort and do not by themselves establish healthy-carrier status for a fully penetrant early-onset disorder.

Still missing

• A well-phenotyped healthy-adult cohort (e.g., a cancer-free elderly control series) would be needed to ever support BS2 for this gene; none is present in this case.

Sources

Reasoning

BS3 requires well-established functional studies showing no damaging effect on protein function or splicing. No such functional assay exists for c.3828-24dup in the collected evidence. SpliceAI predicts minimal splice impact (max delta score 0.056, below the 0.1 significance threshold), but this is an in silico prediction, not a well-established functional study, and is insufficient on its own for BS3 under the generic ACMG/AMP framework. The high population frequency (gnomAD v4.1 AF ~1.3%, 8 homozygotes) is contextual but is population-based evidence adjudicated under population criteria (BS1/BA1), not functional evidence for BS3. With no RNA- or protein-based assay demonstrating preserved function/splicing, BS3 is not met.

Evidence considered

• SpliceAI (in silico) predicts no significant splice impact for c.3828-24dup (max delta score 0.056, all sub-scores <0.1); evaluated for BS3 but insufficient because in silico prediction is not a well-established functional study.
• PMID:25741868 (ACMG/AMP 2015) defines BS3 as requiring well-established functional studies showing no damaging effect on protein function or splicing.
• gnomAD v4.1 AF 1.297% (16359/1261538 alleles, 8 homozygotes) and gnomAD v2.1 AF ~1.0% (2 homozygotes) noted as context only; population frequency is not functional evidence for BS3.

Still missing

• No RT-PCR/cDNA or minigene splicing assay demonstrating normal splicing for c.3828-24dup.
• No protein-level functional assay demonstrating preserved MLH3 function.

Sources

Reasoning

Per group policy, BP6 applies only for an exact-variant ClinVar expert-panel classification of Benign or Likely benign. Although ClinVar VCV001697684 is labeled Likely benign, the classification comes from a single ordinary clinical laboratory (SCV002551390, 1 star, 'criteria provided, single submitter'); the submission audit reports expert_panel_submissions = 0 and is_expert_panel = false. Ordinary laboratory submissions and non-expert assertions must never trigger BP6, so BP6 is not met.

Rule

BP6 (Richards et al. 2015): reputable source recently reports variant as benign, but the evidence is not available to the laboratory to perform an independent evaluation. Pipeline restriction: only an exact-variant ClinVar expert-panel Benign/Likely benign classification may trigger BP6.

Evidence considered

• ClinVar VCV001697684 (source key clinvar) is labeled Likely benign (api_classification), but review_status = 'criteria provided, single submitter' with review_stars = 1 and number_of_submissions = 1.
• Submission audit (prefetch.json / clinvar_submission_audit): the sole submission (SCV002551390, Center for Genomic Medicine, Rigshospitalet, Copenhagen University Hospital) has is_expert_panel = false, trust_tier = single_submitter; expert_panel_submissions = 0. Per group policy, ordinary laboratory submissions must never trigger BP6.
• BP6 rule per Richards et al. 2015 (PMID:25741868) with pipeline restriction: BP6 is triggered only by an exact-variant ClinVar expert-panel Benign/Likely benign classification; the existing Likely benign label is a non-expert laboratory assertion, so BP6 is not met.

Still missing

• If a ClinGen/InSiGHT or other expert panel were to classify this exact variant Benign/Likely benign with 3-star review status, BP6 could be reconsidered; no such classification exists as of the ClinVar data in this case (date_last_updated 2025-03-16).

Sources

Reasoning

PS2 requires a documented de novo occurrence of the variant in a proband with paternity and maternity confirmed by parental testing. No proband, parental testing, or trio data exists anywhere in this case bundle: the ClinVar record (VCV001697684, SCV002551390) contains no proband phenotype or parental testing results and lists no applied criteria; the only retrieved publication (PMID:25741868) does not mention MLH3 or c.3828-24dup. There is therefore no evidence to support or refute a de novo origin, and PS2 cannot be evaluated.

Rule

PS2: De novo (paternity and maternity confirmed) - pathogenic strong (Richards et al. 2015, PMID:25741868)

Evidence considered

• Generic ACMG/AMP 2015 (PMID:25741868) defines PS2 as a de novo variant in the proband with paternity and maternity confirmed; this rule governs because no ClinGen CSPEC/VCEP specification exists for MLH3 (vcep_materials.json: cspec found=false, generic_acmg_fallback).
• ClinVar record VCV001697684 / SCV002551390 (Center for Genomic Medicine, Rigshospitalet, Copenhagen University Hospital; Likely benign, single submitter, condition 'not specified') was inspected for proband and parental testing information: none is reported, and the submission's criteria list is empty, so no de novo assertion exists for c.3828-24dup.
• The single fetched paper, PMID:25741868 (ACMG/AMP standards), does not mention MLH3 or c.3828-24dup (paper_extracts.json: variant_mentioned=false); it was consulted only as the rule-defining source for PS2.
• No proband or parental testing data is present in case_summary.json, evidence.json, or literature_pass.json; the case bundle contains no family/trio section at all.

Still missing

• No proband clinical/family report is available in this case; no parental testing (maternity/paternity confirmation) results exist to document a de novo occurrence.

Sources

Reasoning

PM3 (ACMG/AMP 2015, PMID 25741868) requires, for a recessive disorder, that the variant be detected in trans with a pathogenic variant in an affected proband. No such observation exists in any case source: there is no affected-proband or family segregation/phase data, no pathogenic MLH3 variant reported in trans or cis with this variant, and the only clinical record (ClinVar VCV001697684, single submitter, Likely benign, condition 'not specified') reports no criteria or phase information. No MLH3 ClinGen CSPEC/VCEP exists, so the generic ACMG/AMP 2015 definition applies, and a recessive disease association for MLH3 is not established in the case materials. Population frequency (gnomAD v4.1 AF 1.30%, 8 homozygotes) is contextually inconsistent with a rare recessive disease allele but provides no phase information. Evidence is insufficient to assess PM3.

Rule

PM3: For recessive disorders, detected in trans with a pathogenic variant (ACMG/AMP 2015, PMID 25741868).

Evidence considered

• PMID 25741868 (ACMG/AMP 2015) defines PM3: 'For recessive disorders, detected in trans with a pathogenic variant'; this generic framework applies because no MLH3 ClinGen CSPEC/VCEP exists.
• ClinVar VCV001697684 (SCV002551390, Center for Genomic Medicine, Rigshospitalet): Likely benign, single submitter (1 star), condition 'not specified', no criteria provided; contains no phase, segregation, or affected-proband observations.
• gnomAD v4.1: AF 0.01297 (16,359/1,261,538 alleles), 8 homozygotes, grpmax FAF 0.02498; population-level homozygosity is not an affected-proband trans observation.
• gnomAD v2.1: AF 0.00998 (1,246/124,796 alleles), 2 homozygotes, grpmax FAF 0.03699; consistent common-variant context.

Still missing

• No affected-proband observations or family segregation data are available for this variant.
• No phase (trans/cis) information exists for this variant relative to any pathogenic MLH3 variant.
• No pathogenic MLH3 variant is reported in the case materials to test trans configuration.
• Inheritance mode / recessive disease association for MLH3 is not established in the available evidence.

Sources

Reasoning

PM6 requires an assumed de novo occurrence in a proband without confirmation of paternity and maternity. As with PS2, there is no proband or parental testing information in the case bundle whatsoever, so an assumed de novo origin cannot be documented. PM6 is not assessable from available evidence.

Rule

PM6: Assumed de novo, but without confirmation of paternity and maternity - pathogenic moderate (Richards et al. 2015, PMID:25741868)

Evidence considered

• Generic ACMG/AMP 2015 (PMID:25741868) defines PM6 as an assumed de novo variant without confirmation of paternity and maternity; this rule governs because no ClinGen CSPEC/VCEP specification exists for MLH3.
• ClinVar record VCV001697684 contains no proband description or parental testing data and lists no applied criteria, providing no basis for an assumed de novo assertion.
• PMID:25741868 does not mention MLH3 or c.3828-24dup (paper_extracts.json: variant_mentioned=false); consulted only as the rule-defining source for PM6.
• No de novo assumption is reported in any case-bundle artifact (case_summary.json, evidence.json, literature_pass.json).

Still missing

• No proband report or parental testing results exist to support an assumed de novo occurrence (e.g., affected proband, unaffected parents not tested).

Sources

Reasoning

PP1 requires co-segregation of the variant with disease in multiple affected family members (meioses observed) in a gene definitively known to cause the disease. No affected family members, segregation data, or meioses are reported anywhere in the case bundle, so PP1 cannot be evaluated.

Rule

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

Evidence considered

• Generic ACMG/AMP 2015 (PMID:25741868) defines PP1 as co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease; this rule governs because no ClinGen CSPEC/VCEP specification exists for MLH3.
• ClinVar record VCV001697684 (single submitter, Likely benign, condition 'not specified') contains no family/segregation assertions and no applied criteria, so no segregation evidence is contributed.
• PMID:25741868 does not mention MLH3 or c.3828-24dup (paper_extracts.json: variant_mentioned=false); consulted only as the rule-defining source for PP1.
• No segregation, affected-relative testing, or meioses data is present in case_summary.json, evidence.json, or literature_pass.json.

Still missing

• No segregation data: no affected relatives tested for the variant, no meioses counted, no family pedigree available.

Sources

Reasoning

PP4 requires a phenotype or family history highly specific for a disease with a single genetic etiology. No proband phenotype, clinical indication, or family history data are present anywhere in the case bundle; the single ClinVar submission (SCV002551390) carries condition label 'not specified'. With no phenotype information, phenotype specificity cannot be evaluated, so PP4 is not assessed rather than declared met or not met.

Rule

PP4 (Richards et al. 2015): phenotype or family history is highly specific for a disease with a single genetic etiology.

Evidence considered

• No proband phenotype, clinical indication, or family-history data are present in review_manifest.json, case_summary.json, prefetch.json, or evidence.json; the only ClinVar submission (SCV002551390) reports condition label 'not specified' (source key clinvar).
• PP4 rule per Richards et al. 2015 (PMID:25741868): the variant's phenotype or family history is highly specific for a disease with a single genetic etiology. Without any phenotype information, PP4 cannot be evaluated (not_assessed).

Still missing

• Proband phenotype, clinical indication, and family history are missing from the case bundle; these must be supplied (e.g., from the test requisition or clinical notes) before PP4 can be adjudicated.

Sources

Reasoning

BS4 requires observed non-segregation: affected family members who do not carry the variant, or unaffected carriers, indicating the variant does not track with disease. No family members were tested and no non-segregation observations exist in the case bundle, so BS4 cannot be evaluated. Population frequency evidence (relevant to BA1/BS1) is handled by the population group, not BS4.

Rule

BS4: Non-segregation with disease - benign strong (Richards et al. 2015, PMID:25741868)

Evidence considered

• Generic ACMG/AMP 2015 (PMID:25741868) defines BS4 as non-segregation with disease (absence of the variant in affected family members); this rule governs because no ClinGen CSPEC/VCEP specification exists for MLH3.
• ClinVar record VCV001697684 contains no family testing data and lists no applied criteria, contributing no non-segregation evidence.
• PMID:25741868 does not mention MLH3 or c.3828-24dup (paper_extracts.json: variant_mentioned=false); consulted only as the rule-defining source for BS4.
• No non-segregation observations (affected relatives without the variant) are present in any case-bundle artifact.

Still missing

• No non-segregation data: no affected family members tested negative for the variant and no unaffected carriers reported.

Sources

Reasoning

BP2 (ACMG/AMP 2015, PMID 25741868) requires observation of the variant in trans with a pathogenic variant for a fully dominant disease/mechanism, or in cis with a pathogenic variant in any inheritance pattern. No co-occurrence or phase observation exists in any case source: ClinVar VCV001697684 provides no phase or co-occurrence data, no pathogenic variant is reported in cis or trans with this variant, and the reviewed literature (PMID 25741868) contains no variant-specific data. A fully dominant MLH3 disease mechanism is also not established in the case materials. Population homozygosity in gnomAD (8 homozygotes in v4.1) is not an affected-individual cis/trans observation. Evidence is insufficient to assess BP2.

Rule

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

Evidence considered

• PMID 25741868 (ACMG/AMP 2015) defines 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'; no variant-specific phase data are provided by this paper.
• ClinVar VCV001697684 (SCV002551390): Likely benign, single submitter (1 star), condition 'not specified'; no co-occurrence or phase information is reported.
• gnomAD v4.1: 8 homozygotes among 16,359 alleles; population homozygosity reflects the same variant on both alleles in unselected individuals and is not an affected-proband observation of cis/trans phase with a pathogenic variant.

Still missing

• No observation of this variant in trans with a pathogenic variant for a fully dominant MLH3 disease/mechanism.
• No observation of this variant in cis with a pathogenic variant in any inheritance pattern.
• No co-occurrence or phase data provided by the single ClinVar submission.
• Fully dominant disease mechanism for MLH3 not established in the available evidence.

Sources

Reasoning

BP5 requires the variant to be found in a case with an alternate molecular basis for the disease. No proband-level data exist to evaluate this: there is no multi-gene panel result, no second pathogenic variant, no alternate genetic diagnosis, and no phenotype context in any fetched source. The criterion is therefore not assessed; there is neither evidence for nor against an alternate molecular explanation.

Rule

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

Evidence considered

• No evidence of an alternate molecular basis was identified in any fetched source: no proband-level multi-gene panel result, no second pathogenic variant, and no alternate genetic diagnosis are documented in review_manifest.json, case_summary.json, prefetch.json, or evidence.json.
• BP5 rule per Richards et al. 2015 (PMID:25741868): variant found in a case with an alternate molecular basis for disease. Insufficient proband data to evaluate, so BP5 is not assessed.

Still missing

• Proband genetic-testing results beyond this variant (e.g., multi-gene panel or exome findings documenting an alternate molecular diagnosis) and phenotype context are required to adjudicate BP5; none are present in the case bundle.

Sources