RUNX1 c.707dup, p.(Met236IlefsTer25)

Reasoning

NM_001754.4:c.707dupT is a single-nucleotide duplication in exon 7 (c.614-805) of the ClinGen Myeloid Malignancy VCEP default transcript NM_001754.4 (RUNX1c isoform), producing an out-of-frame frameshift predicted as p.(Met236IlefsTer25); the wild-type protein is 481 aa and the predicted mutant protein is 260 aa, with the premature termination codon at codon ~260 (c.781-783). Germline loss of function is an established mechanism of disease for RUNX1 (FPD/AML, autosomal dominant) per the MM-VCEP spec. The PTC is >50-55 nt upstream of the final exon-exon junction (exon 8/9 at c.967), so the mutant transcript is predicted to undergo nonsense-mediated mRNA decay; the variant lies well downstream of c.98, outside the VCEP's 'before c.98' region where PVS1 is not applicable, and is not a splice-consensus variant (SpliceAI max delta 0.036). Per the modified RUNX1 PVS1 decision tree, an NMD-predicted truncating variant in the default isoform meets PVS1 at Very Strong; the VCEP's PVS1_strong tier applies to C-terminal truncating variants NOT predicted to undergo NMD, and PVS1_moderate applies to exon 1-3 deletions affecting only isoform c - neither applies here.

Rule

Per modified RUNX1 PVS1 decision tree for SNVs and CNVs (MM-VCEP v3.1): frameshift variant in the default transcript NM_001754.4 (RUNX1c), downstream of c.98, predicted to undergo NMD, in a gene with established LoF mechanism -> PVS1 (Very Strong).

Evidence considered

• Mutalyzer/VariantValidator normalization (prefetch.json) confirms NM_001754.4:c.707dupT as a duplication of T at c.707 with predicted protein NP_001745.2:p.(Met236IlefsTer25); wild-type protein length 481 aa, predicted mutant length 260 aa, frameshift start at position 235.
• VariantValidator places the variant in exon 7 of NM_001754.4 (start_exon/end_exon = 7) for GRCh37, GRCh38 and NG_011402.2; exon 7 spans c.614-805, so c.707 is internal to the exon and not near a splice junction.
• The ClinGen Myeloid Malignancy VCEP RUNX1 v3.1 (cspec, doc 135637580) designates NM_001754.4 (RUNX1c) as the default transcript and states that RUNX1 LOF variants are a common mechanism of disease in FPD/AML; PVS1 is applied per a modified RUNX1 PVS1 decision tree, with C-terminal truncating variants not predicted to undergo NMD classified as PVS1_strong and exon 1-3 deletions (presumably affecting only isoform c) as PVS1_moderate.
• The VCEP PVS1 instructions note that nonsense/frameshift variants before c.98 only affect one functional RUNX1 transcript and PVS1 is not applicable there; c.707dup is at c.707, well downstream of c.98, so this exclusion does not apply.
• PTC position analysis: the premature termination codon falls at codon ~260 (c.781-783) in exon 7; the final exon-exon junction is the exon 8/9 boundary at c.967, so the PTC is >50-55 nt upstream of the last exon-exon junction and the frameshift transcript is predicted to undergo NMD, consistent with the SVI PVS1 framework (pvs1_generic_framework, PMC6185798) criterion for full-strength PVS1.
• SpliceAI reports max delta score 0.036 (DS_AL 0.036, DS_DG 0.031, DS_DL 0.027, DS_AG 0.002), below the 0.20 threshold, indicating no predicted splicing impact; the variant is a pure coding frameshift with no splice-confounding effect.
• pvs1_gene_context.json confirms germline LoF is an established disease mechanism for RUNX1 (official source: cspec_vcep) and the PVS1 gene gate is eligible; pvs1_variant_assessment.json classifies the consequence as frameshift and suggests default strength PVS1 with no downgrade triggers met (transcript relevance confirmed for NM_001754.4; truncation is not in a non-critical distal region).

Still missing

• No RNA-based confirmation of NMD (e.g., patient RNA with an NMD inhibitor) is available; NMD is inferred from PTC position relative to the final exon-exon junction.
• The full MM-VCEP PVS1 decision-tree and splicing-effects table text was not included in the case bundle; strength assignment relies on the PVS1 instructions present in the CSPEC rule text plus the generic SVI PVS1 framework (PMC6185798).

Sources

Reasoning

The MM-VCEP RUNX1 v3.1 specification defines PM2_Supporting as MAF <= 0.00005 with at least 2,000 alleles tested around and 20x coverage at the position, recommending evaluation with the GrpMax FAF when available in gnomAD v4.1.0 and otherwise requiring all subpopulations to meet the threshold. The variant is absent from gnomAD v2.1 (exome), gnomAD v4.1 (joint exome+genome), and gnomAD-Canada v1.0 (genomes): MAF = 0 in every subpopulation, trivially below 0.00005 and satisfying the all-subpopulations/GrpMax FAF caveat; the datasets far exceed the 2,000-allele minimum. PM2_Supporting is met.

Rule

MM-VCEP RUNX1 v3.1 PM2_Supporting: Minor allele frequency ≤ 0.00005 with at least 2000 alleles tested around and 20x coverage at the position (use GrpMax FAF when available in gnomAD v4.1.0, otherwise all subpopulations must meet the threshold)

Evidence considered

• MM-VCEP RUNX1 v3.1 PM2 rule (cspec): 'PM2_Supporting: Minor allele frequency ≤ 0.00005 with at least 2000 alleles tested around and 20x coverage at the position'; caveat: evaluate using the GrpMax FAF when available in gnomAD v4.1.0, otherwise require all subpopulations to meet the threshold; the mean coverage of RUNX1 in the population database used should be at least 20x.
• Variant absent from gnomAD v2.1 (exome, GRCh37; 21-36206804-C-CA), gnomAD v4.1 (joint exome+genome, GRCh38; chr21-34834507-C-CA), and gnomAD-Canada v1.0 (genomes; 21-34834507-C-CA): MAF = 0 in every dataset and every continental subpopulation, satisfying the <= 0.00005 threshold; each dataset contains far more than the required 2,000 alleles (gnomAD v2.1 exomes alone exceed 250,000 alleles).
• Ancestry consideration: no continental population shows any allele at this position, so the 'all subpopulations meet the threshold' fallback of the VCEP caveat is satisfied and no ancestry-specific excess exists.
• Homozygote consideration: variant absent everywhere, so no homozygotes are observed; consistent with an autosomal dominant LoF disease model.
• Indel/threshold-artifact consideration: the ACMG/AMP PM2 caveat notes population data for indels may be poorly called by NGS; uniform absence across two exome/genome call sets and an independent genome-only cohort (gnomAD-Canada) supports genuine absence rather than a systematic calling artifact.
• Per-position >=20x coverage could not be directly confirmed from the bundle (see gaps_remaining); this does not change the PM2 decision because the variant is absent from the entire dataset.

Still missing

• Per-position sequencing coverage (>= 20x at c.707 / chr21:34834507) was not explicitly captured in the evidence bundle (gnomAD source_data fields are empty for these absent records); uniform absence across three independent datasets makes the coverage requirement very likely satisfied, but it was not directly verified.
• Explicit GrpMax FAF values were not retrieved for this variant; because the variant is absent from the entire dataset, the GrpMax FAF and all subpopulation frequencies are 0, satisfying the VCEP caveat either way.

Sources

Reasoning

MM-VCEP v3.1 PM5_Supporting states: 'PM5_Supporting is also applied to nonsense/frameshift variants that are downstream of c.98 (in transcript NM_001754.4).' c.707dupT is a frameshift variant (p.Met236IlefsTer25) located downstream of c.98 (exon 7, c.707). Both PM5 caveats are satisfied: (1) the variant must not impact splicing based on RNA assay or SpliceAI <= 0.20 - SpliceAI max delta score is 0.036 (<= 0.20) and no conflicting RNA assay is available; (2) PM5 cannot be used if PM1 was applied at any strength - PM1 is not applied here. PM5 is therefore met at Supporting strength.

Rule

PM5_Supporting: Missense change at an AA residue where a different missense change which has been determined to be likely pathogenic before (after accounting for Grantham scores). PM5_Supporting is also applied to nonsense/frameshift variants that are downstream of c.98 (in transcript NM_001754.4). Caveats: variant must not impact splicing based on RNA assay or SpliceAI <= 0.20; PM5 cannot be used if PM1 was applied at any strength level.

Evidence considered

• MM-VCEP v3.1 (CSPEC doc 135637580) PM5_Supporting rule text: 'PM5_Supporting is also applied to nonsense/frameshift variants that are downstream of c.98 (in transcript NM_001754.4)', with caveats that the variant must not impact splicing (RNA assay or SpliceAI <= 0.20) and that PM5 cannot be used if PM1 was applied at any strength level.
• SpliceAI lookup for NM_001754.4:c.707dup (variant 21-36206804-C-CA): max delta score = 0.036 (DS_AL 0.036, DS_DG 0.031, DS_DL 0.027, DS_AG 0.002), i.e. <= 0.20; no significant splice impact predicted.
• pm5_candidates.json confirms the variant consequence class is frameshift (not missense), so classic same-residue missense PM5 search was flagged not applicable; consistent with applying the VCEP's explicit frameshift-downstream-of-c.98 extension instead.
• PM1 is not applied for this variant, so the 'PM5 cannot be used if PM1 was applied' caveat is satisfied.

Still missing

• No RNA assay data available for this variant; the no-splicing-impact caveat rests on SpliceAI (max delta 0.036 <= 0.20) only.
• No same-residue missense comparator exists (variant absent from ClinVar), but this does not block PM5_Supporting, which the MM-VCEP applies categorically to nonsense/frameshift variants downstream of c.98.

Sources

Reasoning

MM-VCEP v3.1 restricts PM1 to the Runt Homology Domain (RHD): PM1_strong for variants affecting residues R107, K110, A134, R162, R166, S167, R169, G170, K194, T196, D198, R201, R204, and PM1_Supporting for variants affecting other residues within AA 89-204. The affected residue of this variant is p.Met236 (frameshift start), which lies outside the RHD (AA 89-204) defined by the VCEP. CancerHotspots reports no statistically significant hotspot at this residue. PM1 is therefore not met.

Rule

PM1_strong: Variant affecting one of the following 13 AA residues within the RHD: R107, K110, A134, R162, R166, S167, R169, G170, K194, T196, D198, R201, R204. PM1_Supporting: Variant affecting one of the other AA residues 89-204 within the RHD.

Evidence considered

• MM-VCEP v3.1 (CSPEC doc 135637580) PM1 rule: PM1_strong for the 13 listed RHD residues (R107, K110, A134, R162, R166, S167, R169, G170, K194, T196, D198, R201, R204); PM1_Supporting for other AA residues 89-204 within the RHD.
• VariantValidator/Mutalyzer placed the frameshift at p.Met236IlefsTer25 (NP_001745.2, RUNX1c); residue 236 is not within the VCEP-defined RHD interval AA 89-204.
• CancerHotspots search for 'RUNX1 M236' found no statistically significant hotspot (evidence.json hotspots result, found=false).

Still missing

• No evidence places residue 236 within the RUNX1 RHD (AA 89-204) or within a defined mutational hotspot; CancerHotspots query 'RUNX1 M236' returned no hotspot row.

Sources

Reasoning

Variant NM_001754.4:c.707dupT is a frameshift duplication in exon 7 (p.(Met236IlefsTer25)) and is not a missense, synonymous, or intronic variant. Under the MM-VCEP RUNX1 v3.1 specification (cspec), PP3 for missense requires REVEL >= 0.88 OR SpliceAI >= 0.38, and for synonymous/intronic variants requires SpliceAI >= 0.38; a residual MM-VCEP note permits other predictors in agreement only when REVEL/SpliceAI scores are unavailable. The only computational line available here is SpliceAI: max delta score = 0.036 (DS_AG 0.002, DS_AL 0.036, DS_DG 0.031, DS_DL 0.027), far below the 0.38 threshold, so the splice-impact path does not support PP3. REVEL and BayesDel are unavailable for this indel (prefetch skipped: 'not snv or no coords'), and no calibrated alternative predictor scores are present, so no other in-silico line can substitute. The VCEP pilot spreadsheet contains no entry for c.707dup (no pre-assigned code) and its precedent rows apply PP3 only to missense variants, never frameshift. PP3 is therefore not met.

Rule

PP3: For missense variants, REVEL >= 0.88 OR SpliceAI >= 0.38 (including cryptic novel splice sites); for synonymous and intronic (intron 4-8) variants, SpliceAI >= 0.38. MM-VCEP note: other predictors in agreement may substitute only when REVEL/SpliceAI scores are unavailable. Do not use for canonical splice-site variants.

Evidence considered

• MM-VCEP RUNX1 v3.1 (cspec) PP3 rule: 'For missense variants: PP3: REVEL score >= 0.88 or SpliceAI >= 0.38, including the creation of cryptic novel splice sites. For synonymous and intronic (intron 4-8) variants: PP3: SpliceAI >= 0.38, including the creation of cryptic novel splice sites.' with MM-VCEP note 'For some variant types that REVEL or SpliceAI scores are not available, multiple other predictors in agreements can be used in PP3' and caveat 'Do not use for canonical splice site variants.'
• SpliceAI Lookup (spliceai) scores for NM_001754.4:c.707dup (21-36206804-C-CA, hg37): DS_AG 0.002, DS_AL 0.036, DS_DG 0.031, DS_DL 0.027; max delta score 0.036; variant consequence 'sequence_variant'. Evidence sentence: 'SpliceAI predicts no significant splice impact for this variant (max delta score = 0.04).'
• The VCEP establishes the SpliceAI PP3 threshold (>= 0.38, 90% sensitivity) and BP4/BP7 threshold (<= 0.20, 90% specificity) internally, based on 202 variants in genes associated with inherited hematologic malignancies (per cspec BP4/PP3 instructions-to-use).
• REVEL and BayesDel were skipped for this variant in prefetch/evidence collection with reason 'not snv or no coords'; no REVEL or BayesDel score exists for this indel, and BayesDel has no verified published threshold available to this pipeline.
• Variant consequence: frameshift duplication c.707dupT in exon 7 (exon 7 spans c.614-805), predicted protein NP_001745.2:p.(Met236IlefsTer25); not a canonical splice-site variant; not synonymous and not intronic.
• Myeloid-Malignancy-VCEP-RUNX1-Pilot-Results.xlsx contains no row for c.707dup (checked all 969 variant rows; no exact HGVS match), so no pre-assigned PP3/BP4/BP7 code applies; in the pilot, PP3 appears only on missense variants (e.g., p.Arg162Gly, p.Arg162Lys, p.Asp198Asn, p.Gly199Ala, p.His105Pro, p.Ala156Thr) and never on frameshift variants.
• SpliceAI max delta 0.036 < 0.38: the splice-impact prediction is not consistent with PP3, so the only available computational line affirmatively fails the PP3 threshold.

Still missing

• REVEL score unavailable for this indel (frameshift), so the missense in-silico arm of the VCEP PP3 rule could not be independently evaluated; only the SpliceAI splice-impact arm was testable.
• No other calibrated computational predictors (e.g., CADD, VEST) are present in the case evidence bundle, and BayesDel is treated as insufficiently calibrated (no verified published threshold), so the MM-VCEP residual PP3 clause ('multiple other predictors in agreement') cannot be exercised.
• No RNA/minigene data available to assess splicing experimentally (would be relevant to PVS1_variable/BP7_variable RNA pathways, not PP3 per VCEP caveat).

Sources

Reasoning

The governing framework is the ClinGen Myeloid Malignancy VCEP (MM-VCEP) RUNX1 specification v3.1 (cspec doc 135637580), which defines BA1 as MAF >= 0.0015 (0.15%) in any general continental population dataset with >= 2,000 alleles tested and variant present in >= 5 alleles. NM_001754.4:c.707dupT (p.Met236IlefsTer25) is absent from gnomAD v2.1 (21-36206804-C-CA), gnomAD v4.1 (chr21-34834507-C-CA), and gnomAD-Canada v1.0; observed MAF is 0 in every dataset and every continental subpopulation, far below the 0.0015 stand-alone threshold. BA1 is not met.

Rule

MM-VCEP RUNX1 v3.1 BA1: Minor allele frequency ≥ 0.0015 (0.15%) in any general continental population dataset with ≥ 2,000 alleles tested and variant present in ≥ 5 alleles

Evidence considered

• MM-VCEP RUNX1 v3.1 BA1 rule (cspec): 'BA1: Minor allele frequency ≥ 0.0015 (0.15%) in any general continental population dataset with ≥ 2,000 alleles tested and variant present in ≥ 5 alleles'; the threshold was derived with the Whiffin/Ware allele-frequency calculator using FPD/AML prevalence 1 in 40, a conservative unascertained penetrance estimate of 85%, allelic heterogeneity 100%, and maximum genetic heterogeneity 10%.
• Variant absent from gnomAD v2.1 (GRCh37, exome; evidence_sentence 'Absent from gnomAD v2.1.'), gnomAD v4.1 (GRCh38, joint exome+genome; 'Absent from gnomAD v4.1.'), and gnomAD-Canada v1.0 (genomes; 'Absent from gnomAD-Canada v1.0.'); observed MAF = 0 in all continental populations, so no population reaches the 0.0015 BA1 threshold.
• Threshold-artifact check: this is a single-nucleotide duplication (indel), a class for which the ACMG/AMP PM2 caveat notes population data may be poorly called by NGS; however, for BA1 the concern is false-negative calling (would bias toward absence), and absence here is concordant across three independent datasets including a genome-only cohort, which supports genuine rarity rather than artifact.

Sources

Reasoning

The MM-VCEP RUNX1 v3.1 specification defines BS1 as MAF between 0.00015 (0.015%) and 0.0015 (0.15%) in any general continental population dataset with >= 2,000 alleles tested and variant present in >= 5 alleles, derived from the Whiffin/Ware calculator with prevalence 1 in 40, penetrance 85%, allelic heterogeneity 100%, and genetic heterogeneity 1%. The variant is absent from gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0 (MAF = 0), which is below the lower bound of the BS1 range (0.00015). BS1 is not met.

Rule

MM-VCEP RUNX1 v3.1 BS1: Minor allele frequency between 0.00015 (0.015%) and 0.0015 (0.15%) in any general continental population dataset with ≥ 2,000 alleles tested and variant present in ≥ 5 alleles

Evidence considered

• MM-VCEP RUNX1 v3.1 BS1 rule (cspec): 'BS1: Minor allele frequency between 0.00015 (0.015%) and 0.0015 (0.15%) in any general continental population dataset with ≥ 2,000 alleles tested and variant present in ≥ 5 alleles'; the range was derived using the Whiffin/Ware calculator with FPD/AML prevalence 1 in 40, penetrance 85%, allelic heterogeneity 100%, and genetic heterogeneity 1% (one magnitude lower than for BA1), 95% CI, and the MM-VCEP notes the variant can be classified likely benign on BS1 alone if no contradictory evidence supports pathogenicity.
• Variant absent from gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0 (MAF = 0 across all continental subpopulations); an observed MAF of 0 falls below the BS1 range lower bound of 0.00015, so the frequency is not 'greater than expected for the disorder' under the VCEP threshold.
• Disease-prevalence/penetrance context (from the VCEP, not an independent source): FPD/AML thrombocytopenia prevalence ~1 in 40 with near-complete (up to 85%) penetrance in carriers; these parameters are already baked into the VCEP BA1/BS1 thresholds, and no founder mutations are known in RUNX1.

Sources

Reasoning

BP2 is applicable under the MM-VCEP RUNX1 v3.1 specification (supporting strength), but no evidence satisfies it: the variant was not observed in trans with a pathogenic variant for this fully penetrant dominant disorder, not observed in cis with a pathogenic variant, and not observed in the homozygous state in individuals without FPD/AML phenotype. NM_001754.4:c.707dupT is absent from ClinVar (no submissions) and absent from gnomAD v2.1, v4.1, and gnomAD-Canada v1.0; the case bundle contains no proband/family/phase data, and none of the three available full-text publications mention this variant. BP2 is therefore not met.

Rule

MM-VCEP RUNX1 v3.1 BP2 (supporting): observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder, or observed in cis with a pathogenic variant in any inheritance pattern

Evidence considered

• MM-VCEP RUNX1 v3.1 (cspec) BP2 entry: applicable at supporting strength - 'Observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder or observed in cis with a pathogenic variant in any inheritance pattern.' MM-VCEP notes: biallelic pathogenic RUNX1 variants have never been reported in FPD/AML patients; a variant in trans with a known pathogenic variant or observed in the homozygous state in individuals without FPD/AML phenotype can be considered supporting benign evidence.
• ClinVar searches (RUNX1[GENE] AND c.707dup[VARNAME]; NM_001754.4:c.707dupT) returned no results - variant absent from ClinVar, so no ClinVar submissions document in-trans or in-cis observations with a pathogenic variant.
• gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0 all report the variant absent, so no homozygous carriers in population data exist that would support BP2.
• No case-observation literature for this variant was identified: literature_pass.json lists only three gene-level functional papers (PMID:15386419, PMID:15864279, PMID:17394134) with empty candidate_criteria, and targeted full-text checks confirmed none mention c.707dup, p.Met236IlefsTer25, or any duplication; no in-trans/in-cis proband observations could be assessed.

Still missing

• No parental/phase testing or family segregation data are available. If the variant were ever observed in trans with a known pathogenic RUNX1 variant (or homozygous in an individual without FPD/AML phenotype), BP2 would be met at supporting strength and should be re-adjudicated.

Sources

Reasoning

Governing framework is the ClinGen Myeloid Malignancy VCEP (MM-VCEP) specification for RUNX1 v3.1 (cspec), which defines PS2_Supporting as 1 proven de novo occurrence (both maternity and paternity confirmed) in a patient with FPD/AML phenotype and PS2_Moderate as >=2 such occurrences; confirmation of paternity only is explicitly insufficient, and 'no family history' requires absence of the variant and of all RUNX1-phenotypic criteria in first- and second-degree relatives. The case bundle contains no proband genotype, no parental testing results, and no family history assessment for c.707dupT. ClinVar has no submissions for this variant (variant absent), and none of the three retrieved publications (all functional-themed) mention c.707dup/c.707dupT/p.(Met236IlefsTer25) or report any germline de novo RUNX1 occurrence. No evidence exists to apply PS2 at any strength, so the criterion is not assessed rather than met.

Rule

PS2_Supporting: 1 proven de novo occurrence (both maternity and paternity confirmed) in a patient with FPD/AML phenotype. PS2_Moderate: >=2 proven de novo occurrences (both maternity and paternity confirmed) in patients with FPD/AML phenotype. Confirmation of paternity only is insufficient.

Evidence considered

• Governing framework: MM-VCEP RUNX1 v3.1 (cspec doc 135637580) specifies PS2_Supporting = 1 proven de novo occurrence (both maternity and paternity confirmed) in a patient with FPD/AML phenotype; PS2_Moderate = >=2 such occurrences; paternity-only confirmation is insufficient (egg donation, surrogacy, embryo-transfer errors can cause non-maternity).
• MM-VCEP PS2 requires that the proband meet at least one RUNX1-phenotypic criterion (mild-moderate thrombocytopenia with normal platelet size/volume in absence of other causes; platelet ultrastructural/functional defects; or hematologic malignancy such as AML/MDS/T-ALL) and that there be no family history (variant and RUNX1-phenotypic criteria absent in first- and second-degree relatives).
• ClinVar search (RUNX1[GENE] AND c.707dup[VARNAME] plus genomic HGVS terms) returned no result: 'This variant is absent from ClinVar', so no ClinVar-reported de novo occurrence exists.
• Literature screening: the only retrieved papers (PMID:15386419, PMID:15864279, PMID:17394134) are functional-themed (gastric cancer RUNX expression, RUNX gain/loss-of-function review, 8p11 myeloproliferative syndrome FGFR1 fusions); full-text search of all three found no mention of c.707dup/c.707dupT/p.(Met236IlefsTer25) and no germline de novo RUNX1 occurrence data.
• No proband or parental genotype data of any kind is present in review_manifest.json, case_summary.json, evidence.json, or prefetch.json for this case.

Still missing

• No proband-parent trio testing data for this variant (maternity and paternity confirmation required for PS2)
• No documented proven de novo occurrence of c.707dupT in any patient meeting >=1 RUNX1-phenotypic criterion
• No family history assessment showing absence of the variant and RUNX1-phenotypic criteria in first- and second-degree relatives
• No ClinVar submissions and no literature case report documenting de novo status of this variant

Sources

Reasoning

Governing framework is the ClinGen Myeloid Malignancy VCEP (MM-VCEP) RUNX1 specification v3.1 (cspec), which is applicable to PS3/BS3. MM-VCEP PS3 (Strong) requires validated transactivation assays demonstrating altered transactivation (<20% of wt and/or reduced to levels similar to well-established pathogenic variants such as R201Q or R166Q) AND data from a secondary assay demonstrating altered function; PS3_Moderate requires altered transactivation OR >=2 altered secondary assays; PS3_Supporting requires enhanced transactivation (>115% of wt). No variant-specific functional assay data of any kind exists for NM_001754.4:c.707dupT (p.(Met236IlefsTer25)): no transactivation assay, no secondary assay (DNA binding, CBFB heterodimerization, localization, colony-forming, xenotransplant), and no RNA-based functional studies in the case evidence bundle; the variant is absent from ClinVar and COSMIC. OncoKB provides a curated 'Likely Loss-of-function' biological-effect annotation for p.(Met236Ilefs*25), but this is a database curation label in a somatic oncology context, not a validated assay result, and a criterion cannot be inferred from another source's final classification. The three retrieved functional-themed publications (PMID:15386419, PMID:15864279, PMID:17394134) do not mention this variant in abstract or full text. Additionally, the MM-VCEP PS3 rule states PS3 is not applicable if the variant meets PVS1; this frameshift (p.(Met236IlefsTer25), exon 7) is a PVS1 candidate per the PVS1 assessment scaffold, so PS3 application is contingent on the PVS1 group's determination. Evidence is insufficient to assess PS3.

Rule

PS3 (Strong): Transactivation assays demonstrating altered transactivation (<20% of wt, and/or reduced to levels similar to well established pathogenic variants such as R201Q or R166Q) AND data from a secondary assay demonstrating altered function. Not applicable if variant meets PVS1. If variant meets PVS1_strong, upgrade to PVS1. PS3_Moderate: altered transactivation (<20% of wt and/or reduced to levels similar to R201Q/R166Q) OR >=2 secondary assays demonstrating altered function. PS3_Supporting: transactivation assays demonstrating enhanced transactivation (>115% of wt).

Evidence considered

• Governing framework: ClinGen Myeloid Malignancy VCEP (MM-VCEP) RUNX1 specification v3.1 (cspec); PS3 requires transactivation assays (<20% of wt and/or reduced to levels similar to well-established pathogenic variants such as R201Q or R166Q) AND a secondary assay showing altered function for Strong; altered transactivation OR >=2 secondary assays for Moderate; enhanced transactivation (>115% of wt) for Supporting.
• No variant-specific functional assay data found in the case evidence bundle for c.707dupT: no transactivation assay, no secondary assay, no RNA functional studies; variant absent from ClinVar and COSMIC.
• OncoKB entry for p.(Met236Ilefs*25) reviewed: curated biological-effect annotation 'Likely Loss-of-function' and curated oncogenicity 'Likely Oncogenic' in a somatic oncology context; treated as a database curation label, not a validated functional assay result, and not used as PS3 evidence (criteria must not be inferred from another source's final classification).
• Full texts of PMID:15386419, PMID:15864279, and PMID:17394134 searched for c.707dupT / p.Met236IlefsTer25; the variant is not mentioned, so none provide variant-specific functional evidence for PS3.
• MM-VCEP PS3 rule states PS3 is not applicable if the variant meets PVS1; c.707dupT is a frameshift predicted to truncate the protein (p.(Met236IlefsTer25), exon 7) and the PVS1 assessment scaffold suggests default PVS1 strength, so PS3 application is contingent on the PVS1 group's determination; PS3 left not_assessed.

Still missing

• No transactivation assay data for c.707dupT (required for PS3/PS3_Moderate/PS3_Supporting per MM-VCEP)
• No secondary functional assay data (DNA binding/EMSA, CBFB heterodimerization, subcellular localization, colony-forming potential, xenotransplantation)
• No RNA-based functional/splicing assay data for this variant
• PVS1 determination pending (per MM-VCEP, PS3 is not applicable if the variant meets PVS1)

Sources

Reasoning

Governing framework is the ClinGen Myeloid Malignancy VCEP RUNX1 v3.1 (cspec), which makes PS4 applicable but scored by proband count in a quasi-case-control model: PS4 (strong) requires >=4 probands, PS4_Moderate 2-3 probands, PS4_Supporting 1 proband, each meeting at least one RUNX1-phenotypic criterion (OR 127.1 / 63.5-95.3 / 31.8), with the variant absent from gnomAD or present only once. The specification explicitly states there is no published RUNX1 case-control study. For this variant (NM_001754.4:c.707dupT, p.Met236IlefsTer25) no proband observations were found in any consulted source: the variant is absent from ClinVar, absent from gnomAD v2.1/v4.1/gnomAD-Canada, absent from COSMIC, and none of the three available full texts in this case mention the variant. gnomAD absence satisfies only the population precondition, not the proband-count requirement. Evidence is therefore insufficient to score PS4 at any strength; the criterion is left not_assessed pending proband-level observation data.

Rule

PS4: >=4 probands meeting at least one RUNX1-phenotypic criterion (OR 127.1); PS4_Moderate: 2-3 probands (OR 63.5-95.3); PS4_Supporting: 1 proband (OR 31.8). MM-VCEP quasi-case-control: proband must meet the RUNX1-phenotypic criteria and variant must be absent from gnomAD or present only once.

Evidence considered

• cspec (MM-VCEP RUNX1 v3.1) defines PS4 as a quasi-case-control code: >=4 probands meeting at least one RUNX1-phenotypic criterion (OR 127.1) = strong; 2-3 probands (OR 63.5-95.3) = moderate; 1 proband (OR 31.8) = supporting; the proband must meet the RUNX1-phenotypic criteria and the variant must be absent from gnomAD or present only once. The spec notes there is currently no published RUNX1 case-control study.
• No proband observations for c.707dupT were found: the variant is absent from ClinVar (evidence.json clinvar search_status=no_result) and absent from gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0.
• The three full texts available in this case (PMID:15386419, PMID:15864279, PMID:17394134) were screened for the variant (c.707dupT / p.Met236IlefsTer25 / M236); none mention it, so no FPD/AML proband reports were available from the literature.
• OncoKB curates M236Ifs*25 as 'Likely Oncogenic' with biological effect 'Likely Loss-of-function' in a somatic cancer context; this is neither case-control enrichment nor a RUNX1-phenotypic proband observation and was not used for PS4.

Still missing

• No probands meeting the RUNX1-phenotypic criteria were identified for c.707dupT in any source in this case (ClinVar absent, no case reports/case series found; the three screened full texts do not mention the variant).
• No published RUNX1 case-control study exists (per MM-VCEP note), so case-control enrichment data are unavailable for this variant.

Sources

Reasoning

Under the governing MM-VCEP RUNX1 v3.1 specification (cspec), PM6_Supporting requires 2-3 assumed de novo occurrences (without confirmation of maternity and paternity) and PM6 requires >=4 such occurrences, in patients with FPD/AML phenotype; the combined PS2+PM6 contribution is capped at 1 point (one moderate or two supporting rules). The case bundle contains no assumed de novo occurrences: there is no parental testing data, no ClinVar submissions for this variant, and no retrieved publication reporting c.707dupT in a proband with unconfirmed parental status. PM6 is therefore not assessed due to absent evidence.

Rule

PM6_Supporting: 2 or 3 assumed de novo occurrences (without confirmation of maternity and paternity) in patients with FPD/AML phenotype. PM6: >=4 assumed de novo occurrences. Combined PS2+PM6 maximum allowable value is 1 point (one moderate or two supporting rules).

Evidence considered

• Governing framework: MM-VCEP RUNX1 v3.1 (cspec doc 135637580) specifies PM6_Supporting = 2 or 3 assumed de novo occurrences (without confirmation of maternity and paternity) in patients with FPD/AML phenotype; PM6 = >=4 such occurrences; maximum allowable combined PS2+PM6 value is 1 point (one moderate or two supporting rules).
• MM-VCEP PM6 requires that probands meet at least one RUNX1-phenotypic criterion and that no family history be present (absent variant and phenotypic criteria in first/second-degree relatives), mirroring PS2 requirements.
• ClinVar has no submissions for this variant ('This variant is absent from ClinVar'), so no assumed de novo occurrence is reported in ClinVar.
• Full-text screening of the three retrieved publications (PMID:15386419, PMID:15864279, PMID:17394134) found no mention of this variant and no assumed de novo occurrence data; the only 'de novo' text in PMID:15864279 refers to de novo childhood AML somatic FLT3 mutations and RUNX3 expression, not germline RUNX1 variant occurrence.
• No parental testing or proband inheritance data exists anywhere in the case bundle.

Still missing

• No assumed de novo occurrences (2-3 needed for PM6_Supporting; >=4 for PM6) documented for c.707dupT in FPD/AML-phenotype patients
• No case reports in the retrieved literature or ClinVar describing this variant in a proband without parental confirmation
• No proband-level data (phenotype and inheritance testing results) present in the case bundle

Sources

Reasoning

Under the governing MM-VCEP RUNX1 v3.1 specification (cspec), PP1 (supporting) requires 3-4 meioses of cosegregation, PP1_Moderate 5-6, and PP1_Strong >=7, observed within one or across multiple families; only genotype- and phenotype-positive individuals and obligate carriers are counted, and affected individuals must meet at least one RUNX1-phenotypic criterion. The case bundle contains no pedigree, no segregation data, and no multi-generation family testing results for c.707dupT. ClinVar has no submissions for this variant and the retrieved literature contains no segregation data. PP1 is therefore not assessed due to absent evidence.

Rule

PP1: 3 or 4 meioses observed within one or across multiple families. PP1_Moderate: 5 or 6 meioses. PP1_Strong: >=7 meioses. Only genotype- and phenotype-positive individuals and obligate carriers are counted.

Evidence considered

• Governing framework: MM-VCEP RUNX1 v3.1 (cspec doc 135637580) specifies PP1 = 3 or 4 meioses, PP1_Moderate = 5 or 6 meioses, PP1_Strong = >=7 meioses observed within one or across multiple families; only genotype- and phenotype-positive individuals and obligate carriers are counted; affected individuals must show at least one RUNX1-phenotypic criterion; the VCEP waived the requirement to demonstrate cosegregation in more than one family.
• ClinVar has no submissions for this variant ('This variant is absent from ClinVar'), so no ClinVar-reported segregation observations exist.
• Literature_pass.json criteria_index maps the three retrieved PMIDs (15386419, 15864279, 17394134) to BS3/PM1/PS3 only - no paper was triaged to PP1 or segregation, and full-text screening confirmed no segregation or meiosis data and no mention of this variant.
• No family/pedigree information of any kind is present in the case bundle.

Still missing

• No pedigree or segregation data available for c.707dupT
• No informative meioses counted (3-4 required for PP1; 5-6 for PP1_Moderate; >=7 for PP1_Strong)
• No multi-generational family testing results in the case bundle, ClinVar, or retrieved literature
• No affected relatives genotyped for this variant who meet RUNX1-phenotypic criteria

Sources

Reasoning

Governing framework is the ClinGen Myeloid Malignancy VCEP (MM-VCEP) RUNX1 specification v3.1 (cspec). BS3 (Strong) requires transactivation assays demonstrating normal transactivation (80-115% of wt) AND data from a secondary assay demonstrating normal function; BS3_Supporting requires normal transactivation (80-115% of wt) alone. No functional studies exist for NM_001754.4:c.707dupT demonstrating normal function: there is no transactivation assay result, no secondary assay result, and no RNA functional data in the case evidence bundle. OncoKB offers only a curated 'Likely Loss-of-function'/oncogenic annotation (somatic context), which is not normal-function assay evidence. The predicted protein consequence of this duplication is a frameshift truncation (p.(Met236IlefsTer25); wild-type RUNX1c is 481 aa), which is biologically inconsistent with the 'no damaging effect' premise of BS3, but no assay data is available to formally assess the criterion. Evidence is insufficient to assess BS3.

Rule

BS3 (Strong): Transactivation assays demonstrating normal transactivation (80-115% of wt) AND data from a secondary assay demonstrating normal function. BS3_Supporting: Transactivation assays demonstrating normal transactivation (80-115% of wt).

Evidence considered

• Governing framework: MM-VCEP RUNX1 v3.1 (cspec); BS3 requires transactivation assays demonstrating normal transactivation (80-115% of wt) AND a secondary assay demonstrating normal function for Strong; BS3_Supporting requires normal transactivation (80-115% of wt) alone.
• No functional studies demonstrating normal function exist for c.707dupT in the case evidence bundle (no transactivation assay, no secondary assay, no RNA functional data).
• OncoKB reviewed: provides only a curated 'Likely Loss-of-function'/'Likely Oncogenic' annotation in a somatic context; not normal-function assay evidence and not used for BS3.
• Predicted protein consequence is a frameshift truncation (p.(Met236IlefsTer25); wild-type RUNX1c 481 aa), inconsistent with the normal-function premise of BS3; criterion not assessable without assay data.
• Full texts of PMID:15386419, PMID:15864279, and PMID:17394134 searched for c.707dupT / p.Met236IlefsTer25; variant not mentioned, so no normal-function evidence from these papers.

Still missing

• No transactivation assay data demonstrating normal transactivation (80-115% of wt) for c.707dupT
• No secondary assay data demonstrating normal function
• Predicted frameshift truncation (p.(Met236IlefsTer25)) makes a normal-function assay result biologically implausible; BS3 would require actual assay data to the contrary

Sources

Reasoning

Under the governing MM-VCEP RUNX1 v3.1 specification (cspec), BS4 (benign strong) applies when the variant is observed in >=2 informative meioses showing lack of segregation, and per MM-VCEP notes it should only be applied for genotype-negative, phenotype-positive family members. No affected family members have been genotyped for c.707dupT and no informative meioses demonstrating non-segregation are reported anywhere in the case bundle, in ClinVar (variant absent), or in the retrieved literature. BS4 is therefore not assessed due to absent evidence.

Rule

BS4: Applicable when observed in >=2 informative meioses. Per MM-VCEP notes, this code should only be applied for genotype-negative, phenotype-positive family members.

Evidence considered

• Governing framework: MM-VCEP RUNX1 v3.1 (cspec doc 135637580) specifies BS4 as 'Applicable when observed in >=2 informative meioses', with the MM-VCEP note that the code should only be applied for genotype-negative, phenotype-positive family members.
• ClinVar has no submissions for this variant ('This variant is absent from ClinVar'), so no non-segregation observations are reported.
• Full-text screening of the three retrieved publications (PMID:15386419, PMID:15864279, PMID:17394134) found no family segregation or non-segregation data and no mention of this variant.
• No family testing results of any kind are present in the case bundle.

Still missing

• No genotyped affected family members (genotype-negative, phenotype-positive) for this variant
• No informative meioses demonstrating lack of segregation (>=2 required for BS4)
• No family testing data in the case bundle, ClinVar, or retrieved literature

Sources