NF1 c.4836-2A>G, p.?

Reasoning

NM_001042492.2:c.4836-2A>G disrupts the canonical -2 position of the intron 36 splice-acceptor consensus (variant_exonic_positions marks start/end exon as '36i', i.e. intronic, flanking exon 36). The NF1 CSPEC (ClinGen Neurofibromatosis and Schwannomatosis EP, doc 1553923915) does not provide a structured PVS1 decision-tree payload for this case (final_classification_framework.framework_complete=false, rule_payload={}), so the generic ClinGen SVI PVS1 recommendations (PMC6185798) were applied per pvs1_variant_assessment, which explicitly classifies this variant as consequence_class='canonical_splice' with canonical_splice_consensus=true and flags suggested_default_strength='PVS1'. Germline loss-of-function is an established NF1 disease mechanism: NF1 causes autosomal dominant neurofibromatosis type 1 via haploinsufficiency/loss of neurofibromin GAP activity, and pvs1_gene_context confirms germline LoF mechanism support (lof_mechanism_supported=true, pvs1_gene_gate='eligible') from targeted literature review. SpliceAI strongly supports a splice-altering effect at this position (DS_AL=0.989, DS_AG=0.468, max_delta_score=0.989), consistent with acceptor-site disruption, though this splice-effect prediction evidence is used here only to corroborate the canonical-consensus classification (not stacked as separate PP3 evidence per the downgrade_considerations note in pvs1_variant_assessment). Under the PVS1 canonical-splice decision path, disruption of the -1/-2 acceptor dinucleotide is presumed to abolish or significantly alter normal splicing at exon 36, a coding exon roughly mid-transcript position (c.4836 of the NM_001042492.2 coding sequence), consistent with either exon skipping producing a frameshift or intron retention producing a premature termination codon predicted to trigger nonsense-mediated decay (NMD) rather than escape into the final exon or last 50 bp of the penultimate exon. No case-specific minigene/RT-PCR functional data confirming the exact splicing outcome (exon skip vs. intron retention, in-frame vs. frameshift, exact PTC position relative to the NMD boundary) were found in this evidence bundle, and no repeat-region or alternative-transcript-rescue context was documented that would argue for a weaker strength. Applying full-strength PVS1 (very strong) is consistent with the deterministic suggested_default_strength from pvs1_variant_assessment and generic ClinGen SVI guidance for canonical +/-1,2 splice-site variants in a gene with established LoF mechanism, pending confirmation of exact splicing/NMD outcome.

Rule

canonical_splice_generic_pvs1_PMC6185798

Evidence considered

• pvs1_variant_assessment.json classifies this variant as consequence_class='canonical_splice', canonical_splice_consensus=true, variant_bucket='canonical_splice', and states 'This variant affects a canonical +/-1,2 splice consensus position. Under PMC6185798, canonical splice variants are evaluated with the generic PVS1 framework once germline loss of function is established for the gene...' with suggested_default_strength='PVS1'.
• prefetch.json variant_exonic_positions for NC_000017.10/NC_000017.11/NG_009018.1 all report start_exon='36i' and end_exon='36i', confirming the variant lies in the intron flanking exon 36 (consistent with a canonical acceptor-site position, c.4836-2).
• pvs1_gene_context.json states lof_mechanism_supported=true and pvs1_gene_gate='eligible', with mechanism_rationale: 'Targeted germline literature review identified disease-focused publications supporting NF1 loss of function as a germline disease mechanism, so generic PVS1 framework assessment is eligible.'
• final_classification_framework (cspec_ruleset) for NF1 has framework_complete=false and empty rule_payload, confirming no structured VCEP PVS1 decision tree is available for this case, justifying fallback to the generic ClinGen SVI PVS1 framework (PMC6185798).
• SpliceAI evidence_sentence: 'SpliceAI predicts possible splice impact for this variant (max delta score = 0.99)' with source_data scores DS_AG=0.468, DS_AL=0.989, DS_DG=0.0, DS_DL=0.195, corroborating acceptor-site disruption at this canonical position; used only to support the canonical-consensus classification per the pvs1_variant_assessment instruction not to double-count this same prediction as separate PP3 evidence.
• No case-specific functional splicing assay (RT-PCR/minigene) or documented total-exon-count/NMD-boundary mapping for exon 36 was found in the evidence bundle, literature_pass.json, or the retrieved full-text publications, so the NMD/frameshift outcome is inferred from canonical-splice-consensus disruption rather than confirmed empirically.

Still missing

• No RT-PCR/minigene splicing assay for this specific variant was found in the retrieved evidence to confirm whether the acceptor-site disruption produces exon 36 skipping (and whether that skip is in-frame) versus intron 36 retention with a premature termination codon.
• Exact total exon count/position of exon 36 relative to the final exon boundary and the NMD-relevant last-50bp-of-penultimate-exon rule was not explicitly documented in the bundle, so the NMD-competent assumption, while consistent with the mid-transcript cDNA position (c.4836), was not independently exon-mapped against a documented total exon count.
• No confirmation that exon 36 lies outside any characterized non-critical or population-variation-enriched region of NF1.

Sources

Reasoning

PM2 is met at moderate strength under the generic ACMG/AMP framework. The exact variant is absent from gnomAD v2.1 and v4.1, independent releases queried using matched GRCh37 and GRCh38 genomic representations. No NF1-specific PM2 modification was available in the retrieved CSPEC material.

Rule

Generic ACMG/AMP (Richards et al., 2015): PM2 is moderate evidence when a variant is absent from a large general population or race-matched control cohort; adequate site coverage should be confirmed.

Evidence considered

• gnomAD v2.1 (GRCh37 17-29652836-A-G), gnomAD v4.1 (GRCh38 chr17-31325818-A-G), and gnomAD-Canada v1.0 each report the exact variant as absent.
• The retrieved NF1 CSPEC material has no population-criterion rule payload, so no applicable NF1-specific PM2 strength modification was available.
• Richards et al. state that absence from a large general population or control cohort can be considered moderate pathogenic evidence, while noting that database read depth should be confirmed.

Still missing

• The retrieved gnomAD summaries do not provide per-ancestry allele numbers or site-level coverage metrics; the underlying gnomAD records should be reviewed if a coverage audit is required.

Sources

Reasoning

PM6_supporting is met as assumed de novo evidence only. The exact ClinVar record labels one clinical-testing submission as de novo, but parental results and parentage confirmation are absent, precluding PS2 and requiring review of the underlying laboratory record.

Rule

Generic ACMG/AMP PM6: assumed de novo occurrence without confirmation of maternity and paternity; applied at supporting strength only.

Evidence considered

• ClinVar variation 2573197 exactly matches NF1 c.4836-2A>G. SCV004015019, a single-submitter clinical-testing record, lists origin as de novo; no parental genotype or parentage-confirmation data are supplied in the case bundle.
• The retrieved NF1 CSPEC record has no criterion-specific content for PM6; generic ACMG/AMP PM6 logic was used for assumed, unconfirmed de novo occurrence.

Still missing

• Obtain both parental test results and biological parentage confirmation to determine whether PS2 can replace PM6.
• Obtain proband phenotype details to confirm clinical consistency with NF1.
• Confirm that the ClinVar origin field reflects testing of this exact variant in the proband rather than an unverified submitter annotation.

Sources

Reasoning

The exact ClinVar record has no expert-panel submission. Its pathogenic and likely pathogenic assertions are from ordinary clinical-laboratory submitters and cannot support PP5 under the required exact-variant expert-panel rule.

Rule

PP5 is permitted only for an exact-variant ClinVar expert-panel Pathogenic or Likely pathogenic classification.

Evidence considered

• ClinVar variation 2573197 exactly matches NM_001042492.3(NF1):c.4836-2A>G and contains four single-submitter clinical-testing records; the submission audit reports zero expert-panel submissions.

Still missing

• An exact-variant ClinVar Pathogenic or Likely pathogenic assertion reviewed by an expert panel would be required.

Sources

Reasoning

BA1 is not met. The variant is absent from gnomAD v2.1 and v4.1; therefore, its observed allele frequency is 0 and cannot meet the generic ACMG/AMP stand-alone BA1 threshold (>5%).

Rule

Generic ACMG/AMP (Richards et al., 2015): BA1 requires a population allele frequency >5%.

Evidence considered

• gnomAD v2.1 (GRCh37 17-29652836-A-G) and gnomAD v4.1 (GRCh38 chr17-31325818-A-G) each report the exact variant as absent.
• Richards et al. state that a control-population allele frequency over 5% is stand-alone benign evidence (BA1).

Sources

Reasoning

BS1 is not met. The exact variant is absent in gnomAD v2.1 and v4.1, so no allele frequency exceeds that expected for NF1. No applicable NF1-specific BS1 threshold was retrieved from the available CSPEC materials.

Rule

Generic ACMG/AMP (Richards et al., 2015): BS1 requires an allele frequency greater than expected for the disorder.

Evidence considered

• The retrieved NF1 CSPEC record contained no population-criterion rule payload or local VCEP materials.
• gnomAD v2.1 and gnomAD v4.1 each report the exact variant as absent.
• Richards et al. define BS1 as a control-population allele frequency greater than expected for the disorder.

Still missing

• No NF1-specific BS1 numerical threshold was available in the retrieved CSPEC material; this does not affect the conclusion because the observed frequency is 0 in both queried gnomAD releases.

Sources

Reasoning

BS2 is not met. There are no reported population observations of the variant, and thus no well-documented healthy adult heterozygote or homozygote whose observation could be evaluated against NF1 age of onset and penetrance.

Rule

Generic ACMG/AMP (Richards et al., 2015): BS2 requires observation in a well-documented healthy adult that is inconsistent with disease penetrance.

Evidence considered

• gnomAD v2.1 and gnomAD v4.1 each report the exact variant as absent; consequently, neither dataset supplies a carrier, homozygote, or healthy-adult observation for BS2.
• Richards et al. describe BS2 as requiring observation in a well-documented healthy adult when disease penetrance and age of onset make that observation inconsistent with disease.

Still missing

• No healthy adult carrier observation with phenotype, age, zygosity, and ancestry information is available.

Sources

Reasoning

SpliceAI predicts a strong splice-altering effect (max delta score 0.989, DS_AL=0.989) for this canonical acceptor-site variant, which is the opposite direction of a BP4 benign-supporting computational prediction. There is no in-silico signal here that supports a benign interpretation. The missense in-silico pathway (BayesDel/REVEL/aGVGD) is not applicable because this is a splice-site variant with no missense protein consequence (protein predicted as p.?). Separately, the only missense-style predictor score present in the bundle, BayesDel=0.235399, is disregarded for calibration purposes per this pipeline's evidence standard: no verified published threshold/PMID is available for BayesDel, so it cannot be used to support or refute BP4 regardless of its numeric value. REVEL was not found (found=false) so it does not factor in either.

Rule

generic_acmg_bp4_spliceai_no_benign_signal

Evidence considered

• SpliceAI (spliceai) max delta score = 0.989 (DS_AL acceptor-loss = 0.989), indicating a damaging splice-altering prediction, not a benign one.
• BayesDel score 0.235399 is present (source: local BayesDel_170824_noAF chr17 lookup) but is treated as insufficiently calibrated for PP3/BP4 because no verified published threshold/PMID for this BayesDel build is available to this pipeline; the score is not used to support BP4.
• REVEL score not available (found=false), so no REVEL-based benign signal exists to support BP4.
• Variant has no missense protein consequence (NP_001035957.1:p.?), so the missense in-silico BP4 pathway does not apply; assessment rests on the splice-impact pathway only, which shows a damaging (non-benign) prediction.

Sources

Reasoning

The exact ClinVar record has no expert-panel submission and no Benign/Likely benign expert-panel assertion. BP6 is therefore not met.

Rule

BP6 is permitted only for an exact-variant ClinVar expert-panel Benign or Likely benign classification.

Evidence considered

• ClinVar variation 2573197 exactly matches NM_001042492.3(NF1):c.4836-2A>G; its four listed submissions are single-submitter clinical-laboratory assertions, and the submission audit reports zero expert-panel submissions.

Still missing

• An exact-variant ClinVar Benign or Likely benign assertion reviewed by an expert panel would be required.

Sources

Reasoning

The exact ClinVar record includes one submission with origin recorded as de novo, but the case bundle contains no parental genotypes, confirmation of maternity and paternity, phenotype details, or laboratory documentation. Confirmed de novo evidence required for PS2 cannot be established.

Rule

Generic ACMG/AMP PS2: confirmed de novo occurrence requires parental testing with maternity and paternity confirmed; not assessable from the available record.

Evidence considered

• ClinVar variation 2573197 is an exact match for NF1 c.4836-2A>G. One single-submitter clinical-testing record (SCV004015019) lists origin as de novo, but supplies no parental genotypes, parentage confirmation, phenotype, or case-level laboratory evidence in the bundle.
• The retrieved NF1 CSPEC record has no criterion-specific content for PS2; generic ACMG/AMP de novo logic was therefore used.

Still missing

• Obtain both parental test results for c.4836-2A>G.
• Obtain documentation confirming biological maternity and paternity.
• Obtain the proband phenotype and its consistency with NF1.

Sources

Reasoning

No validated functional assay data (e.g. RT-PCR/RNA analysis, minigene splicing assay, or protein-function assay) for NM_001042492.2:c.4836-2A>G is present in this case's evidence bundle. The NF1 CSPEC (cspec) rule_payload retrieved for this case is empty/unstructured (framework_complete=false, no PS3-specific criteria text), so no VCEP-specific PS3 rule could be applied; generic ACMG/AMP PS3 (Richards et al. 2015, PMID:25741868) requires well-established functional studies showing a damaging effect, which are absent here. literature_pass.json flags PMID 28492532 and PMID 33210232 as candidate PS3 sources, but neither is present in source_registry.json or fulltext_map.json (no full text retrieved, and 28492532's title/metadata could not even be resolved), so their content cannot be verified as containing an actual functional assay for this variant and they cannot be cited. SpliceAI in-silico prediction (max delta 0.989) is a computational splice prediction, not laboratory functional assay evidence, and is out of scope for PS3.

Still missing

• No RNA/RT-PCR, minigene splicing assay, or protein functional assay data retrieved for this specific variant.
• Candidate PS3 papers (PMID 28492532, PMID 33210232) were flagged by literature triage but were never fetched into source_registry/fulltext_map for this case, so grounding to the exact variant could not be confirmed.
• NF1 CSPEC framework retrieved for this case has an empty rule_payload; no VCEP-specific PS3 assay-strength calibration is available.

Reasoning

No case-control study, affected-case count with an appropriate control comparison, or enrichment statistic for the exact NF1 variant was provided. PS4 cannot be assessed from the ClinVar submission count or aggregate classification.

Rule

Generic ACMG/AMP PS4: significant enrichment of the exact variant in affected individuals compared with controls is required.

Evidence considered

• The NF1 CSPEC record is present but its retrieved rule payload contains no criterion-specific PS4 guidance; generic ACMG/AMP PS4 evidence requirements therefore apply.
• ClinVar records four clinical-laboratory submissions for the exact variant, but provides no case-control comparison, odds ratio, or prevalence analysis.

Still missing

• Variant-specific affected and control allele counts or a validated case-control enrichment analysis are required.

Sources

Reasoning

No informative relatives, genotypes, phenotypes, meioses, or cosegregation data are present in the case bundle. PP1 cannot be evaluated.

Rule

Generic ACMG/AMP PP1: requires cosegregation of the variant with disease in informative affected relatives; no such data are available.

Evidence considered

• The consolidated case-evidence bundle contains no family segregation or meiosis data for this variant.

Still missing

• Provide pedigree-based segregation data with each relative's genotype and NF1 phenotype.
• Provide the number of informative meioses and any age-dependent penetrance considerations.

Reasoning

No individual-level phenotype for the tested person or variant-bearing affected cohort was provided. NF1 is associated with neurofibromatosis type 1, but gene-level disease association alone does not establish a highly specific phenotype for this individual.

Rule

Generic ACMG/AMP PP4: the patient's phenotype or family history must be highly specific for a disease with a single genetic etiology.

Evidence considered

• The retrieved NF1 CSPEC record identifies neurofibromatosis type 1 as the associated disease but contains no usable PP4 specification.
• ClinVar lists conditions associated with submissions but does not provide the tested individual's phenotype or demonstrate phenotype specificity.

Still missing

• Clinical phenotype and diagnostic evaluation for the variant-bearing individual, with assessment of phenotype specificity for NF1, are required.

Sources

Reasoning

No functional assay data of any kind (damaging or non-damaging) for NM_001042492.2:c.4836-2A>G is present in the case evidence bundle, so there is no basis to assert a well-established functional study showing no damaging effect (generic ACMG/AMP BS3, PMID:25741868). The NF1 CSPEC retrieved for this case has an empty rule_payload, so no VCEP-specific BS3 assay calibration exists to apply. ClinVar classification for this variant is Pathogenic (3 labs) / Likely pathogenic (1 lab), consistent with absence of benign functional evidence, but final classifications from other sources were not used to infer this criterion.

Still missing

• No functional/splicing assay data of any kind was retrieved for this variant, so a benign functional signal cannot be assessed one way or the other.
• NF1 CSPEC framework retrieved for this case has an empty rule_payload; no VCEP-specific BS3 rule is available.

Reasoning

No informative non-segregation evidence is present. The bundle does not document unaffected carriers, affected non-carriers, or phenotype-genotype discordance within a family.

Rule

Generic ACMG/AMP BS4: requires lack of segregation in affected family members; no informative family data are available.

Evidence considered

• The consolidated case-evidence bundle contains no familial non-segregation observations for this variant.

Still missing

• Provide pedigree and testing data documenting informative non-segregation, including relatives' phenotypes, ages, and genotypes.
• Assess age-dependent penetrance and variable expressivity before using any unaffected carrier as benign evidence.

Reasoning

No affected-proband co-occurrence, second pathogenic variant, or phase result is present in the consolidated evidence bundle. Thus there is no observation of c.4836-2A>G in trans with a pathogenic variant that could support BP2 in the autosomal-dominant NF1 context.

Rule

Generic ACMG/AMP BP2 requires an observed trans co-occurrence with a pathogenic variant in the dominant-disease context; no qualifying co-occurrence or phase evidence is available.

Evidence considered

• The retrieved NF1 ClinGen Neurofibromatosis and Schwannomatosis specification identifies neurofibromatosis type 1 as autosomal dominant and supplies no criterion-specific BP2 rule.
• The consolidated bundle contains no individual-level report of a second pathogenic variant or phase relationship for c.4836-2A>G.
• The generic ACMG/AMP framework states that, in dominant disorders, a variant found in trans with a pathogenic variant can support BP2; no such observation is available here.

Still missing

• A documented co-occurring pathogenic NF1 variant in the same individual and validated cis/trans phase are required to assess BP2.

Sources

Reasoning

No clinical record documents an alternate molecular diagnosis that fully explains the relevant phenotype in a person carrying this NF1 variant.

Rule

Generic ACMG/AMP BP5: a variant is found in a case with an alternate molecular basis for disease.

Evidence considered

• The case bundle provides variant-level database evidence but no patient-level alternate diagnosis or phenotype-explanation analysis.
• The retrieved NF1 CSPEC record contains no usable BP5-specific rule.

Still missing

• Individual-level phenotype and testing results establishing an alternate diagnosis that fully explains the phenotype are required.

Sources