ATRX c.4699+2T>C, p.?

Reasoning

No ATRX CSPEC/VCEP exists (cspec found=false, gene_vcep_dir=null), so the generic ACMG/AMP 2015 PVS1 rule applies, implemented per the ClinGen SVI PVS1 recommendations (pvs1_generic_framework, PMC6185798) as directed by pvs1_variant_assessment (apply_generic_pvs1_framework=true, variant_bucket='canonical_splice'). Gene gate: germline loss of function is an established ATRX disease mechanism (pvs1_gene_context: lof_mechanism_supported=true, pvs1_gene_gate='eligible'; ATRX germline LoF causes ATR-X syndrome, X-linked). Variant consequence: NM_000489.5:c.4699+2T>C alters the +2 position of the canonical donor splice consensus of intron 16 (VariantValidator variant_exonic_positions '16i'; canonical_splice_consensus=true), satisfying the PVS1 null-variant class 'canonical +/-1/2 splice site'. SpliceAI strongly predicts donor loss (max delta 0.99; DS_DL 0.99, DS_DG 0.82), supporting a splicing-mediated null consequence (exon 16 skipping or intron 16 retention with frameshift/premature termination). NMD: the variant lies in intron 16 of the 36-exon NM_000489.5 transcript, far upstream of the terminal exon-exon junction, so any resulting premature termination codon is predicted to undergo NMD; the NMD-escape downgrade (last exon / last 50 bp of penultimate exon) does not apply. Exon importance/population LoF: the variant is absent from gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0, providing no signal of population LoF tolerance at this position; no evidence identifies exon 16 as a non-critical or alternative-splicing-rescued exon. Protein length change: the disruption occurs at ~amino acid 1567 of the 2492-residue NP_000480.3 protein, truncating approximately the C-terminal third of the protein (loss of >900 residues) rather than a small distal fragment. Repeat-region context: no repeat-region annotation for ATRX exon 16/intron 16 was available in the case bundle; no evidence of a benign repetitive region. None of the PMC6185798 downgrade triggers (NMD escape, non-canonical transcript, non-critical exon, population LoF enrichment, distal non-critical region) is met, so full PVS1 (Very Strong) is applied. Per the SVI framework note in pvs1_variant_assessment, the same splice-prediction evidence must not be double-counted as PP3.

Rule

PVS1 (Very Strong): null variant (nonsense, frameshift, canonical +/-1/2 splice site, initiation codon, single/multi-exon deletion) in a gene where loss of function is a known mechanism of disease (ACMG/AMP 2015, PMID:25741868), applied per ClinGen SVI PVS1 recommendations (PMC6185798) as generic fallback because no ATRX CSPEC/VCEP exists

Evidence considered

• pvs1_variant_assessment: variant classified canonical_splice, canonical_splice_consensus=true, variant_bucket='canonical_splice', apply_generic_pvs1_framework=true, suggested_default_strength='PVS1'
• pvs1_gene_context: germline LoF mechanism supported for ATRX (ATR-X syndrome), pvs1_gene_gate='eligible'
• VariantValidator normalization (prefetch.json): c.4699+2T>C maps to intron 16 ('16i') of the 36-exon NM_000489.5 transcript; RefSeq Select transcript (refseq_select=true)
• spliceai: max delta score 0.99 with DS_DL 0.99 (donor loss) and DS_DG 0.82 (donor gain) - strong predicted splice disruption at the canonical donor
• gnomad_v2, gnomad_v4, gnomad_canada: variant absent - no population LoF tolerance signal
• Exon position analysis: intron 16 donor is well upstream of the terminal exon, so NMD is predicted; truncation would remove ~37% of the 2492-residue protein (NP_000480.3)

Still missing

• No RNA, minigene, or other functional assay evidence exists for this exact variant confirming the actual splicing outcome (exon 16 skipping vs intron 16 retention) or NMD; the null consequence is inferred from the canonical +2 donor position and SpliceAI (max delta 0.99).
• No ClinVar record, published case, or population observation for c.4699+2T>C is available for cross-validation.

Sources

Reasoning

PM2 (Richards et al. 2015, PMID:25741868) is met when the variant is absent or at extremely low frequency in population databases; under the generic ACMG/AMP 2015 fallback, the non-VCEP gnomAD PM2 threshold is <0.1%. NM_000489.5:c.4699+2T>C is absent from gnomAD v2.1 exomes, gnomAD v4.1 exomes, and gnomAD-Canada v1.0 genomes (AF=0, well below 0.1%), in an X-linked gene (ATRX) with an established germline loss-of-function disease mechanism (ATR-X syndrome; pvs1_gene_context lof_mechanism_supported=true), and no hemizygotes were observed. Absence across three independent population datasets makes a sequencing artifact unlikely. PM2 is applied at supporting strength per current ClinGen SVI population-criteria convention, consistent with the supporting-strength application used for the other evidence criteria in this case.

Rule

generic_acmg_2015_pm2: af_lt_0.1%_non_vcep_met_absent

Evidence considered

• gnomAD v2.1 (GRCh37 X-76891404-A-G): variant absent (AF effectively 0, <0.1% threshold)
• gnomAD v4.1 (GRCh38 chrX-77635913-A-G): variant absent (AF effectively 0, <0.1% threshold)
• gnomAD-Canada v1.0: AC=0, AN=0, AF=0.0
• ATRX germline LoF mechanism supported for ATR-X syndrome (pvs1_gene_context: lof_mechanism_supported=true), an X-linked disorder
• Generic ACMG/AMP 2015 fallback with non-VCEP PM2 threshold <0.1%

Sources

Reasoning

PP3 (Richards et al. 2015, PMID:25741868) is met when multiple lines of computational evidence support a deleterious effect on the gene or gene product (including splicing impact). No ATRX CSPEC/VCEP or gene-specific PP3/BP4 lookup spreadsheet exists (vcep_materials.json: 'no cspec url available', files_found=[], summaries={}; final_classification_framework.json: precedence_rank=3, generic_acmg_fallback), so generic ACMG/AMP 2015 governs and would cap PP3 at supporting strength. The splice-impact line is decisive in direction: SpliceAI predicts near-certain splice disruption (max delta score = 0.99; DS_DL 0.99 donor loss at position 2, DS_DG 0.82 donor gain, DS_AG 0.0, DS_AL 0.04). However, NM_000489.5:c.4699+2T>C is a variant at a canonical +2 splice-donor consensus position (consequence_class=canonical_splice, canonical_splice_consensus=true). Under the ClinGen SVI generic PVS1 framework (PMC6185798, source_registry key pvs1_generic_framework), canonical +/-1,2 splice variants are evaluated with PVS1 once germline loss of function is established, and the case's pvs1_variant_assessment (suggested_default_strength=PVS1) explicitly directs: 'For canonical splice variants, do not stack PP3 for the same splice-effect prediction evidence.' The SpliceAI prediction is exactly the same splice-effect evidence that underlies the PVS1 assignment; awarding PP3 here would double-count that prediction. BayesDel noAF = 0.83 is in the pathogenic direction but is a missense-calibrated predictor whose thresholds are flagged PLACEHOLDER/unverified in the pipeline SVCv4 insilico ladder (ACMG_V4/svc4/insilico.py; BayesDel is not the production-verified tool), and it is not an appropriate independent line for a canonical splice variant; in any event it would only duplicate the same deleterious prediction already captured by PVS1. PP3 is therefore not awarded to avoid double-counting the same splice-effect prediction.

Rule

Generic ACMG/AMP 2015 PP3 (Richards et al., PMID:25741868; source_registry key generic_acmg_combination_rules): 'Multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.).' Supporting-strength ceiling under the generic framework. PP3 withheld per ClinGen SVI PVS1 guidance (PMC6185798, source_registry key pvs1_generic_framework) applied through pvs1_variant_assessment: for canonical +/-1,2 splice variants, the same splice-effect prediction evidence used for PVS1 must not be double-counted as PP3.

Evidence considered

• SpliceAI Lookup max delta score 0.99 (DS_DL 0.99 donor loss at +2, DS_DG 0.82 donor gain, DS_AG 0.0, DS_AL 0.04; variant_consequence=sequence_variant; source_registry key spliceai) - predicts near-certain disruption of the canonical donor splice site; this is the same splice-effect prediction captured by the PVS1 framework for this canonical +2 variant
• pvs1_variant_assessment (case artifact): consequence_class=canonical_splice, canonical_splice_consensus=true, apply_generic_pvs1_framework=true, suggested_default_strength=PVS1, downgrade consideration 'For canonical splice variants, do not stack PP3 for the same splice-effect prediction evidence' - the framework basis for withholding PP3
• pvs1_generic_framework (ClinGen SVI PVS1 recommendations, PMC6185798, source_registry key pvs1_generic_framework) - canonical +/-1,2 splice variants are evaluated with the generic PVS1 framework once germline LoF is established for the gene; PP3 must not be double-counted for the same splice prediction evidence
• BayesDel noAF score 0.83 (local lookup, BayesDel_170824_noAF chrX; source_registry key bayesdel) - pathogenic-direction score, but missense-calibrated with PLACEHOLDER/unverified cutoffs in the pipeline SVCv4 ladder and not the appropriate tool for a canonical splice variant; not counted toward PP3
• Generic ACMG/AMP 2015 PP3 definition (PMID:25741868, source_registry key generic_acmg_combination_rules): 'Multiple lines of computational evidence support a deleterious effect on the gene or gene product'

Still missing

• If the PVS1 group downgrades or does not apply PVS1 for this variant (e.g., exon 16 not biologically relevant, NMD escape, or transcript relevance concern under PMC6185798), a human reviewer should revisit whether PP3 should be awarded from the SpliceAI max delta 0.99 prediction
• REVEL was not retrievable for this case (prefetch.json error: 'No REVEL ZIP found for chrX:77635913'), so the missense in-silico path is empty and provides no corroborating or contradicting line
• BayesDel noAF 0.83 uses PLACEHOLDER/unverified thresholds in the pipeline SVCv4 ladder and is missense-calibrated (off-label for a canonical splice variant); it was not counted toward PP3

Sources

Reasoning

PP5 is applied only for an exact-variant ClinVar expert-panel classification of Pathogenic/Likely Pathogenic. ClinVar was searched exhaustively (multiple query terms: 'ATRX[GENE] AND c.4699+2T>C[VARNAME]', 'NM_000489.5:c.4699+2T>C[VARNAME]', genomic HGVS forms, etc.) and the variant is absent from ClinVar entirely (variation_ids: [], review_status: '', evidence_sentence: 'This variant is absent from ClinVar.'). No ClinVar record exists, therefore no expert-panel classification exists, and PP5 cannot be met. Ordinary laboratory submissions, aggregate labels, or non-expert assertions are not present either, so there is no basis of any kind to trigger PP5.

Rule

PP5: Reputable source recently reports variant as pathogenic, but evidence available to the laboratory is insufficient to perform an independent evaluation; per this pipeline, applied only for an exact-variant ClinVar expert-panel (3-star) Pathogenic/Likely Pathogenic classification.

Evidence considered

• ClinVar search results: exact_match false, variation_ids [], submissions [] (prefetch.json and evidence.json)

Still missing

• No ClinVar record exists for this variant; an expert-panel classification could appear in the future and would need re-review

Sources

Reasoning

BA1 (Richards et al. 2015, PMID:25741868) requires an allele frequency in large population cohorts above the benign threshold; under the generic ACMG/AMP 2015 fallback used here (no ATRX VCEP/CSPEC available), the non-VCEP gnomAD BA1 threshold is >1%. NM_000489.5:c.4699+2T>C is absent from gnomAD v2.1 exomes (GRCh37 X-76891404-A-G), gnomAD v4.1 exomes (GRCh38 chrX-77635913-A-G), and gnomAD-Canada v1.0 genomes (AC=0, AN=0, AF=0), so the allele frequency is 0 and the variant cannot approach the BA1 threshold; no threshold artifact applies to an absent variant. BA1 is not met.

Rule

generic_acmg_2015_ba1: af_gt_1%_non_vcep_not_met

Evidence considered

• gnomAD v2.1 (GRCh37 X-76891404-A-G): variant absent, search_status=absent, no allele counts recorded
• gnomAD v4.1 (GRCh38 chrX-77635913-A-G): variant absent, search_status=absent, no allele counts recorded
• gnomAD-Canada v1.0: AC=0, AN=0, AF=0.0
• Generic ACMG/AMP 2015 fallback with non-VCEP BA1 threshold >1% (per governing gnomAD cutoffs)

Sources

Reasoning

BS1 (Richards et al. 2015, PMID:25741868) requires an allele frequency greater than expected for the disorder; under the generic ACMG/AMP 2015 fallback, the non-VCEP gnomAD BS1 threshold is >0.3%. ATRX-related ATR-X syndrome is a very rare X-linked disorder, and NM_000489.5:c.4699+2T>C is absent from gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0 (AF=0), so the observed frequency is far below the BS1 threshold and cannot exceed the expected disease prevalence. BS1 is not met.

Rule

generic_acmg_2015_bs1: af_gt_0.3%_non_vcep_not_met

Evidence considered

• gnomAD v2.1 (GRCh37 X-76891404-A-G): variant absent
• gnomAD v4.1 (GRCh38 chrX-77635913-A-G): variant absent
• gnomAD-Canada v1.0: AC=0, AN=0, AF=0.0
• ATR-X syndrome is a rare X-linked disorder (gene context), so expected-frequency threshold is low; AF=0 cannot exceed it
• Generic ACMG/AMP 2015 fallback with non-VCEP BS1 threshold >0.3%

Sources

Reasoning

BS2 (Richards et al. 2015, PMID:25741868) requires observation of the variant in a healthy adult individual in the homozygous state for recessive disorders or the hemizygous state for X-linked disorders. ATRX is X-linked, so BS2 would require observation of hemizygous males or homozygous females in population cohorts; NM_000489.5:c.4699+2T>C is absent from gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0 with no homozygotes or hemizygotes observed, so the variant is not observed in any healthy carrier state. BS2 is not met.

Rule

generic_acmg_2015_bs2: healthy_hom_hemizygote_observation_not_met

Evidence considered

• gnomAD v2.1 (GRCh37 X-76891404-A-G): variant absent, no hemizygotes/homozygotes recorded
• gnomAD v4.1 (GRCh38 chrX-77635913-A-G): variant absent, no hemizygotes/homozygotes recorded
• gnomAD-Canada v1.0: AC=0, AN=0, AF=0.0, hom=0
• No healthy-individual observation exists for this variant in any population cohort

Sources

Reasoning

BP4 (Richards et al. 2015, PMID:25741868) requires multiple lines of computational evidence suggesting no impact on the gene or gene product (conservation, evolutionary, splicing impact, etc.). Under the generic ACMG/AMP 2015 framework (no ATRX VCEP/CSPEC or local gene framework exists), the only available computational lines both predict an impact: SpliceAI max delta score = 0.99 (DS_DL 0.99 donor loss, DS_DG 0.82 donor gain) predicts strong splice disruption, and BayesDel noAF = 0.83 is in the pathogenic direction (missense-calibrated, PLACEHOLDER/unverified thresholds, so not treated as decisive, but it does not support no-impact either). There is no benign-direction computational line of any kind, and the variant sits at a canonical +2 splice-donor position where a no-impact call is not plausible. BP4 is not met.

Rule

Generic ACMG/AMP 2015 BP4 (Richards et al., PMID:25741868; source_registry key generic_acmg_combination_rules): 'Multiple lines of computational evidence suggest no impact on the gene or gene product (conservation, evolutionary, splicing impact, etc.).' Not met: every available computational line (SpliceAI max delta 0.99; BayesDel 0.83) predicts an impact and no benign-direction line exists.

Evidence considered

• SpliceAI Lookup max delta score 0.99 (DS_DL 0.99 donor loss at +2, DS_DG 0.82 donor gain; source_registry key spliceai) - strong predicted splice impact; directly contradicts a no-impact call
• BayesDel noAF score 0.83 (local lookup, source_registry key bayesdel) - pathogenic-direction score; contradicts no-impact (placeholder calibration and off-label for a canonical splice variant, so weighted only as absence of any benign-direction support)
• Generic ACMG/AMP 2015 BP4 definition (PMID:25741868, source_registry key generic_acmg_combination_rules): 'Multiple lines of computational evidence suggest no impact on the gene or gene product'

Still missing

• No benign-direction computational line exists for this variant; SpliceAI 0.99 and BayesDel 0.83 are both in the deleterious direction
• No RNA-based splicing assay (functional) evidence was captured; BP4 would rest only on in silico splice prediction even if a benign-direction line existed

Sources

Reasoning

BP6 is applied only for an exact-variant ClinVar expert-panel classification of Benign/Likely Benign. The variant is absent from ClinVar entirely (no record, no submissions, no review status; evidence_sentence: 'This variant is absent from ClinVar.'). No expert-panel benign classification exists, so BP6 cannot be met. Population-absence data (gnomAD v2.1, v4.1, gnomAD-Canada) are informative for PM2/BS1/BA1 in the population group but are not an expert-panel benign classification and must not trigger BP6.

Rule

BP6: Reputable source recently reports variant as benign, but evidence available to the laboratory is insufficient to perform an independent evaluation; per this pipeline, applied only for an exact-variant ClinVar expert-panel (3-star) Benign/Likely Benign classification.

Evidence considered

• ClinVar search results: variant absent, zero submissions (prefetch.json and evidence.json)

Still missing

• No ClinVar record exists for this variant; a future expert-panel benign classification would need re-review

Sources

Reasoning

No de novo evidence is available for this variant. The case bundle contains no proband phenotype description, no parental testing results, and no parentage-confirmation information for NM_000489.5:c.4699+2T>C. The variant is absent from ClinVar (no submissions at all), so no ClinVar-submitted de novo occurrence exists. The generic ACMG/AMP 2015 framework (Richards et al., PMID:25741868) requires a documented de novo occurrence with confirmed parentage to apply PS2; that evidence is entirely absent from this case, so PS2 cannot be applied or refuted.

Rule

Generic ACMG/AMP 2015 (Richards et al. PMID:25741868) PS2: de novo occurrence with confirmed parentage (strong). No ATRX-specific VCEP/CSPEC specification exists (cspec found=false; framework_mode=generic_acmg).

Evidence considered

• clinvar: This variant is absent from ClinVar (no variation IDs, no submissions), providing no de novo occurrence record.
• generic_acmg_combination_rules: Generic ACMG/AMP 2015 PS2 definition (de novo with confirmed parentage) used as governing rule since no ATRX CSPEC/VCEP exists.

Still missing

• No proband clinical/phenotype data is present in the case bundle to establish the variant arose de novo.
• No parental testing (biological parents) results are available.
• No parentage confirmation (e.g., STR/single-nucleotide marker panel) results are available.
• Variant absent from ClinVar, so no external de novo submission record exists.

Sources

Reasoning

No functional assay evidence exists for NM_000489.5:c.4699+2T>C in any evidence source. PS3 requires well-established functional studies (e.g., RNA/minigene splicing assays, RT-PCR on patient samples, protein expression/activity assays) demonstrating a deleterious effect. The variant is absent from ClinVar, the literature pass returned 0 PMIDs, and fulltext_map.json is empty. SpliceAI (max delta = 0.99) and BayesDel (0.83) are in silico predictors and, per ACMG/AMP 2015 (PMID:25741868), cannot substitute for functional assay evidence in PS3/BS3 calibration; the in silico splice prediction is handled by other criteria groups (PVS1/PP3), not PS3. The five full-text articles on disk (output/pvs1_articles/) are ATRX gene-context/PVS1 LoF-mechanism papers; none mention this variant or any functional assay, so they are not citable for PS3 under strict variant-grounding rules. PS3 is therefore not assessed due to absent evidence; a human review is flagged because RNA-based studies for this canonical +2 donor variant may exist in the literature outside the automated retrieval scope and would strengthen the classification if found.

Rule

ACMG/AMP 2015 (PMID:25741868, source_registry key generic_acmg_combination_rules): PS3 = well-established functional studies show a deleterious effect; in silico predictors (SpliceAI/BayesDel) do not qualify as functional evidence.

Still missing

• No RNA-based functional study (minigene splicing assay, RT-PCR on patient-derived RNA, or transcriptome data) demonstrating aberrant splicing of c.4699+2T>C.
• No protein-level functional assay (e.g., ATRX expression, chromatin-remodeling activity, DAXX interaction) for this variant.
• No ClinVar submission or case publication reporting this variant with functional follow-up (literature_pass total_pmids = 0).

Sources

Reasoning

PS4 requires the variant to show statistically significant enrichment in affected individuals versus controls (case-control or cohort data) or to be reported in multiple unrelated probands with the same phenotype. For NM_000489.5:c.4699+2T>C (ATRX), no such data exist: the variant is absent from ClinVar (no submissions of any kind), the literature pass returned zero PMIDs across all evidence sources (literature_pass.json: 'No PMIDs found across evidence sources.'), the fulltext map is empty, and no case reports or cohort studies mentioning this exact variant are present in the workspace. Without any affected-individual or control observations, PS4 cannot be evaluated; this is an evidence-absence gap, not a negative finding.

Rule

PS4: The prevalence of the variant in affected individuals is significantly increased compared with controls (case-control or cohort enrichment; also supported by multiple unrelated probands with the same phenotype under the generic ACMG/AMP 2015 framework, PMID 25741868).

Evidence considered

• ClinVar query: variant absent, zero submissions (evidence.json clinvar record)
• literature_pass.json: total_pmids = 0, no high/medium priority papers
• fulltext_map.json: empty; publications/ directory contains no files

Still missing

• No case-control or cohort enrichment data for c.4699+2T>C (no published studies found)
• No probands carrying this exact variant reported in literature or ClinVar submissions
• No phenotype-matched multiple-proband observations to support PS4

Sources

Reasoning

No de novo evidence of any kind is available for this variant. PM6 under generic ACMG/AMP 2015 applies to a de novo occurrence without confirmed parentage; the case bundle contains no proband data, no parental testing, and no record of the variant being observed as de novo in any individual. The variant is absent from ClinVar and from all literature sources identified (literature_pass total_pmids=0), so there is no unconfirmed-parentage de novo observation to evaluate.

Rule

Generic ACMG/AMP 2015 (Richards et al. PMID:25741868) PM6: de novo without confirmed parentage (moderate). No ATRX-specific VCEP/CSPEC specification exists.

Evidence considered

• clinvar: Absence of any ClinVar submission for this variant means no de novo observation (confirmed or unconfirmed) is on record.
• generic_acmg_combination_rules: Generic ACMG/AMP 2015 PM6 definition (de novo without confirmed parentage, moderate) used as governing rule.

Still missing

• No proband or family testing data present to identify a possible de novo occurrence.
• No parental testing results (even without parentage confirmation) are available.
• No literature reports of this variant as de novo (no PMIDs found for the variant anywhere in the evidence bundle).

Sources

Reasoning

No segregation data is available. There is no pedigree, no affected-family-member genotyping, and no meiotic-segregation information for NM_000489.5:c.4699+2T>C anywhere in the case bundle. The generic ACMG/AMP 2015 PP1 criterion requires the variant to cosegregate with disease in multiple affected family members; with zero family data and zero literature reports of this variant (literature_pass total_pmids=0; fulltext_map empty; publications directory empty), PP1 cannot be scored. Note the variant is in an X-linked gene (ATRX); a gene-specific segregation framework is not available, so generic counting rules would apply if data existed.

Rule

Generic ACMG/AMP 2015 (Richards et al. PMID:25741868) PP1: cosegregation with disease in multiple affected family members (supporting). No ATRX-specific VCEP/CSPEC specification exists.

Evidence considered

• clinvar: No ClinVar submission exists, so no segregation observations are reported through ClinVar.
• generic_acmg_combination_rules: Generic ACMG/AMP 2015 PP1 definition (cosegregation with disease in affected family members, supporting) used as governing rule.

Still missing

• No pedigree or family structure information in the case bundle.
• No genotype data for affected or unaffected family members.
• No meiosis/segregation counts available.
• No literature segregation reports for this variant (no PMIDs identified).

Sources

Reasoning

PP4 requires the proband's phenotype or family history to be highly specific for a disease caused by variants in the gene (here, ATR-X syndrome, a neurodevelopmental disorder associated with ATRX). The case bundle contains no proband phenotype information whatsoever: case_summary.json has no phenotype, clinical presentation, or family-history fields, and no referring-clinician or clinical-data source was captured. Phenotype specificity cannot be scored without the proband's phenotype, so PP4 is not assessed. The ATRX germline LoF disease mechanism is supported at gene level (pvs1_gene_context), but gene-level mechanism does not substitute for proband-level phenotype evidence under PP4.

Rule

PP4: Variant's phenotype or family history is highly specific for a disease with a single genetic etiology (generic ACMG/AMP 2015 framework, PMID 25741868).

Evidence considered

• case_summary.json: no phenotype, clinical, or family-history fields present
• Generic ACMG/AMP 2015 PP4 rule consulted (PMID 25741868)

Still missing

• No proband phenotype/clinical presentation data provided in the case bundle
• No family history data provided
• No documented phenotype-to-genotype specificity evaluation for this proband

Sources

Reasoning

No functional assay evidence exists that could establish a lack of damaging effect for NM_000489.5:c.4699+2T>C. BS3 requires well-established functional studies demonstrating no damaging effect on protein function or splicing. No RNA/minigene studies, no protein-level assays, no ClinVar entry, and no variant-specific literature (0 PMIDs) are available. The only directional signal is in silico: SpliceAI predicts strong splice impact (max delta = 0.99, DS_DL 0.99, DS_DG 0.82) and BayesDel scores 0.83, both pointing away from benign, but these are computational predictions and do not constitute functional assay evidence for BS3 either way. BS3 is therefore not assessed due to absent functional data.

Rule

ACMG/AMP 2015 (PMID:25741868, source_registry key generic_acmg_combination_rules): BS3 = well-established functional studies show no damaging effect on protein function or splicing; in silico predictors do not qualify as functional evidence.

Still missing

• No functional assay (minigene, RT-PCR, or protein assay) showing preservation of normal splicing or protein function for c.4699+2T>C.
• No published evidence of this variant in healthy/control individuals with functional follow-up.

Sources

Reasoning

No non-segregation evidence is available. BS4 under generic ACMG/AMP 2015 requires observation of the variant in an affected family member who does not carry it (or absence of segregation in affected relatives); the case bundle contains no family genotyping data whatsoever for NM_000489.5:c.4699+2T>C. Absence of data cannot support a benign non-segregation finding, and no literature or ClinVar source reports non-segregation for this variant.

Rule

Generic ACMG/AMP 2015 (Richards et al. PMID:25741868) BS4: absence of segregation in affected family members (strong benign). No ATRX-specific VCEP/CSPEC specification exists.

Evidence considered

• clinvar: Absence of any ClinVar submission means no non-segregation observation is on record.
• generic_acmg_combination_rules: Generic ACMG/AMP 2015 BS4 definition (non-segregation with disease, strong benign) used as governing rule.

Still missing

• No affected-family-member genotyping data to evaluate non-segregation.
• No unaffected-carrier or affected-non-carrier observations available.
• No literature reports of non-segregation for this variant (no PMIDs identified).

Sources

Reasoning

BP2 (generic ACMG/AMP 2015, PMID 25741868) has two prongs: (a) observed in trans with a pathogenic variant for a fully penetrant dominant disorder, or (b) observed in cis with a pathogenic variant in any inheritance pattern. Prong (a) is structurally inapplicable because ATRX-associated disease is X-linked (gene on Xq21.1; ATR-X syndrome), not a fully penetrant autosomal dominant disorder. Prong (b) is inheritance-agnostic and could in principle apply, but there is no observation data of any kind: c.4699+2T>C is absent from ClinVar (no variation IDs, no submissions), absent from gnomAD v2.1/v4.1/Canada, absent from COSMIC, and literature_pass identified 0 PMIDs. No proband, family, phase, segregation, or second-variant data exist in the evidence bundle, so the cis/trans observation required by BP2 cannot be established or refuted. Status is therefore not_assessed with the data gap identified; no benign evidence is derived from BP2.

Rule

BP2 (Richards et al. 2015, PMID 25741868): 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

• clinvar
• gnomad_v2
• gnomad_v4
• gnomad_canada
• generic_acmg_combination_rules

Still missing

• No observations of c.4699+2T>C in cis or in trans with any pathogenic variant (variant absent from ClinVar and the literature)
• No proband or family-level genotype data (phase, segregation) available in this case
• No second ATRX variant reported to evaluate cis/trans configuration
• If any future allelic observation (cis or trans) is reported, BP2 must be re-adjudicated

Sources

Reasoning

BP5 requires identification of an alternative molecular basis for the disease in the proband (e.g., a pathogenic variant in another gene fully explaining the phenotype). The case bundle contains no proband-level testing data: no other variants, no prior genetic testing results, no alternative-diagnosis documentation. Without any proband molecular data, BP5 cannot be evaluated; this is an evidence-absence gap.

Rule

BP5: Variant found in a case with an alternate molecular basis for disease (generic ACMG/AMP 2015 framework, PMID 25741868).

Evidence considered

• case_summary.json: no proband testing data or alternative-diagnosis fields present
• Generic ACMG/AMP 2015 BP5 rule consulted (PMID 25741868)

Still missing

• No proband-level genetic testing data documenting an alternative molecular cause
• No information on whether the proband carries a second, explanatory variant in another gene

Sources