Five-step scale from benign to pathogenic. This variant is likely benign. Select a step to read what it means.
This variantScenario
BenignPathogenic
Likely benign.A likely benign classification under the recorded framework.
In plain language
Gene-level background from the recorded interpretation.
Words with a dotted underline have a short definition. Hover or tap them.
About ATM
ATM encodes a kinase that acts as a master controller of the cellular DNA damage response: it detects double-strand breaks and coordinates DNA repair, cell cycle arrest, or apoptosis through downstream targets including p53, BRCA1, and CHK2. Loss-of-function mutations in both copies of ATM cause ataxia telangiectasia, an autosomal recessive disorder characterized by neurological problems, immune deficiency, and cancer predisposition. ATM functions as a tumor suppressor: individuals with ataxia telangiectasia are prone to childhood lymphomas, leukemias, and breast cancer, while carriers of a single altered copy have increased risk of breast, pancreatic, prostate, and other cancers. Somatic ATM mutations also occur in lymphoid malignancies and solid tumors, and ATM-deficient cancers are often especially sensitive to DNA-damaging treatments.
LYFE Sciences is an AI system, and it can make mistakes.
Criteria may be applied incorrectly, sources may be misread, and a
confident-looking classification can still be wrong.
Double-check every criterion and its underlying evidence
before relying on any call.
01The reasoning
How it was classified
The recorded framework is ClinGen ATM VCEP v1.5. The criteria and strengths shown below come from the saved assessment; this page does not recalculate the classification.
Recorded applied criteria
The applied criteria and their strengths are shown as assessed under the stated framework.
Met (Supporting): gnomAD v4.1 grpmax filtering AF 7.39e-06, below the 0.001% threshold. Flagged for human review: the rule specifies gnomAD v4 frequency without stating whether grpmax or a single-population estimate applies.
Benign criteria applied · 2
BP4Supporting
BP4 · supporting
Met (Supporting): SpliceAI maximum delta score 0.00, at or below the <=0.1 no-splice-impact threshold.
BP7Supporting
BP7 · supporting
Met (Supporting): the variant lies 11 nucleotides into intron 17, beyond the +7 donor-site deep-intronic threshold.
Recorded classification: Likely benign
Likely Benign: BP4_Supporting (SpliceAI maximum delta 0.00) and BP7_Supporting (donor +11, beyond +7) satisfy ATM v1.5 Rule19, with PM2_Supporting not treated as conflicting evidence.
02The sources
Evidence at a glance
Evidence summaries and captured source records, where available. Missing entries should not be interpreted as negative findings.
gnomAD v4.1
gnomAD v2.1
Population frequency
7 / 1,612,814 alleles
gnomAD v4.1: 7 / 1,612,814 alleles. Recorded AF 4.34024e-06. Recorded homozygote count 0.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 3.97997e-06; MAF= 0.00040%, 1/251258 alleles, homozygotes = 0) and has highest observed frequency in the East Asian population (AF= 5.4401e-05; MAF= 0.00544%, 1/18382 alleles, homozygotes = 0).
ACMG/AMP criteria under ClinGen ATM VCEP v1.5, with the data and sources behind each one.
28 criteria with recorded statuses: 3 met · 8 not met · 5 not assessed · 12 not applicable. Open any row for the full reasoning, the rule and the evidence behind it.
All three consequence/LOF criteria (PVS1, PM4, BP3) are non-applicable to ATM NM_000051.4:c.2638+11A>G: it is an intronic substitution 11 nt into intron 17 (exonic position '17i'), outside the canonical donor/acceptor +/-1,2 dinucleotides, with SpliceAI max delta score 0.00 and no RNA or protein data available.
PVS1 is not applicable under ATM VCEP v1.5 because the variant is not a null variant and predicted splice defects outside canonical +/-1,2 sites require observed RNA evidence (PVS1_Strength(RNA)), which does not exist for this variant.
PM4 is not applicable because the ATM VCEP restricts it to stop-loss variants (and in-frame indels of at least one exon) and this variant produces no protein length change; BP3 is explicitly disabled ('Do not use') by the ATM VCEP.
The splice-directed evidence for this variant (SpliceAI <=0.1, deep-intronic position beyond +7) is the province of the in-silico splicing group (BP4/BP7), not of PVS1/PM4/BP3; no classification from any external source was used to derive these calls.
The codon/residue/domain group contributes no pathogenic evidence for NM_000051.4:c.2638+11A>G: PS1 is not met, PM5 is not met, and PM1, PP2 and BP1 are Not Applicable under the governing ClinGen HBOP ATM VCEP v1.5 specification.
The variant is a deep intronic +11 change with no amino-acid substitution (p.?) and SpliceAI max delta 0.00, so neither the same-amino-acid-change mechanism (PS1) nor the same-residue/truncation mechanisms (PM5) can be satisfied; there is also no reference P/LP splice-region comparator at the same nucleotide or donor motif in the collected evidence.
No authoritative ATM critical-domain table exists in the case materials (vcep_materials.json domain_tables is empty) and the VCEP explicitly disables PM1 for ATM, so no hotspot/domain-based pathogenic evidence can be generated.
ClinVar (variation 384288) lists the variant as Likely benign from three clinical laboratories with no expert-panel submission; this was recorded as context only and no criterion was inferred from that final classification.
Governing framework is the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer (HBOP) Expert Panel Specifications to the ACMG/AMP Variant Interpretation Guidelines for ATM Version 1.5 (cspec, doc 639508985), used in preference to generic ACMG/AMP; the framework mode for this case is 'vcep'.
PS2 is not_applicable by explicit ATM VCEP v1.5 text ('Do not use for AD or AR disease: Informative de novo occurrences have not yet been observed and de novo AR conditions are unlikely to be informed by phase'); all strength tiers are marked Not applicable.
PM6 is not_applicable by the same ATM VCEP v1.5 instruction, which excludes both de novo criteria for this gene.
BS4 is not_applicable by explicit ATM VCEP v1.5 text: AD co-segregation is uninformative in a low-penetrance gene (PMID 32773770), and informative lack of co-segregation in A-T families is too rare to be weighted, with such observations instead routed to BP2.
PP1 is the only criterion in this group that the ATM VCEP v1.5 declares applicable (AR condition only, requiring both variants identified in the proband), but the case bundle contains no proband, pedigree, affected-relative genotypes, or meioses, so PP1 is returned as not_assessed with the missing evidence identified rather than not_met.
No source in the registry reports c.2638+11A>G as de novo, as segregating, or as showing lack of segregation; ClinVar final classifications (Likely benign, 3 single-submitter entries) were not used to infer any criterion in this group.
Framework: ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel ATM specification v1.5 (CSPEC doc 639508985, framework_mode=vcep) governs PS3/BS3; generic ACMG/AMP 2015 (PMID 25741868) was consulted only as background on what constitutes well-established functional evidence.
Variant: NM_000051.4:c.2638+11A>G is an intronic substitution in ATM intron 17 (VariantValidator exon 17i; non-canonical splice position, canonical splice consensus = false; protein consequence p.?). It is therefore not assessable by the VCEP's three approved protein assays, all of which transfect intron-free ATM cDNA constructs into ATM-null LCLs (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335).
No variant-specific functional evidence exists anywhere in the bundle: the approved-assay papers were not retrieved, the variant has no row in Sun et al. 2025 (PMID 40580951) Table S1 (searched in full, 29,374 rows), and no RNA-based assay (minigene, patient cDNA/RT-PCR with or without NMD inhibition) is reported.
SpliceAI max delta 0.00 predicts no splice impact, but it is an in silico prediction (PP3/BP4 domain) and cannot substitute for the functional assay PS3 or BS3 requires; under the ATM VCEP, RNA-level functional evidence is routed to PVS1_Strength(RNA) (pathogenic) and BP7_RNA (benign), which are adjudicated by other groups.
ClinVar's Likely benign classification (3 laboratories, no expert panel) was not used to infer PS3 or BS3, per the rule that a criterion must never be inferred from another source's final classification.
Both PS3 and BS3 are returned as not_assessed for insufficient variant-specific functional evidence; the decisive gap is the absence of any approved-assay or RNA-based functional result for this intronic variant.
Group scope: PS4, PP4, PP5, BP5, BP6 (phenotype prevalence assertions) for NM_000051.4(ATM):c.2638+11A>G, assessed under the ClinGen HBOP ATM VCEP v1.5 specification (framework_mode=vcep).
PS4 (VCEP Applicable, Strong) is not met: no case-control or cohort enrichment study exists for this variant, and the VCEP prohibits the PS4_Moderate proband-counting path for ATM; the ClinVar record and gnomAD contain no affected-versus-control comparison.
PP4 and BP5 are Not applicable under the ATM VCEP v1.5 specification for all strengths; phenotype specificity is folded into the VCEP's PM3/BP2 points table (PP4) and BP5 is disallowed for this low-penetrance, genetically heterogeneous setting.
PP5 and BP6 are 'Not Applicable for this VCEP' by the ClinGen SVI VCEP Review Committee; adjudicated under the lab's exact-variant ClinVar expert-panel override, neither is triggered because ClinVar 384288 has zero expert-panel submissions (three single-submitter clinical-laboratory submissions only).
Net contribution of this group to the final classification: zero pathogenic and zero benign points/evidence; no criterion in this group is met.
Governing framework is the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer (HBOP) Expert Panel Specifications to the ACMG/AMP Variant Interpretation Guidelines for ATM Version 1.5 (cspec, doc 639508985), including the dedicated ATM_PM3_BP2_1.5.pdf points table; framework_mode is 'vcep', so generic ACMG/AMP PM3/BP2 is not used.
Both PM3 and BP2 in this specification are points-based and require a concrete carrier observation: PM3 needs an unrelated A-T proband with c.2638+11A>G plus a co-occurring ATM variant (P/LP or VUS) or homozygosity; BP2 needs an unaffected adult (>18 years) carrier with a P/LP ATM variant in trans, or a homozygote from a phenotyped cohort.
The case bundle contains no A-T proband, no unaffected-adult carrier, no co-occurring ATM variant, and no phase or zygosity data for c.2638+11A>G: literature triage found no variant-specific paper, the three ClinVar submissions (GeneDx, Color Health, Labcorp/Invitae) carry no proband/phase evidence and cite only generic methodology papers (PMID:25741868, PMID:28492532), and the screened full texts (PMID:25741868, PMID:24418350, PMID:25394175) do not mention the variant.
Rarity does not create either code but does clear PM3's eligibility ceiling: gnomAD v4.1 grpmax filtering AF 7.39e-06 (0.00074%) is far below the VCEP's 0.01% limit, and no homozygotes are reported in gnomAD v2.1 or v4.1.
The BP2 splicing prerequisite (footnote 3) is satisfied - SpliceAI max delta score 0.00, Pangolin SG 0.001 - so c.2638+11A>G is not disqualified from BP2 on predicted-splicing grounds; the disqualifying factor is the absence of a qualifying unaffected-adult co-occurrence observation.
Net allelic contribution: PM3 = 0.0 points (below the 1.0-point PM3_Supporting threshold) and BP2 = 0.0 points (above the -1.0 BP2_Supporting threshold); both are returned as not_assessed because absence of observations is not affirmative evidence against PM3 nor affirmative evidence for BP2, and the missing evidence is itemised in gaps_remaining.
ClinVar's final Likely benign classification (3 single-submitter entries, no expert panel) was not used to infer PM3 or BP2; no submission carried phase, zygosity, or co-occurrence evidence signals.
In-silico/splicing group contributes two benign-supporting codes - BP4 (splicing sub-path, SpliceAI 0.00 <= 0.1) and BP7 (positional deep-intronic rule, donor +11, beyond +7) - and no pathogenic computational evidence: PP3 is not met because SpliceAI 0.00 is below the VCEP threshold of >=0.2.
BP4 and BP7 are distinct evidence lines, not double-counting: BP7 is the ATM VCEP v1.5 positional deep-intronic code (further than +7 at the donor site), whereas BP4 is the SpliceAI no-impact prediction; the VCEP explicitly permits BP4 in addition to BP7 ('May also apply BP4 to achieve Likely Benign').
No missense computational path applies (intronic variant; REVEL and BayesDel unavailable/not applicable), and the ATM VCEP supplementary Table S1 from Sun et al. 2025 (PMID 40580951) contains no row for c.2638+11A>G, so no secondary computational corroboration was available.
All splice evidence rests on SpliceAI, which is the single predictor the ATM VCEP specifies for PP3/BP4; no RNA or minigene data exist for this variant to support a BP7(RNA) strength modification or to independently confirm absence of an aberrant transcript.
Population-frequency evidence for NM_000051.4(ATM):c.2638+11A>G (NC_000011.10:g.108267353A>G) was assessed under the ClinGen HBOP ATM VCEP v1.5 specification (source key cspec), which is the governing framework for this gene; no generic ACMG frequency thresholds were needed because the VCEP supplies explicit BA1, BS1 and PM2 rules.
The variant is ultra-rare: gnomAD v4.1 grpmax filtering AF 7.39e-06 (0.000739%), total AF 4.34e-06 (7/1,612,814 alleles), 0 homozygotes; highest subpopulation frequency East Asian AF 4.46e-05 (2/44,846 alleles). gnomAD v2.1 is concordant (total AF 3.98e-06, 1/251,258 alleles, 0 homozygotes).
BA1 (grpmax FAF >0.5%) and BS1 (grpmax FAF >0.05%) are both not met by a wide margin (~676-fold and ~68-fold below their respective ceilings).
BS2 is Not Applicable under ATM VCEP v1.5.
PM2 is met at Supporting strength (PM2_Supporting) because grpmax FAF and total AF are both <=0.001% in gnomAD v4.1; the VCEP explicitly permits PM2_Supporting when a variant is seen only once in a single subpopulation, and this variant is seen twice in East Asian individuals with no homozygotes.
Disease prevalence and penetrance were considered: A-T is a rare recessive disorder and ATM heterozygosity confers only reduced-penetrance cancer risk; the VCEP's frequency ceilings already account for this, and the variant remains far below all benign-frequency thresholds.
No threshold artifact (small subpopulation AN, founder effect, or homozygote clustering) was identified that would move the variant toward BA1 or BS1; the main open question is the grpmax-versus-popmax interpretation of the PM2 rule, which is flagged for human review.
No paper identified through the literature pass reports population-frequency data for this specific variant; ClinVar's Likely-benign classification by three laboratories was not used to infer any population criterion.
No criteria match.
Met · 3
PM2SupportingMet (Supporting): gnomAD v4.1 grpmax filtering AF 7.39e-06, below the 0.001% threshold. Flagged for human review: the rule specifies gnomAD v4 frequency without stating whether grpmax or a single-population estimate applies.
Details ▾Hide ▴
Reasoning
Met at Supporting strength (PM2_Supporting). ATM VCEP v1.5 specifies PM2_Supporting when frequency is <=0.001% in gnomAD v4, and explicitly states that a single observation in one subpopulation is still sufficiently rare. gnomAD v4.1 gives grpmax filtering AF 7.39e-06 (0.000739%) and total AF 4.34e-06 (7/1,612,814 alleles), both <=0.001%, with 0 homozygotes; the highest subpopulation frequency is East Asian AF 4.46e-05 (2/44,846 alleles, 0 homozygotes). gnomAD v2.1 corroborates (total AF 3.98e-06, 1/251,258 alleles, 0 homozygotes). The VCEP specifies PM2_Supporting rather than moderate, and notes PM2 is not treated as conflicting evidence for variants that otherwise appear benign. Human review is flagged because the VCEP PM2 rule text names only 'frequency <=0.001% in gnomAD v4' without explicitly specifying grpmax, whereas the BA1/BS1 rules do; under the grpmax FAF the criterion is met, but under a popmax point-estimate reading (East Asian AF 0.00446%) it would not be.
Rule
Frequency <=.001% in gnomAD v4 dataset (ATM VCEP v1.5) - observed grpmax FAF 7.39e-06 (0.000739%), total AF 4.34e-06; PM2_Supporting
Evidence considered
gnomAD v4.1 (gnomad_v4): grpmax FAF 7.39e-06 (0.000739%), exome grpmax FAF 8.35e-06, total AF 4.34024e-06 (AC=7, AN=1,612,814), 0 homozygotes.
The ATM VCEP v1.5 PM2 rule is 'Frequency <=.001% in gnomAD v4 dataset. If n=1 in a single sub population, that is sufficiently rare and PM2_supporting would apply.' Both the grpmax FAF (0.000739%) and the total AF (0.000434%) satisfy the <=0.001% threshold.
Subpopulation detail: East Asian AF 4.45971e-05 (AC=2, AN=44,846, 0 homozygotes); Remaining individuals AF 1.60e-05 (1/62,466, 0 homozygotes); African/African American, Admixed American, Finnish, Amish, Ashkenazi Jewish and other groups all 0.
gnomAD v2.1 (gnomad_v2): total AF 3.97997e-06 (1/251,258 alleles), 0 homozygotes; highest subpopulation East Asian AF 5.44e-05 (1/18,382); corroborates ultra-rarity.
The VCEP specifies use as PM2_Supporting (not moderate) and states PM2 is not considered a conflicting piece of evidence for variants that otherwise are likely benign/benign.
Ancestry review: no ancestry group shows frequency at or above 0.001% by grpmax FAF, and no homozygotes were observed; the two East Asian observations do not reach the VCEP's rarity ceiling on the grpmax metric.
Interpretive caveat flagged for human review: applying the popmax point estimate instead of the grpmax filtering AF would place the East Asian frequency (0.00446%) above the 0.001% PM2 ceiling.
Still missing
Interpretive ambiguity: VCEP PM2 text does not specify grpmax FAF, unlike the BA1/BS1 rules. grpmax FAF (0.000739%) is <=0.001% (met), but the East Asian popmax point estimate (0.00446%) exceeds 0.001% (would be not met).
gnomAD Canada (HostSeq) registered as a source but no result retrieved, so the frequency assessment rests on gnomAD v4.1 and v2.1 only.
No homozygote observations in either dataset, so no homozygote-based rarity/penetrance signal was available; expected homozygote count at this frequency is effectively zero.
BP4SupportingMet (Supporting): SpliceAI maximum delta score 0.00, at or below the <=0.1 no-splice-impact threshold.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 BP4 splicing sub-path is applicable and is met: SpliceAI maximum delta score for NM_000051.4:c.2638+11A>G is 0.00, which is <=0.1, the VCEP's no-predicted-splice-impact threshold. This is a non-missense intronic variant (donor +11, outside the canonical +/-1,2 positions) for which splicing is the primary plausible mechanism, so the lab's BP4 missense/splicing policy directs use of the splicing sub-path; that policy does not change the outcome here because the VCEP's literal text also permits BP4 from the splice sub-path for such variants. BP4 is applied once, for the splicing mechanism only (REVEL is inapplicable and unavailable for an intronic variant). The VCEP restriction on combining BP4 with BP7_Variable(RNA) is not triggered (no RNA data exist), and the VCEP explicitly permits BP4 alongside BP7 ('May also apply BP4 to achieve Likely Benign'); BP7 is assigned here from the positional deep-intronic rule, not by reusing the SpliceAI prediction, so the two codes do not double-count one prediction.
ClinGen HBOP ATM VCEP v1.5 (source_registry key 'cspec'; CSPEC doc 639508985) specifies BP4_Supporting for no predicted splice impact at SpliceAI <=0.1 (and missense REVEL <=0.249) and states that BP4 may also be applied with BP7 for synonymous/deep intronic variants.
SpliceAI Lookup for NM_000051.4:c.2638+11A>G (source_registry key 'spliceai') returned max delta score 0.00 (all four delta scores 0.00; Pangolin SG 0.001, SL -0.001), meeting the VCEP BP4 splicing threshold of <=0.1.
The variant is intronic at donor +11 (outside the canonical +/-1,2 sites), so splicing is the primary plausible mechanism and the lab's BP4 missense/splicing policy in vcep_materials gene_notes directs use of the splicing sub-path; this does not change the outcome from the VCEP's literal text, which also allows the splice sub-path for intronic variants outside donor/acceptor 1,2.
Threshold provenance: the SpliceAI 0.1 cut-off is the ATM VCEP v1.5 specification's own value (CSPEC doc 639508985), which states the VCEP uses SpliceAI (Jaganathan et al., 2019, Cell; PMID 30661751) with the SVI-recommended splice thresholds (Walker et al., 2023, Am J Hum Genet) for non-canonical splice variants.
Still missing
BP4 rests on a single predictor (SpliceAI), which is however the VCEP-specified sole predictor for splice evidence; no RNA data exist to corroborate absent aberrant splicing.
BP7SupportingMet (Supporting): the variant lies 11 nucleotides into intron 17, beyond the +7 donor-site deep-intronic threshold.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 directly specifies BP7 at supporting strength for deep intronic variants located further than (but not including) +7 at a donor site. NM_000051.4:c.2638+11A>G is 11 nt into intron 17 downstream of the exon 17 donor site, i.e. beyond +7, so it is BP7-eligible and BP7_Supporting is met. No RNA study is available to support a BP7_Variable(RNA) strength modification, so the default supporting strength is used. BP7 here derives from the VCEP's positional deep-intronic rule, not from the SpliceAI prediction used for BP4; the SpliceAI score of 0.00 is concordant with (but not the basis of) BP7. The VCEP notes BP7 is not considered a conflicting piece of evidence against a body of evidence supporting a pathogenic splice defect.
Rule
ATM VCEP v1.5 BP7_Supporting: use for synonymous and deep intronic variants further than (but not including) +7 at donor and -21 at acceptor sites; BP7(RNA) strength depends on observed lack of aberrant RNA defect and assay quality.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (source_registry key 'cspec'; CSPEC doc 639508985) specifies BP7_Supporting for synonymous and deep intronic variants defined as further than (but not including) +7 at donor and -21 at acceptor sites, and reserves BP7_Strong/Moderate(RNA) for observed lack of an aberrant RNA defect at curator-discretion weight.
Variant context: NM_000051.4:c.2638+11A>G is intronic at donor +11 of intron 17 (VariantValidator genome_context_intronic_sequence NC_000011.10(NM_000051.4):c.2638+11A>G; variant_exonic_positions 17i), beyond the +7 boundary, so the VCEP deep-intronic BP7 rule applies.
No RNA/minigene data for this variant were found in the case bundle, so BP7 is applied at the default supporting strength with no BP7(RNA) modification.
SpliceAI Lookup (source_registry key 'spliceai') reports max delta score 0.00 for this variant, concordant with the absence of a predicted splice impact; this score is used for BP4, while BP7 is assigned from the VCEP's positional deep-intronic eligibility rule.
Still missing
No variant-specific RNA assay is available to assess BP7_Variable(RNA) or to confirm absence of an aberrant transcript; BP7 is assigned on the VCEP positional deep-intronic rule only.
PS1Not met: no pathogenic or likely pathogenic ATM variant is reported at the same nucleotide or donor motif.
Details ▾Hide ▴
Reasoning
The variant under assessment is the non-coding intronic change NM_000051.4:c.2638+11A>G (11 bp into the intron following the last coding base of exon 17), so it creates no amino-acid substitution and the VCEP ATM v1.5 standard missense PS1 sub-path ('same amino acid change as a previously established pathogenic variant') cannot apply. The remaining, applicable sub-path is the VCEP ATM v1.5 PS1 splicing table, which is available here (vcep_atm_ps1_1_5). That table requires a reference pathogenic or likely pathogenic variant, classified under VCEP specifications, located at the same nucleotide as the VUA or within the same donor/acceptor motif, AND requires the VUA's predicted splice event to precisely match the reference variant's predicted event with equal or higher prediction strength. Neither prerequisite is satisfied: no P/LP ATM variant at c.2638+11 (or within the same donor motif of intron 17) is present anywhere in the collected evidence, and SpliceAI predicts no splice impact for this variant (max delta 0.00; DS_AG 0.0, DS_AL 0.0, DS_DG 0.0, DS_DL 0.0), i.e. there is no predicted splice event to match against a reference. PS1 is therefore not met. ClinVar classifies the variant Likely benign without expert-panel review, which is noted only as context and was not used to infer any criterion.
Rule
ClinGen HBOP ATM VCEP v1.5 PS1: use for missense changes as long as splicing is ruled out for both alterations; use the ATM PS1 Splicing table for splicing variants with similar predictions or observations of splice defect. Splicing-table prerequisite: the VUA's predicted event must precisely match the reference P/LP variant's predicted event, and the VUA's prediction strength must be equal to or higher than the reference's.
Evidence considered
VUA is the intronic change NM_000051.4:c.2638+11A>G, with no amino-acid change (normalized protein consequence NP_000042.3:p.?), so the VCEP ATM v1.5 PS1 missense sub-path ('same amino acid change as a previously established pathogenic variant') is inapplicable by construction.
VCEP ATM v1.5 PS1 guide (ATM_PS1_1.5.pdf) splicing table: for a VUA located outside the donor/acceptor +/-1,2 dinucleotide positions with baseline PP3, a reference P variant at the same nucleotide gives PS1 and an LP reference gives PS1_Moderate; a reference within the same donor/acceptor motif gives PS1_Moderate (P reference) or PS1_Supporting (LP reference). Prerequisite for all rows: the predicted event of the VUA must precisely match the predicted event of the known (likely) pathogenic variant, and the VUA's prediction strength must be similar or higher than the reference's.
No P/LP ATM variant at c.2638+11 or at the same donor motif of intron 17 was found in the collected evidence; the only variant at nucleotide position 2638 in the bundle is the exonic missense c.2638G>A (p.Gly880Ser) listed in Suppl_TableS1_PMID 40580951, which is a different variant and not a splice-region comparator.
SpliceAI lookup for NM_000051.4:c.2638+11A>G reports max delta score 0.00 (DS_AG 0.0, DS_AL 0.0, DS_DG 0.0, DS_DL 0.0; DP_DL -11), i.e. no predicted loss or gain of a donor/acceptor site, so the VUA has no predicted splice event that could match a reference variant's event.
ClinVar variation 384288 lists NM_000051.4(ATM):c.2638+11A>G as Likely benign with three clinical-laboratory submissions and no expert-panel submission; used only to confirm no same-position P/LP comparator was available.
Still missing
No pathogenic/likely pathogenic ATM variant at c.2638+11 or within the same donor/acceptor motif of intron 17 was included in the evidence bundle, so the same-position comparator arm of the PS1 splicing table could not be populated; however this does not change the outcome because the VUA has no predicted splice event (SpliceAI 0.00) to match.
No RNA/protein functional data were collected for this variant; such data would be needed to establish any splice defect that could participate in the PS1 splicing path.
PS4Not met: no case-control study reports this variant, which is present in only 7 of 1,612,814 gnomAD alleles.
Details ▾Hide ▴
Reasoning
ATM VCEP v1.5 (ClinGen HBOP) lists PS4 as Applicable at Pathogenic Strong with a single quantitative rule: case-control studies with p-value <=.05 AND (odds ratio, hazard ratio, or relative risk >=2 OR lower 95% CI >=1.5); the VCEP explicitly states PS4_Moderate (proband counting) must not be used for ATM because it would require population-specific calibration for a gene causing a common disorder. No case-control or cohort study reporting c.2638+11A>G in affected versus control individuals was identified anywhere in the evidence bundle, and no OR/HR/RR, p-value, or affected-vs-control frequency comparison exists for this variant. All ClinVar-linked publications are gene-level guidelines or methodology papers (ACMG/AMP 2015 standards; EFNS/ENS chronic ataxia consensus; ACMG/NSGC cancer-referral guideline; Sherloc) and none reports this variant; the two retrieved full texts mention ATM only at gene level. ClinVar record 384288 contains no case-enrichment signal: three single-submitter clinical-laboratory submissions, all Likely benign, review status 'criteria provided, single submitter', zero expert-panel submissions. Population data are also uninformative for the rare-variant moderate path: gnomAD v4.1 shows 7/1,612,814 alleles (AF 4.34e-06) and gnomAD v2.1 shows 1/251,258 alleles, with no affected probands carrying this variant reported in any retrieved source. PS4 is therefore not met.
Rule
ATM VCEP v1.5 PS4 (Applicable, Pathogenic Strong): Case-control studies; p-value <=.05 AND (Odds ratio, hazard ratio, or relative risk >=2 OR lower 95% CI >=1.5). PS4_Moderate: Do not use.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 marks PS4 Applicable (default Pathogenic Strong) and gives exactly one rule: 'Case-control studies; p-value <=.05 AND (Odds ratio, hazard ratio, or relative risk >=2 OR lower 95% CI >=1.5).' The same specification states 'PS4_Moderate: Do not use. Proband counting for genes causing a common disorder need to be calibrated in a population-specific way before use.'
ClinVar record 384288 for NM_000051.4(ATM):c.2638+11A>G holds three submissions (GeneDx, SCV000525025; Color Diagnostics, SCV000682066; Labcorp Genetics/formerly Invitae, SCV002395369), all 'Likely benign', all 'criteria provided, single submitter', all is_expert_panel=false; submission audit counts expert_panel_submissions=0. No PS4-type case-enrichment or case-control signal is present in the record.
No retrieved publication reports a case-control, cohort, or proband series for this variant. The extracted full texts of PMID:24418350 (EFNS/ENS chronic ataxia consensus) and PMID:25394175 (ACMG/NSGC cancer-referral practice guideline) mention ATM only at gene level (ataxia telangiectasia as the second most frequent recessive ataxia; ATM mutations in 2.4% of familial pancreatic cancer) and never c.2638+11A>G; PMID:25741868 (ACMG/AMP 2015 standards) contains no variant or gene-level ATM data relevant to prevalence enrichment.
Control/population frequency data: gnomAD v4.1 reports 7/1,612,814 alleles (AF 4.34024e-06; grpmax FAF 7.39e-06) and gnomAD v2.1 reports 1/251,258 alleles (AF 3.97997e-06), homozygotes 0 in both. These frequencies are too low to demonstrate a case-control difference, and no affected carriers were identified to compute an odds ratio.
The 7 ClinVar-linked PMIDs (25394175, 20301790, 20301317, 20050888, 24418350, 25741868, 28492532) were screened for case-observation/enrichment content via the pre-computed paper extracts and the literature pass; none provides affected-individual counts for this exact variant.
Still missing
No case-control or cohort study reporting c.2638+11A>G in affected individuals versus controls was found; PS4 cannot be satisfied without such enrichment data.
The VCEP prohibits the PS4_Moderate proband-counting path for ATM, so absence of case-control data cannot be offset by counting reported probands.
ClinVar-linked PMIDs 20301317, 20301790, 20050888, and 28492532 were not retrieved as full text, so any variant-specific case observation in them could not be verified; their titles and abstracts identify them as gene-level reviews/guidelines with no expected variant mention.
PM5Not met: the variant is deep intronic with no predicted splice impact and introduces no premature termination codon.
Details ▾Hide ▴
Reasoning
Under the governing ClinGen HBOP ATM VCEP v1.5, PM5 is applicable only as PM5_Supporting and only for (a) frameshifting or truncating variants with a premature termination codon upstream of p.Arg3047, or (b) splice variants with a PTC upstream of p.Arg3047 where PVS1_VS(RNA) is applied on the basis of high-quality observed splicing impact and the transcript must be NMD-prone; the VCEP further states 'Do not use for missense changes' and 'Do not use for start-loss variants'. The variant under assessment, c.2638+11A>G, is a non-canonical deep intronic +11 change: SpliceAI predicts no splice impact (max delta 0.00), no RNA data were collected, no PTC is introduced, and the variant is not truncating. Neither PM5 branch's precondition is met. The classic same-residue missense PM5 mechanism is unavailable both because the variant is non-missense and because the VCEP explicitly forbids missense PM5 for ATM; consistently, the PM5 comparator collection returned zero candidates and could not even parse a residue (p.?). PM5 is therefore not met.
Rule
ClinGen HBOP ATM VCEP v1.5 PM5_Supporting: apply to frameshifting or truncating variants with premature termination codons upstream of p.Arg3047; apply to splice variants as with premature termination codons upstream of p.Arg3047 where PVS1_VS(RNA) is applied based on high quality observed splicing impact and must be NMD prone; do not use for missense changes or start-loss variants.
Evidence considered
CSPEC ATM v1.5 PM5 rule: 'Apply to frameshifting or truncating variants with premature termination codons upstream of p.Arg3047' and 'Apply to splice variants as with premature termination codons upstream of p.Arg3047 where PVS1_VS(RNA) is applied based on high quality observed splicing impact and must be NMD prone'; use as PM5_Supporting (not moderate); 'Do not use for start-loss variants'; 'Do not use for missense changes'.
The VUA is intronic (c.2638+11A>G) and is not a frameshift/truncating variant; it introduces no premature termination codon, and no PTC position upstream of p.Arg3047 is involved.
SpliceAI lookup reports max delta score 0.00 for this variant (no predicted donor/acceptor loss or gain), so there is no high-quality predicted or observed splice defect to support the PVS1_VS(RNA) prerequisite of the splice arm.
pm5_candidates.json reports found=false with 0 candidates seen and residue context unparseable (NP_000042.3:p.?), i.e. no same-residue comparator exists to evaluate.
The VCEP PVS1 decision guide is relevant only as the source of the PVS1_VS(RNA)/NMD condition referenced by the PM5 splice arm; no PVS1_VS(RNA) evidence was collected for this variant.
Still missing
No RNA-based splice assay or NMD assessment is available for this intronic variant; without an observed splice defect PVS1_VS(RNA) (and hence the splice arm of the ATM PM5 rule) cannot be triggered.
No same-residue comparator is applicable because the variant is intronic and the VCEP disallows missense PM5 for ATM.
PP3Not met: SpliceAI maximum delta score 0.00, well below the >=0.2 PP3 threshold.
Details ▾Hide ▴
Reasoning
NM_000051.4:c.2638+11A>G is intronic, 11 nt into intron 17 downstream of the exon 17 donor site, i.e. outside the canonical donor/acceptor +/-1,2 dinucleotides, so the ATM VCEP v1.5 PP3 splicing sub-path is the applicable mechanism (the missense/REVEL sub-path does not apply to an intronic variant). The VCEP applies PP3_Supporting for SpliceAI >=0.2 on this sub-path. The SpliceAI Lookup result for this exact variant is a maximum delta score of 0.00 (DS_AG 0.00, DS_AL 0.00, DS_DG 0.00, DS_DL 0.00; Pangolin SG 0.001, SL -0.001), well below the 0.2 threshold, so PP3 is not met. The VCEP restriction that splice-based PP3 may not be combined with PVS1/PVS1_Variable(RNA) is moot because PP3 is negative.
Rule
ATM VCEP v1.5 PP3_Supporting: missense REVEL >0.7333; splicing SpliceAI >=0.2 for silent, missense/in-frame and intronic variants outside donor and acceptor 1,2 sites.
Evidence considered
Governing framework: ClinGen HBOP ATM VCEP v1.5 (source_registry key 'cspec'; CSPEC doc 639508985) specifies PP3_Supporting for missense REVEL >0.7333 and, on the splicing sub-path, for SpliceAI >=0.2 in silent, missense/in-frame and intronic variants outside the donor and acceptor 1,2 sites.
Variant context: NM_000051.4:c.2638+11A>G is intronic (VariantValidator genome_context_intronic_sequence NC_000011.10(NM_000051.4):c.2638+11A>G; variant_exonic_positions 17i), 11 nt into intron 17 from the exon 17 donor site, so it is outside the canonical +/-1,2 splice positions and the VCEP splicing sub-path is the applicable mechanism.
SpliceAI Lookup for NM_000051.4:c.2638+11A>G (source_registry key 'spliceai') returned max delta score 0.00 (DS_AG 0.00, DS_AL 0.00, DS_DG 0.00, DS_DL 0.00; Pangolin SG 0.001, SL -0.001), below the VCEP PP3 threshold of >=0.2; PP3 is therefore not met.
Threshold provenance: the SpliceAI 0.2/0.1 cut-offs are the ATM VCEP v1.5 specification's own values, taken verbatim from CSPEC doc 639508985 (source_registry key 'cspec'), which states the VCEP uses SpliceAI (Jaganathan et al., 2019, Cell; PMID 30661751) as sole predictor with the SVI-recommended splice thresholds (Walker et al., 2023, Am J Hum Genet) for non-canonical splice variants.
The missense sub-path is not applicable to this intronic variant and no REVEL score was retrievable (REVEL lookup found=false, score null); the ATM VCEP supplementary Table S1 (PMID 40580951) contains no row for c.2638+11A>G, so no secondary computational corroboration is available.
Still missing
No RNA/minigene data for this variant are available; a negative SpliceAI result cannot exclude a splice effect with certainty, but the VCEP calibrates PP3 solely on SpliceAI and the score is 0.00.
PP5Not met: ClinVar contains zero expert-panel submissions, and all three laboratory submissions classify the variant Likely benign.
Details ▾Hide ▴
Reasoning
ATM VCEP v1.5 marks PP5 'Not Applicable for this VCEP' ('This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee'). Under the lab's PP5/BP6 override, PP5 may only be triggered by an exact-variant ClinVar expert-panel (3-star) Pathogenic/Likely pathogenic classification; ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions never trigger it. For c.2638+11A>G, ClinVar record 384288 has zero expert-panel submissions: all three submissions are single-submitter clinical laboratories (GeneDx, Color Diagnostics, Labcorp/formerly Invitae), each with review status 'criteria provided, single submitter', and the aggregate label is 'Likely benign' (i.e. in any case the wrong direction for PP5). No reputable-source expert-panel assertion of pathogenicity exists, so PP5 is not met.
Rule
ATM VCEP v1.5 PP5: Not Applicable for this VCEP (ClinGen SVI VCEP Review Committee). Lab override: trigger PP5 (Supporting) only on an exact-variant ClinVar expert-panel (3-star) Pathogenic/Likely pathogenic classification; ordinary lab submissions and aggregate labels never trigger it.
Evidence considered
ATM VCEP v1.5 PP5 entry: applicability 'Not Applicable for this VCEP'; instructionsToUse 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee.'
ClinVar record 384288 (NM_000051.4(ATM):c.2638+11A>G): review_status 'criteria provided, single submitter'; submission_summary only 'Likely benign - 3 clinical laboratories'; source_data.expert_panels is empty; submission audit counts expert_panel_submissions=0 and marks all three SCVs (SCV000525025 GeneDx, SCV000682066 Color Diagnostics, SCV002395369 Labcorp Genetics/formerly Invitae) is_expert_panel=false.
No expert-panel assertion (3-star) exists for this variant, so the exact-variant expert-panel trigger for PP5 is absent regardless of the ATM VCEP's own PP5 not-applicable status.
Still missing
No ClinGen/VCEP expert-panel classification of this exact variant exists in ClinVar; the record is 1-star single-submitter only. PP5 cannot be triggered unless an expert-panel P/LP assertion is added.
BA1Not met: gnomAD v4.1 grpmax filtering AF 0.000739%, roughly 676-fold below the 0.5% BA1 ceiling.
Details ▾Hide ▴
Reasoning
Not met. ATM VCEP v1.5 specifies BA1 at grpmax filtering AF >0.5% in gnomAD v4. gnomAD v4.1 gives grpmax FAF 7.39e-06 (0.000739%; exome grpmax FAF 8.35e-06), total AF 4.34e-06 (7/1,612,814 alleles), 0 homozygotes; the highest subpopulation frequency is East Asian AF 4.46e-05 (2/44,846 alleles, 0 homozygotes). The observed frequency is roughly 676-fold below the 0.5% ceiling, and gnomAD v2.1 independently corroborates extreme rarity (total AF 3.98e-06, 1/251,258 alleles, 0 homozygotes; EAS AF 5.44e-05). No ancestry, homozygote, prevalence/penetrance or threshold-artifact consideration brings this variant near BA1.
Rule
Grpmax Filtering AF >.5% in gnomAD v4 dataset (ATM VCEP v1.5) - observed grpmax FAF 7.39e-06 (0.000739%); threshold not met
Evidence considered
gnomAD v4.1 (source key gnomad_v4) reports grpmax filtering AF 7.39e-06 (0.000739%), exome grpmax FAF 8.35e-06, total AF 4.34e-06 (AC=7, AN=1,612,814), 0 homozygotes; highest subpopulation frequency is East Asian AF 4.45971e-05 (2/44,846 alleles, 0 homozygotes), with the remaining-individuals group at 1.60e-05 (1/62,466) and all other ancestry groups 0.
The ATM VCEP v1.5 BA1 rule is 'Grpmax Filtering AF >.5% in gnomAD v4 dataset'; the observed grpmax FAF of 0.000739% is ~676-fold below that ceiling, so BA1 is not met.
gnomAD v2.1 (source key gnomad_v2) corroborates extreme rarity: total AF 3.97997e-06 (1/251,258 alleles), 0 homozygotes, highest subpopulation East Asian AF 5.44e-05 (1/18,382 alleles, 0 homozygotes); grpmax FAF was null in this dataset.
Ancestry/homozygote review: the only ancestry group with more than one observed allele is East Asian (AC=2 of 44,846), with no homozygotes in any group in either dataset, so there is no ancestry-specific or homozygote signal that could support benign stand-alone frequency evidence.
Disease prevalence/penetrance context: ATM-related Ataxia-Telangiectasia is a rare recessive disorder (the EFNS/ENS consensus describes A-T as the second most frequent recessive ataxia globally), and ATM heterozygotes confer only reduced-penetrance cancer risk; the VCEP's 0.5% grpmax FAF ceiling already accommodates that phenotype spectrum, and this variant remains far below it.
No threshold artifact identified: the modest East Asian AN (44,846) does not inflate the estimate above the BA1 ceiling, and the grpmax FAF is the conservative metric the VCEP names for BA1.
Still missing
gnomAD Canada (HostSeq) is registered as a source (source key gnomad_canada) but no result was retrieved in this case; no additional population dataset was available to cross-check.
gnomAD v2.1 grpmax filtering AF was not computed/available (null), so BA1 was assessed on the v4.1 grpmax FAF the VCEP names plus v2.1 total/subpopulation AF.
BS1Not met: grpmax filtering AF 0.000739%, about 68-fold below the 0.05% BS1 ceiling.
Details ▾Hide ▴
Reasoning
Not met. ATM VCEP v1.5 specifies BS1 at grpmax filtering AF >0.05% in gnomAD v4. gnomAD v4.1 grpmax FAF is 7.39e-06 (0.000739%), ~68-fold below the 0.05% ceiling; total AF 4.34e-06 (7/1,612,814 alleles) with 0 homozygotes, and the highest subpopulation frequency is East Asian AF 4.46e-05 (2/44,846). gnomAD v2.1 corroborates (total AF 3.98e-06, 1/251,258, 0 homozygotes). The variant is far too rare for benign-strong frequency evidence under the VCEP specification.
Rule
Grpmax Filtering AF >.05% in gnomAD v4 dataset (ATM VCEP v1.5) - observed grpmax FAF 7.39e-06 (0.000739%); threshold not met
Evidence considered
gnomAD v4.1 (gnomad_v4): grpmax FAF 7.39e-06 (0.000739%), total AF 4.34e-06 (AC=7, AN=1,612,814), 0 homozygotes; East Asian AF 4.45971e-05 (2/44,846, 0 homozygotes).
The ATM VCEP v1.5 BS1 rule is 'Grpmax Filtering AF >.05% in gnomAD v4 dataset'; the observed grpmax FAF of 0.000739% is ~68-fold below this threshold, so BS1 is not met.
gnomAD v2.1 (gnomad_v2): total AF 3.97997e-06 (1/251,258 alleles), 0 homozygotes, East Asian AF 5.44e-05 (1/18,382); independently confirms the variant sits far below the BS1 ceiling.
Disease prevalence/penetrance context: BS1's 0.05% grpmax FAF ceiling is calibrated for the ATM hereditary breast/ovarian/pancreatic cancer and A-T phenotype spectrum; this variant's frequency is orders of magnitude lower and therefore provides no benign-frequency evidence.
No homozygote observations (0 in both v4.1 and v2.1) and no ancestry-specific enrichment approaching the BS1 threshold were identified.
Still missing
gnomAD Canada (HostSeq) registered but no result retrieved; no independent population dataset available to cross-check.
The VCEP's BS1 threshold is set for the ATM HBOP phenotype spectrum; no ATM-specific lower alternative threshold was supplied, so the published 0.05% grpmax FAF rule was applied as written.
BP6Not met: ClinVar contains zero expert-panel submissions, and the aggregate Likely benign label comes only from single laboratories.
Details ▾Hide ▴
Reasoning
ATM VCEP v1.5 marks BP6 'Not Applicable for this VCEP' ('This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee'). Under the lab's PP5/BP6 override, BP6 may only be triggered by an exact-variant ClinVar expert-panel (3-star) Benign/Likely benign classification; ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions never trigger it. For c.2638+11A>G, ClinVar record 384288 contains three submissions from clinical laboratories (GeneDx, Color Diagnostics, Labcorp/formerly Invitae), all with review status 'criteria provided, single submitter' and is_expert_panel=false; the submission audit counts expert_panel_submissions=0. Although the aggregate ClinVar label is Likely benign, that aggregate non-expert label cannot trigger BP6 under the exact-variant expert-panel rule, so BP6 is not met.
Rule
ATM VCEP v1.5 BP6: Not Applicable for this VCEP (ClinGen SVI VCEP Review Committee). Lab override: trigger BP6 (Supporting) only on an exact-variant ClinVar expert-panel (3-star) Benign/Likely benign classification; ordinary lab submissions and aggregate labels never trigger it.
Evidence considered
ATM VCEP v1.5 BP6 entry: applicability 'Not Applicable for this VCEP'; instructionsToUse 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee.'
ClinVar record 384288: three submissions, all 'Likely benign' with review_status 'criteria provided, single submitter'; source_data.expert_panels empty; audit expert_panel_submissions=0; all three SCVs flagged is_expert_panel=false (GeneDx SCV000525025; Color Diagnostics SCV000682066; Labcorp Genetics/formerly Invitae SCV002395369).
The aggregate 'Likely benign - 3 clinical laboratories' label is a non-expert aggregate and is explicitly excluded from triggering BP6 under the exact-variant expert-panel rule.
Still missing
No ClinGen/VCEP expert-panel classification of this exact variant exists in ClinVar; the record is 1-star single-submitter only. BP6 cannot be triggered unless an expert-panel B/LB assertion is added.
PS3Not assessed: no variant-specific functional assay exists, and the intronic change cannot be tested by the approved protein-rescue assays.
Details ▾Hide ▴
Reasoning
Governing framework is the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel (HBOP) ATM specification v1.5 (CSPEC doc 639508985), under which PS3 is applicable but explicitly may not be used at Strong ('Do not use as strong'); Moderate requires failure to rescue both an ATM-specific feature (e.g. phosphorylation of ATM-specific targets) AND radiosensitivity, and Supporting requires failure to rescue an ATM-specific feature only, with radiosensitivity-only results receiving no weight. The index variant NM_000051.4:c.2638+11A>G is an intronic substitution in intron 17 (exon 17i, non-canonical splice position, +11), so it produces no protein change and cannot be interrogated by the VCEP's three approved protein assays, all of which are stable transfection of ATM cDNA constructs (which contain no introns) into ATM-null LCLs: Mitui 2009 (PMID 18634022), Barone 2009 (PMID 19431188) and Scott 2002 (PMID 11805335). None of those three papers is present in this case's retrieved full-text set (fulltext_map.json contains only PMID 25741868, 25394175 and 24418350), so the variant cannot be confirmed to have been tested in any approved assay. The prime-editing saturation screen of Sun et al. 2025 (PMID 40580951, Table S1) was searched in full (all 29,374 rows) and contains no row for c.2638+11A>G, so it supplies no functional measurement for this variant. SpliceAI max delta score 0.00 (DS_AG/DS_AL/DS_DG/DS_DL all 0.0) is an in silico prediction, not a functional study, and cannot support PS3. No RNA-level functional data (minigene, patient cDNA/RT-PCR, aberrant splice product) exist in the bundle, and the VCEP routes RNA functional evidence to PVS1_Strength(RNA) rather than PS3. With no variant-specific functional assay of any kind, PS3 cannot be assessed in either direction and is returned as not_assessed.
Rule
ClinGen HBOP ATM VCEP v1.5 PS3: applicable; do not use as Strong. Moderate = variant fails to rescue both an ATM-specific feature (e.g. phosphorylation of ATM-specific targets) AND radiosensitivity; Supporting = variant fails to rescue an ATM-specific feature only (radiosensitivity-only results get no weight). Protein evidence must come from the VCEP-approved ATM assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335; kinase activity PS3_Supporting, radiosensitivity PS3_Moderate only if combinable with kinase activity); RNA-level functional evidence is routed to PVS1_Strength(RNA).
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) PS3 rule: applicable, but 'Do not use as strong'; Moderate = 'Use when a variant fails to rescue both an ATM specifc feature (e.g. phosphorylation of ATM-specific targets) AND radiosensitivity'; Supporting = 'Use when a variant fails to rescue an ATM specifc feature, only ... Do not use for radiosensitivity-only as that is not a feature specific to ATM deficiency'. The same CSPEC instructs: 'For protein, see detailed notes on ATM-specific assays; For RNA use code PVS1_Strength(RNA) and modulate strength based on assay quality and quantity (curator discretion).'
VCEP-approved assay table (clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx): all three approved kinase-activity assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335) use 'Stable transfection of ATM-based cDNA constructs in an ATM null cell line' with Western-blot readouts of ATM substrate phosphorylation; the radiosensitivity sheet adds Mitui 2009 and Scott 2002 (PMID 11805335). Because these constructs are cDNA-based they contain no intron 17, so an intronic +11 substitution is outside the assay's measurable scope, and the table itself states it is 'NOT a per-variant results table' and that the variant under assessment must be confirmed in the paper itself before any PS3/BS3 weight is taken.
None of PMID 18634022, PMID 19431188 or PMID 11805335 is present in this case's retrieved publications (fulltext_map.json lists only PMID 25741868, 25394175, 24418350), so the variant could not be confirmed to appear in an approved assay; no PS3_Supporting/PS3_Moderate weight can therefore be read off the VCEP table.
Sun et al. 2025 (PMID 40580951, vcep_suppl_tables1_pmid_40580951) Table S1 was searched directly across all 29,374 data rows: there is no row for c.2638+11A>G. Only exonic/adjacent-position rows exist at hg38_pos 108267342-108267343 (c.2638G>A/p.G880S, c.2638G>C/p.G880R, c.2638G>T/p.G880C, c.2638+1G>A, c.2638+1G>C), which are different variants; the index intronic variant was therefore not functionally measured. Even a directly measured (DeepATM_predicted=No) result would not qualify as a VCEP-approved assay, since this paper postdates and is not part of the VCEP v1.5 approved-assay table.
SpliceAI (spliceai) for NM_000051.4:c.2638+11A>G: max delta score 0.00 (DS_AG 0.0, DS_AL 0.0, DS_DG 0.0, DS_DL 0.0; Pangolin SG 0.001, SL -0.001) - no significant splice impact predicted. This is a computational prediction (PP3/BP4 domain) and does not constitute the well-established functional study PS3 requires.
PMID:25741868 (ACMG/AMP 2015, Richards et al.) consulted as background: it states that although splice-site variants are predicted to lead to a null effect, 'confirmation of impact requires functional analysis by either RNA or protein analysis' - the exact functional evidence that is absent here for this intronic variant.
Still missing
No variant-specific PS3 assay result for c.2638+11A>G in any VCEP-approved ATM assay (kinase activity or radiosensitivity); none of the three approved-assay papers (PMID 18634022, 19431188, 11805335) was retrieved in this case bundle, and their intron-free cDNA-construct design cannot interrogate an intron 17 +11 substitution.
No RNA-based functional assay (minigene splicing assay, or patient-derived cDNA/RT-PCR with and without NMD inhibition) demonstrating a splice defect for this variant; the VCEP routes such RNA evidence to PVS1_Strength(RNA), which is adjudicated elsewhere.
No directly measured functional result in Sun et al. 2025 (PMID 40580951) Table S1: the table was searched in full and contains no c.2638+11A>G row (DeepATM_predicted status therefore not applicable), and that paper is not one of the VCEP's approved assays in any case.
PM3Not assessed: no affected proband carries this variant, so the recessive case-point total is 0.0 against the 1.0-point threshold.
Details ▾Hide ▴
Reasoning
Governing framework is the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel Specifications to the ACMG/AMP Variant Interpretation Guidelines for ATM Version 1.5 (cspec, doc 639508985; ATM_PM3_BP2_1.5.pdf), which is applicable to PM3 and supersedes generic ACMG/AMP. Under this specification PM3 for the recessive condition (Ataxia Telangiectasia, A-T) is points-based per unrelated A-T proband: a 'confident' A-T phenotype scores 4.0 points when the co-occurring P/LP ATM variant is confirmed in trans, 2.0 when phase is unknown, 1.0 when the second variant is unidentified or is a VUS, and 2.0 for a homozygous proband (max 2 individuals); a 'consistent' phenotype scores half those values; the point total converts to PM3_Supporting (1.0), PM3_Moderate (2.0), PM3_Strong (4.0) or PM3_VeryStrong (>=8.0). The frequency prerequisite is satisfied (gnomAD v4.1 grpmax filtering AF 7.39e-06 = 0.00074%, well under the VCEP's 0.01% ceiling; gnomAD v2.1 AF 3.98e-06; no homozygotes in either dataset). However, the case bundle contains no A-T-affected proband carrying c.2638+11A>G, no co-occurring P/LP ATM variant, and no phase information: the literature triage identified no variant-specific paper, the three ClinVar submissions for VCV000384288 (GeneDx, Color Health, Labcorp/Invitae) contain no proband, zygosity, or phase data, and the screened full texts (PMID:25741868, PMID:24418350, PMID:25394175) neither mention the variant nor report ATM in-trans co-occurrences. Because no qualifying observation exists, the PM3 point total is 0.0 (below the 1.0-point PM3_Supporting threshold) and the criterion cannot be affirmatively adjudicated from the available data; it is therefore returned as not_assessed rather than not_met, with the missing evidence identified. Note that A-T is recessive, so a single heterozygous carrier observation would not generate PM3 points even if one existed.
Rule
ATM VCEP v1.5 PM3/BP2 table (PM3): points per unrelated A-T proband - phenotype confident: 4.0 confirmed in trans / 2.0 phase unknown / 1.0 second variant unidentified or VUS / 2.0 homozygous (max 2 individuals); phenotype consistent: half those values. PM3_Supporting = 1.0, PM3_Moderate = 2.0, PM3_Strong = 4.0, PM3_VeryStrong >= 8.0. Variant may not exceed general population frequency >0.01%; co-occurring P/LP variant must be classified under VCEP specifications; multiple unrelated cases are additive.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 PM3 specification (cspec doc 639508985; ATM_PM3_BP2_1.5.pdf): PM3 is a points-based criterion for the recessive A-T phenotype, scored per unrelated A-T proband - confident phenotype 4.0 points if the co-occurring P/LP variant is confirmed in trans, 2.0 if phase unknown, 1.0 if the second variant is unidentified or a VUS, 2.0 if homozygous (max 2 individuals); consistent phenotype scores half those values. Conversion: PM3_Supporting 1.0, PM3_Moderate 2.0, PM3_Strong 4.0, PM3_VeryStrong >=8.0. General considerations: the variant may not exceed a general population frequency of 0.01%; multiple unrelated cases are additive; co-occurring P/LP variants must themselves be classified under VCEP specifications; in-trans stipulation can be met by genotyping at least one parent, assumed with >=2 different P/LP co-occurrences, or inferred for homozygotes from consanguinity or bilineal ATM-related cancer family history.
Frequency eligibility for PM3 is satisfied but is not itself PM3 evidence: gnomAD v4.1 reports 7/1,612,814 alleles (AF 4.34e-06; grpmax filtering AF 7.39e-06 = 0.00074%), well below the VCEP's 0.01% ceiling, and gnomAD v2.1 reports 1/251,258 alleles (AF 3.98e-06); neither dataset reports a homozygote, and gnomAD-Canada v1.0 returned no result (connection reset).
No A-T proband or any co-occurring ATM variant was found for c.2638+11A>G: the three ClinVar submissions for VCV000384288 (GeneDx SCV000525025; Color Diagnostics SCV000682066; Labcorp Genetics/Invitae SCV002395369) contain no proband, zygosity, or phase data and cite only generic methodology papers (PMID:25741868, PMID:28492532).
The full-text extracts screened for this group (PMID:25741868, PMID:24418350, PMID:25394175) neither mention c.2638+11A>G nor report ATM in-trans co-occurrence or A-T proband data; they are a general ACMG/AMP guideline, an EFNS/ENS ataxia guideline, and an ACMG/NSGC cancer-referral guideline respectively.
ClinVar's final Likely benign classification was not used to infer PM3 status; only the absence of proband-level phase/co-occurrence evidence in the submission trails was considered.
Still missing
No A-T-affected proband carrying c.2638+11A>G has been reported in any retrieved source; at least one qualifying proband (or a homozygous affected observation) is required before any PM3 points can be scored.
No phase information (trans vs cis) exists for any carrier of c.2638+11A>G and no co-occurring P/LP ATM variant has been identified; parental genotyping, or co-occurrence with at least two different VCEP-classified P/LP variants, would be needed to award confirmed-in-trans points.
No A-T phenotype details (confident vs consistent, clinical features, laboratory results) are available for any carrier, so the per-proband point tier cannot be determined even hypothetically.
gnomAD-Canada v1.0 could not be queried (connection reset), so a Canadian-cohort homozygote/co-occurrence check was not completed.
PP1Not assessed: no pedigree, affected relative, or countable meiosis was available for segregation analysis.
Details ▾Hide ▴
Reasoning
Under the governing ATM VCEP v1.5 (cspec), PP1 is applicable only for the autosomal recessive condition (Ataxia-Telangiectasia) and requires that affected relatives carry both variants identified in the proband; strength is Supporting for segregation in 1 affected relative, Moderate for 2, and Strong for >=3. The AD (cancer predisposition) condition is explicitly not to be used for PP1 because co-segregation analysis in lower-penetrance genes can lead to false positives (PMID 32773770). The case bundle contains no pedigree, no family history, no affected relative genotyped for c.2638+11A>G, no second ATM variant phased in a proband, and no countable meioses; the three ClinVar submissions for VCV000384288 likewise carry no segregation data. PP1 therefore cannot be assessed rather than being met or refuted.
Rule
AR Condition: segregation in 1 affected relative = PP1_Supporting, 2 affected relatives = PP1_Moderate, >=3 affected relatives = PP1_Strong; affected relatives must have both variants identified in the proband; do not use for the AD condition (ATM VCEP v1.5).
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) PP1 entry: applicable; 'AR Condition: Affected relatives must have both variants identified in proband'; 'AD Condition - Do not use: Co-segregation analysis in lower-penetrance genes can lead to false positive results (PMID 32773770)'; strength rules AR Condition: Segregation in >=3 affected relatives = Strong, in 2 affected relatives = Moderate, in 1 affected relative = Supporting.
Case-evidence bundle contains no pedigree, family history, segregation testing, or affected-relative genotypes for c.2638+11A>G; no second ATM variant phase information for any proband.
ClinVar record VCV000384288 (NM_000051.4(ATM):c.2638+11A>G; 3 submissions: GeneDx, Color Health, Labcorp Genetics) contains no segregation or family data; the only criterion signal recorded is an in-silico PP3-type statement, so nothing in ClinVar addresses PP1.
None of the fetched full texts (PMID:25741868, PMID:24418350, PMID:25394175) mentions the variant or reports ATM family segregation data (paper_extracts.json).
Still missing
No family history or pedigree data in the case bundle for c.2638+11A>G
No affected relatives genotyped for the variant, and no documentation that affected relatives carry both ATM variants identified in the proband (AR PP1 prerequisite)
No meioses countable for the VCEP PP1 strength tiers (1 affected relative = Supporting, 2 = Moderate, >=3 = Strong)
BS3Not assessed: no variant-specific functional result exists, and the approved protein-rescue assays cannot assess an intronic change.
Details ▾Hide ▴
Reasoning
Under the governing ClinGen HBOP ATM v1.5 specification, BS3 is applicable and is met at Moderate when a variant rescues both an ATM-specific feature (e.g. phosphorylation of ATM-specific targets) AND radiosensitivity, or at Supporting when it rescues either an ATM-specific feature OR radiosensitivity; radiosensitivity-only and conflicting kinase/radiosensitivity results get no weight. Both VCEP paths are protein-rescue assays performed by stable transfection of ATM cDNA constructs into ATM-null LCLs (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335), which cannot assess the intronic variant NM_000051.4:c.2638+11A>G (intron 17, exon 17i, non-canonical +11 position; no protein change). No variant-specific rescue/kinase/radiosensitivity result exists in this bundle, none of the three approved-assay papers was retrieved (fulltext_map.json lists only PMID 25741868, 25394175, 24418350), and the variant has no row in Sun et al. 2025 (PMID 40580951) Table S1 (searched in full, 29,374 rows). SpliceAI max delta score 0.00 predicts no splice impact but is in silico only and cannot substitute for a functional assay. For RNA, the VCEP directs benign functional evidence to BP7_RNA rather than BS3, and no RNA assay data (minigene, patient cDNA/RT-PCR) are present. BS3 is therefore returned as not_assessed for insufficient variant-specific functional evidence.
Rule
ClinGen HBOP ATM VCEP v1.5 BS3: Moderate = variant rescues both an ATM-specific feature (e.g. phosphorylation of ATM-specific targets) AND radiosensitivity; Supporting = variant rescues either an ATM-specific feature OR radiosensitivity. Protein evidence must come from the VCEP-approved ATM assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335); radiosensitivity-only and conflicting kinase/radiosensitivity results get no weight; RNA-level benign functional evidence is routed to BP7_RNA.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) BS3 rule: Moderate = 'Use when a variant rescues both an ATM specifc feature (e.g. phosphorylation of ATM-specific targets) AND radiosensitivity'; Supporting = 'Use when a variant rescues EITHER an ATM specifc feature OR rescues radiosensitivity'. The CSPEC instructs: 'For protein, see detailed notes on ATM-specific assays; For RNA use code BP7_RNA and modulate strength based on assay quality and quantity (curator discretion).'
VCEP-approved assay table (clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx): approved assays are 'Stable transfection of ATM-based cDNA constructs in an ATM null cell line' (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335) for kinase activity, with radiosensitivity approved only for Mitui 2009 (PMID 18634022) and requiring combination with a kinase result ('No weight should be applied if only radiosensitivity is available'). These cDNA constructs contain no introns, so intron 17 c.2638+11A>G lies outside the assay's scope, and the table is not a per-variant results table.
No variant-specific BS3 result could be read from the VCEP table: PMID 18634022, PMID 19431188 and PMID 11805335 are absent from this case's full-text set (fulltext_map.json: only PMID 25741868, 25394175, 24418350), so the variant's presence in an approved assay is unconfirmed.
Sun et al. 2025 (PMID 40580951) Table S1 searched in full (29,374 rows): no row for c.2638+11A>G, so no directly measured benign/functional result exists from the prime-editing screen; only different variants at hg38_pos 108267342-108267343 (c.2638G>A/C/T, c.2638+1G>A/C) are present. The paper is not a VCEP-approved assay regardless.
SpliceAI (spliceai): max delta score 0.00 for c.2638+11A>G (all DS components 0.0), i.e. no predicted splice impact; this is computational and, per the VCEP, benign splice evidence for RNA must come from a laboratory assay (BP7_RNA), not from in silico prediction, so it does not satisfy BS3.
PMID:25741868 (ACMG/AMP 2015) consulted as background for the BS3 definition ('well-established in vitro or in vivo functional studies show no damaging effect on protein function or splicing') and for the principle that assay validity depends on how closely the assay reflects the biological environment; it reports no data on this variant.
Still missing
No variant-specific BS3 rescue/kinase-activity or radiosensitivity result for c.2638+11A>G in any VCEP-approved ATM assay; the approved assays use intron-free ATM cDNA constructs and none of PMID 18634022, 19431188 or 11805335 was retrieved in this bundle to confirm the variant was tested.
No RNA-based functional assay (minigene, patient-derived cDNA/RT-PCR with and without NMD inhibition, allelic imbalance measurement) demonstrating absence of a splice defect; the VCEP routes such benign RNA evidence to BP7_RNA, which is adjudicated elsewhere.
No directly measured result for this variant in Sun et al. 2025 (PMID 40580951) Table S1 (searched in full; no c.2638+11A>G row), and that source is not a VCEP-approved functional assay in any case.
BP2Not assessed: no unaffected adult carrier was reported, giving 0.0 points against the -1.0 threshold.
Details ▾Hide ▴
Reasoning
Governing framework is the ClinGen HBOP ATM VCEP v1.5 (cspec, doc 639508985; ATM_PM3_BP2_1.5.pdf), which lists BP2 as applicable and points-based. BP2 points are assigned per unaffected (non-A-T) adult proband older than 18 years who carries the variant under assessment together with a P/LP ATM variant: -4.0 points if confirmed in trans, -2.0 if phase unknown, and -2.0 (laboratory setting) or -1.0 (database setting) for homozygotes (max -2.0). Conversion: BP2_Supporting = -1.0, BP2_Moderate = -2.0, BP2_Strong <= -4.0. Footnote 3 additionally requires that the variant under assessment not be bioinformatically predicted or experimentally proven to have a clinically important effect on protein or mRNA splicing; that prerequisite is satisfied here (SpliceAI max delta score 0.00; Pangolin SG 0.001, SL -0.001), so the variant is not disqualified from BP2 on splicing grounds. However, no qualifying observation exists: no unaffected adult carrier of c.2638+11A>G with a P/LP ATM variant in trans (or in cis) is reported in any retrieved source, and no homozygous carrier is reported (gnomAD v2.1 and v4.1 both show 0 homozygotes; gnomAD-Canada v1.0 was unavailable). VCEP footnote 4 further excludes homozygotes observed in gnomAD frequency data from BP2 scoring (they are already captured by BA1/BS1), and age/A-T status of the gnomAD carriers is unknown. With 0.0 BP2 points (above the -1.0 BP2_Supporting threshold), the criterion cannot be affirmatively adjudicated from the available data and is returned as not_assessed rather than not_met, with the missing evidence identified.
Rule
ATM VCEP v1.5 PM3/BP2 table (BP2): points per unaffected (non-A-T) adult proband >18 years - confirmed in trans -4.0 / phase unknown -2.0 / homozygous -2.0 laboratory or -1.0 database setting (max -2.0). BP2_Supporting = -1.0, BP2_Moderate = -2.0, BP2_Strong <= -4.0. Footnote 3: the variant under assessment should not be predicted or proven to have a clinically important effect on protein or mRNA splicing. Footnote 4: database homozygotes count only for phenotyped individuals from clinical/research cohorts, not frequency datasets.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 BP2 specification (cspec doc 639508985; ATM_PM3_BP2_1.5.pdf): BP2 is points-based per unaffected (non-A-T) adult proband >18 years carrying the variant under assessment with a P/LP ATM variant - confirmed in trans -4.0, phase unknown -2.0, homozygous -2.0 (laboratory setting) or -1.0 (database setting), maximum -2.0. Conversion: BP2_Supporting -1.0, BP2_Moderate -2.0, BP2_Strong <= -4.0. Footnote 3: the variant under assessment should not be predicted or proven to have a clinically important effect on protein or mRNA splicing. Footnote 4: homozygous observations apply only to phenotyped individuals from clinical or research cohorts, not to data used for frequency-based codes (gnomAD homozygotes are already captured by BA1/BS1). Footnote 3 of the PM3/BP2 table also notes that individuals assessed for downgrade should be 18 years or older with no evidence of A-T.
The BP2 footnote-3 prerequisite is met on its face: SpliceAI predicts no significant splice impact for NM_000051.4:c.2638+11A>G (max delta score 0.00; DS_AG/DS_AL/DS_DG/DS_DL all 0.0; Pangolin SG 0.001, SL -0.001), so the variant is not excluded from BP2 by a predicted splicing effect; this alone does not supply BP2 points.
No qualifying unaffected-adult co-occurrence was found: gnomAD v4.1 reports 7/1,612,814 alleles (AF 4.34e-06) and gnomAD v2.1 reports 1/251,258 alleles (AF 3.98e-06), both with 0 homozygotes; gnomAD-Canada v1.0 returned no result (connection reset). The VCEP excludes gnomAD frequency-dataset homozygotes from BP2 scoring in any case.
The three ClinVar submissions for VCV000384288 (GeneDx, Color Health, Labcorp/Invitae) provide no unaffected-adult co-occurrence, zygosity, or phase data; ClinVar's final Likely benign classification was not used to infer BP2.
No publication screened for this group (PMID:25741868, PMID:24418350, PMID:25394175) mentions c.2638+11A>G or reports ATM in-trans observations in unaffected individuals.
Still missing
No unaffected (non-A-T) adult carrier of c.2638+11A>G with a P/LP ATM variant in trans (or in cis) has been reported; at least one qualifying observation is required before any BP2 points can be scored.
No homozygous carrier of c.2638+11A>G is reported in gnomAD v2.1 or v4.1, and homozygotes from frequency datasets are excluded from BP2 scoring by VCEP footnote 4 regardless.
Age (>=18 years) and A-T status are unknown for the individuals carrying this variant in gnomAD, so they cannot be used as BP2 probands even if a co-occurring variant were present.
gnomAD-Canada v1.0 lookup failed (connection reset); COSMIC and OncoKB returned no usable data, so additional population/co-occurrence sources could not be checked for BP2-relevant observations.
Not applicable · 12· recorded as not applicableShow reasons ▾Hide reasons ▴PVS1N/ANot applicable: an intronic +11 change outside the canonical +/-1,2 donor sites, with no null effect predicted.
Details ▾Hide ▴
Reasoning
c.2638+11A>G is an intronic substitution 11 nucleotides into intron 17 of ATM (VariantValidator exonic position '17i'; c.2638 is the last exonic nucleotide of exon 17, so the variant lies just outside the canonical donor dinucleotide). The governing ClinGen HBOP ATM VCEP v1.5 PVS1 specification restricts PVS1 to null variants: nonsense, frameshift, canonical +/-1,2 splice sites, initiation codon, and single/multi-exon deletion. This variant is none of these, and no null effect is predicted: SpliceAI reports DS_AG 0.00, DS_AL 0.00, DS_DG 0.00, DS_DL 0.00 (max delta score 0.00; Pangolin SG 0.001, SL -0.001), with DP_DL = -11 confirming the variant is 11 nt from the donor site and does not disrupt it. The ATM VCEP allows PVS1 for predicted splice defects only at canonical +/-1,2 positions; for non-canonical splice-region variants it requires observed RNA evidence (PVS1_Strength(RNA)), and no RNA or protein data exist for this variant in any source reviewed. PVS1 is therefore not applicable to this variant; the relevant splice evidence (SpliceAI <=0.1) belongs to the BP4/BP7 in-silico splicing group, not to PVS1.
Rule
ClinGen HBOP ATM VCEP v1.5 PVS1: null variants only (nonsense, frameshift, canonical +/-1,2 splice sites, initiation codon, single/multi-exon deletion); predicted splice defects outside canonical +/-1,2 require observed RNA evidence (PVS1_Strength(RNA)); no RNA evidence available -> PVS1 not applicable.
Evidence considered
ATM VCEP v1.5 (cspec doc 639508985) PVS1 definition: 'Null variant (nonsense, frameshift, canonical +/-1 or 2 splice sites, initiation codon, single or multi-exon deletion) in a gene where loss of function (LOF) is a known mechanism of disease', with the VCEP instruction 'Use ATM PVS1 Decision Tree' and the ATM PVS1/PVS1(RNA) guide (adapted from PMID 30192042).
VariantValidator annotation (prefetch/evidence bundle): NM_000051.4:c.2638+11A>G maps to NC_000011.10:g.108267353A>G (chr11:108267353 A>G, hg38), variant_exonic_positions start/end exon '17i', predicted protein consequence NP_000042.3:p.? - i.e. an intronic variant with no predicted amino-acid change.
SpliceAI lookup for NM_000051.4:c.2638+11A>G: DS_AG 0.0, DS_AL 0.0, DS_DG 0.0, DS_DL 0.0, max_delta_score 0.00, DP_DL -11, Pangolin SG 0.001 / SL -0.001; SpliceAI predicts no significant splice impact for this variant.
ClinGen SVI PVS1 recommendations (PMC6185798, the basis of the ATM VCEP PVS1 guide adapted from PMID 30192042): PVS1 for predicted splice effects is limited to canonical +/-1,2 sites; non-canonical splice-region variants require observed RNA evidence before a null effect can be assigned.
ACMG/AMP 2015 standards (PMID:25741868), PVS1 section: 'Although splice site variants are predicted to lead to a null effect, confirmation of impact requires functional analysis by either RNA or protein analysis.' No such functional analysis is available for this variant.
No full-text paper reviewed for this case (PMID:25741868, PMID:25394175, PMID:24418350) names c.2638+11A>G or reports RNA/protein data for it, so the PVS1_Strength(RNA) path is unavailable.
Still missing
No RNA or protein functional study for c.2638+11A>G was identified, so an observed splice defect (PVS1_Strength(RNA)) cannot be evaluated; absent such data PVS1 remains non-applicable rather than met at any strength.
The ATM VCEP PVS1 decision-tree figure pages in the retrieved PDF are image-only, so the tree's finer branches could not be text-verified; the top-level eligibility (null-variant classes plus canonical +/-1,2 for predicted splice defects) is unambiguous from the specification text and the cspec entry.
PS2VCEPNot applicable: de novo occurrences are not informative for ATM, and no parental testing was available.
Details ▾Hide ▴
Reasoning
Governing framework is the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel Specifications to the ACMG/AMP Variant Interpretation Guidelines for ATM Version 1.5 (cspec). The VCEP lists PS2 with applicability 'Not applicable' and the instruction 'Do not use for AD or AR disease: Informative de novo occurrences have not yet been observed and de novo AR conditions are unlikely to be informed by phase.' All strength tiers (Supporting, Moderate, Strong, Very Strong, Stand Alone) are likewise marked 'Not applicable' with no rule text, so PS2 cannot be applied to an ATM variant under this specification regardless of case data. Independently, the case bundle contains no proband, no parental testing (maternity/paternity), and no de novo observation for NM_000051.4:c.2638+11A>G.
VCEP rule
Not applicable - ATM VCEP v1.5: 'Do not use for AD or AR disease: Informative de novo occurrences have not yet been observed and de novo AR conditions are unlikely to be informed by phase.'
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) PS2 entry: applicability 'Not applicable'; instructionsToUse 'Do not use for AD or AR disease: Informative de novo occurrences have not yet been observed and de novo AR conditions are unlikely to be informed by phase'; every strength tier (Supporting, Moderate, Strong, Very Strong, Stand Alone) is also marked 'Not applicable' with an empty rule.
The case-evidence bundle (review_manifest.json, case_summary.json, prefetch.json, evidence.json, literature_pass.json, paper_extracts.json) contains no proband section, no parental testing, no family history, and no assertion of de novo occurrence for c.2638+11A>G.
Generic ACMG/AMP 2015 (PMID:25741868) defines PS2 as a de novo occurrence with both maternity and paternity confirmed in a patient with the disease and no family history, and notes paternity confirmation alone is insufficient; because the ATM VCEP explicitly excludes PS2 for this gene, the generic definition does not restore applicability.
PM1VCEPNot applicable: ATM has no approved critical-domain table, and pathogenic and benign variants occur within the same domains.
Details ▾Hide ▴
Reasoning
The governing ClinGen HBOP ATM VCEP v1.5 specification explicitly marks PM1 as Not Applicable, with the instruction: 'Do not use: Benign and pathogenic variants are known to occur within the same domains and germline mutational hotspots are not well defined at this time.' In addition, vcep_materials.json contains an empty domain_tables object ({}), i.e. the governing VCEP supplies no approved critical-functional-domain table for ATM, so there is no authoritative domain list against which a position could be checked; this was verified before concluding PM1 is not met/not applicable. The variant is also intronic (c.2638+11A>G) and encodes no protein residue, so no domain/hotspot mapping is possible. PM1 is therefore not applicable to this gene/variant under the applicable specification.
VCEP rule
ClinGen HBOP ATM VCEP v1.5 PM1: Not Applicable. 'Do not use: Benign and pathogenic variants are known to occur within the same domains and germline mutational hotspots are not well defined at this time.'
Evidence considered
CSPEC ATM v1.5 lists PM1 with applicability 'Not applicable' and the instruction 'Do not use: Benign and pathogenic variants are known to occur within the same domains and germline mutational hotspots are not well defined at this time'.
vcep_materials.json 'domain_tables' key is an empty object ({}); there is no ATM entry, so the VCEP provides no authoritative, complete list of approved critical functional domains for this gene to check the variant against.
The variant NM_000051.4:c.2638+11A>G is intronic with protein consequence p.?, so it has no codon/residue or domain position in the first place.
Still missing
None required: PM1 is disabled by the governing VCEP specification for ATM and the VCEP provides no domain table (domain_tables is empty).
PM4N/ANot applicable: the variant is intronic and causes no protein length change, which the criterion requires.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 PM4 entry is Applicable to the gene but explicitly narrows the criterion: 'Do not use for in-frame insertions or deletions less than a single exon; Use for stop-loss variants, only.' c.2638+11A>G is a single-nucleotide intronic substitution with no in-frame deletion/insertion and no stop-loss; VariantValidator predicts no protein consequence (NP_000042.3:p.?) and no protein length change is present or predicted. Because the criterion's defining event (a protein length change) is absent, PM4 is not applicable to this variant.
Rule
ClinGen HBOP ATM VCEP v1.5 PM4: 'Do not use for in-frame insertions or deletions less than a single exon; Use for stop-loss variants, only.' Variant is neither -> not applicable.
Evidence considered
ATM VCEP v1.5 (cspec doc 639508985) PM4 entry: original ACMG summary 'Protein length changes due to in-frame deletions/insertions in a non-repeat region or stop-loss variants'; instruction 'Do not use for in-frame insertions or deletions less than a single exon; Use for stop-loss variants, only.'; strength rule 'Use for stop-loss variants.'
VariantValidator annotation for NM_000051.4:c.2638+11A>G: intronic position (exon '17i'), hgvs_predicted_protein_consequence NP_000042.3:p.? - no protein length change; the variant is a substitution, not an in-frame indel or stop-loss.
ACMG/AMP 2015 (PMID:25741868) defines PM4 as protein length changes due to in-frame deletions/insertions in a non-repeat region or stop-loss variants; neither mechanism is present here.
SpliceAI max delta score 0.00 argues against any splice-mediated protein-length change (e.g. in-frame exon skipping) that could otherwise bring PM4 into scope.
Still missing
PM4 would become evaluable only if an in-frame protein-length-altering event (stop-loss, or an in-frame single/multi-exon deletion or insertion) were demonstrated for ATM; none is present or predicted for c.2638+11A>G.
PM6VCEPNot applicable: assumed de novo occurrences are not informative for ATM, and none was observed.
Details ▾Hide ▴
Reasoning
Under the governing ATM VCEP v1.5 (cspec), PM6 (assumed de novo without confirmation of maternity and paternity) is listed with applicability 'Not applicable' and the same instruction as PS2: 'Do not use for AD or AR disease: Informative de novo occurrences have not yet been observed and de novo AR conditions are unlikely to be informed by phase.' No strength tier is applicable. PM6 is therefore not a usable criterion for ATM regardless of evidence; additionally, no assumed de novo observation of c.2638+11A>G exists in the case bundle.
VCEP rule
Not applicable - ATM VCEP v1.5 excludes PM6 for AD and AR disease.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) PM6 entry: applicability 'Not applicable'; instructionsToUse 'Do not use for AD or AR disease: Informative de novo occurrences have not yet been observed and de novo AR conditions are unlikely to be informed by phase'; all strength tiers marked 'Not applicable' with an empty rule.
Case bundle contains no observation of c.2638+11A>G as an assumed de novo event in any proband, publication, or ClinVar submission.
PP2VCEPNot applicable: ATM has no defined low rate of benign missense variation, and this variant is not missense.
Details ▾Hide ▴
Reasoning
The governing ClinGen HBOP ATM VCEP v1.5 specification explicitly marks PP2 as Not Applicable with the instruction: 'Do not use: ATM does not have a defined low rate of missense benign variation.' Additionally, the variant under assessment is an intronic +11 change (c.2638+11A>G) and is not a missense variant, so PP2's premise (a missense variant in a gene with a low rate of benign missense variation where missense is a common disease mechanism) does not arise. PP2 is therefore not applicable.
VCEP rule
ClinGen HBOP ATM VCEP v1.5 PP2: Not Applicable. 'Do not use: ATM does not have a defined low rate of missense benign variation.'
Evidence considered
CSPEC ATM v1.5 lists PP2 with applicability 'Not applicable' and the instruction 'Do not use: ATM does not have a defined low rate of missense benign variation'.
The variant NM_000051.4:c.2638+11A>G is intronic (non-missense), so the criterion's missense-variant premise is not satisfied.
Still missing
None required: PP2 is disabled by the governing VCEP specification for ATM and the variant is non-missense.
PP4VCEPNot applicable: phenotype specificity is already built into the ATM PM3/BP2 table rather than scored separately.
Details ▾Hide ▴
Reasoning
The governing ATM VCEP v1.5 (ClinGen HBOP) specification marks PP4 'Not applicable' for every strength level, with the explicit instruction: for autosomal dominant disease 'do not use as breast cancer is a disease with multiple genetic etiology (genetic heterogeneity) and there are no features that can readily distinguish hereditary from sporadic causes'; for autosomal recessive disease 'do not use as a separate line of evidence. Such evidence is built into the Ataxia Telangiectasia PM3|BP2 table'. Because ATM phenotype specificity is already accounted for inside the VCEP's PM3/BP2 points table, PP4 is not scored as an independent line of evidence under this framework. Independently, no proband-level phenotype or family-history data for this variant exist in the bundle that could be evaluated even under a generic ACMG/AMP reading.
VCEP rule
ATM VCEP v1.5 PP4: Not applicable for all strengths (AD: genetic heterogeneity, no features distinguishing hereditary from sporadic; AR: evidence built into the ATM PM3/BP2 points table).
Evidence considered
ATM VCEP v1.5 PP4 entry: applicability 'Not applicable'; instructionsToUse 'Autosomal Dominant: do not use as breast cancer is a disease with multiple genetic etiology (genetic heterogeneity) and there are no features that can readily distinguish hereditary from sporadic causes. Autosomal Recessive: do not use as a separate line of evidence. Such evidence is built into the Ataxia Telangiectasia PM3|BP2 table'; all strength_rules (Stand Alone, Very Strong, Strong, Moderate, Supporting) are marked 'Not applicable'.
No proband-level phenotype/family-history description tied to c.2638+11A>G was found in the bundle; ClinVar condition labels for the exact record are 'not specified' (SCV000525025) or empty (SCV000682066, SCV002395369), so no phenotype-specificity assessment is possible even outside the VCEP framework.
BS2VCEPNot applicable: the ATM specification disables BS2, and population data are covered under BA1, BS1, and PM2.
Details ▾Hide ▴
Reasoning
Not applicable. The ATM VCEP v1.5 specification marks BS2 as 'Not applicable' for this VCEP (no strength rules supplied), so BS2 cannot be applied to this variant regardless of the population data. The variant's population-frequency information is fully captured under BA1, BS1 and PM2.
VCEP rule
Not Applicable per ATM VCEP v1.5
Evidence considered
ATM VCEP v1.5 (source key cspec) lists BS2 with applicability 'Not applicable' and an empty strength-rule set, i.e. the criterion is switched off for ATM by the expert panel.
No BS2-type observation (variant found in unaffected homozygous or compound-heterozygous adults) was available in the case bundle; the population data available (gnomAD v4.1/v2.1) are already assessed under BA1/BS1/PM2.
BS4VCEPNot applicable: lack of co-segregation in A-T families is too rare to carry weight and is instead routed to BP2.
Details ▾Hide ▴
Reasoning
Under the governing ATM VCEP v1.5 (cspec), BS4 is listed with applicability 'Not applicable' and no applicable strength tier. The VCEP instructions state: 'AD Condition: Co-segregation analysis in low penetrance genes can lead to false positive results (PMID 32773770); AR Condition: Informative instances of lack of co-segregation in A-T families are too rare to be considered for weight at this time and can also be considered for BP2 if biallelic unaffected patients are observed in an A-T family.' BS4 is therefore excluded for ATM under both inheritance models; documented lack of co-segregation in an A-T family is instead routed to BP2, and no such observation exists in the case bundle in any event.
VCEP rule
Not applicable - ATM VCEP v1.5 excludes BS4; lack of co-segregation in A-T families may be considered under BP2 instead.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) BS4 entry: applicability 'Not applicable'; instructionsToUse 'AD Condition: Co-segregation analysis in low penetrance genes can lead to false positive results (PMID 32773770); AR Condition: Informative instances of lack of co-segregation in A-T families are too rare to be considered for weight at this time and can also be considered for BP2 if biallelic unaffected patients are observed in an A-T family.'; all strength tiers (Supporting through Stand Alone) marked 'Not applicable'.
Case-evidence bundle contains no family non-segregation data (no biallelic unaffected individuals, no affected relative lacking the variant), and in any case the specification routes such observations to BP2 rather than BS4.
BP1VCEPNot applicable: missense pathogenic variants are known for ATM, and this variant is intronic rather than missense.
Details ▾Hide ▴
Reasoning
The governing ClinGen HBOP ATM VCEP v1.5 specification explicitly marks BP1 as Not Applicable with the instruction: 'Do not use: Missense pathogenic variants are known for ATM.' Additionally, the variant under assessment is an intronic +11 change (c.2638+11A>G) and is not a missense variant, so BP1's premise (a missense variant in a gene where primarily truncating variants cause disease) does not arise. BP1 is therefore not applicable.
VCEP rule
ClinGen HBOP ATM VCEP v1.5 BP1: Not Applicable. 'Do not use: Missense pathogenic variants are known for ATM.'
Evidence considered
CSPEC ATM v1.5 lists BP1 with applicability 'Not applicable' and the instruction 'Do not use: Missense pathogenic variants are known for ATM'.
The variant NM_000051.4:c.2638+11A>G is intronic (non-missense), so the criterion's missense-variant premise is not satisfied.
Still missing
None required: BP1 is disabled by the governing VCEP specification for ATM and the variant is non-missense.
BP3VCEPNot applicable: the ATM specification disables BP3, and no in-frame deletion or insertion is present.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 specification disables BP3 entirely: applicability 'Not applicable' with the instruction 'Do not use.' Independently, the generic BP3 premise (an in-frame deletion/insertion within a repetitive region without known function) is absent here, because c.2638+11A>G is a single-nucleotide intronic substitution with no in-frame indel; ATM's HEAT-repeat-domain context is therefore never reached. BP3 is not applicable.
ATM VCEP v1.5 (cspec doc 639508985) BP3 entry: applicability 'Not applicable', instruction 'Do not use.', with the original ACMG summary 'In frame-deletions/insertions in a repetitive region without a known function.'
VariantValidator annotation for NM_000051.4:c.2638+11A>G: single-nucleotide intronic substitution (exon '17i') with predicted protein consequence NP_000042.3:p.?; no in-frame deletion or insertion is present, so the criterion's defining event is absent independent of the VCEP's 'Do not use' instruction.
BP5VCEPNot applicable: ATM's low penetrance makes co-occurring pathogenic variants too common to be informative.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 specification marks BP5 'Not applicable' for all strengths and instructs: 'Do not use: Cases with multiple pathogenic variants have been observed with no noticeable difference in phenotype (e.g. BRCA1 and BRCA2). In addition, ATM has low penetrance and will naturally occur with other pathogenic variants more frequently due to higher tolerance/presence in the general population.' BP5 is therefore not scored as an independent benign line of evidence under the governing framework. No case-level data for this variant (no proband with an alternate molecular basis of disease) exist in the bundle in any event.
VCEP rule
ATM VCEP v1.5 BP5: Not applicable for all strengths (do not use; multiple pathogenic variants occur without phenotype difference and ATM is low-penetrance with higher general-population tolerance).
Evidence considered
ATM VCEP v1.5 BP5 entry: applicability 'Not applicable'; instructionsToUse 'Do not use: Cases with multiple pathogenic variants have been observed with no noticeable difference in phenotype (e.g. BRCA1 and BRCA2). In addition, ATM has low penetrance and will naturally occur with other pathogenic variants more frequently due to higher tolerance/presence in the general population.'
No case-level data linking this variant to a proband with an alternate molecular basis for disease were found in the bundle; the ClinVar submissions carry no case detail beyond a Likely benign classification.
AF 4.3e-06 · 7 / 1,612,814 alleles · 0 hom · FAF 7.4e-06
East Asian2 / 44,846
4.5e-05
Remaining individuals1 / 62,466
1.6e-05
European (non-Finnish)4 / 1,178,994
3.4e-06
0 alternate alleles recorded in 7 other groups: Admixed American, European (Finnish), Amish, Middle Eastern, South Asian, Ashkenazi Jewish, African/African American.
gnomAD v2.1
AF 4e-06 · 1 / 251,258 alleles · 0 hom
East Asian1 / 18,382
5.4e-05
0 alternate alleles recorded in 7 other groups: African/African American, Admixed American, Ashkenazi Jewish, European (Finnish), European (non-Finnish), Remaining individuals, South Asian.
Statements
gnomAD v4.1
This variant is present in gnomAD v4.1 (AF= 4.34024e-06; MAF= 0.00043%, 7/1612814 alleles, homozygotes = 0) and has highest observed frequency in the East Asian population (AF= 4.45971e-05; MAF= 0.00446%, 2/44846 alleles, homozygotes = 0); grpmax FAF= 7.39e-06.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 3.97997e-06; MAF= 0.00040%, 1/251258 alleles, homozygotes = 0) and has highest observed frequency in the East Asian population (AF= 5.4401e-05; MAF= 0.00544%, 1/18382 alleles, homozygotes = 0).
gnomAD-Canada
Not available in gnomAD-Canada v1.0.
SpliceAI and Pangolin
SpliceAI acceptor lossat +166 bp
0.000
SpliceAI acceptor gainat +188 bp
0.000
SpliceAI donor lossat -11 bp
0.000
SpliceAI donor gainat -415 bp
0.000
Pangolin splice lossmagnitude shown
-0.001
Pangolin splice gain
0.001
Other predictors
REVEL
Not available
BayesDel
Not available
How the predictions were used
BP4SupportingMet (Supporting): SpliceAI maximum delta score 0.00, at or below the <=0.1 no-splice-impact threshold.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 BP4 splicing sub-path is applicable and is met: SpliceAI maximum delta score for NM_000051.4:c.2638+11A>G is 0.00, which is <=0.1, the VCEP's no-predicted-splice-impact threshold. This is a non-missense intronic variant (donor +11, outside the canonical +/-1,2 positions) for which splicing is the primary plausible mechanism, so the lab's BP4 missense/splicing policy directs use of the splicing sub-path; that policy does not change the outcome here because the VCEP's literal text also permits BP4 from the splice sub-path for such variants. BP4 is applied once, for the splicing mechanism only (REVEL is inapplicable and unavailable for an intronic variant). The VCEP restriction on combining BP4 with BP7_Variable(RNA) is not triggered (no RNA data exist), and the VCEP explicitly permits BP4 alongside BP7 ('May also apply BP4 to achieve Likely Benign'); BP7 is assigned here from the positional deep-intronic rule, not by reusing the SpliceAI prediction, so the two codes do not double-count one prediction.
ClinGen HBOP ATM VCEP v1.5 (source_registry key 'cspec'; CSPEC doc 639508985) specifies BP4_Supporting for no predicted splice impact at SpliceAI <=0.1 (and missense REVEL <=0.249) and states that BP4 may also be applied with BP7 for synonymous/deep intronic variants.
SpliceAI Lookup for NM_000051.4:c.2638+11A>G (source_registry key 'spliceai') returned max delta score 0.00 (all four delta scores 0.00; Pangolin SG 0.001, SL -0.001), meeting the VCEP BP4 splicing threshold of <=0.1.
The variant is intronic at donor +11 (outside the canonical +/-1,2 sites), so splicing is the primary plausible mechanism and the lab's BP4 missense/splicing policy in vcep_materials gene_notes directs use of the splicing sub-path; this does not change the outcome from the VCEP's literal text, which also allows the splice sub-path for intronic variants outside donor/acceptor 1,2.
Threshold provenance: the SpliceAI 0.1 cut-off is the ATM VCEP v1.5 specification's own value (CSPEC doc 639508985), which states the VCEP uses SpliceAI (Jaganathan et al., 2019, Cell; PMID 30661751) with the SVI-recommended splice thresholds (Walker et al., 2023, Am J Hum Genet) for non-canonical splice variants.
Still missing
BP4 rests on a single predictor (SpliceAI), which is however the VCEP-specified sole predictor for splice evidence; no RNA data exist to corroborate absent aberrant splicing.
BP7SupportingMet (Supporting): the variant lies 11 nucleotides into intron 17, beyond the +7 donor-site deep-intronic threshold.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 directly specifies BP7 at supporting strength for deep intronic variants located further than (but not including) +7 at a donor site. NM_000051.4:c.2638+11A>G is 11 nt into intron 17 downstream of the exon 17 donor site, i.e. beyond +7, so it is BP7-eligible and BP7_Supporting is met. No RNA study is available to support a BP7_Variable(RNA) strength modification, so the default supporting strength is used. BP7 here derives from the VCEP's positional deep-intronic rule, not from the SpliceAI prediction used for BP4; the SpliceAI score of 0.00 is concordant with (but not the basis of) BP7. The VCEP notes BP7 is not considered a conflicting piece of evidence against a body of evidence supporting a pathogenic splice defect.
Rule
ATM VCEP v1.5 BP7_Supporting: use for synonymous and deep intronic variants further than (but not including) +7 at donor and -21 at acceptor sites; BP7(RNA) strength depends on observed lack of aberrant RNA defect and assay quality.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (source_registry key 'cspec'; CSPEC doc 639508985) specifies BP7_Supporting for synonymous and deep intronic variants defined as further than (but not including) +7 at donor and -21 at acceptor sites, and reserves BP7_Strong/Moderate(RNA) for observed lack of an aberrant RNA defect at curator-discretion weight.
Variant context: NM_000051.4:c.2638+11A>G is intronic at donor +11 of intron 17 (VariantValidator genome_context_intronic_sequence NC_000011.10(NM_000051.4):c.2638+11A>G; variant_exonic_positions 17i), beyond the +7 boundary, so the VCEP deep-intronic BP7 rule applies.
No RNA/minigene data for this variant were found in the case bundle, so BP7 is applied at the default supporting strength with no BP7(RNA) modification.
SpliceAI Lookup (source_registry key 'spliceai') reports max delta score 0.00 for this variant, concordant with the absence of a predicted splice impact; this score is used for BP4, while BP7 is assigned from the VCEP's positional deep-intronic eligibility rule.
Still missing
No variant-specific RNA assay is available to assess BP7_Variable(RNA) or to confirm absence of an aberrant transcript; BP7 is assigned on the VCEP positional deep-intronic rule only.
PP3Not met: SpliceAI maximum delta score 0.00, well below the >=0.2 PP3 threshold.
Details ▾Hide ▴
Reasoning
NM_000051.4:c.2638+11A>G is intronic, 11 nt into intron 17 downstream of the exon 17 donor site, i.e. outside the canonical donor/acceptor +/-1,2 dinucleotides, so the ATM VCEP v1.5 PP3 splicing sub-path is the applicable mechanism (the missense/REVEL sub-path does not apply to an intronic variant). The VCEP applies PP3_Supporting for SpliceAI >=0.2 on this sub-path. The SpliceAI Lookup result for this exact variant is a maximum delta score of 0.00 (DS_AG 0.00, DS_AL 0.00, DS_DG 0.00, DS_DL 0.00; Pangolin SG 0.001, SL -0.001), well below the 0.2 threshold, so PP3 is not met. The VCEP restriction that splice-based PP3 may not be combined with PVS1/PVS1_Variable(RNA) is moot because PP3 is negative.
Rule
ATM VCEP v1.5 PP3_Supporting: missense REVEL >0.7333; splicing SpliceAI >=0.2 for silent, missense/in-frame and intronic variants outside donor and acceptor 1,2 sites.
Evidence considered
Governing framework: ClinGen HBOP ATM VCEP v1.5 (source_registry key 'cspec'; CSPEC doc 639508985) specifies PP3_Supporting for missense REVEL >0.7333 and, on the splicing sub-path, for SpliceAI >=0.2 in silent, missense/in-frame and intronic variants outside the donor and acceptor 1,2 sites.
Variant context: NM_000051.4:c.2638+11A>G is intronic (VariantValidator genome_context_intronic_sequence NC_000011.10(NM_000051.4):c.2638+11A>G; variant_exonic_positions 17i), 11 nt into intron 17 from the exon 17 donor site, so it is outside the canonical +/-1,2 splice positions and the VCEP splicing sub-path is the applicable mechanism.
SpliceAI Lookup for NM_000051.4:c.2638+11A>G (source_registry key 'spliceai') returned max delta score 0.00 (DS_AG 0.00, DS_AL 0.00, DS_DG 0.00, DS_DL 0.00; Pangolin SG 0.001, SL -0.001), below the VCEP PP3 threshold of >=0.2; PP3 is therefore not met.
Threshold provenance: the SpliceAI 0.2/0.1 cut-offs are the ATM VCEP v1.5 specification's own values, taken verbatim from CSPEC doc 639508985 (source_registry key 'cspec'), which states the VCEP uses SpliceAI (Jaganathan et al., 2019, Cell; PMID 30661751) as sole predictor with the SVI-recommended splice thresholds (Walker et al., 2023, Am J Hum Genet) for non-canonical splice variants.
The missense sub-path is not applicable to this intronic variant and no REVEL score was retrievable (REVEL lookup found=false, score null); the ATM VCEP supplementary Table S1 (PMID 40580951) contains no row for c.2638+11A>G, so no secondary computational corroboration is available.
Still missing
No RNA/minigene data for this variant are available; a negative SpliceAI result cannot exclude a splice effect with certainty, but the VCEP calibrates PP3 solely on SpliceAI and the score is 0.00.
SpliceAI predicts no significant splice impact for this variant (max delta score = 0.00).
7 paper records listed. No publication is explicitly linked to an applied criterion in the recorded metadata. None of the listed paper records explicitly indicates that the variant is named.
Recorded publications
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology.
Richards S et al. · 2015
PMID 25741868 is the ACMG/AMP 2015 standards and guidelines for the interpretation of sequence variants. It is a general classification framework and reports no case, functional, segregation, or population-frequency data for ATM or for c.2638+11A>G (ATM is not named). Its relevant content for this group is generic: splice-site variants may cause exon skipping, shortening, or inclusion of intronic material via alternative or newly created donor/acceptor sites, and although such variants are predicted to have a null effect, confirmation of impact requires RNA or protein functional analysis. It also cautions that in-frame deletions/insertions may retain critical domains, and it supplies the generic definitions of PVS1 (null variant), PM4 (protein length change from in-frame indels in a non-repeat region or stop-loss), and BP3 (in-frame indel in a repeat region without known function).
For splice site variants, the variant may lead to exon skipping, shortening or inclusion of intronic material, due to alternative donor/acceptor site usage or creation of new sites. Although splice site variants are predicted to lead to a null effect, confirmation of impact requires functional analysis by either RNA or protein analysis.
Proposed Criteria for Interpretation of Sequence Variants - PVS1 Null variants section
Hereditary Ataxia Overview.
Adam MP et al. · 1993
GeneReviews chapter 'Hereditary Ataxia Overview'. Identified through the ClinVar citation trail as a candidate case-observation source for PS4/PP5. Full text was not retrieved into this case (screenshot only), so no content could be verified and no variant-specific data are available. As a gene/disorder-level review it would not be expected to report the exact intronic variant c.2638+11A>G.
GeneReviews chapter 'Ataxia-Telangiectasia'. Identified through the ClinVar citation trail as a candidate case-observation source for PS4/PP5. Full text was not retrieved into this case (screenshot only), so no content could be verified and no variant-specific data are available. As a gene/disorder-level review it would not be expected to report the exact intronic variant c.2638+11A>G.
EFNS/ENS Consensus on the diagnosis and management of chronic ataxias in adulthood.
van de Warrenburg BP et al. · 2014
This EFNS/ENS consensus guideline on the diagnosis and management of chronic ataxias in adulthood contains no variant-level or functional data on ATM and never names the ATM gene, c.2638+11A>G, or any ATM variant. At the disorder level it states that ataxia telangiectasia (caused by biallelic ATM mutations) is the second most frequent recessive ataxia globally (except in North Africa, where ataxia with vitamin E deficiency is second) and recommends testing for A-T or AOA2 first when cerebellar ataxia is accompanied by chorea and severe saccadic disturbance. No data relevant to PVS1, PM4, or BP3 were found.
The second most frequent recessive ataxia globally is ataxia telangiectasia, with the exception of North Africa where this is ataxia with vitamin E deficiency [15,16].
Recessive family history section, paragraph 4 (epidemiology of recessive ataxias); related clinical guidance in the same section, paragraph 3
A practice guideline from the American College of Medical Genetics and Genomics and the National Society of Genetic Counselors: referral indications for cancer predisposition assessment.
Hampel H et al. · 2015
This ACMG/NSGC practice guideline covers referral indications for cancer predisposition assessment and is not a variant-specific study. It mentions ATM only once, in the context of familial pancreatic cancer (ATM mutations in 2.4% of 166 familial pancreatic cancer patients and 4.6% of 87 high-risk families, cited from an external study), and provides no information on c.2638+11A>G and no functional, case, segregation, or population data relevant to PVS1, PM4, or BP3.
Some families with familial pancreatic cancer also have mutations in the CDKN2A, PALB2, or ATM genes. PALB2 mutations occur in 0.9-3.7% of pancreatic cancer patients with at least one additional relative affected with pancreatic cancer.48-50 ATM mutations were found in 2.4% (4/166) of patients with familial pancreatic cancer and in 4.6% (4/87) of families with three or more affected individuals.51
Recommendations, Familial pancreatic cancer (OMIM 260350) section
EFNS guidelines on the molecular diagnosis of ataxias and spastic paraplegias.
Gasser T et al. · 2010
EFNS guidelines on the molecular diagnosis of ataxias and spastic paraplegias (background review). Not retrieved as full text for this case; triaged as background only with no candidate criteria. No variant-specific content on c.2638+11A>G is available or expected.
Sherloc: a comprehensive refinement of the ACMG-AMP variant classification criteria.
Nykamp K et al. · 2017
Sherloc: a comprehensive refinement of the ACMG-AMP variant classification criteria (methodology paper). Not retrieved as full text for this case; triaged as background only with no candidate criteria. It is cited by one ClinVar submitter (Labcorp/formerly Invitae) but provides no variant-specific case, case-control, or prevalence data for c.2638+11A>G.
The evidence statements recorded by the interpretation pipeline. They have not been reinterpreted by the page.
gnomAD v4.1
This variant is present in gnomAD v4.1 (AF= 4.34024e-06; MAF= 0.00043%, 7/1612814 alleles, homozygotes = 0) and has highest observed frequency in the East Asian population (AF= 4.45971e-05; MAF= 0.00446%, 2/44846 alleles, homozygotes = 0); grpmax FAF= 7.39e-06.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 3.97997e-06; MAF= 0.00040%, 1/251258 alleles, homozygotes = 0) and has highest observed frequency in the East Asian population (AF= 5.4401e-05; MAF= 0.00544%, 1/18382 alleles, homozygotes = 0).
gnomAD-Canada
Not available in gnomAD-Canada v1.0.
ClinVar
This variant has been reported in ClinVar as Likely benign (3 clinical laboratories). (ClinVarID = 384288)
In silico
SpliceAI predicts no significant splice impact for this variant (max delta score = 0.00).
Somatic
This variant has not previously been reported in somatic cancers (COSMIC).
This variant was interpreted by pipeline 7.2.0.
The current version is 11.0.07.
Re-running queues a fresh interpretation; the result replaces this page when complete.
Re-runs are not free. Blank draws on today's public interpretation allowance; with a code, it uses one of that code's uses.