ATM c.7308-9C>T, p.?

Reasoning

Met, at Supporting strength. ClinGen HBOP ATM VCEP v1.6 specifies PM2 as: frequency <=0.001% (1e-5) in the gnomAD subpopulation with the highest frequency, with an exception that a frequency slightly above 0.001% arising from a single allele (n=1) in one subpopulation, absent elsewhere, is still sufficiently rare for PM2_supporting. This variant's highest subpopulation frequency in gnomAD v4.1 is European (non-Finnish) 8.4857e-07 (1/1,178,460 alleles, 0 homozygotes), about 12-fold below the 1e-5 cutoff; the all-comers total AF is 6.2032e-07 (1/1,612,070 alleles, 0 homozygotes). Every other ancestry subpopulation in the capture carries zero copies, and the variant is reported absent from gnomAD v2.1 (exome), gnomAD-Canada v1.0, and from both non-cancer subsets (v2.1 non-cancer exomes, v3.1 non-cancer genomes) - so the VCEP's n=1/absent-elsewhere exception is independently satisfied as well. The VCEP supplies only a Supporting-strength rule for PM2, so no higher strength is available; the criterion is met at Supporting. The ATM HBOP specification does not designate a non-cancer or exome-only population source for PM2, so the all-comers GNOMAD_V2_1/GNOMAD_V4_1 entries were used as primary, with the non-cancer subsets as concordant corroboration.

Rule

Frequency <=.001% in gnomAD subpopulation with the highest frequency (ClinGen HBOP ATM VCEP v1.6, PM2, Pathogenic Supporting)

Evidence considered

• ClinGen HBOP ATM VCEP v1.6 PM2 rule: 'Frequency <=.001% in gnomAD subpopulation with the highest frequency. Exception: If >0.001% but n=1 in a single subpopulation (not present in any other subpopulation), that is sufficiently rare and PM2_supporting would apply.' Only a Supporting-strength rule is provided.
• gnomAD v4.1 highest subpopulation: European (non-Finnish) AF 8.4857e-07 (ac=1, an=1178460, homozygotes=0) - below the 1e-5 (0.001%) threshold.
• gnomAD v4.1 total: ac=1, an=1612070, AF 6.2032e-07, homozygotes=0; exome ac=1/an=1459862 (AF 6.85e-07) and genome ac=0/an=0.
• All non-NFE gnomAD v4.1 subpopulations are 0 (AMR, FIN, AMI, EAS, MID, SAS, ASJ, AFR, remaining); NFE_XX 0/610786, NFE_XY 1/567674 - consistent with a single, non-recurrent observed allele.
• gnomAD v2.1 (exome) reported absent; gnomAD-Canada v1.0 reported absent; GNOMAD_V2_1_NON_CANCER (v2.1 exomes-only non-cancer) and GNOMAD_V3_1_NON_CANCER (v3.1 genomes-only non-cancer) both reported absent.
• The ATM HBOP VCEP v1.6 provides no ATM-specific direction to use a non-cancer or exome-only population source for PM2, so the all-comers GNOMAD_V2_1/GNOMAD_V4_1 totals govern; the non-cancer subsets were checked and are concordant.
• No homozygote for this variant is reported in any population source, so no recessive-disease prevalence/penetrance argument (which would bear on BS2 rather than PM2) arises from the population data.

Sources

Reasoning

NM_000051.4:c.7308-9C>T is an intronic acceptor-region substitution, so BP4 must be taken from the splice path and not from REVEL. The governing ClinGen HBOP ATM VCEP v1.6 specifies BP4 supporting for no predicted splice impact when SpliceAI is <=0.1. The observed SpliceAI maximum delta is 0.039 (DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001), which satisfies the <=0.1 cutoff, and the independent Pangolin scores (splice gain 0.071, loss -0.002) agree. The variant lies outside the donor and acceptor +/-1,2 dinucleotides and is therefore eligible for the non-canonical-splice BP4 rule, and no other spliceogenic feature is evident at this position. The VCEP further permits BP4 to be applied together with BP7(RNA) when RNA shows no defect; here no variant-specific RNA data exist, so BP4 rests on the SpliceAI prediction alone. The same prediction is not reused to support PP3 (PP3 is not met) and no second predictor is counted as an independent criterion, so there is no double-counting. BP4 is met at supporting strength, the only strength the VCEP defines for this criterion.

Rule

ATM VCEP v1.6 BP4 Supporting: Splicing - no predicted impact via SpliceAI <=0.1.

Evidence considered

• ClinGen HBOP ATM VCEP v1.6 (cspec doc 639508985) BP4 rule: splicing - no predicted impact via SpliceAI <=0.1; applied at supporting strength.
• The case SpliceAI lookup for NM_000051.4:c.7308-9C>T reports a maximum delta of 0.039, which meets the <=0.1 cutoff, with Pangolin splice gain 0.071 and splice loss -0.002 in agreement.
• c.7308-9C>T is intronic, 9 nucleotides from the exon 50 acceptor site and outside the -1,-2 dinucleotide, so the non-canonical/intronic splice path rather than the missense REVEL path applies.
• REVEL was not evaluated and is not available for this variant; the one-path rule assigns intronic variants to SpliceAI only, and the low SpliceAI delta is not used as support for the missense path.
• The exact variant has no entry in Suppl_TableS1_PMID 40580951.xlsx, so no VCEP-provided BP4 code exists to apply directly.

Still missing

• No entry for c.7308-9C>T, c.7308-9, or the corresponding hg38 position 108330205 was found in Suppl_TableS1_PMID 40580951.xlsx (searched as converted full text; the table has no rows between c.7307+5C>T at hg38 108329243 and c.7308-5G>T at hg38 108330209). This is a weaker statement than the evidence being absent.
• No variant-specific RNA data were available; a lack of an observed splice defect (BP7(RNA)) could not be assessed.

Sources

Reasoning

NM_000051.4:c.7308-9C>T is an intronic substitution 9 nt upstream of the exon 50 acceptor site (intron 49 / exon_50 acceptor region; NC_000011.10:g.108330205C>T GRCh38 = NC_000011.9:g.108200932C>T GRCh37; predicted protein consequence NP_000042.3:p.?). The governing ClinGen HBOP ATM VCEP PVS1 criterion (v1.5 and v1.6, adapted from PMID 30192042) is restricted to null variants: nonsense, frameshift, canonical +/-1 or 2 splice sites, initiation codon, and single or multi-exon deletion. c.7308-9C>T lies outside the canonical donor/acceptor +/-1,2 dinucleotide, so it does not enter the decision tree on the canonical splice-site path; the VCEP's own PS1 splicing table classifies variants at positions outside the donor/acceptor +/-1,2 dinucleotide as PP3-baseline rather than PVS1, consistent with this exclusion. The variant is neither nonsense, frameshift, an initiation-codon change, nor an exon deletion, and VariantValidator predicts p.? with no protein-length change. SpliceAI max delta score is 0.039 (DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001) - no predicted acceptor or donor gain or loss - so there is no predicted splice defect that could be scored at any PVS1 tier under the SVI framework. The PVS1_Variable(RNA) path for observed splice defects cannot be entered either, because no RNA/cDNA study of this variant exists and the VCEP requires an observed (not predicted) defect for that route. PVS1 is therefore not met and no strength tier is assigned; because the criterion is not met, no SVI/VCEP downgrade rule is invoked and strength is left null.

Rule

ClinGen HBOP ATM VCEP PVS1 decision guide (v1.5/v1.6, adapted from PMID 30192042): PVS1 requires a null variant (nonsense, frameshift, canonical +/-1 or 2 splice site, initiation codon, single/multi-exon deletion); c.7308-9C>T is a non-canonical intronic substitution with SpliceAI max delta 0.039 and no observed RNA defect, so no PVS1 tier applies.

Evidence considered

• Governing specification: ClinGen HBOP ATM VCEP v1.5/v1.6 PVS1 section - '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 PVS1/PVS1(RNA) decision guide adapted from PMID 30192042; PVS1_Variable(RNA) is reserved for observed splice defects.
• Full-text greps of vcep_db/ATM/fulltext for 'c.7308-9C>T', 'c.7308', '7308-' and 'exon_50' returned no per-variant PVS1 entry: the PVS1 PDFs contain only the criterion text, notes and figure placeholders, with no variant table at all.
• VariantValidator normalisation of the case input: NM_000051.4:c.7308-9C>T, MANE Select transcript, intronic (start/end exon '49i' on both NC_000011.10 and NC_000011.9), genomic GRCh38 NC_000011.10:g.108330205C>T / GRCh37 NC_000011.9:g.108200932C>T, predicted protein consequence NP_000042.3:p.? - no protein-length change.
• SpliceAI Lookup result for NM_000051.4:c.7308-9C>T: max delta score 0.039 (DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001) - no gain or loss of the exon 50 acceptor or donor, i.e. no predicted splice impact at a variant 9 nt from the acceptor and outside the canonical +/-1,2 dinucleotide.
• Suppl_TableS1_PMID 40580951.xlsx (Sun et al. 2025, prime-editing saturation screen of ATM SNVs) contains rows for the neighbouring intronic positions c.7308-5 through c.7308-1 but no row for c.7308-9C>T, so there is no direct functional measurement of this variant to route to any PVS1 tier.
• The three full texts fetched for this case (PMID 15604628, PMID 24366376, PMID 35802134) do not mention ATM c.7308-9C>T (paper_extracts.json records variant_mentioned=false for each), so they provide no RNA or consequence evidence for this group's criteria.
• REVEL is not applicable to this variant: it is intronic with no missense consequence, so no missense in-silico path exists.

Still missing

• Searched ATM_PVS1_1.5.pdf and ATM_PVS1_1.6.pdf in full (vcep_db/ATM/fulltext): neither file contains any per-variant entry for c.7308-9C>T or any pre-assigned PVS1 code - both are generic null-variant decision guides with no variant rows, so there is no governing pre-assigned PVS1 answer for this variant to override the above reasoning.
• Searched Suppl_TableS1_PMID 40580951.xlsx (PMID 40580951) for c.7308-9C>T: no row exists for this exact variant. The nearest intronic rows are c.7308-5G>T/A/C, c.7308-4C>A/T/G, c.7308-3A>G/T/C, c.7308-2A>G/T/C and c.7308-1G>A/C/T; c.7308-9 is not tabulated, so no functional or combined-score measurement is available for it.
• No RNA/cDNA (transcript) study of c.7308-9C>T was identified in the fetched literature or in ClinVar's citations, so an observed splice defect cannot be confirmed or excluded experimentally and PVS1_Variable(RNA) is not assessable.
• The ATM PVS1 decision tree is supplied as a figure that did not convert to machine-readable text in either ATM_PVS1_1.5.pdf or ATM_PVS1_1.6.pdf; the entry criterion (which variant types are PVS1-eligible) was taken from the textual criterion definition and notes. This does not affect the call, because the variant does not qualify as a null variant on any entry path.

Sources

Reasoning

NM_000051.4:c.7308-9C>T is intronic (acceptor region of intron 49, 9 nt upstream of exon 50; NP_000042.3:p.?, no amino-acid change), so the standard missense sub-path of ATM PS1 (same amino-acid change as an established P/LP variant) cannot apply. The governing ATM HBOP VCEP PS1 splicing table carries an explicit prerequisite for every row: the predicted event of the variant under assessment (VUA) must precisely match the predicted event of the known P/LP reference variant, and the VUA's prediction strength must be equal to or higher than the reference's; the table is also entered through the VUA's own baseline computational/predictive code (PP3, PVS1, PVS1_Strong/Moderate/Supporting). Neither condition can be satisfied here. First, the VUA has no predicted splice event and no baseline code: SpliceAI max delta is 0.039 (DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001), below the ATM VCEP PP3 splicing cut-off (SpliceAI >= 0.2) and also below the BP4 cut-off (SpliceAI <= 0.1) - there is nothing to match against a reference variant's predicted event. Second, no known P/LP variant is curated at the same nucleotide or at a comparable position for this boundary: ClinVar 236769, which is this exact variant, is Likely benign (2 clinical laboratories, no expert-panel submission, 2-star review status), and the only P/LP changes in the vicinity in the VCEP functional dataset are the canonical acceptor substitutions c.7308-2A>C and c.7308-1G>C, which sit at different positions, belong to the donor/acceptor +/-1,2 dinucleotide class rather than the VUA's position class, and carry a predicted splice defect the VUA does not share. With no matching predicted event, no comparable-strength reference, and no same-position P/LP comparator, PS1 is not met at any strength (PS1, PS1_Moderate or PS1_Supporting).

Rule

ATM HBOP VCEP v1.5/v1.6 PS1: missense path requires the same amino-acid change as a previously established P/LP variant with splicing ruled out for both; splice-region path uses the ATM PS1 Splicing table (PS1 / PS1_Moderate / PS1_Supporting) subject to the stated prerequisite of an exactly matching predicted event and equal-or-higher prediction strength.

Evidence considered

• ATM PS1 v1.5/v1.6 splicing table, prerequisite line (verbatim): "Prerequisite for all: The predicted event of the VUA must precisely match the predicted event of the known (likely) pathogenic variant (e.g. both predicted to lead to exon A skipping, or both to enhanced use of cryptic site B), AND the strength of the prediction for the VUA must be of similar or higher strength than the strength of the prediction for the known (likely) pathogenic variant."
• ATM PS1 v1.5/v1.6 table rows are gated on the VUA's own "Baseline computational/predictive code applicable to VUA" (PP3, or PVS1/PVS1_Strong/PVS1_Moderate/PVS1_Supporting) and on the "Position of reference variant compared to VUA"; acceptor motif defined in the same document as the first base of the exon plus 20 nucleotides upstream, so c.7308-9 lies within the exon 50 acceptor motif but outside the +/-1,2 dinucleotide.
• VUA SpliceAI result: DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001, max delta 0.039 - below the ATM VCEP PP3 splicing threshold (SpliceAI >= 0.2, same value as Jaganathan et al. 2019, PMID:30661751) and below the BP4 threshold (SpliceAI <= 0.1). No predicted splice event and no applicable baseline predictive code.
• ClinVar 236769 (NM_000051.4(ATM):c.7308-9C>T): Likely benign, 2 submissions (Myriad Genetics; Labcorp Genetics/Invitae), criteria provided multiple submitters no conflicts, no expert-panel submission; conditions Ataxia-telangiectasia syndrome and Familial cancer of breast. No P/LP reference variant at the same position.
• Sun et al. 2025 ATM SNV table (Suppl_TableS1_PMID 40580951) was searched for hg38 108330205 / c.7308-9: no row exists (the nearest rows are c.7308-5G>T/A/C at 108330209, c.7308-4, c.7308-3, then the canonical c.7308-2A>G/A>T/A>C and c.7308-1G>A/G>C/G>T). The only ClinVar P/LP changes in this block are c.7308-2A>C and c.7308-1G>C (both Likely pathogenic); they are different nucleotide positions and a different variant class (canonical acceptor), so they are not same-position PS1 references for this VUA.
• Files searched directly: /Users/gs_agent/LYFESCI_HERMES_v9/vcep_db/ATM/fulltext/ATM_PS1_1.5.pdf.txt (130 lines, read in full) and ATM_PS1_1.6.pdf.txt (differs only in version string and line wrapping) - zero occurrences of 'c.7308', '7308', or 'p.?' keyed to this variant.

Still missing

• ATM_PS1_1.5.pdf (and the near-identical ATM_PS1_1.6.pdf) were searched in full for c.7308-9C>T, p.? and the hg38 position 108330205: no entry keyed to this variant exists in either file. Both files contain only the generic missense PS1 statement and the generic PS1(Splicing) weight table, no per-variant pre-assigned code for this position.
• No ATM P/LP variant is curated at the acceptor c.7308-9 nucleotide in any source available to this case, so the PS1 splicing sub-path could not be exercised against a same-position reference even if the VUA had carried a matching predicted splice event.

Sources

Reasoning

Not met. c.7308-9C>T (NM_000051.4) is an intronic variant in intron 49, nine nucleotides upstream of the exon 50 acceptor site (predicted protein consequence NP_000042.3:p.?). Both VCEP sources declared for PS3 were searched in full and neither contains the variant. (1) clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx is an assay-calibration table, not a per-variant results table: its three approved kinase-activity assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335) all rely on stable transfection of site-directed-mutagenised ATM cDNA constructs into ATM-null LCLs, which cannot interrogate an intronic position 9 nt from the exon boundary; the file contains no entry for this variant and no reported readout to read off. (2) Sun et al. 2025 (PMID 40580951) Table S1 was searched on 'c.7308-9C>T', '7308-9' and the hg38 coordinate chr11:108330205 with zero matching rows. Neighbouring rows are present (c.7308-5G>T, c.7308-4C>T, c.7308-3A>T, c.7308-2A>C, c.7308-1G>C, all in intron 49 upstream of exon 50), but the table's intronic coverage is limited to offsets -5..-1 and +1..+5 (each offset represented by exactly 183-186 rows), so the -9 position was never assayed in the prime-editing saturation screen and has no Combined_score, Classification or DeepATM_predicted call at all. The three full-text papers fetched for this case (PMID 15604628, PMID 24366376, PMID 35802134) contain no ATM functional assay data and do not mention the variant or ATM splicing. Because no study demonstrates a damaging effect on ATM function, the VCEP v1.6 strength ladder (Supporting = fails to rescue an ATM-specific feature such as phosphorylation of ATM-specific targets; Moderate = fails to rescue both an ATM-specific feature and radiosensitivity; Strong explicitly disallowed) cannot be applied, and no strength is assigned.

Rule

ClinGen HBOP ATM VCEP v1.6 PS3: no functional readout for the variant -> criterion not met; Strong explicitly disallowed ('Do not use as strong'), Moderate requires failure to rescue both an ATM-specific feature and radiosensitivity, Supporting requires failure to rescue an ATM-specific feature.

Evidence considered

• Variant identity: NM_000051.4:c.7308-9C>T = NC_000011.10:g.108330205C>T (GRCh37 g.108200932C>T), intronic, last nucleotide region of intron 49 (start/end exon 49i), predicted protein consequence NP_000042.3:p.?, i.e. no amino-acid substitution - so the missense-oriented cDNA Construct kinase assays are structurally inapplicable to this position.
• VCEP HBOP ATM framework (cspec, doc 639508985, v1.6) defines the operative PS3 ladder: Strong = '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, explicitly 'Do not use for radiosensitivity-only as that is not a feature specific to ATM deficiency'. PS3 requires a positive (failing) functional readout before any strength can be assigned.
• The VCEP approved-assay table (clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx) contains only assay methodology and calibration, not per-variant results: three kinase-activity assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335; proposed PS3_Supporting) and a radiosensitivity sheet (Mitui 2009 approved; PS3_Moderate but only combinable with a kinase-activity result, no weight for radiosensitivity alone, no weight for conflicting results). Its validation-control variants are c.5908C>T, c.7636del9/c.8147T>C/c.8546G>C/c.8599G>C/c.7987delGTT and benign controls c.1744T>C/c.2119T>C - c.7308-9C>T is not among them and is not reported anywhere in the file.
• Searched 'Suppl_TableS1_PMID 40580951.xlsx' with grep for 'c.7308-9C>T', '7308-9' and '108330205': no match. The nearest rows are intronic offsets -5 to -1 (e.g. 108330209 G>T c.7308-5G>T; 108330213 G>C c.7308-1G>C, classification Non-functional, DeepATM_predicted=No) and the exonic c.7308A>G/A>C/A>T rows; offset -9 is absent from the file entirely.
• Coverage check on the same file: every intronic cHGVS notation present is one of offsets -5, -4, -3, -2, -1, +1, +2, +3, +4, +5 (183-186 rows per offset). This establishes the absence of c.7308-9C>T as a genuine coverage limit of that dataset (the -9 position was never assayed or modelled) rather than a retrieval or formatting artefact.
• No functional-assay evidence was found in the papers fetched for this case: PMID 15604628 (NSGC cancer risk assessment/counselling guideline), PMID 24366376 (USPSTF BRCA recommendation statement) and PMID 35802134 (ACMG SF v3.1 policy statement) contain no ATM functional assays and do not mention this variant.
• ClinVar (variation 236769, Likely benign, 2 laboratories) and the SpliceAI result (max delta 0.039; DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001) were noted only as context: neither is functional assay evidence, and no classification was inferred from them for this criterion.

Still missing

• c.7308-9C>T has no entry in 'clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx' (searched in full, both 'ATM kinase activity' and 'ATM radiosensitivity' sheets) - expected, since this file calibrates assay strength and lists only which papers/variants were used as validation controls (none of which is this variant).
• c.7308-9C>T has no entry in 'Suppl_TableS1_PMID 40580951.xlsx' (searched on c.7308-9C>T, 7308-9, and chr11:108330205) - the -9 intronic offset was not assayed by that prime-editing screen, so no direct functional measurement and no DeepATM extrapolation exists for this position, rather than one existing and being unfavourable.
• No published functional assay (kinase activity, radiosensitivity, or RNA/minigene splicing assay) testing c.7308-9C>T was identified in any source searched for this case, including the three full-text papers fetched (PMID 15604628, PMID 24366376, PMID 35802134), none of which reports ATM functional data.
• PS3 cannot be lowered to a lower-strength assertion here either: the ATM HBOP VCEP framework requires a positive functional readout for any strength, and no readout exists for this variant.

Sources

Reasoning

Governing specification: ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer (HBOP) Expert Panel Specifications to the ACMG/AMP Guidelines for ATM v1.6 (cspec doc 639508985), which marks PS4 'Applicable' at exactly one strength. The PS4 (Strong) rule, quoted verbatim from that criterion's own text and from no other criterion in this group: 'Case-control studies; p-value <=.05 **and** (Odds ratio, hazard ratio, or relative risk >=2 **or** lower 95% CI >=1.5).' Operator direction, taken from this criterion's own operator: the enrichment statistic strengthens as the value goes UP, so the test is 'value >= 2' (alternatively 'lower 95% CI >= 1.5'), with p <= 0.05 as a co-requirement; the v1.6 release notes confirm the Boolean operators in PS4 were clarified in this version. Explicit substitution: no odds ratio, hazard ratio, relative risk or confidence interval is reported anywhere for ATM c.7308-9C>T, so the inequality has no left-hand side - 'OR >= 2? no value available', 'lower 95% CI >= 1.5? no value available' - and cannot be satisfied. Searching the sources that could carry such data: none of the six PMIDs triaged as PS4/PP5 candidates (PMID:20301425, PMID:15604628, PMID:24366376, PMID:35802134, PMID:42258614, PMID:28492532) is a case-control or cohort enrichment study, and the three retrieved in full text (PMID:24366376, PMID:15604628, PMID:35802134) never mention c.7308-9C>T, p.(?), ATM c.7308-9C>T or the gene in any variant-specific way - they are a USPSTF BRCA recommendation, an NSGC risk-assessment/counselling guideline and the ACMG SF v3.1 policy statement respectively. The generic ACMG fallback route (Richards 2015 PS4 Note 2: prior observation in multiple unrelated patients with absence in controls, moderate strength) is explicitly closed for ATM because the VCEP instructs 'PS4_Moderate: Do not use. Proband counting for genes causing a common disorder need to be calibrated in a population-specific way before use.' The only variant-level quantitative data available are population frequencies with no affected-versus-control contrast: gnomAD v4.1 total AF 6.2032e-07 (1/1,612,070 alleles, 0 homozygotes), absent from gnomAD v2.1 and from gnomAD-Canada v1.0. This is a single-allele, ultra-rare observation for which no enrichment study exists or is likely to exist, and no proband set accompanies the variant in this curation. PS4 is therefore not met.

Rule

PS4 (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 (HBOP ATM VCEP v1.6).

Evidence considered

• HBOP ATM VCEP v1.6 (cspec doc 639508985) PS4 rule quoted verbatim from this criterion's own text: 'Case-control studies; p-value <=.05 **and** (Odds ratio, hazard ratio, or relative risk >=2 **or** lower 95% CI >=1.5).' Direction read from this criterion alone: the statistic strengthens upward, so met requires OR/HR/RR >= 2 or lower 95% CI >= 1.5 together with p <= 0.05.
• Explicit inequality check: 'OR/HR/RR >= 2? no value available' and 'lower 95% CI >= 1.5? no value available' - no case-control estimate of any kind is reported for c.7308-9C>T, so neither branch can be evaluated, let alone satisfied.
• VCEP instruction attached to PS4: 'PS4_Moderate: Do not use. Proband counting for genes causing a common disorder need to be calibrated in a population-specific way before use.' This removes the generic ACMG multiple-patient fallback for ATM.
• All six PMIDs triaged as PS4 candidates were reviewed: none is a case-control or cohort enrichment study. Full texts available for PMID:24366376 (USPSTF BRCA recommendation statement), PMID:15604628 (NSGC cancer risk assessment/counselling recommendations) and PMID:35802134 (ACMG SF v3.1 policy statement); none of the three mentions c.7308-9C>T or reports any ATM variant-level data.
• PMID:20301425 (BRCA1/BRCA2-associated HBOC overview), PMID:42258614 (GeneReviews: ATM-Related Cancer Predisposition - autosomal dominant, moderately increased risk of female breast, pancreatic, prostate and gastric cancer) and PMID:28492532 (Sherloc, an ACMG/AMP criteria refinement) are gene- or framework-level sources with no case-control content for this variant.
• Variant-level rarity without any case/control contrast: gnomAD v4.1 total AF 6.2032e-07 (1/1,612,070 alleles, 0 homozygotes; highest subpopulation European non-Finnish AF 8.48565e-07, 1/1,178,460), absent from gnomAD v2.1 and absent from gnomAD-Canada v1.0.
• ClinVar exact match (VCV000236769, variation ID 236769) carries only Likely benign assertions from two ordinary clinical laboratories (Labcorp Genetics/Invitae SCV001674612; Myriad Genetics SCV005084900); no affected-case series or enrichment data are provided by either submission.

Still missing

• No case-control, cohort or case-series enrichment study of ATM c.7308-9C>T was identified; the variant remains a single-allele gnomAD v4.1 observation (1/1,612,070) with no affected-versus-control data anywhere.
• VCEP PS4_Moderate is explicitly disabled for ATM ('Do not use'), so the generic ACMG multiple-unrelated-patient fallback cannot be used to substitute for the missing case-control estimate.
• No proband-level phenotype, cancer history or family history accompanies this variant in the case record, so no genotype-phenotype enrichment could be assembled locally.

Sources

Reasoning

Governing specification: the ClinGen HBOP ATM VCEP PM3/BP2 points table (ATM_PM3_BP2_1.5.pdf; re-issued textually unchanged as v1.6, which is the version governing this case). PM3 for ATM is not the classic 'detected in trans with a pathogenic variant' code applied once; it is a points-accumulation code whose total converts to a strength: PM3_Supporting = 1 point, PM3 = 2, PM3_Strong = 4, PM3_VeryStrong = 8. Points are earned per unrelated A-T proband carrying the variant under assessment, at 4.0 (phenotype confident, confirmed in trans), 2.0 (confident, phase unknown), 1.0 (confident, second variant unidentified or VUS), or 2.0 (homozygous, max 2 individuals), with the 'consistent phenotype' rows worth half of those values. The table's own eligibility gate requires the variant not to exceed a general population frequency of 0.01%; this variant passes that gate (gnomAD v4.1 total AF 6.2032e-07, 1/1,612,070 alleles, 0 homozygotes; absent from gnomAD v2.1 and gnomAD-Canada), so the table is applicable and the null result is not an artefact of frequency disqualification. Against that table, zero points were identified. No A-T proband carrying c.7308-9C>T is reported anywhere in the assembled evidence: the six fetched papers contain no ATM mention at all (three of them full-text searched for c.7308-9C>T, p.?, and ATM), and ClinVar variation 236769 carries two Likely benign submissions (Myriad Genetics; Labcorp/Invitae) with zero expert-panel submissions, zero criterion-level leads, and no validated PMID trail. There is therefore no proband observation of any phenotype class (confident or consistent), no homozygous A-T individual, and no documented in-trans co-occurrence with a VCEP-classified P/LP ATM variant. With 0 points accumulated, the variant falls below the 1-point PM3_Supporting floor, so PM3 is not met. Because the finding is a null result rather than a near-threshold one, no strength is assigned. Note that the small residual possibility of an unpublished or unfetched A-T proband report is recorded as a gap below, not as support.

Rule

ATM VCEP PM3/BP2 points table - PM3_Supporting requires >=1 point accumulated from unrelated A-T proband observations; 0 points were accumulated, so PM3 is not met.

Evidence considered

• ATM PM3/BP2 points table (v1.5 and v1.6, identical in substance): PM3_Supporting = 1 point, PM3 = 2, PM3_Strong = 4, PM3_VeryStrong = 8; points are assigned 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).
• Table eligibility gate: 'Variant under assessment (VUA) must be sufficiently rare (not meeting a benign population evidence code)' and 'Variant may not exceed general population frequency >0.01%'. gnomAD v4.1 shows total AF 6.2032e-07 (1/1,612,070 alleles, 0 homozygotes), highest subpopulation European (non-Finnish) AF 8.48565e-07 (1/1,178,460 alleles, 0 homozygotes); absent from gnomAD v2.1 and from gnomAD-Canada v1.0. The rarity gate is satisfied, so the table is applicable to this variant.
• Table footnote 2 requires any co-occurring P or LP variant to be classified under VCEP specifications rather than assumed, and permits inferred in-trans only for a homozygous A-T-affected patient from a consanguineous union or where both parental lineages show ATM-related cancer family history. None of these conditions is documented for this case: no co-occurring ATM variant, no homozygous A-T individual, and no parental-phase genotyping data appear in any assembled source.
• ClinVar variation 236769 (NM_000051.4(ATM):c.7308-9C>T): classification Likely benign, review status 'criteria provided, multiple submitters, no conflicts' (2 stars), conditions Ataxia-telangiectasia syndrome and Familial cancer of breast; two submissions, both Likely benign - SCV005084900 (Myriad Genetics, 2024-06-13, origin unknown) and SCV001674612 (Labcorp Genetics/formerly Invitae, 2024-01-04, germline). Zero expert-panel submissions. No A-T proband, no in-trans observation, and no homozygous affected individual is described in the record.
• ClinVar submission audit for this case: 2 submissions total, 2 usable for classification, but 0 usable for criterion leads, 0 expert-panel submissions, 0 validated PMIDs, 1 rejected PMID (28492532, 'pubmed_metadata_unresolved'), 0 criterion-lead submissions, and an empty criterion_leads list - i.e. no submission supplies a criterion-level observation that could be converted into PM3 points.
• Literature packet for this case: 6 PMIDs (15604628, 20301425, 24366376, 35802134, 42258614, 28492532), all retrieved from ClinVar citations, none mentioning ATM at variant level. Grep of the case evidence bundle for 'proband', 'in trans', 'homozyg', 'phase', and 'ataxia' returns only the VCEP table text itself - no case-level observation of this variant in an A-T proband anywhere in the assembled evidence.
• SpliceAI for this variant: max delta score 0.039 (DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001; Pangolin SG 0.071), i.e. no predicted splice effect. Recorded only as variant context for the allelic assessment; it is not PM3 evidence and does not substitute for a proband observation.

Still missing

• ATM_PM3_BP2_1.5.pdf (declared governing file for PM3/BP2, together with its v1.6 re-issue) was searched in full for 'c.7308-9C>T', 'p.?', and '7308' and contains no entry for this variant by any notation; it is a per-proband points table, not a per-variant lookup table, so its governing content is the point-to-strength conversion rather than a pre-assigned code. A pre-assigned code therefore does not exist for this variant in this file.
• Suppl_TableS1_PMID 40580951.xlsx (Sun et al. 2025) was searched and has no row for c.7308-9C>T: it jumps from chr11:108329243 (c.7307+5C>G) to chr11:108330209 (c.7308-5G>T), so the intron interior including the c.7308-9 position is not covered by that table. This does not affect PM3, which is not a functional-assay code.
• No variant-specific literature search was performed for this case. The literature packet was derived from ClinVar citations only (6 PMIDs, none of which mention ATM), so an A-T proband report carrying c.7308-9C>T could in principle exist in literature that was never retrieved. The PM3 verdict rests on the assembled sources (ClinVar, gnomAD, the VCEP table, and the fetched corpus), all of which are negative for A-T probands.
• ClinVar submitter-level evidence codes for SCV001674612 and SCV005084900 are not exposed by the available ClinVar capture, so the reasoning behind the two Likely benign calls could not be inspected for an A-T proband observation; the submission audit records zero criterion leads and no validated PMIDs (PMID:28492532 unresolved).

Sources

Reasoning

The governing ATM VCEP specification restricts PM4 in this gene to a single variant class: 'Use for stop-loss variants.' (Pathogenic Moderate, ClinGen HBOP ATM VCEP v1.6). NM_000051.4:c.7308-9C>T is an intronic substitution 9 nt upstream of the exon 50 acceptor site; VariantValidator predicts protein consequence NP_000042.3:p.?, i.e. it alters no termination codon and produces no protein-length change (no stop-loss, no in-frame indel). The generic ACMG definition of PM4 (protein length change due to in-frame indel or stop-loss) is therefore also not satisfied. PM4 is not met and no strength tier is assigned.

Rule

ClinGen HBOP ATM VCEP v1.6 PM4: applicable to stop-loss variants only; c.7308-9C>T is intronic with p.? and creates neither a stop-loss nor an in-frame indel.

Evidence considered

• ATM VCEP v1.6 PM4 entry, default strength 'Pathogenic Moderate', strength rule: 'Use for stop-loss variants.' - the VCEP limits PM4 to a single variant class rather than the generic in-frame-indel/stop-loss definition.
• Variant consequence for NM_000051.4:c.7308-9C>T: intronic substitution in the exon 50 acceptor region, predicted protein consequence NP_000042.3:p.? (VariantValidator), splice prediction max delta 0.039 (SpliceAI) - no stop-loss and no in-frame indel, therefore no protein-length change to score.

Still missing

• Searched the ATM VCEP file set (ATM_PVS1_1.5.pdf, ATM_PVS1_1.6.pdf, ATM_PS1_*.pdf, ATM_PM3_BP2_*.pdf, Suppl_TableS1_PMID 40580951.xlsx, clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx) for c.7308-9C>T: no per-variant entry exists in any of them, so there is no pre-assigned governing PM4 code for this variant.
• No RNA/cDNA study of c.7308-9C>T was identified, so a potential splice-mediated change in protein length cannot be assessed experimentally; the in-silico prediction (SpliceAI max delta 0.039) shows no predicted splice effect, and no stop-loss is created by the substitution itself.

Sources

Reasoning

The governing ATM HBOP VCEP (CSPEC v1.6) rewrites PM5 as a truncation-based rule and restricts its only applicable weight to PM5_Supporting: it applies to NMD-prone truncating variants receiving PVS1 at Very Strong with a premature termination codon upstream of p.Leu3048, and for splice variants only to observed (not predicted) NMD-prone truncating consequences receiving PVS1(RNA) at Very Strong with high-quality RNA data. The rule text itself states this path "is not applicable for predicted splice impact without RNA data or PVS1(RNA) at reduced strength", and the gene-level instruction adds "Do not use for missense changes". None of the required elements are present for c.7308-9C>T: it is a non-truncating intronic acceptor-region substitution with no amino-acid change (NP_000042.3:p.?), it does not receive PVS1 (SpliceAI max delta 0.039, below the >= 0.2 SpliceAI PP3 cut-off, i.e. no predicted splice impact at all, and it is not a canonical +/-1,2 variant), no RNA study is available in this case, and there is no premature termination codon. Independently, the classical PM5 reading (a novel missense change at a residue where a different missense change is pathogenic) is structurally impossible for a p.? variant, and pm5_candidates.json records found = false with zero same-residue candidates, so no comparator exists. PM5 is not met.

Rule

ATM HBOP CSPEC v1.6 PM5 = PM5_Supporting only: NMD-prone truncating variants receiving PVS1 at Very Strong with a PTC upstream of p.Leu3048, or (splice) observed NMD-prone truncating variants receiving PVS1(RNA) at Very Strong with high-quality RNA data; not applicable for predicted splice impact without RNA data; do not use for missense changes.

Evidence considered

• ATM HBOP CSPEC v1.6, PM5 entry: applicability "Applicable"; instructionsToUse "Based on location of the most C-terminal known pathogenic variant, p.Arg3047*. Use as PM5_Supporting (not moderate). Do not use for start-loss variants. Do not use for missense changes: Multiple amino acid substitutions at the same residue can be pathogenic or benign and bioinformatic tools cannot yet confidently distinguish them."
• ATM HBOP CSPEC v1.6, PM5_Supporting strength rule (verbatim): "DNA Variants: Apply to all NMD-prone, truncating variants receiving PVS1 (must be at VeryStrong) with premature termination codons upstream of p.Leu3048. Splice Variants: Apply only to observed (not predicted) NMD-prone, truncating variants receiving PVS1(RNA) (must be at VeryStrong with high-quality RNA data) with premature termination codons upstream of p.Leu3048. This is not applicable for predicted splice impact without RNA data or PVS1(RNA) at reduced strength."
• VUA SpliceAI result: max delta 0.039 (DS_AG 0.039) - no predicted splice event and below the ATM VCEP PP3 splicing threshold (SpliceAI >= 0.2); the variant is intronic and non-truncating, so no PVS1 and no premature termination codon.
• pm5_candidates.json: found = false, eligible_for_classic_pm5_search = false, total_candidates_seen = 0, same_residue_candidates = 0; recommendation level "not_applicable" with instruction to review whether any same-residue comparator satisfies the governing PM5 rule. No comparator exists for a p.? variant.

Still missing

• No RNA-based splice study and no observed NMD-prone/truncating consequence is available for c.7308-9C>T in this case, so the VCEP's RNA-dependent PM5 path could not be evaluated positively; that path is explicitly excluded for predicted-only splice impact without RNA data.

Sources

Reasoning

NM_000051.4:c.7308-9C>T is an intronic variant 9 nucleotides upstream of the exon 50 acceptor site (between the canonical -20..-1 acceptor region and the exon), so the applicable computational path is the splice path (SpliceAI), not REVEL; no REVEL score exists for a non-coding substitution and none is used. The governing ClinGen HBOP ATM VCEP v1.6 specifies PP3 supporting for predicted splice impact when SpliceAI is >=0.2 for silent, missense/in-frame, and intronic variants outside the donor and acceptor +/-1,2 sites (Walker et al. 2023, as adopted by the VCEP). The observed SpliceAI maximum delta is 0.039 (DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001), well below the 0.2 threshold; the independent Pangolin scores (splice gain 0.071, splice loss -0.002) are likewise low and do not change that conclusion. The variant is outside the donor and acceptor +/-1,2 dinucleotides, so it is in scope and was evaluated rather than excluded. No on-record VCEP material pre-assigns a PP3 code for this variant: the exact change c.7308-9C>T was not found in Suppl_TableS1_PMID 40580951.xlsx (see gaps_remaining), and the VCEP-approved assay table clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx is a PS3/BS3 strength-calibration table that assigns no PP3/BP4 codes. PP3 is therefore not met.

Rule

ATM VCEP v1.6 PP3 Supporting: Splicing - predicted impact via SpliceAI >=0.2 for silent, missense/in-frame, and intronic variants outside donor and acceptor 1,2 sites.

Evidence considered

• ClinGen HBOP ATM VCEP v1.6 (cspec doc 639508985) PP3 rule: splicing - predicted impact via SpliceAI >=0.2 for silent, missense/in-frame, and intronic variants outside donor and acceptor +/-1,2 sites; the VCEP uses SpliceAI as its sole splice predictor.
• The case SpliceAI lookup for NM_000051.4:c.7308-9C>T reports DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001, i.e. a maximum delta of 0.039, plus Pangolin splice gain 0.071 and splice loss -0.002.
• The case normalization places c.7308-9C>T in intron 49, 9 nucleotides upstream of the exon 50 acceptor site, outside the acceptor -1,-2 dinucleotide, so the VCEP splice path applies.
• The regulatory SpliceAI thresholds quoted in the generic SVI calibration (moderate >=0.5, supporting >=0.2; PMID:30661751) are consistent with, and not more permissive than, the VCEP's own >=0.2 cutoff, so applying the VCEP rule governs.
• No PP3 code for this variant was pre-assigned anywhere in the VCEP materials; the exact change was not present in Suppl_TableS1_PMID 40580951.xlsx.

Still missing

• No entry for c.7308-9C>T, c.7308-9, or the corresponding hg38 position 108330205 was found in Suppl_TableS1_PMID 40580951.xlsx (searched as converted full text; intron 49 positions c.7307+6 through c.7308-6 are absent from the table, which jumps from c.7307+5C>T at hg38 108329243 to c.7308-5G>T at hg38 108330209). This is a weaker statement than the evidence being absent.
• No variant-specific RNA study was available, so no observed splice effect could be assessed; the VCEP permits PP3 together with BP7(RNA) only when RNA data show no variant-specific impact.

Sources

Reasoning

Under the applied rule PP5 is triggered only by an exact-variant ClinVar expert-panel classification of Pathogenic or Likely pathogenic; ordinary laboratory submissions, aggregate ClinVar labels and non-expert assertions must never trigger it. The exact-match ClinVar record for NM_000051.4:c.7308-9C>T is VCV000236769 (variation ID 236769), classified Likely benign with review status 'criteria provided, single submitter' (1 star) and no expert-panel submissions (expert_panel_submissions = 0, expert-panels map empty, usable_for_criterion_leads = 0). Both submissions are ordinary clinical-testing laboratories - Labcorp Genetics/Invitae (SCV001674612) and Myriad Genetics (SCV005084900), each flagged is_expert_panel = false - and neither asserts Pathogenic or Likely pathogenic for this variant. The governing HBOP ATM VCEP additionally lists PP5 as 'Not Applicable for this VCEP' ('This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee'); the lab's expert-panel override precondition (an exact-variant expert-panel P/LP classification) is absent, so it is the absence of any expert-panelled assertion, not the VCEP exclusion, that decides this call. PP5 is therefore not met.

Rule

Not Applicable for this VCEP (SVI VCEP Review Committee). Lab override: apply PP5 at Supporting only when an exact-variant ClinVar expert-panel classification is Pathogenic or Likely pathogenic.

Evidence considered

• cspec: PP5 applicability 'Not Applicable for this VCEP' - 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee.'
• Applied rule: PP5 requires an exact-variant ClinVar expert-panel classification of Pathogenic/Likely pathogenic; ordinary laboratory submissions, aggregate ClinVar labels and non-expert assertions never trigger it.
• ClinVar VCV000236769 (variation ID 236769, exact transcript-HGVS match NM_000051.4(ATM):c.7308-9C>T, match_score 16, matched_by transcript_hgvs): classification Likely benign; review status 'criteria provided, single submitter' (1 star); 2 submissions from 2 clinical laboratories.
• Both submissions are ordinary clinical laboratories (Labcorp Genetics/Invitae SCV001674612; Myriad Genetics SCV005084900); is_expert_panel = false for both; expert-panel submissions = 0; expert-panels map empty; criterion-lead submissions = 0.
• No Pathogenic or Likely pathogenic assertion for this exact variant exists from any expert panel, so the PP5 trigger condition is absent.
• None of the six reviewed PMIDs supplies an expert-panel variant classification for c.7308-9C>T: PMID:28492532 (Sherloc) is a generic ACMG/AMP criteria-refinement framework and the remaining papers never mention the variant.

Still missing

• If an expert panel later issues an exact-variant Pathogenic or Likely pathogenic classification for c.7308-9C>T, PP5 would then apply at Supporting strength and this call should be revisited.

Sources

Reasoning

Not met. The governing specification is the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel ATM specification v1.6 (cspec doc 639508985), which sets BA1 (Benign Stand-Alone) at a Grpmax Filtering AF >0.5% in gnomAD. This variant, NM_000051.4:c.7308-9C>T (NC_000011.10:g.108330205C>T / NC_000011.9:g.108200932C>T), is present in gnomAD v4.1 at a total AF of 6.20e-07 (1/1,612,070 alleles, 0 homozygotes; exome component 1/1,459,862 = 6.85e-07, genome component 0/0 in this capture). The highest observed subpopulation frequency is European (non-Finnish) at 8.49e-07 (1/1,178,460 alleles, 0 homozygotes); every other ancestral subpopulation in v4.1 is 0, and NFE_XX is 0/610,786. The variant is reported absent from gnomAD v2.1 (exome) and from gnomAD-Canada v1.0. That highest subpopulation AF is roughly 5,900-fold below the 0.5% stand-alone cutoff, far beyond any plausible effect of the capture gap in grpmax filtering AF (which was returned null), so BA1 cannot be met under any reasonable reading of the data. The VCEP provides no ATM-specific guidance directing a non-cancer or exome-only population source (its release notes only state that additional gnomAD guidance was provided, without specifying a filtered dataset), so the all-comers GNOMAD_V2_1/GNOMAD_V4_1 entries govern here; the non-cancer subsets were also queried and report the variant absent, which is concordant rather than decisive.

Rule

Grpmax Filtering AF >.5% in gnomAD (ClinGen HBOP ATM VCEP v1.6, BA1, Benign Stand Alone)

Evidence considered

• ClinGen HBOP ATM VCEP v1.6 BA1 rule: 'Grpmax Filtering AF >.5% in gnomAD' (Benign Stand Alone).
• gnomAD v4.1 (GRCh38, chr11-108330205-C-T): total_ac=1, total_an=1612070, total_af=6.2032e-07, total_hom=0; exome_ac=1, exome_an=1459862, exome_af=6.85e-07, exome_hom=0; genome_ac=0, genome_an=0.
• Highest-frequency subpopulation in gnomAD v4.1: European (non-Finnish), ac=1, an=1178460, af=8.4857e-07, homozygotes=0. All other subpopulations (AMR, FIN, AMI, EAS, MID, SAS, ASJ, AFR, remaining) are 0 in the capture; NFE_XX is 0/610786 and NFE_XY is 1/567674.
• gnomAD v2.1 (GRCh37, 11-108200932-C-T, exome): search_status 'absent', source_data empty.
• gnomAD v4.1 grpmax filtering AF fields (exome_grpmax_faf, genome_grpmax_faf, joint_grpmax_faf, grpmax_faf) were all null in the captured source_data; the 0.5% stand-alone threshold is exceeded by neither the total AF nor the highest ancestry-specific AF.
• ATM HBOP VCEP v1.6 release notes: 'Provided additional guidance on use of gnomAD data'; no ATM-specific instruction names a non-cancer or exome-only dataset, so the all-comers GNOMAD_V2_1/GNOMAD_V4_1 entries were used as the primary source.

Still missing

• The gnomAD v4.1 scrape returned grpmax filtering AF as null (joint, exome and genome grpmax all null), so BA1 was evaluated against the highest observed ancestry-specific allele frequency (NFE 8.49e-07) rather than a true grpmax filtering AF; this cannot change the verdict because that value is ~5,900-fold below the 0.5% threshold.

Sources

Reasoning

Not met. Under the ClinGen HBOP ATM VCEP v1.6, BS1 (Benign Strong) requires a Grpmax Filtering AF >0.05% in gnomAD. The highest observed allele frequency for NM_000051.4:c.7308-9C>T in any gnomAD subpopulation is European (non-Finnish) 8.49e-07 (1/1,178,460 alleles, 0 homozygotes) in gnomAD v4.1; the overall v4.1 AF is 6.20e-07 (1/1,612,070 alleles, 0 homozygotes), and the variant is absent from gnomAD v2.1 (exome) and gnomAD-Canada v1.0. The highest subpopulation frequency is about 590-fold below the 0.05% strong-benign threshold. Because the clinical question for ATM is a rare, high-penetrance cancer-predisposition allele, a single observed allele in ~1.18 million NFE chromosomes is far below any frequency compatible with a benign polymorphism, so BS1 is not met and cannot be rescued by the un-captured grpmax filtering AF value.

Rule

Grpmax Filtering AF >.05% in gnomAD (ClinGen HBOP ATM VCEP v1.6, BS1, Benign Strong)

Evidence considered

• ClinGen HBOP ATM VCEP v1.6 BS1 rule: 'Grpmax Filtering AF >.05% in gnomAD' (Benign Strong).
• gnomAD v4.1 total AF 6.2032e-07 (1/1,612,070 alleles, 0 homozygotes); highest subpopulation European (non-Finnish) AF 8.4857e-07 (1/1,178,460 alleles, 0 homozygotes).
• gnomAD v2.1 (exome) reports the variant absent; gnomAD-Canada v1.0 reports the variant absent; the gnomAD v2.1 non-cancer (exome-only) and gnomAD v3.1 non-cancer (genome-only) subsets also report the variant absent.
• The magnitude of the gap to the 0.05% threshold (~590-fold) makes the BS1 verdict insensitive to the missing grpmax filtering AF and to any ancestry-subpopulation ascertainment noise.

Still missing

• As for BA1, the gnomAD v4.1 grpmax filtering AF was null in the capture; the verdict rests on the highest ancestry-specific AF (NFE 8.49e-07), which is ~590-fold below the 0.05% BS1 threshold.

Sources

Reasoning

Not met. BS3 requires an assay showing that the variant does not damage ATM function (rescue of an ATM-specific feature and/or radiosensitivity under the VCEP v1.6 ladder: Supporting = rescues either; Moderate = rescues both). No functional data exist for c.7308-9C>T (NM_000051.4), an intronic variant 9 nt upstream of the exon 50 acceptor (NP_000042.3:p.?). The VCEP-approved assay table (clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx) is a calibration document whose three approved kinase-activity assays (PMID 18634022, PMID 19431188, PMID 11805335) express site-directed-mutagenised ATM cDNA and therefore cannot test an intronic -9 position; it reports no result for this variant, and neither radiosensitivity sheet lists it. Sun et al. 2025 (PMID 40580951) Table S1, the only SNV-level functional dataset available, contains no row for c.7308-9C>T (searched c.7308-9C>T, '7308-9', chr11:108330205), because its intronic coverage is restricted to offsets -5..-1 and +1..+5, so the variant was neither assayed nor modelled. The full-text papers in this case (PMID 15604628, PMID 24366376, PMID 35802134) report no ATM functional data. Absence of a functional defect is therefore unevidenced, not demonstrated, and BS3 is not met at any strength.

Rule

ClinGen HBOP ATM VCEP v1.6 BS3: requires documented rescue of an ATM-specific feature and/or radiosensitivity (Supporting = either, Moderate = both); no assay result exists for this variant -> criterion not met.

Evidence considered

• VCEP HBOP ATM framework (cspec doc 639508985, v1.6) defines the operative BS3 ladder: Moderate = 'Use when a variant rescues both an ATM specific feature (e.g. phosphorylation of ATM-specific targets) AND radiosensitivity'; Supporting = 'Use when a variant rescues EITHER an ATM specific feature OR rescues radiosensitivity'. Both require an actual rescue readout, which does not exist for this variant.
• The VCEP approved-assay table (clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx) covers only three cDNA-transfection kinase-activity assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335) plus a radiosensitivity sheet (Mitui 2009 approved); it lists the variants used as positive/negative validation controls and c.7308-9C>T is not among them, so it offers no BS3 support or refutation for this position.
• Searched 'Suppl_TableS1_PMID 40580951.xlsx' in full for 'c.7308-9C>T', '7308-9' and '108330205' with no match; the file's only intronic offsets are -5..-1 and +1..+5, confirming this -9 variant was outside the prime-editing saturation screen's scope and could not have generated a rescue/non-rescue result.
• The three full-text papers fetched for this case (PMID 15604628, PMID 24366376, PMID 35802134) were checked for ATM functional content and contain none, nor any mention of this variant.
• Surrounding-position context from the same table is not transferable evidence: the intronic neighbours c.7308-1G>C (Non-functional) and c.7308-5G>T (no call modelled) concern different nucleotides at different offsets and were not used to infer anything about c.7308-9C>T.
• SpliceAI max delta score 0.039 (DS_AG 0.039, DS_AL 0.001, DS_DG 0.003, DS_DL 0.001) was reviewed only as context; a computational splice prediction is not functional assay evidence for BS3.

Still missing

• c.7308-9C>T has no entry in 'clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx' (searched in full); the file calibrates assay strength and its validation variants do not include this position, so it provides no rescue/non-rescue readout for BS3.
• c.7308-9C>T has no entry in 'Suppl_TableS1_PMID 40580951.xlsx' (searched on c.7308-9C>T, 7308-9, chr11:108330205); intronic offsets -6 and beyond are absent from that dataset, so there is no Functional/Intermediate/Non-functional call and no DeepATM_predicted value for this variant.
• No RNA-based splicing assay (minigene, patient RNA, RT-PCR) or ATM kinase/radiosensitivity assay testing c.7308-9C>T was found in any source searched for this case; BS3 could only be considered if such a result were produced.
• Absence of functional evidence here is a coverage limit of the available datasets, not evidence of a benign functional effect - BS3 must not be inferred from the lack of an assay result.

Sources

Reasoning

Governing specification: the ClinGen HBOP ATM VCEP PM3/BP2 points table (ATM_PM3_BP2_1.5.pdf; re-issued textually unchanged as v1.6, the version governing this case). BP2 for ATM is likewise a points-accumulation code, and its benign direction is defined narrowly: points are earned only per unaffected (non-A-T) individual carrying the variant under assessment together with a P/LP ATM variant, at -4.0 (confirmed in trans), -2.0 (phase unknown), -2.0 (homozygous, laboratory setting) or -1.0 (homozygous, database setting, maximum -2.0). The v1.6 text adds the standard BP2 definition and the explicit requirement that, when assessing homozygous or in-trans variants with a P/LP ATM variant for possible downgrade in an unaffected individual, the individual must be 18 years or older with no evidence of A-T. Strength conversion: BP2_Supporting = -1 point, BP2_Moderate = -2 points, BP2_Strong = <= -4 points. Footnote 3 additionally bars the BP2 points route for a variant that is bioinformatically predicted or experimentally proven to have a clinically important effect on protein or mRNA splicing - a bar that would not exclude this variant (SpliceAI max delta 0.039), but which is irrelevant here because the point-generating observations are absent. Against that table, zero BP2 points were identified. No unaffected, non-A-T adult carrying c.7308-9C>T together with a VCEP-classified P/LP ATM variant is reported in any assembled source; there is no documented trans co-occurrence, no homozygous carrier in an unaffected adult, and no proband-level observation of any kind. The gnomAD v4.1 data show exactly one heterozygous allele in 1,612,070 (0 homozygotes), and gnomAD v2.1 and gnomAD-Canada show none, so no database-setting homozygous point accrues either. The ClinVar record (variation 236769) is Likely benign on two independent clinical-laboratory submissions, but that is a classification-level statement, not the per-individual observation the table requires, and the submission-level evidence codes are unavailable; no points can be derived from it. With 0 points, the variant falls short of the -1 point BP2_Supporting floor, so BP2 is not met. No strength is assigned, since no points were accrued in either direction.

Rule

ATM VCEP PM3/BP2 points table - BP2_Supporting requires <= -1 point accumulated from unaffected (non-A-T) adult individuals carrying the variant with a P/LP ATM variant (or homozygous, database setting); 0 points were accrued, so BP2 is not met.

Evidence considered

• ATM PM3/BP2 points table (v1.5 and v1.6): points per unaffected (non-A-T) individual carrying the variant under assessment with a P/LP ATM variant = -4.0 (confirmed in trans), -2.0 (phase unknown), -2.0 (homozygous, laboratory setting), -1.0 (homozygous, database setting, max -2.0); conversion BP2_Supporting = -1 point, BP2_Moderate = -2 points, BP2_Strong = <= -4 points.
• VCEP condition for the unaffected-carrier route: 'When assessing homozygous or in trans variants (with a likely pathogenic or pathogenic ATM variant) for possible downgrade in an unaffected individual, the individual should be 18 years or older with no evidence of A-T.' No such individual is documented for this variant in any assembled source.
• Table footnote 4: the homozygous database-setting point applies only to phenotyped individuals from clinical or research cohorts, NOT to frequency-based observations; gnomAD homozygote counts (0 in v4.1, absent in v2.1) therefore contribute no BP2 points in either direction.
• Table footnote 3: the unaffected-carrier BP2 route is unavailable for a variant bioinformatically predicted or experimentally proven to have a clinically important splicing/protein effect. SpliceAI max delta for this variant is 0.039 (no predicted impact), so this exclusion does not apply - the reason BP2 is not met is the absence of point-generating observations, not the splicing exclusion.
• gnomAD v4.1 for chr11-108330205-C-T: total AC 1, total AN 1,612,070, total AF 6.2032e-07, 0 homozygotes; exome AC 1/1,459,862; genome AC 0. gnomAD v2.1: variant absent. gnomAD-Canada v1.0: variant absent. No unaffected adult homozygous carrier is recorded in any population dataset.
• ClinVar variation 236769: Likely benign, 2 stars ('criteria provided, multiple submitters, no conflicts'), conditions Ataxia-telangiectasia syndrome and Familial cancer of breast, last updated 2026-06-27; two submissions (Myriad Genetics, origin unknown; Labcorp/Invitae, germline), both 'criteria provided, single submitter', no expert-panel review. This is a classification-level benign call, not the per-individual unaffected-carrier observation required for BP2 points under the VCEP table.
• Case literature packet: 6 PMIDs cited by ClinVar, none of which reports this variant or, in the three full-text-available papers (15604628, 24366376, 35802134), mentions ATM at all; grep of the evidence bundle for 'unaffected', 'in trans', 'homozyg', and 'phase' returns only VCEP specification text.

Still missing

• ATM_PM3_BP2_1.5.pdf (declared governing file for PM3/BP2, together with its v1.6 re-issue) was searched in full for 'c.7308-9C>T', 'p.?', and '7308' and contains no entry for this variant by any notation; the file provides the point-per-individual rules and the point-to-strength conversion, not a pre-assigned BP2 result for this variant. No pre-assigned code exists there.
• ClinVar submitter-level evidence codes for the two Likely benign submissions (SCV001674612 Labcorp/Invitae; SCV005084900 Myriad Genetics) are not available in the captured record, so it could not be determined whether either laboratory based its call on an unaffected adult carrying this variant in trans with a P/LP ATM variant - the observation that would actually generate BP2 points. The submission audit found 0 criterion-level leads from these submissions.
• No variant-specific literature search was performed; the 6-PMID literature packet derives from ClinVar citations and none of the papers mentions ATM. An unaffected-carrier observation in unfetched literature would therefore not have been seen.
• No parental-phase or family data (genotyping of parents, family history of ATM-related cancer in both lineages) are present in any assembled source, so a homozygous unaffected adult or an inferred in-trans configuration could not be established even had the variant been seen.

Sources

Reasoning

Under the applied rule BP6 is triggered only by an exact-variant ClinVar expert-panel classification of Benign or Likely benign; ordinary laboratory submissions, aggregate ClinVar labels and non-expert assertions must never trigger it. The exact-match record for NM_000051.4:c.7308-9C>T is VCV000236769, reviewed as 'criteria provided, single submitter' (1 star) - i.e. a single non-expert laboratory submission, which is the lowest ClinVar evidence tier and not an expert-panel assertion. Both submissions (Labcorp Genetics/Invitae SCV001674612 and Myriad Genetics SCV005084900) are flagged is_expert_panel = false, the expert-panel submission count is 0, the expert-panels map is empty, and no expert-panel record exists for this variant; the ClinVar record also carries no usable criterion-level evidence ('usable_for_criterion_leads' = 0). Although the classification text happens to read Likely benign, the applied rule keys strictly on who asserted it, so an ordinary two-laboratory Likely benign entry does not trigger BP6. The governing HBOP ATM VCEP in addition lists BP6 as 'Not Applicable for this VCEP' (SVI VCEP Review Committee recommendation) and defines no numeric threshold for the criterion. BP6 is therefore not met.

Rule

Not Applicable for this VCEP (SVI VCEP Review Committee). Lab override: apply BP6 at Supporting only when an exact-variant ClinVar expert-panel classification is Benign or Likely benign.

Evidence considered

• cspec: BP6 applicability 'Not Applicable for this VCEP' - 'This criterion is not for use as recommended by the ClinGen Sequence Variant Interpretation VCEP Review Committee.'
• Applied rule: BP6 requires an exact-variant ClinVar expert-panel classification of Benign/Likely benign; ordinary laboratory submissions and aggregate ClinVar labels never trigger it. No numeric threshold is defined for BP6 in the ATM VCEP text.
• ClinVar VCV000236769 (variation ID 236769, exact transcript-HGVS match): classification Likely benign; review status 'criteria provided, single submitter' (1 star).
• Submission audit: 2 submissions from 2 clinical laboratories, both is_expert_panel = false (Labcorp Genetics/Invitae SCV001674612; Myriad Genetics SCV005084900); expert-panel submissions = 0; expert-panels map empty; usable_for_criterion_leads = 0; no expert-panel record exists for this variant.
• The Likely benign label originates solely from ordinary clinical-testing laboratories with no validated supporting literature trail (both submissions flagged as lacking a validated PubMed trail), so it is not an expert-panel benign assertion and does not trigger BP6.

Still missing

• If an expert panel later classifies this exact variant as Benign or Likely benign, BP6 would apply at Supporting strength and this call should be revisited.
• No independent benign functional, segregation or population evidence for c.7308-9C>T was located beyond the two non-expert ClinVar assertions (gnomAD v4.1 AF 6.2032e-07, SpliceAI max delta 0.039, absent from COSMIC).

Sources

Reasoning

The governing ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel specification for ATM (Version 1.6, source key 'cspec') keeps PP1 applicable but only for the autosomal recessive condition (ataxia-telangiectasia), and does not permit its use in the autosomal dominant setting because co-segregation analysis in a lower-penetrance gene can yield false-positive results. The recessive rule is graded by the number of affected relatives carrying the variant alongside the second ATM allele identified in the proband: segregation in 1 affected relative gives Supporting strength, 2 affected relatives Moderate, and 3 or more affected relatives Strong. This case is a variant-only curation of NM_000051.4:c.7308-9C>T; it contains no proband, no affected relatives, no family genotypes, no phase information, and no documented informative meioses, co-segregation or lack of co-segregation. The three full-text papers retrieved for this case are BRCA1/BRCA2- and cancer-risk counselling guidance documents and none reports this variant, ATM, or any family segregation data. With zero informative affected relatives the VCEP's minimum Supporting threshold of one segregating affected relative cannot be reached, and there is no evidence either way on segregation, so PP1 is not_assessed rather than met or not met; family-level data would be needed to score it.

Rule

ATM VCEP PP1 (autosomal recessive condition): Supporting for segregation in 1 affected relative; Moderate for 2; Strong for 3 or more.

Evidence considered

• ClinGen HBOP ATM VCEP PP1 rule (AR condition): Supporting = segregation in 1 affected relative, Moderate = 2 affected relatives, Strong = 3 or more affected relatives; affected relatives must have both variants identified in the proband, and the AD condition is not to be used.
• No proband, affected relative, family genotype, parental testing result, phase information, informative meiosis count, or segregation/non-segregation finding is present anywhere in the evidence assembled for NM_000051.4:c.7308-9C>T.
• The three retrieved full texts (PMID:15604628, PMID:24366376, PMID:35802134) are cancer-risk/genetic-counselling guidance documents that do not name c.7308-9C>T, ATM segregation, or any family study, so they contribute no segregation evidence.

Still missing

• Pedigree data are required: the number and phenotype of affected relatives, their ATM genotypes, and confirmation that affected relatives carry both variants identified in the proband.
• Parental and other family testing results are absent, so phase, informative meiosis count, co-segregation and non-segregation cannot be evaluated.

Sources