ATM c.3332T>C, p.(Leu1111Pro)

Reasoning

PM2_Supporting is met. The ATM VCEP v1.5 specifies PM2_Supporting for frequency <=0.001% in gnomAD v4.1, including when n=1 in a single subpopulation. This variant has AF 3.09843e-06 overall, highest subpopulation AF 4.237877e-06, grpmax FAF 1.24e-06, and no observed homozygotes, all below the threshold.

Rule

Frequency <=0.001% in gnomAD v4 dataset; if n=1 in a single subpopulation, PM2_Supporting applies

Evidence considered

• ClinGen HBOP ATM VCEP v1.5 specifies PM2_Supporting at frequency <=0.001% in gnomAD v4.1 and states that n=1 in a single subpopulation is sufficiently rare.
• gnomAD v4.1 reports 5/1,613,720 alleles overall (AF 3.09843e-06; 0.000309843%), a highest population AF of 4.237877e-06 in European non-Finnish individuals (5/1,179,836), grpmax FAF 1.24e-06, and zero homozygotes.
• The variant is absent from gnomAD-Canada v1.0, providing additional population rarity context without changing the VCEP threshold assessment.

Sources

Reasoning

ATM VCEP v1.5 PS1 guide (ATM_PS1_1.5.pdf) applies PS1 for missense changes when the variant under assessment produces the same amino acid change as an established pathogenic/likely pathogenic variant, provided splicing is ruled out for both. NP_000042.3:p.(Leu1111Pro) is the amino-acid change under assessment. The case's own PM5/PS1 comparator search (pm5_candidates.json, summarized in prefetch.json) found zero candidate variants at this residue at all (total_candidates_seen=0, same_residue_candidates=0) - there is no established P/LP variant reported with the same p.Leu1111Pro change, nor any other classified variant at codon 1111, in the collected evidence. Without an established P/LP comparator producing this exact amino-acid change, the PS1 prerequisite is not met.

Rule

ATM VCEP v1.5 PS1: same amino acid change as an established P/LP variant (missense path), splicing ruled out for both.

Evidence considered

• ATM VCEP v1.5 PS1 guide: standard missense PS1 path requires the variant under assessment to produce the same amino acid change as an established P/LP variant, with splicing ruled out for both alterations.
• pm5_candidates.json comparator collection (summarized in prefetch.json compact_evidence.pm5_candidates_summary) reports total_candidates_seen=0 and same_residue_candidates=0 for this variant, meaning no P/LP comparator at codon 1111 (same or alternate amino acid change) was found.

Still missing

• No ClinVar-confirmed pathogenic/likely-pathogenic variant with the identical p.Leu1111Pro change was identified in the collected evidence; if a comparator surfaces in a future evidence refresh, PS1 should be re-evaluated with splicing ruled out for both variants per the VCEP guide.

Sources

Reasoning

Variant is NM_000051.4:c.3332T>C / NP_000042.3:p.(Leu1111Pro), a missense substitution. ATM VCEP v1.5 PP3 requires, for missense variants, REVEL > 0.7333; the local REVEL lookup returns 0.644, below this threshold, so the missense sub-path does not meet PP3. The VCEP's PP3 text also allows a splicing sub-path (SpliceAI >= 0.2) for missense/in-frame/silent/eligible intronic variants; SpliceAI here gives a max delta score of 0.007 (DS_AG 0.007, DS_AL 0.0, DS_DG 0.003, DS_DL 0.001), far below the 0.2 threshold, so the splicing sub-path also does not meet PP3. BayesDel score (0.101685) is available from the local lookup but this pipeline has no verified published calibration/threshold for BayesDel PP3/BP4 use, so it is treated as not_available and was not used to support or refute PP3. Neither sub-path reaches the VCEP threshold, so PP3 is not met.

Rule

ATM VCEP v1.5 PP3: Missense REVEL > 0.7333 (observed 0.644, not met); Splicing SpliceAI >= 0.2 (observed max 0.007, not met).

Evidence considered

• Local REVEL lookup: score 0.644 for NC_000011.9:g.108150265T>C (source_registry key revel), below the ATM VCEP v1.5 PP3 missense threshold of REVEL > 0.7333.
• SpliceAI scores: DS_AG=0.007, DS_AL=0.0, DS_DG=0.003, DS_DL=0.001, max_delta_score=0.007, well below the ATM VCEP v1.5 PP3 splicing threshold of SpliceAI >= 0.2 for missense/silent/eligible-intronic variants.
• ATM VCEP v1.5 (cspec) PP3 rule text: Missense REVEL > 0.7333; Splicing: SpliceAI >= 0.2 for silent, missense/in-frame and intronic variants outside donor/acceptor +/-1,2 sites.

Still missing

• BayesDel score exists (0.101685) but has no verified published threshold available to this pipeline and could not be used to corroborate or refute PP3.
• VCEP-provided Suppl_TableS1 (PMID 40580951) computational-predictor columns (CADD, boostDM, EVE, AlphaMissense) were not extracted for this specific variant in the evidence bundle; not used.

Sources

Reasoning

BA1 is not met. The ATM VCEP v1.5 threshold is grpmax filtering AF >0.5% in gnomAD v4.1; the observed grpmax FAF is 1.24e-06 (0.000124%), far below the threshold.

Rule

Grpmax Filtering AF >0.5% in gnomAD v4 dataset

Evidence considered

• ClinGen HBOP ATM VCEP v1.5 specifies BA1 at grpmax filtering AF >0.5% in gnomAD v4.
• gnomAD v4.1 reports 5/1,613,720 alleles overall, AF 3.09843e-06, and grpmax FAF 1.24e-06; these values are substantially below the BA1 threshold.

Sources

Reasoning

BS1 is not met. The ATM VCEP v1.5 threshold is grpmax filtering AF >0.05% in gnomAD v4.1; the observed grpmax FAF is 1.24e-06 (0.000124%), below the threshold.

Rule

Grpmax Filtering AF >0.05% in gnomAD v4 dataset

Evidence considered

• ClinGen HBOP ATM VCEP v1.5 specifies BS1 at grpmax filtering AF >0.05% in gnomAD v4.
• gnomAD v4.1 reports a maximum population AF of 4.237877e-06 in the European non-Finnish population and grpmax FAF of 1.24e-06, both below 0.05%.

Sources

Reasoning

For this missense variant, the governing ATM VCEP v1.5 BP4 text nominally offers a missense sub-path (REVEL <= 0.249) and a splicing sub-path (SpliceAI <= 0.1). Applying the curator override recorded in this case's vcep_materials gene_notes: for missense variants, BP4 must be evaluated ONLY via the missense (REVEL) sub-path, never via the splicing sub-path, because a clean SpliceAI score only rules out a splicing mechanism and does not clear the variant's protein-level effect. REVEL = 0.644 falls in the 0.249-0.7333 gray zone (neither <= 0.249 required for BP4 nor > 0.7333 as for PP3), so per the override policy BP4 is not met for this variant regardless of the SpliceAI value (0.007, which would otherwise satisfy the literal splicing sub-path text). This override changes the outcome from what the VCEP's literal text alone would produce: read literally, the VCEP's splicing sub-path text ("any variant has the potential to impact splicing") would allow BP4 via SpliceAI <= 0.1 even for a missense variant, which would call BP4 met; the curator override blocks that route for missense variants and BP4 is therefore not_met here. BayesDel (0.101685) was also available but is not usable for BP4 because this pipeline has no verified published threshold for it.

Rule

ATM VCEP v1.5 BP4 (as modified by curator override in this case's vcep_materials): Missense variants assessed via REVEL <= 0.249 sub-path only; splicing sub-path (SpliceAI <= 0.1) not applied to missense variants. REVEL observed 0.644 (gray zone, not <= 0.249) => not_met.

Evidence considered

• Local REVEL lookup: score 0.644, in the 0.249-0.7333 gray zone, not <= 0.249 required for the ATM VCEP v1.5 BP4 missense sub-path.
• SpliceAI scores: max_delta_score=0.007 (<=0.1), which would satisfy the VCEP's literal splicing sub-path text, but is not applied to this missense variant under the curator override recorded in vcep_materials gene_notes (BP4 missense/splicing policy), which restricts missense variants to the REVEL sub-path only.
• vcep_materials gene_notes curator override: for missense variants, evaluate BP4 via the missense sub-path (REVEL) only; do not apply the splicing sub-path (SpliceAI <= 0.1) even though the VCEP's own text permits it; if REVEL is in the 0.249-0.7333 gray zone or unavailable, BP4 is not_met/not_assessed regardless of SpliceAI.
• Local BayesDel lookup: score 0.101685, not usable for BP4 because no verified published calibration threshold is available to this pipeline for BayesDel.

Still missing

• REVEL 0.644 sits in the VCEP's uncalibrated gray zone (0.249-0.7333); no VCEP-specified rule resolves BP4 status in this range beyond not_met/not_assessed.
• BayesDel score exists (0.101685) but has no verified published threshold available to this pipeline and could not be used to support BP4.

Sources

Reasoning

ATM VCEP v1.5 calibrates PS3/BS3 strength ONLY for three named approved assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335), all stable-transfection ATM cDNA rescue assays in an ATM-null LCL line scored by kinase-activity Western blot (+/- radiosensitivity). The VCEP methodology table (clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx) lists only its own named P/LP and B/LB validation-control variants for each of the three papers, and NM_000051.4:c.3332T>C (p.Leu1111Pro) does not appear among them, so no VCEP-approved-assay result exists for this variant. The only functional data source found for this variant is Sun et al. 2025 (PMID 40580951, Suppl_TableS1_PMID 40580951.xlsx), a prime-editing saturation genome editing + deep-learning screen. Its row for c.3332T>C shows Classification="Functional", Confidence="High", DeepATM_predicted="No" (i.e. a directly-measured screen result, not a model extrapolation) with Combined_score -0.392. A "Functional" (normal-function) direct readout argues against, not for, a damaging effect, so this source does not support PS3 for this variant. Additionally the VCEP's own index guidance explicitly flags that PMID 40580951 is not one of its three approved PS3/BS3 assays and may postdate v1.5, so even a direct measurement from it is not yet VCEP-calibrated evidence and should route to human review rather than an assigned strength.

Rule

ATM VCEP v1.5 PS3: requires a positive (loss-of-function) result in one of the three approved assays (Mitui 2009, Barone 2009, Scott 2002); Strong is explicitly disallowed by the VCEP ("Do not use as strong"), Moderate requires failure to rescue BOTH ATM-specific kinase feature AND radiosensitivity, Supporting requires failure to rescue the ATM-specific kinase feature alone.

Evidence considered

• clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx (ATM kinase activity and ATM radiosensitivity sheets) lists validation-control P/LP and B/LB variants for each of its three approved papers (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335); c.3332T>C is not among any of the listed validation variants for either sheet, indicating this variant was not tested in one of the VCEP-approved PS3/BS3 assays.
• Suppl_TableS1_PMID 40580951.xlsx row for c.3332T>C (hg38 108279538 T>C, Protein_change L1111P, exon_23): Classification='Functional', Confidence='High', DeepATM_predicted='No', Combined_score=-0.392332281518224 — a directly-measured (not model-predicted) normal-function result from the prime-editing saturation genome editing screen.
• ATM VCEP v1.5 index guidance explicitly instructs that PMID 40580951 is not one of the three named approved PS3/BS3 assays and may postdate the current specification, so even a direct-measurement result from it should not be silently assigned a PS3/BS3 strength and should instead flag needs_human_review.

Still missing

• No VCEP-approved PS3/BS3 assay (Mitui 2009, Barone 2009, or Scott 2002) has tested c.3332T>C/p.Leu1111Pro per the VCEP's own validation-control listings.
• The only direct functional data available (Sun et al. 2025, PMID 40580951) reports a normal-function ("Functional") result for this variant, which is directionally inconsistent with PS3 and does not have a VCEP-calibrated strength assignment.
• PMID 40580951 full text was not retrieved for this case (only its supplementary Table S1 data extract), so assay quality/controls for the specific screen result could not be independently reviewed.

Sources

Reasoning

The ATM HBOP VCEP specifies PS4 as applicable only for a qualifying case-control study with p-value <=0.05 and an odds ratio, hazard ratio, or relative risk >=2 or lower 95% confidence interval >=1.5. The case bundle provides no case-control counts, effect estimate, confidence interval, or p-value for this exact variant, so PS4 cannot be assigned.

Rule

Case-control studies; p-value <=0.05 AND (odds ratio, hazard ratio, or relative risk >=2 OR lower 95% CI >=1.5).

Evidence considered

• The ATM HBOP VCEP v1.5 framework states that PS4 requires case-control studies with p-value <=0.05 and an odds ratio, hazard ratio, or relative risk >=2 or lower 95% confidence interval >=1.5.
• The consolidated case-evidence bundle contains no exact-variant case-control effect estimate, confidence interval, or p-value for NM_000051.4:c.3332T>C (p.Leu1111Pro).

Still missing

• No qualifying exact-variant case-control study with analyzable affected-case and control data was identified.

Sources

Reasoning

The ATM VCEP PM3 specification applies to ataxia-telangiectasia and requires qualifying affected-proband evidence involving a pathogenic or likely pathogenic allele in trans, with strength assigned from phenotype and phase points. The consolidated case-evidence bundle contains no affected-proband observation, second pathogenic/likely pathogenic ATM variant, parental testing, phase result, homozygous affected individual, or documented cis/trans relationship for NM_000051.4:c.3332T>C (p.Leu1111Pro). PM3 therefore cannot be assigned.

Rule

ATM VCEP v1.5 PM3 points: confident A-T phenotype 4.0 confirmed in trans, 2.0 phase unknown, 1.0 second variant unidentified or VUS, 2.0 homozygous (maximum two individuals); consistent phenotype values are half these. PM3 Supporting=1.0, Moderate=2.0, Strong=4.0, Very Strong=8.0 points.

Evidence considered

• ClinGen HBOP ATM VCEP v1.5 specifies points-based PM3 for ataxia-telangiectasia: 4 points for a confident affected proband with a confirmed-in-trans pathogenic variant, 2 points when phase is unknown, 1 point when the second variant is unidentified or a VUS, and 2 points for a qualifying homozygous observation subject to the VCEP limit; 1, 2, 4, and 8 points map to supporting, moderate, strong, and very strong PM3, respectively.
• The case bundle identifies the exact variant and ClinVar variation 230062, but its ClinVar submission audit reports classifications and citation trails only; it provides no affected-proband case, second pathogenic/likely pathogenic ATM variant, or phase information for this variant.
• No PM3-bearing publication was available in the case bundle with an auditable affected-proband and phase observation for this exact variant.

Still missing

• Affected-proband phenotype and relationship to the variant are not documented.
• No qualifying pathogenic or likely pathogenic ATM variant in trans is documented.
• No parental or other testing establishes phase, and no qualifying homozygous or inferred-trans observation is documented.

Sources

Reasoning

PP1 is applicable only through the ATM VCEP autosomal-recessive segregation path, which requires affected relatives to have both variants identified in the proband. The case-evidence bundle contains no documented affected relatives, parental testing, family genotypes, meioses, phenotype-segregation table, or other informative segregation observations for NM_000051.4:c.3332T>C (p.Leu1111Pro). No PP1 strength can therefore be assigned.

Rule

AR condition: affected relatives must have both variants identified in the proband; segregation in 1 affected relative is Supporting, 2 affected relatives is Moderate, and at least 3 affected relatives is Strong.

Evidence considered

• ATM VCEP v1.5 allows PP1 for autosomal recessive disease when affected relatives have both variants identified in the proband.
• The consolidated case-evidence bundle provides no variant-specific family segregation, parental-testing, affected-relative, or meiosis data.

Still missing

• Documented affected relatives with phenotype consistency and confirmed genotypes
• Parental or other family testing establishing that each affected relative carries both variants identified in the proband
• Informative meioses and phase information sufficient to count segregation

Sources

Reasoning

As with PS3, ATM VCEP v1.5 calibrates BS3 strength only from the same three approved assays (Mitui 2009, Barone 2009, Scott 2002), and c.3332T>C/p.Leu1111Pro does not appear in that methodology table's listed validation variants, so no VCEP-approved-assay result exists for this variant. The Sun et al. 2025 (PMID 40580951) Table S1 row for c.3332T>C does show a directly-measured (DeepATM_predicted=No, high confidence) "Functional" (normal-function) classification, which is directionally consistent with BS3 (rescue of ATM function). However, per the VCEP's own index guidance, PMID 40580951 is not one of the three approved BS3 assays and may postdate the v1.5 specification, so this pipeline cannot assign it a VCEP-calibrated BS3 strength; it must instead be flagged for human review rather than silently mapped to BS3_Supporting/Moderate.

Rule

ATM VCEP v1.5 BS3: default strength Benign Strong, but the VCEP's own rules only grant Moderate (rescue of both ATM-specific kinase feature AND radiosensitivity) or Supporting (rescue of ATM-specific kinase feature alone) using the three approved assays; radiosensitivity-only rescue is not ATM-specific and gets no weight.

Evidence considered

• clingen_hbop_atm_supplementary_tables_1_and_2_v1.xlsx (ATM kinase activity and ATM radiosensitivity sheets) lists validation-control P/LP and B/LB variants for each of its three approved papers (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335); c.3332T>C is not among any of the listed validation variants for either sheet, indicating this variant was not tested in one of the VCEP-approved PS3/BS3 assays.
• Suppl_TableS1_PMID 40580951.xlsx row for c.3332T>C (hg38 108279538 T>C, Protein_change L1111P, exon_23): Classification='Functional', Confidence='High', DeepATM_predicted='No', Combined_score=-0.392332281518224 — a directly-measured normal-function result from the prime-editing saturation genome editing screen, directionally consistent with BS3 but not from a VCEP-approved assay.
• ClinVar submission SCV007637234 (Copenhagen University Hospital, Rigshospitalet) evidence signal cites PMID 40580951 stating 'functional analysis indicates that the variant does not affect ATM function', corroborating the direction (normal function) of the Sun et al. 2025 data but not itself a VCEP-approved-assay source.
• ATM VCEP v1.5 index guidance explicitly instructs that PMID 40580951 is not one of the three named approved PS3/BS3 assays and may postdate the current specification, so even a direct-measurement result from it should not be silently assigned a PS3/BS3 strength and should instead flag needs_human_review.

Still missing

• No VCEP-approved PS3/BS3 assay (Mitui 2009, Barone 2009, or Scott 2002) has tested c.3332T>C/p.Leu1111Pro per the VCEP's own validation-control listings.
• Sun et al. 2025 (PMID 40580951) reports a directly-measured normal-function result for this variant (Classification='Functional', Confidence='High', DeepATM_predicted='No') that is directionally supportive of BS3, but this assay/paper is not on the VCEP's approved list and has no VCEP-calibrated strength, so it cannot be scored under the current cspec without human adjudication.
• PMID 40580951 full text was not retrieved for this case (only its supplementary Table S1 data extract), so assay quality/controls for the specific screen result could not be independently reviewed.

Sources

Reasoning

The ATM VCEP BP2 specification requires unaffected, non-ataxia-telangiectasia individuals carrying the assessed variant with a pathogenic or likely pathogenic ATM variant in trans, scored by confirmed or unknown phase and by laboratory versus database setting. The consolidated case-evidence bundle contains no unaffected carrier, co-occurring pathogenic/likely pathogenic ATM variant, or cis/trans phase information for NM_000051.4:c.3332T>C (p.Leu1111Pro). BP2 therefore cannot be assigned.

Rule

ATM VCEP v1.5 BP2 points: unaffected non-A-T carrier with a pathogenic/likely pathogenic variant in trans -4.0 confirmed trans, -2.0 phase unknown, or homozygous values of -2.0 in a laboratory setting and -1.0 in a database setting (maximum -2.0); BP2 Supporting=-1.0, Moderate=-2.0, Strong≤-4.0.

Evidence considered

• ClinGen HBOP ATM VCEP v1.5 specifies BP2 points for an unaffected non-A-T individual carrying the assessed variant with a pathogenic or likely pathogenic variant in trans: -4.0 for confirmed trans, -2.0 when phase is unknown, and homozygous observations are limited to a maximum of -2.0 with laboratory/database-specific values; -1.0 maps to BP2 Supporting, -2.0 to Moderate, and ≤-4.0 to Strong.
• The case bundle identifies ClinVar variation 230062 and reports seven clinical-laboratory submissions, but it contains no unaffected-carrier co-occurrence record or phase data that meet the ATM BP2 definition.
• No BP2-bearing publication was available in the case bundle with an auditable unaffected co-occurrence and phase observation for this exact variant.

Still missing

• No unaffected non-A-T carrier observation is documented.
• No co-occurring ATM pathogenic or likely pathogenic variant is documented in such an individual.
• No phase result establishes the variants in trans or supports the applicable laboratory/database point value.

Sources