Five-step scale from benign to pathogenic. This variant is pathogenic. Select a step to read what it means.
This variantScenario
BenignPathogenic
Pathogenic.A pathogenic 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: the nonsense change c.5692C>T creates a premature stop (p.Arg1898Ter) in exon 38 of 63, predicted to trigger nonsense-mediated decay.
PM2Supporting
PM2 · supporting
Met: the allele is essentially absent from population databases (gnomAD v4.1 AF 3.72e-06, 6/1,613,522 alleles), below the 0.001% rarity threshold.
PM3Strong
PM3 · strong
Met: two unrelated ataxia-telangiectasia probands carry this allele in trans with other null variants (6.0 points on the VCEP table, strong).
PM5Supporting
PM5 · supporting
Met: the truncation at p.Arg1898Ter lies upstream of the VCEP's p.Arg3047 boundary, meeting the gene-specific PM5 rule.
PP5Supporting
PP5 · supporting
Met: the ClinGen HBOP expert panel (3-star) classified this exact variant as Pathogenic.
Benign criteria applied · 0
No benign criterion marked as applied in the recorded assessment.
Recorded classification: Pathogenic
Pathogenic: PVS1 (very strong, truncating null variant) plus PM3 (strong, 6.0 points from two A-T probands) satisfies the ATM VCEP's Pathogenic rule.
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
6 / 1,613,522 alleles
gnomAD v4.1: 6 / 1,613,522 alleles. Recorded AF 3.71857e-06. Recorded homozygote count 0.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 7.96794e-06; MAF= 0.00080%, 2/251006 alleles, homozygotes = 0) and has highest observed frequency in the South Asian population (AF= 3.26819e-05; MAF= 0.00327%, 1/30598 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: 5 met · 6 not met · 2 not assessed · 15 not applicable. Open any row for the full reasoning, the rule and the evidence behind it.
Variant consequence: NM_000051.4:c.5692C>T -> NP_000042.3:p.(Arg1898Ter), a nonsense null variant in exon 38 of 63.
PVS1 met at Very Strong: nonsense variant, NMD-prone (PTC upstream of the p.Arg3047 boundary and of the last exon; directly demonstrated for this allele by cycloheximide NMD assay in PMID 25077176), truncating the entire FAT/PI3-K/FATC catalytic module; no extreme-3'-end caveat.
PM4 not applicable: VCEP restricts PM4 to stop-loss variants; this is a stop-gain.
BP3 not applicable: VCEP marks BP3 Not Applicable for ATM, and the variant is not an in-frame repeat-region indel.
ClinVar's Pathogenic classification (including the expert-panel submission) was not used to derive any criterion in this group; all criteria rest on variant consequence, VCEP rules, and primary literature (PMID 17124347, PMID 25077176).
The governing framework is the ClinGen HBOP ATM VCEP v1.5 specification.
This is a nonsense/truncating variant at p.Arg1898Ter; PS1, PM1, PP2, and BP1 are not applicable under the ATM specification, while the ATM-specific PM5 truncation rule is met at Supporting strength because the termination codon is upstream of p.Arg3047.
Governing framework: ClinGen HBOP ATM VCEP v1.5 (cspec doc 639508985) - used in preference to generic ACMG/AMP. The variant under assessment is ATM c.5692C>T, p.(Arg1898Ter) (NM_000051.4), a nonsense variant.
PS2 and PM6 are declared Not Applicable by the VCEP for both ATM disease indications (AD ATM-related cancer predisposition and AR ataxia-telangiectasia); de novo evidence is not used for this gene, and no de novo occurrence is reported for this variant.
BS4 is declared Not Applicable by the VCEP; lack-of-co-segregation in A-T families is not weighted under the specification.
PP1 is applicable (AR segregation rule) but not met: the variant appears only in unrelated probands (AT50RM, PMID 17124347; proband LCL AT213LA in trans with c.6200C>A, PMID 25077176). No affected-relative segregation is documented in any fetched source.
This group contributes no pathogenic or benign evidence: PS2/PM6/BS4 not applicable, PP1 not met.
The ClinGen HBOP ATM VCEP v1.5 (cspec) is the governing framework for this case, so PS3/BS3 are adjudicated strictly against its approved-assay methodology and strength calibration rather than generic ACMG.
Both PS3 and BS3 are returned as not_assessed: the framework requires confirmation that the variant under assessment was actually tested in one of the three VCEP-approved functional assays (Mitui 2009, Barone 2009, Scott 2002), and none of those papers was retrieved for this case, so no approved-assay readout for c.5692C>T exists in the evidence bundle.
The only variant-matched functional result available (Sun et al. 2025, Table S1: 'Non-functional') is inadmissible for PS3/BS3 because DeepATM_predicted=Yes / Confidence='Computational prediction' marks it as the DeepATM model's extrapolation rather than a direct prime-editing screen measurement, and the paper is not among the VCEP-approved assays.
Papers that mention c.5692C>T (PMID 17124347, PMID 25077176) provide variant identification only (nonsense R1898X; NMD-susceptible allele); no functional assay or rescue experiment is reported, so neither PS3 nor BS3 can be met or refuted on functional grounds. The absence of assay evidence should not be conflated with a benign functional readout.
If the three VCEP-approved assay papers or a direct functional measurement for c.5692C>T become available, PS3/BS3 should be re-adjudicated; until then the gap is a missing evidence source, not a resolved negative.
Governing framework: ClinGen HBOP ATM VCEP v1.5 (cspec), framework_mode = vcep; generic ACMG/AMP not used because an applicable specification exists.
PS4: not met. No case-control enrichment study exists for c.5692C>T; only case-level A-T observations (PMID:17124347 heterozygous proband AT50RM; PMID:25077176 second biallelic allele in LCL AT213LA) are documented, which cannot satisfy the VCEP PS4 statistical rule (p <= 0.05 AND OR/HR/RR >= 2 or lower 95% CI >= 1.5).
PP4: not applicable per ClinGen HBOP ATM VCEP v1.5 (cspec).
PP5: met at supporting strength via the exact-variant ClinVar expert-panel Pathogenic classification (ClinGen HBOP VCEP, 3-star, SCV005627280), applied per the global PP5/BP6 rule despite the VCEP 'Not Applicable' flag; laboratory submissions excluded.
BP5: not applicable per ClinGen HBOP ATM VCEP v1.5 (cspec).
BP6: not met. The only expert-panel classification of the exact variant is Pathogenic (supports PP5); no expert-panel Benign/Likely benign classification exists, and the VCEP declares BP6 Not Applicable.
ATM is evaluated under the ClinGen HBOP ATM VCEP v1.5 specification (cspec, vcep_atm_pm3_bp2_1_5), which provides a points-based PM3/BP2 table for the recessive Ataxia Telangiectasia phenotype; generic ACMG/AMP is not used because the applicable gene-specific specification exists.
PM3: 6.0 points across two unrelated A-T probands - AT213LA (PMID:25077176) with c.5692C>T confirmed in trans with founder variant c.6200C>A (4.0 points) and AT50RM (PMID:17124347) with c.5692C>T phase-unknown in a classical A-T compound heterozygote with IVS37+2T>C (2.0 points) - yielding PM3_Strong (>=4.0, <8.0 points).
BP2: no unaffected (non-A-T) individual carrying c.5692C>T with a P/LP variant in trans was found in any collected source, so BP2 accrues 0 points and is not met.
Neither criterion is inferred from another source's final classification (e.g., ClinVar Pathogenic or OncoKB Likely Oncogenic labels were not used as points evidence); points derive only from primary per-proband observations in the two variant-bearing publications.
Phase assignments were conservative: AT50RM scored as phase unknown because parental/phase testing was not reported; AT213LA scored as confirmed in trans because haplotyping explicitly segregated the co-occurring allele in trans.
NM_000051.4:c.5692C>T is a nonsense (stop-gain) SNV, p.(Arg1898Ter), in exon 38 of ATM. All three computational-evidence criteria owned by this group (PP3, BP4, BP7) are variant-class-dependent under the governing ClinGen HBOP ATM VCEP v1.5 specification.
PP3 is not applicable: the VCEP PP3 rule calibrates a missense sub-path (REVEL > 0.7333) and a splicing sub-path (SpliceAI >= 0.2 for silent, missense/in-frame, intronic variants outside +/-1,2); nonsense variants are covered by neither, and the VCEP disallows PP3 in addition to PVS1 codes for this PVS1-eligible truncating variant.
BP4 is not applicable: the VCEP BP4 missense sub-path (REVEL <= 0.249) requires a missense variant, and the lab's documented BP4 override restricts the splicing sub-path (SpliceAI <= 0.1) to variants whose primary plausible mechanism is splicing (synonymous/intronic outside +/-1,2). Although SpliceAI max delta 0.081 satisfies <= 0.1, applying it to a nonsense variant would treat splice neutrality as clearing a protein-level truncation that is demonstrably NMD-targeted (PMID 25077176); per the override this outcome differs from the VCEP's literal BP4 text and is stated explicitly.
BP7 is not applicable: the VCEP BP7 rule applies to synonymous and deep intronic variants only; a nonsense SNV does not qualify, and no RNA evidence (BP7(RNA)) exists in this case.
BayesDel (0.295247) was retrieved but is treated as not_available for PP3/BP4 because no verified published threshold is available; REVEL is unavailable for this variant; the Sun et al. 2025 (PMID 40580951) Table S1 computational row for this variant is a model extrapolation (DeepATM_predicted = Yes) and provides at most weak, non-calibrating corroboration. None of these change the not_applicable determination.
This group contributes no computational evidence codes to the overall classification; the variant's truncating mechanism is addressed by other groups' criteria (e.g. PVS1, PM3/BP2).
Population criteria adjudicated under the ClinGen HBOP ATM VCEP v1.5 specification (cspec), which governs because it is the applicable gene-specific framework.
BA1 not met: grpmax FAF 6.8e-07 (0.000068%) in gnomAD v4.1 is far below the VCEP threshold of >0.5%.
BS1 not met: grpmax FAF 6.8e-07 (0.000068%) in gnomAD v4.1 is far below the VCEP threshold of >0.05%.
PM2 met at Supporting: gnomAD v4.1 total AF 0.00037% (6/1,613,522 alleles, 0 homozygotes) is <= the 0.001% threshold, with n=1 in single subpopulations (SAS, AFR, Remaining individuals) matching the VCEP's explicit PM2_supporting condition.
BS2 not applicable: the ATM VCEP v1.5 specification declares BS2 Not Applicable.
Cross-dataset consistency: gnomAD v2.1 (AF 0.0008%, 2 alleles, 0 homozygotes) and absence in gnomAD-Canada v1.0 corroborate the rarity; no homozygotes observed in any dataset and all counts are single-digit alleles, so no threshold artifact (e.g., founder/paralog artifact) is apparent.
The variant is a nonsense allele (p.Arg1898Ter) in a recessive disorder; the very low population frequency with zero homozygotes is consistent with a rare pathogenic ATM allele and does not argue for benignity.
No criteria match.
Met · 5
PVS1Very strongMet: the nonsense change c.5692C>T creates a premature stop (p.Arg1898Ter) in exon 38 of 63, predicted to trigger nonsense-mediated decay.
Details ▾Hide ▴
Reasoning
NM_000051.4:c.5692C>T is a nonsense variant (NP_000042.3:p.Arg1898Ter) in exon 38 of 63, i.e. a null variant per the ClinGen HBOP ATM VCEP v1.5 PVS1 decision tree (cspec, vcep_atm_pvs1_1_5). Germline LOF is an established ATM disease mechanism (pvs1_gene_context: PVS1 gene gate eligible). The premature termination codon lies well upstream of the last exon and upstream of the p.Arg3047 NMD boundary the VCEP uses, so the transcript is expected to undergo NMD; this is directly demonstrated for this exact allele by PMID 25077176, where c.5692C>T was only detectable in patient LCL AT213LA RNA after cycloheximide inhibition of NMD. The truncation at residue 1898 removes the entire FAT, PI3-K (kinase) and FATC domains, which the VCEP treats as one contiguous functional domain, so the extreme-3'-end caveat does not apply and no critical C-terminal function is retained. PVS1_VeryStrong per the ATM PVS1 Decision Tree. SpliceAI (max delta 0.081) shows no competing splice effect that would alter this conclusion.
Rule
ATM VCEP v1.5 PVS1 Decision Tree: nonsense null variant with NMD-prone PTC upstream of p.Arg3047 that removes the contiguous FAT/PI3-K/FATC domains -> PVS1_VeryStrong
Evidence considered
ATM VCEP v1.5 (cspec) assigns PVS1 via the ATM PVS1 Decision Tree with strengths from Supporting to Very Strong; the decision guide covers nonsense null variants and its only caveats concern extreme-3'-end LOF variants and splice variants causing in-frame exon skipping that leaves the remainder of the protein intact - neither applies to this variant.
The VCEP decision guide uses default transcript NM_000051.3/ENST00000278616.8; the input annotation NM_000051.4 encodes the identical NP_000042.3 protein, so transcript relevance is confirmed and c.5692C>T maps to the same coding position.
c.5692C>T creates a premature termination codon at residue 1898 of 3056 (exon 38 of 63, per Sun et al. 2025 Table S1 row for c.5692C>T); the PTC is upstream of the p.Arg3047 boundary the VCEP treats as the NMD-prone region, and upstream of the last exon, so NMD is expected.
PMID 25077176: in patient-derived LCL AT213LA (which carries c.6200C>A in trans), the c.5692C>T allele was detected by RNA sequencing only after cycloheximide treatment inhibited nonsense-mediated mRNA decay, directly demonstrating that this allele is NMD-prone; the variant was confirmed by genomic DNA sequencing.
PMID 17124347: DHPLC/sequencing study of 22 classical A-T patients reported c.5692C>T (p.R1898X) in exon 40 of their numbering as a novel nonsense/truncated-protein ATM mutation in patient AT50RM (heterozygous).
The truncation at residue 1898 removes the FAT, PI3-K (catalytic kinase) and FATC domains; the ATM VCEP v1.5 guide states these are treated as one contiguous functional domain (with reference to PMID 28508083), so the variant abolishes the entire catalytic module rather than a dispensable distal region.
SpliceAI predicts no significant splice impact (max delta score 0.081), so there is no competing splicing mechanism to consider for this null variant.
pvs1_gene_context.json (registry key pvs1_gene_context): germline LOF is an established ATM disease mechanism under the cspec/VCEP framework and the PVS1 gene gate is eligible.
Still missing
No variant-specific RNA assay in the index case itself; NMD is inferred from PTC position (exon 38/63, upstream of p.Arg3047) and demonstrated for this allele in a patient-derived LCL (PMID 25077176).
PM2SupportingMet: the allele is essentially absent from population databases (gnomAD v4.1 AF 3.72e-06, 6/1,613,522 alleles), below the 0.001% rarity threshold.
Details ▾Hide ▴
Reasoning
ATM VCEP v1.5 (cspec) PM2 rule: Frequency <= 0.001% in gnomAD v4 dataset; if n=1 in a single subpopulation, that is sufficiently rare and PM2_supporting applies. gnomAD v4.1 total AF is 3.72e-06 (0.00037%, 6/1,613,522 alleles), well below the 0.001% threshold, and the allele count is distributed as n=1 in several single subpopulations (Remaining individuals, SAS, AFR) with n=3 in NFE; homozygotes 0. gnomAD v2.1 (AF 0.0008%, 2/251,006 alleles, highest subpopulation n=1 in SAS) and absence from gnomAD-Canada corroborate rarity. This is a known pathogenic ATM nonsense allele (p.Arg1898Ter) in a recessive disorder, so the observed very low allele frequency with no homozygotes is expected and no threshold artifact is apparent. PM2 met at Supporting strength per the VCEP rule text.
Rule
Frequency <= 0.001% in gnomAD v4 dataset; n=1 in a single subpopulation is sufficiently rare and PM2_supporting applies (ATM VCEP v1.5)
Evidence considered
ATM VCEP v1.5 (cspec) PM2 rule: 'Frequency <=.001% in gnomAD v4 dataset. If n=1 in a single sub population, that is sufficiently rare and PM2_supporting would apply.' The rule row is calibrated at Supporting strength.
gnomAD v4.1 (gnomad_v4): total AF 3.718573406498331e-06 = 0.00037%, which is <= 0.001%; AC 6/1,613,522; per-subpopulation AC: Remaining individuals 1/62,486 (AF 1.60e-05), SAS 1/91,062 (AF 1.10e-05), AFR 1/74,980 (AF 1.33e-05), NFE 3/1,179,726 (AF 2.54e-06); homozygotes 0 in all populations.
gnomAD v2.1 (gnomad_v2): total AF 7.967937021425783e-06 = 0.0008%; AC 2/251,006; n=1 in SAS (1/30,598) and n=1 in NFE (1/113,504); homozygotes 0.
gnomAD-Canada v1.0 (gnomad_canada): variant absent, further supporting rarity.
n=1 in single subpopulations (SAS, AFR, Remaining individuals) satisfies the VCEP's explicit PM2_supporting condition; PM2 met at Supporting strength.
No published literature reporting population frequency for this variant was found; all fetched papers (see publications_reviewed) were screened via paper_extracts.json and none report population-frequency data for c.5692C>T.
PM3StrongMet: two unrelated ataxia-telangiectasia probands carry this allele in trans with other null variants (6.0 points on the VCEP table, strong).
Details ▾Hide ▴
Reasoning
Governing framework: ClinGen HBOP ATM VCEP v1.5 points-based PM3/BP2 table (cspec + vcep_atm_pm3_bp2_1_5), applied to Ataxia Telangiectasia (autosomal recessive). Two unrelated A-T probands carry c.5692C>T (p.Arg1898Ter) in trans/compound-heterozygous configurations: (1) PMID:25077176 - AT213LA, one of 10 Mennonite A-T patients, carries c.5692C>T in trans with the founder A-TWinnipeg variant c.6200C>A (p.A2067D); trans phase confirmed by STR/SNP haplotyping (c.6200C>A haplotype segregated in trans) and heterozygosity of both variants confirmed by gDNA sequencing after NMD inhibition with cycloheximide; clinical diagnosis of A-T (A-TWinnipeg variant phenotype) taken at face value per VCEP instruction = 4.0 points (confident phenotype, confirmed in trans). (2) PMID:17124347 - AT50RM, a classical A-T patient from the Italian A-T Registry (inclusion criteria: progressive cerebellar ataxia, oculocutaneous telangiectasias, elevated AFP, chromosomal instability, radiosensitivity), is heterozygous for c.5692C>T (novel, p.R1898X) and heterozygous for IVS37+2T>C (splicing/truncated protein, previously reported); two null alleles in a classical A-T patient but no parental/phase testing described, so scored conservatively as phase unknown = 2.0 points (confident phenotype). Total = 4.0 + 2.0 = 6.0 points -> PM3_Strong (>=4.0, <8.0 points). The strength call is robust: if AT213LA were counted as 'consistent' rather than 'confident' phenotype (2.0 points), the total would be 4.0 points, still PM3_Strong. VCEP gate 'variant may not exceed general population frequency >0.01%' is met (gnomAD v4.1 AF = 3.72e-06, 0.00037%, grpmax FAF 6.8e-07). Per the VCEP rule, where the variant co-occurs with 2 different P/LP variants (c.6200C>A and IVS37+2T>C), one co-occurrence is weighed as phase unknown and the other assumed in trans, which yields the same 6.0-point total. Co-occurring variants are assumed P/LP from disease context (founder A-T mutation; canonical splice-donor null allele in classical A-T); formal VCEP classification of those variants was not re-derived.
ATM VCEP v1.5 PM3/BP2 points table: points per unrelated A-T proband - confident phenotype 4.0 (confirmed in trans) / 2.0 (phase unknown) / 1.0 (second variant unidentified or VUS) / 2.0 (homozygous, max 2 individuals); consistent phenotype = half those values. Point-to-strength: PM3 Supporting=1.0, Moderate=2.0, Strong=4.0, Very Strong=8.0. Multiple unrelated cases are additive; publications claiming a clinical diagnosis of A-T are taken at face value and granted a confident diagnosis; variant may not exceed general population frequency >0.01%; if the variant has co-occurred with at least 2 different P/LP variants, one co-occurrence is weighed as phase unknown while the remaining can be assumed in trans.
PMID:17124347 (Magliozzi et al. 2006): patient AT50RM from a cohort of 22 classical A-T patients (Italian A-T Registry; inclusion criteria progressive cerebellar ataxia, oculocutaneous telangiectasias, increased alpha-fetoprotein, chromosomal instability, cellular radiosensitivity) is heterozygous for c.5692C>T (p.R1898X, reported as novel nonsense/truncated protein) and heterozygous for IVS37+2T>C (splicing/truncated protein, previously reported). Both are null alleles in a classical A-T patient; phase was not determined by parental testing, so scored as phase unknown: 2.0 points.
PMID:25077176 (A-TWinnipeg paper): AT213LA, one of 10 Mennonite ataxia-telangiectasia patients, is a compound heterozygote for c.6200C>A (p.A2067D) and c.5692C>T (p.R1898*); c.5692C>T was only detectable after cycloheximide inhibition of NMD and was confirmed by genomic DNA sequencing, with STR/SNP haplotyping showing the c.6200C>A-associated haplotype segregated in trans in AT213LA. Confirmed in trans, A-T clinical diagnosis claimed -> 4.0 points. c.6200C>A is the founder A-TWinnipeg variant with functional evidence of loss of ATM kinase activity (no detectable ATM pSer1981 autophosphorylation in patient LCLs; site-directed construct failed to rescue).
gnomAD v4.1: AF = 3.71857e-06 (6/1,613,522 alleles), 0 homozygotes, grpmax FAF = 6.8e-07; well below the VCEP PM3 ceiling of 0.01% general population frequency, so the frequency gate is satisfied.
PM3 point tally: 4.0 (AT213LA, confirmed in trans) + 2.0 (AT50RM, phase unknown) = 6.0 points -> PM3_Strong. Robustness check: if AT213LA were scored at 'consistent' phenotype (2.0 confirmed in trans), total = 4.0 points, still PM3_Strong.
Still missing
Phase for proband AT50RM (PMID:17124347) was not formally determined by parental/offspring testing; scored conservatively as phase unknown (2.0 points) rather than confirmed in trans (4.0 points).
Co-occurring variants c.6200C>A (p.A2067D) and IVS37+2T>C are assumed P/LP from disease context; their classification under ATM VCEP v1.5 specifications was not independently re-derived here.
No additional A-T proband observations beyond PMID:17124347 and PMID:25077176 were available in the collected evidence; a third unrelated proband with confirmed in-trans phase would raise PM3 to Very Strong (>=8.0 points).
PM5SupportingMet: the truncation at p.Arg1898Ter lies upstream of the VCEP's p.Arg3047 boundary, meeting the gene-specific PM5 rule.
Details ▾Hide ▴
Reasoning
Under the ATM VCEP v1.5 gene-specific PM5 rule, PM5 applies to frameshifting or truncating variants with premature termination codons upstream of p.Arg3047. c.5692C>T produces the truncating p.Arg1898Ter termination codon, which is upstream of p.Arg3047; therefore PM5 is met at Supporting strength. The variant's nonsense/truncating consequence is directly reported in the cited clinical literature.
Rule
ATM VCEP v1.5 PM5: apply to frameshifting or truncating variants with premature termination codons upstream of p.Arg3047; Supporting strength.
Evidence considered
The ATM VCEP v1.5 PM5 rule states: apply to frameshifting or truncating variants with premature termination codons upstream of p.Arg3047, with the criterion's listed strength of Supporting.
Magliozzi et al. report c.5692C>T as R1898X in exon 40 and classify it as a nonsense mutation producing a truncated protein (Table 3, patient AT50RM).
A-TWinnipeg reports c.5692C>T as the second mutation in patient-derived AT213LA cells, confirmed by genomic DNA sequencing after inhibition of nonsense-mediated decay; the case normalization identifies this change as p.Arg1898Ter.
Global PP5/BP6 rule applied: exact-variant ClinVar expert-panel Pathogenic classification -> PP5 supporting strength, regardless of the VCEP 'Not Applicable' flag (ClinGen HBOP ATM VCEP v1.5 cspec lists PP5 as Not Applicable).
14 clinical-laboratory Pathogenic submissions (aggregate ClinVar label) were NOT used for PP5; ordinary laboratory submissions and non-expert assertions must never trigger PP5.
PS4Not met: no case-control study of this variant exists; only individual ataxia-telangiectasia case reports, which cannot meet the VCEP's statistical threshold.
Details ▾Hide ▴
Reasoning
Governing framework is the ClinGen HBOP ATM VCEP v1.5 (cspec), which defines PS4 (Strong) as case-control enrichment with p-value <= 0.05 AND (OR, HR, or RR >= 2 OR lower 95% CI >= 1.5). No case-control study of ATM c.5692C>T exists anywhere in the evidence bundle. The exact variant is documented only as case-level observations in ataxia-telangiectasia: heterozygous nonsense mutation R1898X in Italian A-T patient AT50RM (PMID:17124347, Table 3) and as the second biallelic A-T allele in patient-derived LCL AT213LA, in trans with c.6200C>A, detected after cycloheximide NMD inhibition (PMID:25077176). Case reports cannot meet the VCEP PS4 statistical threshold. The two A-T cohort papers available (PMID:21665257, PMID:23807571) do not mention c.5692C>T, so no cohort-level enrichment figure exists for this variant. ClinVar submissions cite population/cohort studies (PMID 30287823, PMID 30982232, PMID 36988593) but those were not fetched into this case and their per-variant statistics are unverified; per the VCEP rule, PS4 is not met on the available evidence.
Rule
Case-control studies; p-value <= .05 AND (Odds ratio, hazard ratio, or relative risk >= 2 OR lower 95% CI >= 1.5) [ClinGen HBOP ATM VCEP v1.5, PS4 Strong]
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) PS4 rule: 'Case-control studies; p-value <= .05 AND (Odds ratio, hazard ratio, or relative risk >= 2 OR lower 95% CI >= 1.5)' with default strength Strong.
PMID:17124347 (Magliozzi et al. 2006) reports the exact variant c.5692C>T (R1898X) as a novel heterozygous nonsense/truncating mutation in Italian A-T patient AT50RM (Table 3); this is a case observation, not a case-control enrichment study.
PMID:25077176 (A-TWinnipeg) identifies c.5692C>T as the second A-T allele in patient-derived lymphoblastoid cell line AT213LA, in trans with c.6200C>A, confirmed by RNA sequencing after cycloheximide NMD inhibition and genomic DNA sequencing; case-level observation only, no enrichment statistics.
PMID:21665257 (Micol et al. 2011, French A-T cohort of 240 patients) was checked via paper_extracts: it does not mention c.5692C>T; it classifies nonsense mutations as class A/null alleles expected to cause complete loss of ATM expression or activity, providing context but no per-variant enrichment data.
PMID:23807571 (Chinese A-T cohort, 12 mutations in 8 patients) was checked via paper_extracts: it does not mention c.5692C>T; no per-variant enrichment data.
No case-control study with per-variant OR/RR and p-value for c.5692C>T was found in the evidence bundle; ClinVar criterion-lead submissions cite population studies (PMID 30287823, PMID 30982232, PMID 36988593) but these were not fetched, so their per-variant results for c.5692C>T could not be verified and were not used.
Still missing
No case-control study with per-variant statistics (p-value, OR/HR/RR, 95% CI) for ATM c.5692C>T is available; only case-report-level observations (PMID:17124347, PMID:25077176) exist, which cannot satisfy the VCEP PS4 rule.
ClinVar-cited population/cohort studies (PMID 30287823 Momozawa; PMID 30982232 Wang; PMID 36988593 Usui) were not fetched into this case bundle; per-variant enrichment data for c.5692C>T from those studies could not be verified.
PP1Not met: no affected relative carrying this variant is reported, so no segregation evidence exists.
Details ▾Hide ▴
Reasoning
PP1 is applicable under the ClinGen HBOP ATM VCEP v1.5 (AR condition: segregation in 1 affected relative = Supporting, 2 = Moderate, >=3 = Strong; affected relatives must have both proband variants identified). However, no segregation of c.5692C>T (p.Arg1898Ter) in affected relatives is reported in any available source. The variant-reporting paper (PMID 17124347, Magliozzi et al.) identified the variant in a single unrelated proband (AT50RM, heterozygous) from a cohort explicitly composed of 22 unrelated A-T families; no parental or affected-relative testing is reported. PMID 25077176 (A-TWinnipeg) identified c.5692C>T as the second allele in trans with c.6200C>A in proband-derived LCL AT213LA - a proband-level biallelic observation (PM3-type evidence, outside this group), not co-segregation in affected relatives. No other fetched paper mentions the variant (paper_extracts variant_mentioned=false). Because zero affected relatives carrying the variant are documented, PP1 cannot be met; the criterion is not_met pending segregation data.
Rule
PP1 (AR Condition): Segregation in >=3 affected relatives = Strong; 2 affected relatives = Moderate; 1 affected relative = Supporting (ClinGen HBOP ATM VCEP v1.5). Affected relatives must have both variants identified in the proband.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 PP1 rules (cspec): AR Condition - segregation in >=3 affected relatives = Strong, 2 affected relatives = Moderate, 1 affected relative = Supporting; instructionsToUse notes 'AR Condition: Affected relatives must have both variants identified in proband.'
PMID 17124347 (Magliozzi et al. 2006): c.5692C>T (p.R1898X) reported in Table 3 for patient AT50RM (het) among 22 unrelated A-T patients/22 families; full text reviewed for family/segregation terms - no affected-relative testing or segregation is reported for this variant.
PMID 25077176 (A-TWinnipeg): c.5692C>T identified as the second mutation in trans with c.6200C>A in proband-derived LCL AT213LA after cycloheximide NMD inhibition; this is a biallelic proband observation, not segregation in affected relatives.
paper_extracts.json entries for all other fetched full-text papers (23807571, 21665257, 27413114, 30348496, 30553448) report variant_mentioned=false for c.5692C>T/p.(R1898*); none provide segregation data.
ClinVar submissions for the variant (17 submissions, VCV000482526) carry classifications only; the submission audit surfaces no segregation or family-study assertions for this variant.
Still missing
No segregation studies in affected relatives are reported for c.5692C>T in any fetched source (PubMed full-text set, ClinVar submission audit). Documented observations are single unrelated probands (AT50RM in PMID 17124347; AT213LA proband cell line in PMID 25077176). Segregation testing of affected relatives - with both proband variants confirmed in each affected relative, per the VCEP PP1 AR-condition note - would be required to apply PP1.
BA1Not met: grpmax filtering allele frequency is 6.8e-07 (0.000068%), about four orders of magnitude below the 0.5% BA1 threshold.
Details ▾Hide ▴
Reasoning
ATM VCEP v1.5 (cspec) BA1 rule: Grpmax Filtering AF > 0.5% in gnomAD v4 dataset. gnomAD v4.1 joint grpmax FAF for 11-108307914-C-T is 6.8e-07 (0.000068%), approximately four orders of magnitude below the 0.5% threshold; total AF is 3.72e-06 (0.00037%, 6/1,613,522 alleles) with 0 homozygotes. gnomAD v2.1 (AF 7.97e-06, 2/251,006 alleles, 0 homozygotes) and absence from gnomAD-Canada v1.0 corroborate that this allele is rare across datasets. BA1 stand-alone benign is not met.
Rule
Grpmax Filtering AF > 0.5% in gnomAD v4 dataset (ATM VCEP v1.5)
Evidence considered
ATM VCEP v1.5 (cspec) BA1 rule: 'Grpmax Filtering AF >.5% in gnomAD v4 dataset'.
gnomAD v4.1 (gnomad_v4): variant 11-108307914-C-T, total AF 3.718573406498331e-06 (0.00037%, 6/1,613,522 alleles), exome AF 4.105646495625434e-06 (6/1,461,402), joint grpmax FAF 6.8e-07 (0.000068%), exome grpmax FAF 7.2e-07, homozygotes 0; per-subpopulation counts are n=1 (Remaining individuals 1/62,486; SAS 1/91,062; AFR 1/74,980) and n=3 (NFE 3/1,179,726).
gnomAD v2.1 (gnomad_v2): variant 11-108178641-C-T, total AF 7.967937021425783e-06 (0.00080%, 2/251,006 alleles), highest subpopulation SAS AF 3.268187463232891e-05 (1/30,598), homozygotes 0.
BS1Not met: grpmax filtering allele frequency 6.8e-07 (0.000068%) is nearly three orders of magnitude below the 0.05% BS1 threshold.
Details ▾Hide ▴
Reasoning
ATM VCEP v1.5 (cspec) BS1 rule: Grpmax Filtering AF > 0.05% in gnomAD v4 dataset. Observed joint grpmax FAF 6.8e-07 (0.000068%) is nearly three orders of magnitude below the 0.05% threshold, and total AF is 0.00037% (6/1,613,522 alleles, 0 homozygotes). gnomAD v2.1 (0.0008%, 2 alleles) and gnomAD-Canada (absent) are consistent. BS1 not met.
Rule
Grpmax Filtering AF > 0.05% in gnomAD v4 dataset (ATM VCEP v1.5)
Evidence considered
ATM VCEP v1.5 (cspec) BS1 rule: 'Grpmax Filtering AF >.05% in gnomAD v4 dataset'.
gnomAD v4.1 (gnomad_v4): total AF 3.718573406498331e-06 (0.00037%, 6/1,613,522 alleles), joint grpmax FAF 6.8e-07 (0.000068%), homozygotes 0.
gnomAD v2.1 (gnomad_v2): total AF 7.967937021425783e-06 (0.00080%, 2/251,006 alleles), homozygotes 0.
Observed grpmax FAF 6.8e-07 (0.000068%) is far below the VCEP BS1 threshold of 0.05%; BS1 not met.
BP2Not met: no unaffected individual carrying this variant with a pathogenic variant in trans was identified (0 points).
Details ▾Hide ▴
Reasoning
Governing framework: ClinGen HBOP ATM VCEP v1.5 PM3/BP2 table - BP2 points are accrued per unaffected (non-A-T) individual carrying the variant under assessment together with a P/LP ATM variant in trans (-4.0 confirmed trans / -2.0 phase unknown / -2.0 lab setting or -1.0 database setting homozygous, max -2.0; BP2 Supporting=-1.0, Moderate=-2.0, Strong<=-4.0; the unaffected individual must be 18 years or older with no evidence of A-T). No unaffected (non-A-T) individual carrying c.5692C>T with a P/LP variant in trans was identified in any collected source: the two literature probands (PMID:17124347 AT50RM, PMID:25077176 AT213LA) are A-T-affected, not unaffected carriers; ClinVar submissions do not provide per-individual unaffected-carrier phase data; gnomAD v4.1 shows only 6 heterozygous carriers with 0 homozygotes and no documented phase with a second P/LP ATM variant. BP2 points = 0 -> BP2 not met. (Note: this variant is a bona fide loss-of-function null allele in A-T probands, so absence of benign BP2 evidence is expected and does not conflict with the pathogenic evidence base.)
Rule
ATM VCEP v1.5 PM3/BP2 table (BP2): negative points from unaffected (non-A-T) individuals carrying the VUA with a P/LP ATM variant in trans; 0 points observed -> not met.
Evidence considered
ATM VCEP v1.5 PM3/BP2 table (BP2 half): points per unaffected (non-A-T) individual carrying the variant with a P/LP ATM variant in trans - confirmed in trans -4.0, phase unknown -2.0, homozygous -2.0 (lab) / -1.0 (database), max -2.0; BP2 Supporting=-1.0, Moderate=-2.0, Strong<=-4.0. When assessing homozygous or in-trans variants for a possible downgrade in an unaffected individual, the individual should be 18 years or older with no evidence of A-T.
Reviewed all full-text papers with variant-specific extraction for evidence of unaffected carriers: PMID:17124347 and PMID:25077176 report the variant only in A-T-affected probands (AT50RM, AT213LA); PMID:21665257, PMID:23807571, PMID:30348496, PMID:30553448, PMID:27413114 do not mention c.5692C>T at all and contain no unaffected-carrier phase observations.
ClinVar (VCV000482526) submissions list the variant as Pathogenic, including under Ataxia-telangiectasia syndrome, but provide no individual-level unaffected-carrier in-trans observations usable for BP2 points.
gnomAD v4.1: 6 heterozygous carriers, 0 homozygotes; population data provide no phase information with a second P/LP ATM variant, so no BP2 points are accrued from population observations.
BP2 point tally: 0 points -> BP2 not met.
Still missing
No observations of unaffected (non-A-T) individuals carrying c.5692C>T in trans with a P/LP ATM variant were available in the collected evidence bundle; such data (e.g., from laboratory carrier testing) would be required to apply BP2.
ClinVar submission-level phase/parental-testing notes were not parsed at the individual level in this case's evidence; if retrievable, they could be reviewed for BP2-relevant unaffected-carrier observations.
BP6Not met: no expert-panel benign classification exists; the only expert-panel call for this variant is Pathogenic.
Details ▾Hide ▴
Reasoning
Under the global PP5/BP6 rule, BP6 is triggered only by an exact-variant ClinVar expert-panel Benign/Likely benign classification. The only expert-panel classification for ATM c.5692C>T is Pathogenic (ClinGen HBOP VCEP, SCV005627280, 3-star), which supports PP5 and cannot support BP6. The ClinGen HBOP ATM VCEP v1.5 additionally declares BP6 'Not Applicable'. No benign or likely-benign expert-panel assertion exists for this variant, and laboratory-level benign submissions (none present) would never trigger BP6.
Rule
Exact-variant ClinVar expert-panel Benign/Likely benign classification -> BP6; not triggered when the expert-panel classification is Pathogenic
Evidence considered
ClinVar VCV000482526: the sole expert-panel submission (SCV005627280, ClinGen HBOP VCEP) is classified Pathogenic, not Benign/Likely benign; therefore the BP6 trigger condition is absent.
ClinGen HBOP ATM VCEP v1.5 (cspec) lists BP6 as 'Not Applicable'; the global BP6 rule is not triggered in the absence of an expert-panel benign classification.
PS3Not assessed: no VCEP-approved functional assay result exists for this variant; the only functional-screen entry was a computational prediction.
Details ▾Hide ▴
Reasoning
PS3 is not assessed because no qualifying functional assay evidence exists for c.5692C>T (p.Arg1898Ter) in this case. The ClinGen HBOP ATM VCEP v1.5 framework requires that a variant be confirmed as actually tested in one of the three VCEP-approved assays (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335 - stable transfection of ATM cDNA constructs into ATM-null LCLs with Western-blot readout of ATM substrate phosphorylation after irradiation) before any strength is assigned; the approved-assay table is a methodology/calibration table, not a per-variant results table, and none of those three papers was retrieved for this case, so no variant-specific readout could be confirmed. The only variant-matched functional-screen entry found (Sun et al. 2025, PMID 40580951, Table S1) labels c.5692C>T 'Non-functional' (Combined_score -3.72), but that row has DeepATM_predicted=Yes and Confidence='Computational prediction', meaning the classification is the DeepATM model's extrapolation rather than a direct measurement from the prime-editing screen, and the paper is not among the three VCEP-approved assays; per the governing index guidance it must not be presented as functional evidence for PS3. Full-text papers that mention the variant (PMID 17124347, PMID 25077176) report the variant's identity (nonsense/truncated, NMD-susceptible allele) but perform no functional assay on it. Without an admissible assay result, no PS3 strength (Supporting/Moderate; Strong is disallowed by the VCEP) can be applied.
Rule
PS3_Moderate: use when a variant fails to rescue both an ATM-specific feature (e.g., phosphorylation of ATM-specific targets) AND radiosensitivity. PS3_Supporting: use when a variant fails to rescue an ATM-specific feature only; do not use for radiosensitivity-only. PS3_Strong: do not use (ClinGen HBOP ATM VCEP v1.5, cspec).
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) governs PS3: PS3_Moderate applies when a variant fails to rescue both an ATM-specific feature (e.g., phosphorylation of ATM-specific targets) AND radiosensitivity; PS3_Supporting applies when a variant fails to rescue an ATM-specific feature only; PS3_Strong is explicitly disallowed ('Do not use as strong'); radiosensitivity-only failure is explicitly not usable ('Do not use for radiosensitivity-only as that is not a feature specific to ATM deficiency').
The VCEP-approved functional assay set (vcep_clingen_hbop_atm_supplementary_tables_1_and_2_v1) consists of Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, and Scott 2002/PMID 11805335 (stable transfection of ATM cDNA constructs into ATM-null LCLs; kinase-activity readout of S1981/SMC1/NBS1/CHK2/p53 phosphorylation by Western blot after irradiation; radiosensitivity readout for Mitui 2009 approved only when combined with a kinase-activity result). This table is NOT a per-variant results table: PS3 can only be applied after confirming the variant under assessment was actually tested in one of these specific papers.
None of the three VCEP-approved assay papers was retrieved for this case, so it could not be confirmed whether c.5692C>T (p.Arg1898Ter) was directly tested in an approved assay; no variant-specific assay readout exists anywhere in the case evidence bundle.
Sun et al. 2025 (vcep_suppl_tables1_pmid_40580951, Table S1) contains an exact-match row for c.5692C>T (R1898*, exon 38): Classification 'Non-functional', Combined_score -3.724965, but DeepATM_predicted=Yes and Confidence='Computational prediction'. Per the governing index guidance, DeepATM_predicted=Yes means the variant was NOT directly assayed in the prime-editing saturation screen and its classification is the trained DeepATM model's extrapolation; it must not be presented as functional evidence and is treated as weak computational corroboration at most (PP3/BP4 territory, not PS3/BS3). The paper also postdates the ATM VCEP v1.5 approved-assay table.
PMID:17124347 (Magliozzi et al., DHPLC screening of ATM in Italian A-T patients) reports c.5692C>T as R1898X, a 'Nonsense/Truncated protein' mutation, in A-T patient AT50RM (heterozygous), and classifies it as novel; no functional assay was performed, so this documents variant identity only and does not satisfy PS3.
PMID:25077176 (A-TWinnipeg) identifies c.5692C>T as the second ATM mutation in patient-derived LCL AT213LA (in trans with c.6200C>A), detected only after cycloheximide inhibition of nonsense-mediated mRNA decay; no functional assay readout attributable to c.5692C>T is reported, so it does not satisfy PS3.
Full texts of PMID 21665257, PMID 30348496, PMID 30553448, PMID 27413114, and PMID 23807571 were checked via paper_extracts; none mentions c.5692C>T and none reports any variant-specific functional assay, so they provide no PS3 evidence (general ATM biology/LOF context only).
Still missing
No VCEP-approved functional assay result for c.5692C>T: none of the three approved-assay papers (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335) was retrieved, so it could not be confirmed whether this variant was tested in an approved assay.
The only variant-matched functional screen (Sun et al. 2025, PMID 40580951, Table S1) classifies c.5692C>T as 'Non-functional' but with DeepATM_predicted=Yes / Confidence='Computational prediction' - a model extrapolation, not a direct functional measurement, and not a VCEP-approved assay; it is therefore inadmissible as PS3 evidence.
No direct measurement of ATM kinase activity, substrate phosphorylation (S1981, SMC1, NBS1, CHK2, p53), or radiosensitivity rescue is reported for c.5692C>T in any source in this case.
BS3Not assessed: no VCEP-approved assay result showing preserved ATM function or normal radiosensitivity is available.
Details ▾Hide ▴
Reasoning
BS3 is not assessed because there is no functional evidence that c.5692C>T preserves ATM function. Under the ClinGen HBOP ATM VCEP v1.5, BS3_Moderate requires a variant to rescue BOTH an ATM-specific feature AND radiosensitivity, and BS3_Supporting requires rescue of EITHER feature, in a VCEP-approved assay; no such rescue result exists for this variant in the case evidence bundle, and none of the three approved-assay papers (Mitui 2009/PMID 18634022, Barone 2009/PMID 19431188, Scott 2002/PMID 11805335) was retrieved to confirm whether the variant was even tested. The only variant-matched functional-screen entry (Sun et al. 2025, Table S1) reports 'Non-functional' with DeepATM_predicted=Yes (computational prediction, not a direct measurement) and is not a VCEP-approved assay, so it cannot support BS3 in any case. The variant is a nonsense change (p.Arg1898Ter); no source in this case reports any assay showing preserved/restored ATM kinase activity or normal radiosensitivity for it.
Rule
BS3_Moderate: use when a variant rescues both an ATM-specific feature (e.g., phosphorylation of ATM-specific targets) AND radiosensitivity. BS3_Supporting: use when a variant rescues EITHER an ATM-specific feature OR radiosensitivity (ClinGen HBOP ATM VCEP v1.5, cspec).
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) governs BS3: BS3_Moderate applies when a variant rescues BOTH an ATM-specific feature (e.g., phosphorylation of ATM-specific targets) AND radiosensitivity; BS3_Supporting applies when a variant rescues EITHER an ATM-specific feature OR radiosensitivity.
The VCEP-approved assay table (vcep_clingen_hbop_atm_supplementary_tables_1_and_2_v1) calibrates BS3_Supporting for the approved kinase-activity assays (Mitui 2009, Barone 2009, Scott 2002) and BS3_Supporting for radiosensitivity (Mitui 2009, combinable only with a kinase-activity result), but it is not a per-variant results table; no rescue result is reported for c.5692C>T.
No evidence of preserved ATM function for c.5692C>T was found in any source: the variant is a nonsense change (p.Arg1898Ter) and no assay showing normal kinase activity, substrate phosphorylation, or radiosensitivity rescue was reported in the case evidence bundle.
Sun et al. 2025 (vcep_suppl_tables1_pmid_40580951, Table S1) exact-match row for c.5692C>T is classified 'Non-functional' with DeepATM_predicted=Yes / Confidence='Computational prediction'; because this is a model extrapolation rather than a direct functional measurement (and not a VCEP-approved assay), it cannot be used as BS3 evidence, and it does not suggest normal function in any event.
PMID:17124347 and PMID:25077176 mention c.5692C>T (R1898X nonsense; NMD-susceptible allele in LCL AT213LA) but report no functional assay and no evidence of normal ATM function, so they do not support BS3.
Full texts of PMID 21665257, PMID 30348496, PMID 30553448, PMID 27413114, and PMID 23807571 were checked via paper_extracts; none mentions c.5692C>T and none reports a variant-specific functional assay relevant to BS3.
Still missing
No functional study demonstrating rescue of an ATM-specific feature (e.g., phosphorylation of ATM-specific targets) or of radiosensitivity for c.5692C>T in any VCEP-approved assay.
The three VCEP-approved assay papers (PMID 18634022, 19431188, 11805335) were not retrieved; whether c.5692C>T was tested in them (with any readout) is unknown.
Sun et al. 2025 (PMID 40580951) row for c.5692C>T is DeepATM_predicted=Yes (computational prediction), so it provides no direct functional measurement that could be weighed for BS3.
Not applicable · 15· recorded as not applicableShow reasons ▾Hide reasons ▴PS1N/ANot applicable: PS1 requires a missense change matching a known pathogenic variant; c.5692C>T is a nonsense change.
Details ▾Hide ▴
Reasoning
The variant is a nonsense change, NM_000051.4:c.5692C>T, p.(Arg1898Ter), not a missense substitution. The ATM VCEP PS1 specification's standard path requires a missense change with the same amino-acid change as an established pathogenic or likely pathogenic variant; its alternative path is for comparable splice predictions. Neither path applies.
Rule
ATM VCEP v1.5 PS1: use for missense changes when splicing is ruled out for both alterations; use the ATM PS1 splicing table for qualifying splice-region comparisons.
Evidence considered
The case normalization identifies NM_000051.4:c.5692C>T as NP_000042.3:p.(Arg1898Ter) (p.(R1898*)).
The ATM VCEP v1.5 PS1 specification limits the standard PS1 path to missense changes and provides a separate splice-region path requiring comparable predicted splice events.
PS2VCEPNot applicable: the ATM VCEP bars de novo evidence for this gene, and no de novo occurrence of this variant is reported.
Details ▾Hide ▴
Reasoning
The governing framework is the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer (HBOP) ATM VCEP v1.5 (cspec doc 639508985), which takes precedence over generic ACMG/AMP. In the cspec rule payload, PS2 applicability is explicitly 'Not applicable', with 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.' This applies to both ATM disease indications (ATM-related cancer predisposition, AD; ataxia-telangiectasia, AR). No de novo occurrence of c.5692C>T (p.Arg1898Ter) is reported in any fetched source, and the VCEP bars PS2 use for ATM regardless. PS2 is therefore not applicable and contributes no evidence.
VCEP rule
PS2 applicability: Not applicable. VCEP 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' (ClinGen HBOP ATM VCEP v1.5).
Evidence considered
ClinGen HBOP ATM VCEP v1.5 cspec rule payload lists PS2 applicability as 'Not applicable' with 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' (criteria entry PS2 in cspec).
case_summary.json cspec section likewise reports PS2 applicability 'Not Applicable' for the ATM v1.5 specification.
No fetched source (ClinVar submission audit, paper_extracts, evidence.json) reports a de novo occurrence of c.5692C>T with confirmed parentage in an affected proband.
PM1VCEPNot applicable: the ATM VCEP marks PM1 inapplicable, and this nonsense variant has no missense domain argument.
Details ▾Hide ▴
Reasoning
The governing ATM VCEP v1.5 framework marks PM1 as not applicable. In addition, this is a nonsense variant rather than a missense variant for which a critical functional-domain argument would ordinarily be considered. The case bundle contains no authoritative ATM domain-table entry requiring a residue-level PM1 assessment.
VCEP rule
ATM VCEP v1.5 PM1: Not Applicable.
Evidence considered
The ATM VCEP v1.5 criterion framework marks PM1 as Not Applicable.
The case's vcep_materials.json has an empty domain_tables object, so there is no gene-specific authoritative domain-table list to apply to residue 1898.
PM4N/ANot applicable: PM4 applies only to stop-loss variants; c.5692C>T is a stop-gain (nonsense) change.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 specification restricts PM4 to stop-loss variants ('Use for stop-loss variants'). c.5692C>T is a stop-gain (nonsense) variant, p.Arg1898Ter, which is adjudicated under PVS1 rather than PM4. PM4 (protein length change without truncation) cannot apply to this variant class, and no stop-loss alternative-transcript interpretation is relevant here.
Rule
ATM VCEP v1.5: PM4 applies only to stop-loss variants
Evidence considered
ATM VCEP v1.5 (cspec) defines PM4 with default strength Pathogenic Moderate and the rule 'Use for stop-loss variants'; a nonsense/stop-gain variant is outside the scope of this criterion.
PM6VCEPNot applicable: the ATM VCEP bars de novo evidence for ATM, and no assumed de novo occurrence is reported.
Details ▾Hide ▴
Reasoning
Under the governing ClinGen HBOP ATM VCEP v1.5 framework, PM6 applicability is explicitly 'Not applicable', with 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 assumed de novo occurrence (without confirmation of paternity/maternity) of c.5692C>T is reported in any source. PM6 is therefore not applicable and contributes no evidence.
VCEP rule
PM6 applicability: Not applicable. VCEP 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' (ClinGen HBOP ATM VCEP v1.5).
Evidence considered
ClinGen HBOP ATM VCEP v1.5 cspec rule payload lists PM6 applicability as 'Not applicable' with 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' (criteria entry PM6 in cspec).
case_summary.json cspec section likewise reports PM6 applicability 'Not Applicable' for the ATM v1.5 specification.
No source reports an assumed or confirmed de novo event for c.5692C>T in an A-T proband.
PP2VCEPNot applicable: the ATM VCEP marks PP2 inapplicable, and missense-predominance reasoning does not apply to a nonsense change.
Details ▾Hide ▴
Reasoning
The governing ATM VCEP v1.5 framework marks PP2 as not applicable. The variant is also a nonsense rather than a missense change, so a gene-level predominance of pathogenic missense variation is not relevant to this criterion assessment.
VCEP rule
ATM VCEP v1.5 PP2: Not Applicable.
Evidence considered
The ATM VCEP v1.5 criterion framework marks PP2 as Not Applicable.
The case variant is normalized as p.Arg1898Ter, a nonsense change rather than a missense variant.
PP3N/ANot applicable: the VCEP's PP3 sub-paths (REVEL, SpliceAI >= 0.2) do not cover nonsense variants; SpliceAI max delta is 0.081.
Details ▾Hide ▴
Reasoning
Variant is NM_000051.4:c.5692C>T, a nonsense (stop-gain) SNV predicted to produce p.(Arg1898Ter) in exon 38. The ClinGen HBOP ATM VCEP v1.5 (cspec) PP3 rule provides exactly two calibrated computational sub-paths: (1) missense: REVEL > 0.7333; (2) splicing: SpliceAI >= 0.2 for silent, missense/in-frame, and intronic variants outside the donor/acceptor +/-1,2 dinucleotides. A nonsense variant is covered by neither sub-path, and the VCEP explicitly states that PP3 for splice predictions may not be applied in addition to PVS1 or PVS1_Variable(RNA) codes (this variant is PVS1-eligible as a truncating variant, assessed by the PVS1 group). SpliceAI max delta is 0.081 (< 0.2) so the splicing sub-path would not be met in any case. REVEL is unavailable (lookup miss; Sun et al. 2025 Table S1 row also reports REVEL = NA). BayesDel score 0.295247 was retrieved but no verified published threshold exists for this pipeline, so it is treated as insufficiently calibrated and not_available. The Sun et al. 2025 (PMID 40580951) Table S1 row for c.5692C>T lists CADD.phred 37 and boostDM 0.9908731 with Classification 'Non-functional' but DeepATM_predicted = Yes / Confidence 'Computational prediction', i.e. the classification is a model extrapolation rather than a functional measurement; per the VCEP file index guidance it can serve only as weak computational corroboration for PP3/BP4 and cannot substitute for a VCEP-calibrated PP3 sub-path, which does not exist for nonsense variants. PP3 is therefore not applicable to this variant class.
Rule
ATM VCEP v1.5 PP3: Missense: REVEL > 0.7333; Splicing: SpliceAI >= 0.2 for silent, missense/in-frame and intronic variants outside donor/acceptor +/-1,2 sites. PP3 may not be applied in addition to PVS1/PVS1_Variable(RNA) codes. No sub-path covers nonsense/truncating variants.
Evidence considered
ATM VCEP v1.5 (cspec) PP3 rule: 'Missense: REVEL >.7333; Splicing: Predicted impact via splicing (SpliceAI >=0.2) for silent, missense/in-frame and for intronic variants outside of donor and acceptor 1,2 sites' - the splicing sub-path enumerates silent, missense/in-frame, and intronic variants only; nonsense variants are not included.
ATM VCEP v1.5 instructions: 'PP3 for splice predictions may not be applied in addition to PVS1 or PVS1_Variable(RNA) codes' - c.5692C>T is a PVS1-eligible truncating variant, so stacking PP3 would be disallowed even if a splice prediction existed.
SpliceAI lookup for NM_000051.4:c.5692C>T: max delta score 0.081 (DS_AG 0.003, DS_AL 0.081, DS_DG 0.001, DS_DL 0.044), below the VCEP PP3 splicing threshold of 0.2.
BayesDel score 0.295247 (source: local BayesDel lookup) retrieved but treated as not_available for PP3/BP4 because no verified published threshold is available to this pipeline; per calibration policy a score without a named published threshold is indistinguishable from a fabricated number.
Sun et al. 2025 Table S1 row for c.5692C>T (GRCh38 108307914 C>T): Nonsense, exon_38, CADD.phred 37, boostDM_score 0.9908731, EVE/REVEL/AlphaMissense NA, Classification 'Non-functional', Confidence 'Computational prediction', DeepATM_predicted 'Yes' - model extrapolation, not a functional measurement; treated as at most weak computational corroboration and cannot establish PP3 for a nonsense variant under the VCEP rule.
Variant class confirmed as nonsense/stop-gain via Mutalyzer/VariantValidator (NP_000042.3:p.(Arg1898Ter)) and corroborated by the Sun et al. Table S1 row (protein R1898*, variant type 'Nonsense').
Still missing
REVEL score unavailable for this variant (lookup returned not found; Sun et al. Table S1 also reports NA) - would be required for the missense sub-path, but the variant is nonsense so no PP3 sub-path applies regardless.
No RNA assay data in this case that could have supported PP3(RNA)-type evidence; not required because the VCEP's computational PP3 rule does not cover nonsense variants.
PP4VCEPNot applicable: the ATM VCEP explicitly declares PP4 inapplicable; phenotype specificity is not scored separately.
Details ▾Hide ▴
Reasoning
ClinGen HBOP ATM VCEP v1.5 (cspec) explicitly declares PP4 'Not Applicable' for ATM. Phenotype specificity (e.g., the A-T phenotype in probands carrying the variant) is not scored as a separate criterion under this specification, so PP4 is not assessed under the governing framework.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) lists PP4 as 'Not Applicable'; phenotype-specificity evidence is not used as a standalone criterion for ATM under this specification.
BS2VCEPNot applicable: the ATM VCEP explicitly marks BS2 inapplicable for this gene.
Details ▾Hide ▴
Reasoning
The ClinGen HBOP ATM VCEP v1.5 specification (cspec) explicitly marks BS2 as Not Applicable for ATM. No BS2 evaluation is performed under the governing VCEP framework; generic ACMG BS2 (variant observed in healthy adult homozygote/heterozygote for a fully penetrant dominant disorder) is not applied because an applicable gene-specific specification exists.
VCEP rule
BS2 declared Not Applicable in ATM VCEP v1.5 (cspec)
Evidence considered
ATM VCEP v1.5 (cspec) lists BS2 with applicability 'Not Applicable'; no BS2 rule is defined by the VCEP for ATM.
No homozygous carriers of c.5692C>T are present in gnomAD v2.1 or gnomAD v4.1 (homozygote count 0 in both), which is consistent with a pathogenic recessive allele but is not needed for the BS2 determination given the VCEP declaration.
BS4VCEPNot applicable: the ATM VCEP does not weight lack-of-segregation evidence, and no unaffected carriers are reported.
Details ▾Hide ▴
Reasoning
Under the governing ClinGen HBOP ATM VCEP v1.5 framework, BS4 applicability is explicitly 'Not applicable'. The VCEP instructionsToUse state: '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.' No unaffected individuals carrying c.5692C>T in an A-T family are reported, and the VCEP does not weight lack-of-segregation evidence for ATM. BS4 is therefore not applicable and contributes no evidence.
VCEP rule
BS4 applicability: Not applicable. VCEP instructionsToUse: 'AR Condition: Informative instances of lack of co-segregation in A-T families are too rare to be considered for weight at this time...' (ClinGen HBOP ATM VCEP v1.5).
Evidence considered
ClinGen HBOP ATM VCEP v1.5 cspec rule payload lists BS4 applicability as 'Not applicable' with instructionsToUse: '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.' (criteria entry BS4 in cspec).
case_summary.json cspec section likewise reports BS4 applicability 'Not Applicable' for the ATM v1.5 specification.
No source reports an unaffected (non-A-T) individual carrying c.5692C>T in an A-T family, which would be the only scenario the VCEP considers for lack-of-segregation weighting.
BP1VCEPNot applicable: the ATM VCEP marks BP1 inapplicable, and this is a truncating rather than missense change.
Details ▾Hide ▴
Reasoning
The governing ATM VCEP v1.5 framework marks BP1 as not applicable. The variant is a truncating nonsense change, not a missense variant in a gene where missense variants are a less common disease mechanism.
VCEP rule
ATM VCEP v1.5 BP1: Not Applicable.
Evidence considered
The ATM VCEP v1.5 criterion framework marks BP1 as Not Applicable.
The case variant is normalized as p.Arg1898Ter, a nonsense change.
BP3VCEPNot applicable: the ATM VCEP marks BP3 inapplicable, and this is a nonsense substitution, not an in-frame indel.
Details ▾Hide ▴
Reasoning
The ClinGen HBOP ATM VCEP v1.5 (cspec) explicitly marks BP3 as 'Not Applicable' for ATM. Additionally, BP3 as defined in ACMG/AMP targets in-frame indels within repeat regions of unknown function, whereas this variant is a single-nucleotide nonsense substitution - it is not an in-frame indel and does not fall in a repeat-region context. BP3 therefore receives no weight under the governing specification.
VCEP rule
ATM VCEP v1.5: BP3 Not Applicable
Evidence considered
ATM VCEP v1.5 (cspec) lists BP3 with applicability 'Not Applicable'; the variant's consequence class (nonsense, not in-frame repeat-region indel) is also outside BP3's generic scope.
BP4N/ANot applicable: a clean SpliceAI score (max delta 0.081) only rules out a splicing effect; the protein-truncating, NMD-triggering effect remains.
Details ▾Hide ▴
Reasoning
Variant is a nonsense (stop-gain) SNV, p.(Arg1898Ter). The ATM VCEP v1.5 BP4 rule provides two sub-paths: (1) missense: REVEL <= 0.249; (2) splicing: no predicted impact via splicing (SpliceAI <= 0.1). Neither sub-path applies to a nonsense variant: REVEL is a missense-calibrated score (unavailable here in any case), and the splicing sub-path exists to clear variants whose primary plausible mechanism is splicing. Per the lab's documented BP4 override (gene_notes, applied in addition to the official VCEP text), exactly one mechanism is evaluated per variant and the splicing sub-path may only be used for non-missense variants for which splicing is the primary plausible mechanism, i.e. synonymous and intronic variants outside canonical +/-1,2 sites. A clean SpliceAI score (max delta 0.081, which does satisfy <= 0.1) only rules out a splicing mechanism; it does not clear the variant's own protein-level damaging effect, which for this variant is a premature termination codon (p.Arg1898Ter) with demonstrated NMD sensitivity (PMID 25077176 identifies c.5692C>T only after cycloheximide NMD inhibition in AT213LA). This override changes the outcome from what the VCEP's literal BP4 splicing text alone would produce, and is stated here explicitly per the override instructions: BP4 is not applicable to this nonsense variant regardless of SpliceAI. REVEL is unavailable (lookup miss; Sun et al. Table S1 REVEL = NA) and BayesDel 0.295247 is uncalibrated (no verified published threshold), so no benign computational sub-path can be met or assessed for this variant class.
Rule
ATM VCEP v1.5 BP4: Missense: REVEL <= 0.249; Splicing: SpliceAI <= 0.1. Lab override: splicing sub-path restricted to non-missense variants whose primary plausible mechanism is splicing (synonymous, intronic outside +/-1,2); one mechanism per variant; a clean SpliceAI score does not clear an unresolved protein-level effect.
Evidence considered
ATM VCEP v1.5 (cspec) BP4 rule: 'Missense: REVEL score <=.249; Splicing: No predicted impact via splicing (SpliceAI <=0.1)' - the missense sub-path requires a missense variant; this variant is nonsense.
Lab curator override (gene_notes, BP4 missense/splicing policy, applied in addition to the official VCEP specification): BP4 evaluates exactly one applicable mechanism per variant; the splicing sub-path applies only to non-missense variants for which splicing is the primary plausible mechanism (synonymous, intronic outside canonical +/-1,2); a clean SpliceAI score only rules out the splicing mechanism and does not clear the variant's own protein-level effect. The override changes the outcome here from the VCEP's literal BP4 splicing text (SpliceAI 0.081 <= 0.1 would nominally satisfy it).
SpliceAI lookup for c.5692C>T: max delta 0.081 (<= 0.1 threshold) - noted but not applied to BP4 because the splicing sub-path is inapplicable to this nonsense variant per the override; the variant is damaging at the protein level (p.Arg1898Ter) regardless of splice neutrality.
PMID 25077176 (A-TWinnipeg, via paper_extracts): c.5692C>T was only detectable after cycloheximide treatment inhibited nonsense-mediated mRNA decay in LCL AT213LA (in trans with c.6200C>A), demonstrating the allele is NMD-targeted - direct evidence that the variant's mechanism is protein-level loss, not a splicing-neutral benign state.
BayesDel score 0.295247 treated as not_available for BP4: no verified published threshold available to this pipeline, so no benign/benign-neutral claim can be made from it.
Sun et al. 2025 Table S1 row for c.5692C>T: REVEL = NA and remaining computational columns are model extrapolations (DeepATM_predicted 'Yes', Confidence 'Computational prediction'); no usable benign computational evidence for BP4.
Still missing
REVEL score unavailable (lookup miss; Sun et al. Table S1 reports NA) - would be required for the missense BP4 sub-path, but the variant is not missense so the sub-path does not apply.
No RNA assay data (e.g. BP7(RNA)-style observed lack of aberrant splicing) are available for this variant; not required for this adjudication because BP4's splicing sub-path is restricted by the lab override to variants whose primary plausible mechanism is splicing (synonymous/intronic), and a nonsense SNV is not such a variant.
BP5VCEPNot applicable: the ATM VCEP explicitly declares BP5 inapplicable for ATM.
Details ▾Hide ▴
Reasoning
ClinGen HBOP ATM VCEP v1.5 (cspec) explicitly declares BP5 'Not Applicable' for ATM. The alternate-molecular-basis-of-disease consideration is not evaluated under this specification, and no alternate-basis evidence was assessed for this variant.
Evidence considered
ClinGen HBOP ATM VCEP v1.5 (cspec) lists BP5 as 'Not Applicable'; no alternate-molecular-basis evaluation is performed under this specification.
BP7N/ANot applicable: BP7 applies only to synonymous and deep intronic variants; c.5692C>T is an exonic nonsense change.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 BP7 rule is restricted to synonymous and deep intronic variants defined as further than (but not including) +7 from the donor site and further than (but not including) -21 from the acceptor site; BP7(RNA) applies to observed lack of aberrant RNA for silent substitutions and intronic variants. c.5692C>T is an exonic nonsense (stop-gain) variant, not a synonymous or deep intronic variant, so BP7 is not applicable. No RNA assay evidence of absent aberrant splicing is present in this case that could support a BP7(RNA) application, and such evidence would in any case be restricted to silent/intronic variants. SpliceAI (max delta 0.081) is noted for completeness but is not determinative for BP7, which is a positional/variant-class criterion rather than a score threshold.
Rule
ATM VCEP v1.5 BP7: Use for synonymous and deep intronic variants defined as further than (but not including) +7 and further than (but not including) -21 at donor and acceptor sites, respectively; BP7(RNA) for observed lack of aberrant RNA in silent/intronic variants.
Evidence considered
ATM VCEP v1.5 (cspec) BP7 rule: 'BP7: Use for synonymous and deep intronic variants defined as further than (but not including) +7 and further than (but not including) -21 at donor and acceptor sites, respectively. BP7(RNA): Observed lack of aberrant RNA defect for silent substitutions and intronic variants.' - c.5692C>T is an exonic stop-gain variant (p.Arg1898Ter, exon 38), not synonymous or deep intronic.
Variant class confirmed as nonsense via Mutalyzer/VariantValidator (NP_000042.3:p.(Arg1898Ter)) and Sun et al. Table S1 (variant type 'Nonsense'), excluding the synonymous/deep-intronic classes to which BP7 applies.
SpliceAI max delta 0.081 noted for completeness; BP7 is a variant-class/positional criterion and no score threshold within this group's remit can make BP7 applicable to a nonsense variant.
Still missing
No RNA-based evidence (BP7(RNA): observed lack of aberrant RNA) is available for this variant; not required because BP7 is restricted to synonymous and deep intronic variants and this is a nonsense SNV.
AF 3.7e-06 · 6 / 1,613,522 alleles · 0 hom · FAF 6.8e-07
Remaining individuals1 / 62,486
1.6e-05
African/African American1 / 74,980
1.3e-05
South Asian1 / 91,062
1.1e-05
European (non-Finnish)3 / 1,179,726
2.5e-06
0 alternate alleles recorded in 6 other groups: Admixed American, European (Finnish), Amish, East Asian, Middle Eastern, Ashkenazi Jewish.
gnomAD v2.1
AF 8e-06 · 2 / 251,006 alleles · 0 hom
South Asian1 / 30,598
3.3e-05
European (non-Finnish)1 / 113,504
8.8e-06
0 alternate alleles recorded in 6 other groups: African/African American, Admixed American, Ashkenazi Jewish, East Asian, European (Finnish), Remaining individuals.
Statements
gnomAD v4.1
This variant is present in gnomAD v4.1 (AF= 3.71857e-06; MAF= 0.00037%, 6/1613522 alleles, homozygotes = 0) and has highest observed frequency in the Remaining individuals population (AF= 1.60036e-05; MAF= 0.00160%, 1/62486 alleles, homozygotes = 0); grpmax FAF= 6.8e-07.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 7.96794e-06; MAF= 0.00080%, 2/251006 alleles, homozygotes = 0) and has highest observed frequency in the South Asian population (AF= 3.26819e-05; MAF= 0.00327%, 1/30598 alleles, homozygotes = 0).
gnomAD-Canada
Absent from gnomAD-Canada v1.0.
SpliceAI and Pangolin
SpliceAI acceptor lossat -17 bp
0.081
SpliceAI acceptor gainat +29 bp
0.003
SpliceAI donor lossat +70 bp
0.044
SpliceAI donor gainat -69 bp
0.001
Pangolin splice lossmagnitude shown
-0.322
Pangolin splice gain
0.006
Other predictors
REVEL
Not available
BayesDel
0.295
How the predictions were used
PP3N/ANot applicable: the VCEP's PP3 sub-paths (REVEL, SpliceAI >= 0.2) do not cover nonsense variants; SpliceAI max delta is 0.081.
Details ▾Hide ▴
Reasoning
Variant is NM_000051.4:c.5692C>T, a nonsense (stop-gain) SNV predicted to produce p.(Arg1898Ter) in exon 38. The ClinGen HBOP ATM VCEP v1.5 (cspec) PP3 rule provides exactly two calibrated computational sub-paths: (1) missense: REVEL > 0.7333; (2) splicing: SpliceAI >= 0.2 for silent, missense/in-frame, and intronic variants outside the donor/acceptor +/-1,2 dinucleotides. A nonsense variant is covered by neither sub-path, and the VCEP explicitly states that PP3 for splice predictions may not be applied in addition to PVS1 or PVS1_Variable(RNA) codes (this variant is PVS1-eligible as a truncating variant, assessed by the PVS1 group). SpliceAI max delta is 0.081 (< 0.2) so the splicing sub-path would not be met in any case. REVEL is unavailable (lookup miss; Sun et al. 2025 Table S1 row also reports REVEL = NA). BayesDel score 0.295247 was retrieved but no verified published threshold exists for this pipeline, so it is treated as insufficiently calibrated and not_available. The Sun et al. 2025 (PMID 40580951) Table S1 row for c.5692C>T lists CADD.phred 37 and boostDM 0.9908731 with Classification 'Non-functional' but DeepATM_predicted = Yes / Confidence 'Computational prediction', i.e. the classification is a model extrapolation rather than a functional measurement; per the VCEP file index guidance it can serve only as weak computational corroboration for PP3/BP4 and cannot substitute for a VCEP-calibrated PP3 sub-path, which does not exist for nonsense variants. PP3 is therefore not applicable to this variant class.
Rule
ATM VCEP v1.5 PP3: Missense: REVEL > 0.7333; Splicing: SpliceAI >= 0.2 for silent, missense/in-frame and intronic variants outside donor/acceptor +/-1,2 sites. PP3 may not be applied in addition to PVS1/PVS1_Variable(RNA) codes. No sub-path covers nonsense/truncating variants.
Evidence considered
ATM VCEP v1.5 (cspec) PP3 rule: 'Missense: REVEL >.7333; Splicing: Predicted impact via splicing (SpliceAI >=0.2) for silent, missense/in-frame and for intronic variants outside of donor and acceptor 1,2 sites' - the splicing sub-path enumerates silent, missense/in-frame, and intronic variants only; nonsense variants are not included.
ATM VCEP v1.5 instructions: 'PP3 for splice predictions may not be applied in addition to PVS1 or PVS1_Variable(RNA) codes' - c.5692C>T is a PVS1-eligible truncating variant, so stacking PP3 would be disallowed even if a splice prediction existed.
SpliceAI lookup for NM_000051.4:c.5692C>T: max delta score 0.081 (DS_AG 0.003, DS_AL 0.081, DS_DG 0.001, DS_DL 0.044), below the VCEP PP3 splicing threshold of 0.2.
BayesDel score 0.295247 (source: local BayesDel lookup) retrieved but treated as not_available for PP3/BP4 because no verified published threshold is available to this pipeline; per calibration policy a score without a named published threshold is indistinguishable from a fabricated number.
Sun et al. 2025 Table S1 row for c.5692C>T (GRCh38 108307914 C>T): Nonsense, exon_38, CADD.phred 37, boostDM_score 0.9908731, EVE/REVEL/AlphaMissense NA, Classification 'Non-functional', Confidence 'Computational prediction', DeepATM_predicted 'Yes' - model extrapolation, not a functional measurement; treated as at most weak computational corroboration and cannot establish PP3 for a nonsense variant under the VCEP rule.
Variant class confirmed as nonsense/stop-gain via Mutalyzer/VariantValidator (NP_000042.3:p.(Arg1898Ter)) and corroborated by the Sun et al. Table S1 row (protein R1898*, variant type 'Nonsense').
Still missing
REVEL score unavailable for this variant (lookup returned not found; Sun et al. Table S1 also reports NA) - would be required for the missense sub-path, but the variant is nonsense so no PP3 sub-path applies regardless.
No RNA assay data in this case that could have supported PP3(RNA)-type evidence; not required because the VCEP's computational PP3 rule does not cover nonsense variants.
BP4N/ANot applicable: a clean SpliceAI score (max delta 0.081) only rules out a splicing effect; the protein-truncating, NMD-triggering effect remains.
Details ▾Hide ▴
Reasoning
Variant is a nonsense (stop-gain) SNV, p.(Arg1898Ter). The ATM VCEP v1.5 BP4 rule provides two sub-paths: (1) missense: REVEL <= 0.249; (2) splicing: no predicted impact via splicing (SpliceAI <= 0.1). Neither sub-path applies to a nonsense variant: REVEL is a missense-calibrated score (unavailable here in any case), and the splicing sub-path exists to clear variants whose primary plausible mechanism is splicing. Per the lab's documented BP4 override (gene_notes, applied in addition to the official VCEP text), exactly one mechanism is evaluated per variant and the splicing sub-path may only be used for non-missense variants for which splicing is the primary plausible mechanism, i.e. synonymous and intronic variants outside canonical +/-1,2 sites. A clean SpliceAI score (max delta 0.081, which does satisfy <= 0.1) only rules out a splicing mechanism; it does not clear the variant's own protein-level damaging effect, which for this variant is a premature termination codon (p.Arg1898Ter) with demonstrated NMD sensitivity (PMID 25077176 identifies c.5692C>T only after cycloheximide NMD inhibition in AT213LA). This override changes the outcome from what the VCEP's literal BP4 splicing text alone would produce, and is stated here explicitly per the override instructions: BP4 is not applicable to this nonsense variant regardless of SpliceAI. REVEL is unavailable (lookup miss; Sun et al. Table S1 REVEL = NA) and BayesDel 0.295247 is uncalibrated (no verified published threshold), so no benign computational sub-path can be met or assessed for this variant class.
Rule
ATM VCEP v1.5 BP4: Missense: REVEL <= 0.249; Splicing: SpliceAI <= 0.1. Lab override: splicing sub-path restricted to non-missense variants whose primary plausible mechanism is splicing (synonymous, intronic outside +/-1,2); one mechanism per variant; a clean SpliceAI score does not clear an unresolved protein-level effect.
Evidence considered
ATM VCEP v1.5 (cspec) BP4 rule: 'Missense: REVEL score <=.249; Splicing: No predicted impact via splicing (SpliceAI <=0.1)' - the missense sub-path requires a missense variant; this variant is nonsense.
Lab curator override (gene_notes, BP4 missense/splicing policy, applied in addition to the official VCEP specification): BP4 evaluates exactly one applicable mechanism per variant; the splicing sub-path applies only to non-missense variants for which splicing is the primary plausible mechanism (synonymous, intronic outside canonical +/-1,2); a clean SpliceAI score only rules out the splicing mechanism and does not clear the variant's own protein-level effect. The override changes the outcome here from the VCEP's literal BP4 splicing text (SpliceAI 0.081 <= 0.1 would nominally satisfy it).
SpliceAI lookup for c.5692C>T: max delta 0.081 (<= 0.1 threshold) - noted but not applied to BP4 because the splicing sub-path is inapplicable to this nonsense variant per the override; the variant is damaging at the protein level (p.Arg1898Ter) regardless of splice neutrality.
PMID 25077176 (A-TWinnipeg, via paper_extracts): c.5692C>T was only detectable after cycloheximide treatment inhibited nonsense-mediated mRNA decay in LCL AT213LA (in trans with c.6200C>A), demonstrating the allele is NMD-targeted - direct evidence that the variant's mechanism is protein-level loss, not a splicing-neutral benign state.
BayesDel score 0.295247 treated as not_available for BP4: no verified published threshold available to this pipeline, so no benign/benign-neutral claim can be made from it.
Sun et al. 2025 Table S1 row for c.5692C>T: REVEL = NA and remaining computational columns are model extrapolations (DeepATM_predicted 'Yes', Confidence 'Computational prediction'); no usable benign computational evidence for BP4.
Still missing
REVEL score unavailable (lookup miss; Sun et al. Table S1 reports NA) - would be required for the missense BP4 sub-path, but the variant is not missense so the sub-path does not apply.
No RNA assay data (e.g. BP7(RNA)-style observed lack of aberrant splicing) are available for this variant; not required for this adjudication because BP4's splicing sub-path is restricted by the lab override to variants whose primary plausible mechanism is splicing (synonymous/intronic), and a nonsense SNV is not such a variant.
BP7N/ANot applicable: BP7 applies only to synonymous and deep intronic variants; c.5692C>T is an exonic nonsense change.
Details ▾Hide ▴
Reasoning
The ATM VCEP v1.5 BP7 rule is restricted to synonymous and deep intronic variants defined as further than (but not including) +7 from the donor site and further than (but not including) -21 from the acceptor site; BP7(RNA) applies to observed lack of aberrant RNA for silent substitutions and intronic variants. c.5692C>T is an exonic nonsense (stop-gain) variant, not a synonymous or deep intronic variant, so BP7 is not applicable. No RNA assay evidence of absent aberrant splicing is present in this case that could support a BP7(RNA) application, and such evidence would in any case be restricted to silent/intronic variants. SpliceAI (max delta 0.081) is noted for completeness but is not determinative for BP7, which is a positional/variant-class criterion rather than a score threshold.
Rule
ATM VCEP v1.5 BP7: Use for synonymous and deep intronic variants defined as further than (but not including) +7 and further than (but not including) -21 at donor and acceptor sites, respectively; BP7(RNA) for observed lack of aberrant RNA in silent/intronic variants.
Evidence considered
ATM VCEP v1.5 (cspec) BP7 rule: 'BP7: Use for synonymous and deep intronic variants defined as further than (but not including) +7 and further than (but not including) -21 at donor and acceptor sites, respectively. BP7(RNA): Observed lack of aberrant RNA defect for silent substitutions and intronic variants.' - c.5692C>T is an exonic stop-gain variant (p.Arg1898Ter, exon 38), not synonymous or deep intronic.
Variant class confirmed as nonsense via Mutalyzer/VariantValidator (NP_000042.3:p.(Arg1898Ter)) and Sun et al. Table S1 (variant type 'Nonsense'), excluding the synonymous/deep-intronic classes to which BP7 applies.
SpliceAI max delta 0.081 noted for completeness; BP7 is a variant-class/positional criterion and no score threshold within this group's remit can make BP7 applicable to a nonsense variant.
Still missing
No RNA-based evidence (BP7(RNA): observed lack of aberrant RNA) is available for this variant; not required because BP7 is restricted to synonymous and deep intronic variants and this is a nonsense SNV.
SpliceAI predicts no significant splice impact for this variant (max delta score = 0.08). BayesDel score = 0.295247.
9 paper records listed. 3 cited in the classification.
Cited in the classification
DHPLC screening of ATM gene in Italian patients affected by ataxia-telangiectasia: fourteen novel ATM mutations.
Magliozzi M et al. · 2006
The paper reports the ATM c.5692C>T variant, corresponding to p.(R1898*) (shown as R1898X), in a patient with classical ataxia-telangiectasia (patient AT50RM, heterozygous). It was classified by the authors as a nonsense mutation with a predicted truncated protein effect and reported as a novel mutation in this study. This establishes the variant as a bona fide germline null allele in an A-T patient.
→ PVS1 · very strong→ PM3 · strong→ PM5 · supportingNames this variant
40 AT50RM (het) 5692C>T R1898X Nonsense/Truncated protein This report
Location Table 3, Mutations detected in the ATM gene
Context DHPLC screening followed by bidirectional direct sequencing of the ATM coding exons in 22 patients with classical ataxia-telangiectasia (21 Italian and 1 Japanese patient).
Morbidity and mortality from ataxia-telangiectasia are associated with ATM genotype.
Micol R et al. · 2011 Aug
The paper does not explicitly mention ATM c.5692C>T or p.(R1898*). In its cohort of A-T patients, the authors classified nonsense mutations as class A/null mutations expected to cause complete loss of ATM expression or activity, including production of a truncated protein or complete absence of protein after nonsense-mediated mRNA decay. It provides general genotype-phenotype context that nonsense alleles behave as complete LOF alleles, consistent with PVS1 application, but contains no variant-specific data.
Class A mutations, which cause production of a truncated product or complete lack of protein after nonsense-mediated mRNA decay, included frameshift mutations that arose from small deletions of nucleotides or insertions in the coding sequence, nonsense mutations, splice-site mutations, or large deletions of the gene.
Context Retrospective French cohort study of 240 patients with ataxia-telangiectasia; ATM mutation analysis was performed in 184 patients using sequencing-based methods and, when needed, MLPA and cDNA/in-silico splice analysis.
A-TWinnipeg: Pathogenesis of rare ATM missense mutation c.6200C>A with decreased protein expression and downstream signaling, early-onset dystonia, cancer, and life-threatening radiotoxicity.
The paper explicitly identifies ATM c.5692C>T as the second mutation in patient-derived lymphoblastoid cell line AT213LA, which also carried c.6200C>A (p.A2067D). The variant was discovered after cycloheximide treatment inhibited nonsense-mediated mRNA decay, allowing the c.5692C>T allele to be detected by RNA sequencing and confirmed by genomic DNA sequencing. This is direct experimental evidence that the c.5692C>T transcript is degraded by NMD, and that the variant acts in trans with another ATM mutation in a patient with an atypical A-T phenotype.
→ PVS1 · very strong→ PM3 · strong→ PM5 · supportingNames this variant
This RNA showed the c.6200C>A variant as a heterozygous change (Fig. 3A, bottom) and we were subsequently able to identify the second mutation in AT213LA as c.5692C>T (Fig. 3B bottom sequencing panel).
Location Results, Measurement of nonsense-mediated mRNA decay, paragraph 1
Context Mutation analysis of patient-derived lymphoblastoid cell-line RNA and genomic DNA, with cycloheximide treatment to inhibit nonsense-mediated mRNA decay; AT213LA carried c.5692C>T in trans with c.6200C>A.
Other recorded publications
ATM Mutations in Cancer: Therapeutic Implications.
The paper does not mention ATM c.5692C>T or NP_000042.3:p.(R1898*). It provides general information about ATM mutations, reporting that germline and somatic mutations occur throughout the gene and its functional domains; in mantle cell lymphoma, ATM mutations are generally truncating or missense changes in the region encoding the PI3K domain. The authors state that the functional impact of each individual somatic mutation is not established.
Somatic mutations in ATM occur in many tumor types, though are more often found in hematologic malignancies. Forty-six percent of cases of T-cell prolymphocytic leukemia have inactivating mutations in both ATM alleles. ATM mutations observed in approximately 45% of mantle cell lymphoma cases; such mutations generally are truncating mutations or missense mutations within the region of the gene encoding the PI3K domain.
Somatic ATM Mutations in Cancer, paragraph 2
Inactive Atm abrogates DSB repair in mouse cerebellum more than does Atm loss, without causing a neurological phenotype.
This paper does not mention ATM c.5692C>T or NP_000042.3:p.(R1898*). It reports that most ATM mutations causing ataxia-telangiectasia are null alleles that truncate the ATM protein, and that truncated ATM is usually unstable, resulting in little or no ATM protein. In conditional mouse models, ATM loss or expression of catalytically inactive ATM impaired DNA double-strand-break responses and repair, with kinase-dead ATM producing a greater repair defect than ATM absence in some CNS contexts.
Most A-T mutations are null alleles that truncate the ATM protein. Since truncated ATM is usually unstable, most A-T patients are typically devoid of ATM [11,28].
Introduction, paragraph 3
Loss of ATM positively regulates Rac1 activity and cellular migration through oxidative stress.
The paper does not report ATM c.5692C>T or NP_000042.3:p.(R1898*). Instead, it investigates ATM loss or loss of ATM kinase activity using ATM-/- human fibroblasts, ATM kinase inhibition, and a kinase-dead ATM construct. ATM-deficient or kinase-inhibited cells showed increased intracellular reactive oxygen species, increased Rac1 activity, enhanced cell spreading, and increased migration. Rac1 knockdown or antioxidant treatment with N-acetyl-L-cysteine reduced Rac1 activity and restored migration toward control levels.
Twelve novel Atm mutations identified in Chinese ataxia telangiectasia patients.
The paper does not explicitly mention ATM c.5692C>T or NP_000042.3:p.(R1898*). It reports 12 ATM mutations identified in eight Chinese patients with ataxia-telangiectasia, including four nonsense mutations, and states that most ATM mutations are frameshift or nonsense variants predicted to truncate the ATM protein. No variant-specific functional assay, segregation result, or population-frequency observation is reported for c.5692C>T/p.(R1898*).
The majority of Atm mutations are frameshift or nonsense mutations (Wright et al. 1996; Concannon and Gatti 1997), which are predicted to truncate the whole ATM protein.
Introduction, paragraph 3
Sherloc: a comprehensive refinement of the ACMG-AMP variant classification criteria.
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= 3.71857e-06; MAF= 0.00037%, 6/1613522 alleles, homozygotes = 0) and has highest observed frequency in the Remaining individuals population (AF= 1.60036e-05; MAF= 0.00160%, 1/62486 alleles, homozygotes = 0); grpmax FAF= 6.8e-07.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 7.96794e-06; MAF= 0.00080%, 2/251006 alleles, homozygotes = 0) and has highest observed frequency in the South Asian population (AF= 3.26819e-05; MAF= 0.00327%, 1/30598 alleles, homozygotes = 0).
gnomAD-Canada
Absent from gnomAD-Canada v1.0.
ClinVar
This variant has been reported in ClinVar as Pathogenic (14 clinical laboratories) and as Pathogenic by ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Variant Curation Expert Panel, ClinGen (expert panel). (ClinVarID = 482526)
In silico
SpliceAI predicts no significant splice impact for this variant (max delta score = 0.08). BayesDel score = 0.295247.
Functional
OncoKB identified variant-specific curated literature and context relevant to functional review; biological-effect context: Likely Loss-of-function; curated oncogenicity label: Likely Oncogenic.
Somatic
This variant does not lie in a statistically significant hotspot. This variant has previously been reported in somatic cancers (COSMIC; COSV53742444, n = 10 times).
Cancer Hotspots
This variant does not lie in a statistically significant hotspot.
This variant was interpreted by pipeline 7.1.5.
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.