Five-step scale from benign to pathogenic. This variant is uncertain. Select a step to read what it means.
This variantScenario
BenignPathogenic
Uncertain.The recorded classification is uncertain significance.
In plain language
Gene-level background from the recorded interpretation.
Words with a dotted underline have a short definition. Hover or tap them.
About CDKN2A
CDKN2A encodes two key proteins, p16(INK4a) and p14(ARF), that help control cell division. p16 blocks the cell cycle by inhibiting CDK4 and CDK6, preventing progression from the G1 to S phase, while p14 stabilizes the tumor suppressor p53 by preventing its degradation. CDKN2A is an important tumor suppressor: it is frequently mutated, deleted, or silenced in many cancers, including melanoma, lymphoma, and pancreatic and lung cancer, and inherited changes in the gene can predispose people to familial melanoma and pancreatic cancer.
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 Not recorded. 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 (moderate): ultra-rare population frequency, gnomAD v4.1 allele frequency 3.12e-06 with 0 homozygotes, more than 10-fold below the 0.1% rarity threshold.
PP3Supporting
PP3 · supporting
Met (supporting): REVEL score 0.796, above the 0.750 threshold predicting a damaging effect.
Benign criteria applied · 0
No benign criterion marked as applied in the recorded assessment.
Recorded classification: Uncertain significance
Variant of Uncertain Significance: with no gene-specific framework available, the generic ACMG/AMP 2015 rules apply, and the only criteria met (PM2 moderate, PP3 supporting) satisfy no classification 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
5 / 1,604,348 alleles
gnomAD v4.1: 5 / 1,604,348 alleles. Recorded AF 3.11653e-06. Recorded homozygote count 0.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 4.3201e-06; MAF= 0.00043%, 1/231476 alleles, homozygotes = 0) and has highest observed frequency in the African/African American population (AF= 6.94444e-05; MAF= 0.00694%, 1/14400 alleles, homozygotes = 0).
OncoKB did not identify variant-specific reviewed functional evidence for this variant; gene-level curated context is available for reviewer follow-up. CDKN2A, which encodes both a cyclin-dependent kinase inhibitor and an MDM2 inhibitor, is altered by mutation and deletion in various cancers.
ACMG/AMP criteria under Not recorded, with the data and sources behind each one.
28 criteria with recorded statuses: 2 met · 5 not met · 15 not assessed · 6 not applicable. Open any row for the full reasoning, the rule and the evidence behind it.
Framework: no ClinGen CSPEC and no VCEP materials were retrieved for CDKN2A (cspec found=false; framework_mode=generic_acmg), so PS3/BS3 were adjudicated under the generic ACMG/AMP 2015 definitions (Richards et al., PMID:25741868).
Functional evidence is entirely absent for c.253G>T p.(Ala85Ser): no validated assay of this variant was found in ClinVar submissions (6 labs, all VUS; one submitter explicitly states functional studies have not been performed), in OncoKB (no variant-specific reviewed functional evidence, 'Unknown Oncogenic Effect'), or in any of the fetched ClinVar-cited papers, all of which concern CDKN2A locus deletions/methylation or gene-level leukemia biology and never mention p.Ala85Ser.
PS3 (deleterious effect shown by well-established functional studies) is not met; BS3 (no damaging effect shown by well-established functional studies) is not met. No strength is assigned because neither criterion is met.
COSMIC's single somatic occurrence (COSV58705440, n=1) and ClinVar's VUS consensus are observational and were not used as functional evidence.
Residual uncertainty: the literature corpus derives entirely from ClinVar-cited PMIDs; if a broader functional-literature search for CDKN2A p.Ala85Ser (e.g., p16 growth-suppression or CDK4-binding assays) is desired, it would be the only route to PS3/BS3-grade data.
Allelic-group (PM3/BP2) contribution for NM_000077.5:c.253G>T (p.Ala85Ser): PM3 is not applicable because CDKN2A-associated melanoma/cancer susceptibility is an autosomal dominant, incompletely penetrant disorder and PM3 is reserved for recessive disorders; the biallelic CDKN2A losses in the retrieved literature are somatic tumor events, not germline recessive disease.
BP2 is not assessed: no observation of the variant in trans with a pathogenic CDKN2A variant (fully penetrant dominant-disorder arm) or in cis with a pathogenic variant has been documented in any ClinVar submission or retrieved publication; familial phase testing data would be required.
Net effect on classification: the allelic group contributes no met or not_met criterion call; it neither supports pathogenicity nor benignity, leaving the classification to be driven by other evidence groups.
PP3 (supporting): REVEL 0.796 > 0.750 generic-framework missense threshold (generic_acmg_combination_rules, ACMG/AMP 2015 PMID 25741868); score also above the SVI PP3_moderate REVEL boundary 0.773 (PMID 36413997) but strength applied as supporting per the governing generic rules.
BP4 (not met): REVEL 0.796 is in the damaging-prediction range; the low SpliceAI max delta (0.071) is not applicable to this missense variant under the exactly-one-mechanism rule, and BayesDel is uncalibrated for this pipeline.
No VCEP or CSPEC for CDKN2A was available, so the generic ACMG fallback with the lab's defined in-silico operating thresholds governed this group's adjudication.
No ClinGen CSPEC or VCEP specification exists for CDKN2A (cspec.found=false, gene_vcep_dir=null), so population criteria BA1, BS1, BS2, and PM2 were adjudicated under the generic ACMG/AMP 2015 framework (Richards et al., PMID:25741868) at default strengths.
NM_000077.5:c.253G>T (p.Ala85Ser) is ultra-rare in reference populations: gnomAD v2.1 AF 4.32e-06 (1/231,476 alleles) and gnomAD v4.1 AF 3.12e-06 (5/1,604,348 alleles), 0 homozygotes in both, grpmax FAF 1.77e-05 in v4.1, absent from gnomAD-Canada v1.0; the maximum AF in any population/sex subset is ~0.012%.
BA1 and BS1 are not met: the maximum observed AF is orders of magnitude below the >5% BA1 threshold and below the >0.3% non-VCEP BS1 operating threshold, and is consistent with the rarity expected for a pathogenic allele in this adult-onset dominant cancer-susceptibility gene.
BS2 is not applicable: CDKN2A melanoma is an autosomal-dominant, adult-onset disease with incomplete penetrance, failing the BS2 precondition of full penetrance at an early age; the 5 heterozygous gnomAD v4.1 carriers therefore carry no benign weight.
PM2 is met (moderate, generic ACMG/AMP 2015 weighting): all observed frequencies, including the grpmax FAF and maximum subpopulation AF (~0.007%), are far below the 0.1% non-VCEP rarity operating threshold.
Net population-group contribution: PM2 (moderate) only; no benign population evidence (BA1/BS1/BS2) applies to this variant.
No criteria match.
Met · 2
PM2ModerateMet (moderate): ultra-rare population frequency, gnomAD v4.1 allele frequency 3.12e-06 with 0 homozygotes, more than 10-fold below the 0.1% rarity threshold.
Details ▾Hide ▴
Reasoning
PM2 (moderate under generic ACMG/AMP 2015) is met: the variant is at extremely low frequency, consistent with absence from controls for a dominant disorder. gnomAD v4.1: overall AF 3.12e-06 (0.00031%, 5/1,604,348 alleles) with grpmax FAF 1.77e-05 (0.0018%) and 0 homozygotes; gnomAD v2.1: single allele in 231,476 (AF 4.32e-06), 0 homozygotes; absent from gnomAD-Canada v1.0. The highest subpopulation AF is ~6.9e-05 (0.0069%, gnomAD v2.1 AFR and v4.1 EAS), and the largest sex-subset AF ~0.012% - all more than 10-fold below the 0.1% non-VCEP PM2 rarity operating threshold used by this pipeline, and far below the carrier frequency expected if this were a common benign polymorphism. The five heterozygous gnomAD v4.1 alleles are compatible with an ultra-rare allele in a reduced-penetrance adult-onset dominant cancer gene. PM2 met at the default moderate strength of the generic ACMG/AMP 2015 framework.
Rule
Generic ACMG/AMP 2015 (PMID:25741868) PM2: absent from controls (or at extremely low frequency if recessive); moderate.
Evidence considered
gnomAD v4.1: total AF 3.12e-06 (0.00031%; 5/1,604,348 joint exome+genome alleles), 0 homozygotes, grpmax FAF 1.77e-05 (0.0018%); distribution limited to East Asian (3), African/African American (1), and 'remaining' (1) with 0 alleles in >1.5 million European/South Asian/other chromosomes.
gnomAD v2.1: total AF 4.32e-06 (0.00043%; 1/231,476 exome alleles), 0 homozygotes.
Absent from gnomAD-Canada v1.0 (HostSeq genomes), an independent concordant observation of absence/ultra-rarity.
Maximum subpopulation AF is 6.94e-05 (gnomAD v2.1 African/African American) and 6.69e-05 (gnomAD v4.1 East Asian); largest sex-subset AF is ~1.17e-04 (0.012%, v2.1 AFR females) - all >10-fold below the 0.1% non-VCEP PM2 rarity operating threshold used by this pipeline.
PM2 rule per Richards et al. 2015 (PMID:25741868): PM2 = absent in controls (or extremely low frequency); for an autosomal-dominant adult-onset cancer gene with reduced penetrance, an AF of ~0.0003% overall and ~0.007% at maximum in any population satisfies the rarity expectation, and PM2 carries moderate weight in the 2015 framework.
Still missing
Rare-allele frequency estimate rests on only 5 observed alleles in gnomAD v4.1 (1 in v2.1); precision of the frequency point estimate is limited, but the maximum observed subpopulation AF (~0.007%) remains >10-fold below the 0.1% rarity threshold, so this does not change the decision.
gnomAD population cohorts are not deeply phenotyped for adult-onset cancer; a small number of carriers could conceivably be affected adults, which would not negate PM2 for a reduced-penetrance gene but is noted for transparency.
PP3SupportingMet (supporting): REVEL score 0.796, above the 0.750 threshold predicting a damaging effect.
Details ▾Hide ▴
Reasoning
Missense variant NM_000077.5:c.253G>T (p.Ala85Ser) in CDKN2A. No ClinGen CSPEC and no gene-specific VCEP materials exist for CDKN2A (cspec found:false; vcep_materials files_found:[]), so the generic ACMG fallback governs. The governing generic in-silico rules (generic_acmg_combination_rules) assign exactly one mechanism per variant type: for missense variants, REVEL only, with REVEL > 0.750 -> PP3 (supporting). The local REVEL v1.3 lookup returns 0.796 for this variant, above the 0.750 threshold, so PP3 is met at supporting strength. For context the score also exceeds the ClinGen SVI-published REVEL PP3_moderate boundary (0.773, Pejaver et al. PMID 36413997), but strength is applied per the governing generic framework. BayesDel score 0.3596 was retrieved but is not usable (no verified published threshold available to the pipeline per the governing rules). The variant's own SpliceAI score (max delta 0.071) is not applied because the splice path is reserved for intronic/synonymous/non-canonical-splice variants and must not clear a missense variant's REVEL-assessed protein-level effect.
Governing framework: no ClinGen CSPEC for CDKN2A and no gene-specific VCEP materials retrieved (case_summary framework_mode = generic_acmg; vcep_materials.json files_found []); generic ACMG/AMP 2015 (PMID 25741868) applies, with the lab's generic in-silico operating thresholds in generic_acmg_classification_rules.md governing PP3/BP4.
generic_acmg_classification_rules.md states: missense variants are evaluated with REVEL only; REVEL > 0.750 -> PP3 (supporting); REVEL < 0.250 -> BP4 (supporting); REVEL in [0.250, 0.750] -> neither met (gray zone); the SpliceAI path applies only to intronic, synonymous, or non-canonical-splice-position variants; a missense variant's own SpliceAI score does not clear its REVEL-assessed protein-level effect.
Local REVEL v1.3 lookup (source_registry key 'revel') returns score 0.796 for NM_000077.5:c.253G>T (chr9:21971106 C>A; p.Ala85Ser), which is > 0.750, meeting the governing generic PP3 (supporting) threshold.
Contextual calibration cross-check: REVEL 0.796 also exceeds the ClinGen SVI PP3_moderate threshold of 0.773 from the published REVEL calibration of Pejaver et al. (PMID 36413997, 'Calibration of computational tools for missense variant classification'); the strength applied here nonetheless follows the governing generic framework assignment of supporting.
BayesDel score 0.3596 was retrieved in prefetch but treated as not usable for PP3/BP4: the governing generic rules state no verified published threshold is available to the pipeline, and no named publication/table threshold can be cited for it.
Variant is a coding missense change (p.Ala85Ser, exon 2, position c.253 is 102 nt from the exon 2 start and not at a canonical splice position), so the missense/REVEL path is the single governing mechanism; SpliceAI (max delta 0.071) was therefore not used to argue no impact.
Still missing
No VCEP/gene-specific PP3/BP4 rule exists for CDKN2A; call rests on the generic framework single-predictor (REVEL) rule.
BayesDel score (0.3596) excluded: no verified published threshold available to the pipeline.
REVEL is the only calibrated missense predictor collected for this case; no independent conservation/other-tool confirmation available.
PS3Not met: no well-established functional study demonstrates a deleterious effect; submitters state functional studies have not been performed for this variant.
Details ▾Hide ▴
Reasoning
No ClinGen CSPEC or VCEP specification exists for CDKN2A (cspec found=false, vcep_materials empty), so PS3 is governed by the generic ACMG/AMP 2015 definition (Richards et al., PMID:25741868): well-established functional studies demonstrating a deleterious effect, default strength strong. No variant-specific functional assay evidence for c.253G>T p.(Ala85Ser) exists anywhere in the retrieved evidence corpus: (1) all 6 ClinVar submissions (variation 532289) classify the variant as Uncertain significance and none cites a functional assay; one submitter (Color Diagnostics, SCV000912414) explicitly states 'to our knowledge, functional studies have not been performed for this variant'; (2) OncoKB lists the A85S entry as 'Unknown Oncogenic Effect' with no variant-specific reviewed functional evidence; (3) COSMIC shows a single somatic occurrence (COSV58705440, n=1), which is observational and not a functional assay; (4) all fetched full texts among the validated ClinVar-cited PMIDs (7718873, 18519632, 28765326) plus the abstract-only paper 16818274 concern CDKN2A locus deletions/methylation and gene-level leukemia biology and never mention p.Ala85Ser (paper_extracts confirm variant_mentioned=false for each fetched paper). With no well-established functional study demonstrating a deleterious effect of p.Ala85Ser, PS3 cannot be met. Absence of assay data does not itself constitute any positive evidence for this criterion.
Rule
Generic ACMG/AMP 2015 (PMID:25741868): PS3 - well-established functional studies show a deleterious effect (strong). Not met because no validated functional studies of p.Ala85Ser were identified.
Evidence considered
Generic ACMG/AMP 2015 framework (Richards et al., PMID:25741868) defines PS3 as well-established functional studies demonstrating a deleterious effect (default strong); no CDKN2A VCEP/CSPEC was retrieved, so this generic definition governs.
ClinVar record 532289: 6 clinical laboratories all classify NM_000077.5(CDKN2A):c.253G>T (p.Ala85Ser) as Uncertain significance; no submission cites functional assay results. The Color Diagnostics submission (SCV000912414) explicitly states 'to our knowledge, functional studies have not been performed for this variant', directly confirming the absence of PS3-grade functional data.
OncoKB CDKN2A A85S entry: classification 'Unknown Oncogenic Effect'; OncoKB did not identify variant-specific reviewed functional evidence for this variant (is_variant_specific=false, not_reviewed=true).
COSMIC entry COSV58705440 documents a single somatic occurrence (n=1) of this change; a somatic observation is not a functional assay and provides no PS3/BS3 evidence.
Full-text review of ClinVar-cited PMIDs 7718873, 18519632 and 28765326 (via paper_extracts.json) and abstract-level review of 16818274: all concern homozygous deletions/methylation or gene-level biology of the CDKN2A locus in leukemias; none performs any functional assay of c.253G>T p.Ala85Ser and the variant is never mentioned (variant_mentioned=false).
PMID:25741868 paper extract confirms the ACMG/AMP guideline itself contains no CDKN2A- or p.Ala85Ser-specific functional evidence; it is cited only as the methodological definition of PS3/BS3.
Still missing
No validated functional assay data exist for p.Ala85Ser (e.g., p16INK4a growth-suppression/colony-formation assays, CDK4-binding/kinase-inhibition assays, or cell-cycle arrest assays) that would support or refute a deleterious effect.
The retrieved literature corpus is limited to ClinVar-cited PMIDs (13 total, 7 validated); no independent functional-literature search (e.g., for CDKN2A p.Ala85Ser functional characterization in melanoma/pancreatic cancer literature) is present in the evidence bundle.
BA1Not met: the highest observed allele frequency, about 0.012%, is roughly 400-fold below the 5% stand-alone benign threshold.
Details ▾Hide ▴
Reasoning
BA1 (stand-alone benign) requires a minor allele frequency greater than 5% in a reference population (Richards et al. 2015, ACMG/AMP). This variant is ultra-rare in every queried reference cohort. gnomAD v2.1 (exome): overall AF 4.32e-06 (1/231,476 alleles), 0 homozygotes, highest subpopulation AF in African/African American 6.94e-05 (1/14,400; AFR female subset 1.17e-04, 1/8,556). gnomAD v4.1 (joint exome+genome): overall AF 3.12e-06 (5/1,604,348 alleles), 0 homozygotes, grpmax FAF 1.77e-05, highest subpopulation AF in East Asian 6.69e-05 (3/44,870). The highest AF observed in any population or subset (~0.012% in gnomAD v2.1 AFR females) is more than 80-fold below a 1% level and roughly 400-fold below the 5% BA1 threshold, so BA1 is unambiguously not met. Two independent gnomAD builds and a separate Canadian cohort show concordant ultra-low frequency, arguing against a frequency-recording artifact.
Rule
Generic ACMG/AMP 2015 (PMID:25741868) BA1: minor allele frequency >5% in a reference population; stand-alone benign.
Evidence considered
gnomAD v2.1 (exome, 231,476 alleles): total AC 1, AF 4.32e-06 (0.00043%), homozygotes 0; only African/African American carries the allele (AF 6.94e-05; 1/14,400); genome allele count returned as 0 in this dataset.
gnomAD v4.1 (joint exome+genome, 1,604,348 alleles): total AC 5, AF 3.12e-06 (0.00031%), homozygotes 0; grpmax FAF 1.77e-05; alleles seen only in East Asian (3/44,870; AF 6.69e-05), African/African American (1/74,908), and 'remaining' (1/62,342); absent in European (non-Finnish, 1,179,578 alleles), South Asian, Admixed American, Ashkenazi Jewish, Finnish, Middle Eastern, and Amish.
Highest observed AF in any gnomAD population or sex subset is ~1.17e-04 (0.0117%, gnomAD v2.1 AFR females, 1/8,556 alleles), still far below any BA1 threshold.
BA1 rule per Richards et al. 2015 (ACMG/AMP standards, PMID:25741868): BA1 applies when minor allele frequency is >5% in a reference population; no population approaches this level.
Absent from gnomAD-Canada v1.0 (HostSeq genomes), consistent with the gnomAD v2.1/v4.1 ultra-low frequencies.
BS1Not met: the maximum population allele frequency, ~0.0069%, sits well below the 0.3% benign threshold for this adult-onset dominant cancer gene.
Details ▾Hide ▴
Reasoning
BS1 (strong) requires an allele frequency greater than expected for the disorder. CDKN2A is an autosomal-dominant, adult-onset melanoma/pancreatic-cancer predisposition gene with reduced penetrance, so the expected carrier frequency of a pathogenic allele is very low. Observed frequencies are far too low to support BS1: gnomAD v4.1 overall AF 3.12e-06 with grpmax FAF 1.77e-05 (0.0018%), and the maximum subpopulation AF is ~6.9e-05 (0.0069%, gnomAD v2.1 African/African American and v4.1 East Asian); the largest sex-subset AF is ~0.012%. These values are roughly 25-40-fold below the 0.3% non-VCEP BS1 operating threshold used by this pipeline and are consistent with the rarity expected for a disease allele in this gene, not with an allele frequency greater than expected. BS1 not met.
Rule
Generic ACMG/AMP 2015 (PMID:25741868) BS1: allele frequency greater than expected for the disorder; strong.
Evidence considered
gnomAD v4.1: total AF 3.12e-06 (5/1,604,348 alleles), grpmax FAF 1.77e-05, 0 homozygotes; East Asian subpopulation carries the highest count (3/44,870; AF 6.69e-05).
gnomAD v2.1: total AF 4.32e-06 (1/231,476 alleles), highest subpopulation AF 6.94e-05 (African/African American, 1/14,400).
The maximum AF observed in any population or sex subset (~1.17e-04 in gnomAD v2.1 AFR females) is ~25-fold below the 0.3% non-VCEP BS1 operating threshold used by this pipeline for dominant-disease genes.
BS1 rule per Richards et al. 2015 (PMID:25741868): allele frequency greater than expected for the disorder; for an adult-onset reduced-penetrance dominant cancer-susceptibility gene the tolerated frequency is low, and the observed frequencies are several orders of magnitude below population-level benign expectations.
Still missing
No CDKN2A-specific disease-prevalence/penetrance model was available to compute an exact 'expected allele frequency' for the disorder; not decision-limiting because the observed maximum AF (~0.012%) is far below any plausible BS1 threshold.
BS3Not met: no functional study demonstrates normal function; submitters report no functional studies have been performed for this variant.
Details ▾Hide ▴
Reasoning
No ClinGen CSPEC or VCEP specification exists for CDKN2A, so BS3 is governed by the generic ACMG/AMP 2015 definition (Richards et al., PMID:25741868): well-established functional studies showing no damaging effect, default strength strong. BS3 requires positive functional evidence of normal (wild-type-like) function. No such evidence exists for p.Ala85Ser: no functional assay of this variant was identified in any retrieved source (ClinVar submissions, including the explicit Color Diagnostics statement that functional studies have not been performed; OncoKB 'Unknown Oncogenic Effect' with no variant-specific functional evidence; no variant-level functional data in any of the fetched ClinVar-cited papers, which cover only CDKN2A locus deletion/methylation biology in leukemias and never mention p.Ala85Ser). The absence of functional data cannot support a benign functional assertion; therefore BS3 is not met.
Rule
Generic ACMG/AMP 2015 (PMID:25741868): BS3 - well-established functional studies show no damaging effect (strong). Not met because no functional studies of p.Ala85Ser demonstrating normal function were identified.
Evidence considered
Generic ACMG/AMP 2015 framework (Richards et al., PMID:25741868) defines BS3 as well-established functional studies showing no damaging effect (default strong); no CDKN2A VCEP/CSPEC was retrieved, so this generic definition governs.
ClinVar record 532289: all 6 submissions classify p.Ala85Ser as Uncertain significance and none reports a functional assay showing normal function; Color Diagnostics (SCV000912414) states 'to our knowledge, functional studies have not been performed for this variant'.
OncoKB CDKN2A A85S entry provides no variant-specific functional evidence and is classified 'Unknown Oncogenic Effect'.
Fetched ClinVar-cited full texts (7718873, 18519632, 28765326) and the abstract-only 16818274 contain only gene-level CDKN2A deletion/methylation biology from leukemia studies; none tests p.Ala85Ser function, so none can support a no-damaging-effect assertion.
PMID:25741868 is cited only as the methodological definition of BS3; its extract confirms it provides no CDKN2A/p.Ala85Ser-specific data.
Still missing
No validated functional assay demonstrating wild-type-like (benign) function of p.Ala85Ser exists in the retrieved corpus; BS3 would require e.g., a validated p16 growth-suppression or CDK4-binding assay showing no damaging effect.
The retrieved literature corpus is limited to ClinVar-cited PMIDs; a dedicated search for functional characterization of CDKN2A p.Ala85Ser could not be confirmed from the evidence bundle.
BP4Not met: REVEL 0.796 is far above the 0.250 no-impact threshold, so the score supports impact rather than no impact.
Details ▾Hide ▴
Reasoning
BP4 requires computational evidence suggesting no impact on the gene/gene product. Under the governing generic framework, the missense in-silico path uses REVEL only, and BP4 requires REVEL < 0.250; the observed REVEL score is 0.796, which instead falls in the PP3 range (> 0.750), so BP4 is not met. The low SpliceAI max delta (0.071) does not apply to this missense variant under the framework's exactly-one-mechanism rule and cannot be used to override the REVEL-assessed protein-level effect. BayesDel (0.3596) is not usable for PP3/BP4 (no verified published threshold). No other computational line suggesting no impact was collected.
Rule
Generic ACMG/AMP (PMID 25741868) BP4 via lab generic in-silico thresholds (generic_acmg_classification_rules.md): BP4 requires REVEL < 0.250 for missense variants. REVEL = 0.796 -> BP4 not met.
Evidence considered
Governing generic framework (generic_acmg_classification_rules.md, based on ACMG/AMP 2015 PMID 25741868) assigns BP4 only when REVEL < 0.250 for a missense variant; REVEL = 0.796 is far above this and above the PP3 threshold of 0.750, so no no-impact signal exists on the governing missense path.
SpliceAI lookup (source_registry key 'spliceai') reports max delta score 0.071 (DS_AG 0.0, DS_AL 0.003, DS_DG 0.002, DS_DL 0.071), i.e. no predicted splice impact, but the variant is a missense substitution and the governing rules reserve the SpliceAI path for intronic/synonymous/non-canonical-splice-position variants and prohibit using it to clear a missense variant's REVEL effect; it was therefore not counted toward BP4.
BayesDel local lookup returns 0.3596 but is not usable for BP4: the governing generic rules state no verified published threshold is available to the pipeline (no named publication/table threshold exists for this pipeline).
ClinVar single-submitter evidence sentences citing 'computational prediction' of deleterious effect (e.g. SCV000912414 internal REVEL threshold, SCV001176765 'predicted to be deleterious by in silico analysis') were noted but are not used as evidence here; no inference is made from another submitter's classification.
Still missing
None for this criterion: REVEL 0.796 unambiguously falls in the damaging-prediction range of the governing rule, excluding BP4.
PS1Not assessed: insufficient evidence was available to establish whether this exact amino-acid change has been reported as pathogenic.
Details ▾Hide ▴
Reasoning
PS1 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
PS2Not assessed: no confirmed de novo occurrence is documented, and no parental or maternity/paternity testing was available.
Details ▾Hide ▴
Reasoning
No case-level evidence of a confirmed de novo occurrence is present for CDKN2A NM_000077.5:c.253G>T (p.Ala85Ser): the bundle contains no proband phenotype/indication, no parental testing results, and no maternity/paternity confirmation. ClinVar (variation 532289) lists 6 submissions all classified Uncertain significance with no de novo assertion, and none of the full-text papers captured in paper_extracts.json mention this variant (variant_mentioned=false for all), so no published de novo observation exists in the case sources. PS2 therefore cannot be met or assessed.
Rule
Generic ACMG/AMP PS2: de novo variant (with confirmed maternity and paternity) in a proband with the phenotype and no family history; requires documented parental testing and a phenotype-consistent proband, and is not assessable from this evidence set.
Evidence considered
No ClinGen CSPEC or VCEP specification was found for CDKN2A (cspec.found=false, gene_vcep_dir=null); generic ACMG/AMP 2015 criteria are the governing fallback.
The consolidated case-evidence bundle contains no proband, trio, parental-testing, or confirmed de novo observation for c.253G>T; searches for proband/parental/de novo content across the bundle returned no clinical hits.
All six full-text papers represented in paper_extracts.json (PMID:7718873, PMID:18519632, PMID:25394175, PMID:25741868, PMID:28765326, PMID:31672839) were screened for variant-specific de novo/family content and none mention c.253G>T or p.(Ala85Ser) (variant_mentioned=false).
ClinVar's Uncertain significance classification (6 clinical laboratories) was not used to infer de novo status.
Still missing
Proband-level documentation: phenotype/indication and its consistency with a CDKN2A-related phenotype (e.g., familial melanoma/pancreatic cancer)
Variant genotype results for both biological parents demonstrating absence of c.253G>T
Confirmation of maternity and paternity
Assessment that the proband has no relevant family history
Any published de novo observation of c.253G>T/p.Ala85Ser
PS4Not assessed: insufficient evidence of case-level enrichment among affected individuals was available.
Details ▾Hide ▴
Reasoning
PS4 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
PM1Not assessed: insufficient evidence was available to determine whether the variant lies in a mutational hotspot.
Details ▾Hide ▴
Reasoning
PM1 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
PM5Not assessed: insufficient evidence was available to identify a different pathogenic missense at the same residue.
Details ▾Hide ▴
Reasoning
PM5 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
PM6Not assessed: no family-level report of a presumed de novo occurrence with unconfirmed parentage was available.
Details ▾Hide ▴
Reasoning
No clinical report or family-level observation documents c.253G>T (p.Ala85Ser) as a presumed de novo occurrence without fully confirmed maternity/paternity. No proband or parental testing information exists in the case-evidence bundle, and no fetched full-text paper mentions the variant. PM6 therefore cannot be assessed.
Rule
Generic ACMG/AMP PM6: presumed de novo variant in an affected individual when maternity and paternity are not confirmed; weaker than PS2 and requires documented case-level observation, neither of which is available here.
Evidence considered
The case-evidence bundle contains no report of c.253G>T as a presumed de novo event and no parental or family testing information.
All full-text papers in paper_extracts.json (PMID:7718873, PMID:18519632, PMID:25394175, PMID:25741868, PMID:28765326, PMID:31672839) fail to mention the variant, so no literature source documents a de novo occurrence.
A documented presumed de novo observation of the variant in an affected proband
Parental testing details, including whether maternity/paternity were confirmed and assay sensitivity for parental mosaicism
Phenotype consistency with a CDKN2A-related phenotype
Number of independent de novo observations, if any
PP1Not assessed: no co-segregation of the variant with disease in affected relatives was documented.
Details ▾Hide ▴
Reasoning
No affected relatives, informative meioses, pedigree, or genotype-phenotype co-segregation data for c.253G>T (p.Ala85Ser) are present in the case-evidence bundle. literature_pass.json assigned no segregation-themed papers for this variant, and all fetched full-text papers (paper_extracts.json) have variant_mentioned=false. Familial segregation cannot be assessed.
Rule
Generic ACMG/AMP PP1: co-segregation of the variant with disease in multiple affected family members, weighted by the number of informative meioses and phenotype consistency; not assessable without pedigree and segregation data.
Evidence considered
The case-evidence bundle provides no pedigree, family-level genotypes, affected-relative observations, or meiosis counts for c.253G>T.
literature_pass.json lists no segregation-themed candidate PMID for this case (criteria_index covers PP3/PP5/PS3/PS4 only), and no paper in paper_extracts.json mentions the variant, so no co-segregation report is available.
ClinVar VUS submissions do not carry segregation assertions and were not used to infer PP1.
Still missing
Pedigree/family structure identifying affected and unaffected relatives
Variant genotype results for affected relatives
Number of informative meioses and number of families
Phenotype definition, specificity, and diagnostic confirmation for each relative
Any published family co-segregation of c.253G>T/p.Ala85Ser
PP2Not assessed: insufficient evidence was available to assess this gene's missense-variant burden.
Details ▾Hide ▴
Reasoning
PP2 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
PP4Not assessed: insufficient phenotype-specific evidence was available.
Details ▾Hide ▴
Reasoning
PP4 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
PP5Not assessed: insufficient evidence was available to support a benign assertion from external sources.
Details ▾Hide ▴
Reasoning
PP5 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
BS4Not assessed: no family data demonstrating non-segregation of the variant with disease were available.
Details ▾Hide ▴
Reasoning
No genotyped affected relatives showing non-segregation, and no tested unaffected relatives informative for a CDKN2A-related phenotype, are documented anywhere in the case-evidence bundle. The absence of family data cannot be treated as evidence of non-segregation, and population frequency observations (gnomAD) do not constitute BS4 evidence. BS4 cannot be assessed.
Rule
Generic ACMG/AMP BS4: non-segregation of the variant with disease in affected family members, or presence in informative unaffected individuals, with interpretation accounting for penetrance, age at onset, and phenocopies; not assessable without family genotype-phenotype data.
Evidence considered
The case-evidence bundle contains no non-segregation observations, genotype-negative affected relatives, or informative unaffected-carrier data for c.253G>T.
gnomAD v2.1 presence (1/231,476 alleles, 0 homozygotes) is a population frequency observation relevant to frequency-based criteria only and cannot substitute for family non-segregation evidence under BS4.
No fetched full-text paper (paper_extracts.json: PMID:7718873, PMID:18519632, PMID:25394175, PMID:25741868, PMID:28765326, PMID:31672839) mentions the variant or reports family segregation results.
Still missing
Pedigree with validated phenotype status for tested relatives
Genotype results for affected relatives demonstrating non-segregation (affected relative without the variant)
Age and clinical follow-up sufficient to interpret unaffected status, considering incomplete penetrance and late onset
Assessment of phenocopies or an alternate genetic cause in the family
BP1Not assessed: insufficient evidence was available to evaluate whether benign missense variation is uncommon in this gene.
Details ▾Hide ▴
Reasoning
BP1 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
BP2Not assessed: no allelic-phase data (variant in trans or in cis with a pathogenic CDKN2A variant) were available.
Details ▾Hide ▴
Reasoning
BP2 (ACMG/AMP 2015, PMID:25741868) requires a documented allelic-relationship observation: the variant observed in trans with a pathogenic variant for a fully penetrant dominant disorder, or observed in cis with a pathogenic variant in any inheritance pattern. No such observation exists anywhere in this case. None of the six ClinVar germline submissions (all 'Uncertain significance', clinical testing) reports phase, parental/family testing, co-occurrence with a second CDKN2A variant, or a pathogenic CDKN2A variant on the other allele; their evidence signals are limited to in-silico prediction, population frequency, and 'functional studies have not been performed' statements. No publication retrieved for this case mentions c.253G>T (p.Ala85Ser) at all (variant_mentioned=false in all paper_extracts), so no proband-level or family-level cis/trans data can be drawn from the literature, which is entirely somatic leukemia CDKN2A-deletion work. Evidence is insufficient to determine whether the criterion is met; BP2 is returned not_assessed pending familial phase data.
Rule
BP2: Observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder or observed in cis with a pathogenic variant in any inheritance pattern (generic ACMG/AMP 2015, PMID:25741868). Not assessed - no phase/co-occurrence observation available.
Evidence considered
The generic ACMG/AMP 2015 framework (PMID:25741868) defines BP2 as 'Observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder or observed in cis with a pathogenic variant in any inheritance pattern' (supporting, benign direction); the local fetched text of PMID:25741868 is an excerpt that does not include the Table 3/4 criterion tables, so the BP2 definition is applied from the published guideline text.
CDKN2A melanoma susceptibility is autosomal dominant (PMID:25394175), so the first arm of BP2 (variant in trans with a pathogenic variant in a fully penetrant dominant disorder) is the theoretically applicable scenario; however, no such observation was found, and CDKN2A melanoma is itself incompletely penetrant.
ClinVar record 532289 (6 germline clinical-laboratory submissions, all 'Uncertain significance', single-submitter review status): no submission contains a cis/trans, phase, parental-origin, or second-variant co-occurrence assertion; extracted evidence signals are only in-silico (PP3-type), population-frequency, and 'functional studies have not been performed' statements.
All six papers with paper_extracts (PMID:25741868, PMID:28765326, PMID:31672839, PMID:7718873, PMID:18519632, PMID:25394175) have variant_mentioned=false for c.253G>T/p.Ala85Ser; they provide no proband or family observations of this variant, and the leukemia-deletion studies (PMID:7718873, PMID:18519632, PMID:28765326) concern somatic tumor genetics, so no cis/trans evidence can be derived from them.
Remaining case evidence is non-contributory to BP2: the single COSMIC somatic occurrence (COSV58705440, n=1) is a tumor event without germline phase information, and gnomAD v2.1/v4.1 homozygote counts of 0 in large control cohorts carry no cis/trans phase content.
Still missing
No phase information: parental or family segregation testing establishing whether c.253G>T lies in trans with a documented pathogenic CDKN2A variant (dominant-disease arm) or in cis with a pathogenic variant has not been reported for any proband.
No ClinVar submission or publication documents co-occurrence of this variant with a second CDKN2A variant (pathogenic or otherwise) with determined phase.
No germline case reports of this exact variant (c.253G>T, p.Ala85Ser) were found in the retrieved literature, so proband-observation-level allelic evidence is entirely absent.
BP5Not assessed: insufficient evidence was available to identify an alternate benign missense at the same residue.
Details ▾Hide ▴
Reasoning
BP5 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
BP6Not assessed: insufficient evidence was available from external benign classifications.
Details ▾Hide ▴
Reasoning
BP6 could not be assessed because its agent did not produce valid output.
Still missing
attempt 1: Hermes timed out after 200s and left no usable output_file
attempt 2: Hermes timed out after 200s and left no usable output_file
attempt 3: Hermes timed out after 200s and left no usable output_file
Not applicable · 6· recorded as not applicableShow reasons ▾Hide reasons ▴PVS1N/ANot applicable: c.253G>T is a missense change (p.Ala85Ser), not a null variant predicted to trigger nonsense-mediated decay.
Details ▾Hide ▴
Reasoning
PVS1 (ClinGen SVI PVS1 recommendations, PMC6185798, pvs1_generic_framework) applies only to null variant classes — nonsense, frameshift, canonical ±1/2 splice-consensus, translation-initiation-codon loss, and stop-loss variants — each expected to trigger nonsense-mediated decay or otherwise abolish normal protein function. NM_000077.5:c.253G>T in CDKN2A is a missense substitution predicted to produce NP_000068.1:p.(Ala85Ser). As a result, no null-variant mechanism (nonsense-mediated decay, truncation, or canonical splice disruption) is triggered by this variant class, so PVS1's structural prerequisite is not met by this variant class. This determination was made deterministically from the variant's consequence class, prior to LLM adjudication, because no structured VCEP criteria-combination framework governs this case (framework_complete=false); the generic ACMG/AMP 2015 fallback applies.
PM3N/ANot applicable: CDKN2A cancer susceptibility is autosomal dominant with incomplete penetrance, so the recessive in-trans criterion does not apply.
Details ▾Hide ▴
Reasoning
PM3 is defined for recessive disorders ('For recessive disorders, detected in trans with a pathogenic variant', ACMG/AMP 2015, PMID:25741868). CDKN2A-associated melanoma/cancer susceptibility (familial melanoma / melanoma-pancreatic cancer syndrome) is an autosomal dominant, incompletely penetrant disorder: germline CDKN2A mutations confer dominant heterozygote melanoma risk, with penetrance of at least 28% and up to ~91% in multi-case families (PMID:25394175). No established germline recessive phenotype is attributed to CDKN2A, and the biallelic CDKN2A/CDKN2B losses in the retrieved literature (PMID:7718873, PMID:18519632, PMID:28765326) are somatic tumor-acquired events, not germline biallelic disease. Additionally, no observation of this variant in trans with a documented pathogenic CDKN2A variant in an affected proband exists in any source: all six ClinVar germline submissions are VUS with no allelic/phase assertions, and none of the fetched publications mentions c.253G>T (p.Ala85Ser) at all (variant_mentioned=false in all paper_extracts). PM3 therefore cannot apply and was not scored.
Rule
PM3: For recessive disorders, detected in trans with a pathogenic variant (generic ACMG/AMP 2015, PMID:25741868). Inapplicable: CDKN2A melanoma/cancer susceptibility is autosomal dominant and no trans-phase observation exists.
Evidence considered
No ClinGen CSPEC or gene-specific VCEP specification was available for CDKN2A (cspec found: false; framework_mode: generic_acmg; gene_vcep_dir: null), so the generic ACMG/AMP 2015 framework (PMID:25741868) governs; it defines PM3 as 'For recessive disorders, detected in trans with a pathogenic variant', i.e., applicable only to a recessive gene-disease mechanism. The local fetched text of PMID:25741868 is an excerpt that contains the PM3/in-trans upgrade guidance but not the Table 3/4 criterion tables.
CDKN2A-associated hereditary melanoma is autosomal dominant with incomplete penetrance: the ACMG/NSGC referral guideline (PMID:25394175) states hereditary melanoma is caused by mutations in CDKN2A/ARF and CDK4, with melanoma penetrance in CDKN2A-mutation families of at least 28% (up to ~91% in families with multiple cases) and CDKN2A mutations in 38% of 466 reviewed families with >=3 melanomas - consistent with dominant heterozygous susceptibility and no recessive germline disorder.
Biallelic CDKN2A loss reported in the retrieved literature is somatic: PMID:7718873 found homozygous p16 gene deletions (semiquantitative multiplex PCR, Southern-blot confirmed) in leukemias (CML lymphoid blast crisis 5/10, ALL 5/19, AML 1/25) and demonstrated by paired chronic-phase/remission/constitutional DNA that the deletions were somatically acquired, not inherited; PMID:18519632 reports predominantly biallelic losses of the whole CDKN2A/CDKN2B cluster in >50% of Ph+ ALL tumor samples; PMID:28765326 detected CDKN2A mutations in AML/MDS tumor NGS where germline DNA was explicitly unavailable. None constitutes germline biallelic (recessive) evidence for this variant.
No affected-proband observation of c.253G>T in trans with a pathogenic CDKN2A variant was identified: the six ClinVar germline clinical-testing submissions (SCV000912414 Color, SCV001176765 Ambry, SCV000760544 Labcorp/Invitae, SCV003800618 LabCorp Women's Health, SCV003842766 GeneDx, SCV004212496 Baylor; all 'Uncertain significance', no expert-panel submissions) contain only in-silico/population/'no functional studies' statements and no allelic-relationship or second-variant assertions.
gnomAD v4.1 lists 5/1,604,348 alleles with 0 homozygotes and gnomAD v2.1 lists 1/231,476 alleles with 0 homozygotes for this variant - consistent with extreme rarity but not an observation of recessive disease, and non-contributory once the dominant mechanism makes PM3 inapplicable.
PM4N/ANot applicable: this missense substitution leaves protein length unchanged, so there is no length-change effect to evaluate.
Details ▾Hide ▴
Reasoning
PM4 (Richards et al. 2015, PMID:25741868) applies to protein length changes: in-frame insertions/deletions in a non-repetitive region, or stop-loss variants that extend the protein past the reference stop codon. NM_000077.5:c.253G>T in CDKN2A is a missense substitution predicted to produce NP_000068.1:p.(Ala85Ser). As a result, no change in protein length occurs, so there is nothing for this criterion to evaluate, so PM4's structural prerequisite is not met by this variant class. This determination was made deterministically from the variant's consequence class, prior to LLM adjudication, because no structured VCEP criteria-combination framework governs this case (framework_complete=false); the generic ACMG/AMP 2015 fallback applies.
BS2N/ANot applicable: CDKN2A disease is adult-onset with reduced penetrance, so the healthy-adult observation criterion does not apply.
Details ▾Hide ▴
Reasoning
BS2 (strong) applies when a variant is observed in a healthy adult for a recessive (homozygous), X-linked (hemizygous), or dominant (heterozygous) disorder with FULL penetrance expected at an EARLY age. CDKN2A-related melanoma/pancreatic cancer is an autosomal-dominant, adult-onset cancer-predisposition syndrome with reduced (age-dependent) penetrance, so the criterion's precondition is not fulfilled and BS2 cannot be applied to any heterozygous carriers. Consistent with this, gnomAD v2.1 and v4.1 report 0 homozygotes and only a handful of heterozygous carriers (1 allele in v2.1; 5 alleles in v4.1 among ~800,000 individuals); these unphenotyped population heterozygotes carry no benign weight for this gene-disease mechanism. BS2 is therefore not_applicable rather than met.
Rule
Generic ACMG/AMP 2015 (PMID:25741868) BS2: observed in a healthy adult for a recessive (homozygous), X-linked (hemizygous), or dominant (heterozygous) disorder with full penetrance expected at an early age; strong.
Evidence considered
gnomAD v2.1 and v4.1 both report 0 homozygotes for this variant; heterozygous carriers number 1 (v2.1, 231,476 alleles) and 5 (v4.1, 1,604,348 alleles).
Heterozygous carriers in gnomAD are drawn from unselected/unscreened population cohorts and cannot be considered phenotypically confirmed healthy adults for an adult-onset cancer predisposition.
BS2 rule per Richards et al. 2015 (PMID:25741868): BS2 requires observation in a healthy adult for a disorder with full penetrance expected at an early age; CDKN2A-associated melanoma is adult-onset with incomplete penetrance, so the criterion is inapplicable and contributes no benign evidence.
Still missing
No clinically phenotyped healthy-adult cohort data (age and cancer status) exist for the heterozygous gnomAD carriers; irrelevant to the decision because the BS2 precondition of an early-onset, fully penetrant dominant phenotype is not met for CDKN2A melanoma.
BP3N/ANot applicable: this is a missense substitution, not an in-frame insertion/deletion in a repetitive region.
Details ▾Hide ▴
Reasoning
BP3 (Richards et al. 2015, PMID:25741868) applies to in-frame insertions/deletions located in a repetitive region without a known function. NM_000077.5:c.253G>T in CDKN2A is a missense substitution predicted to produce NP_000068.1:p.(Ala85Ser). As a result, the variant does not alter protein length within a repeat region, so there is nothing for this criterion to evaluate, so BP3's structural prerequisite is not met by this variant class. This determination was made deterministically from the variant's consequence class, prior to LLM adjudication, because no structured VCEP criteria-combination framework governs this case (framework_complete=false); the generic ACMG/AMP 2015 fallback applies.
BP7N/ANot applicable: this is a missense substitution, not a synonymous variant, so the silent-variant criterion does not apply.
Details ▾Hide ▴
Reasoning
BP7 (Richards et al. 2015, PMID:25741868) is defined for a synonymous (silent) variant for which splicing prediction algorithms predict no impact to the splice consensus sequence and the nucleotide is not highly conserved. NM_000077.5:c.253G>T in CDKN2A is a missense substitution predicted to produce NP_000068.1:p.(Ala85Ser). As a result, the encoded protein sequence is altered, so the criterion's silent-variant premise does not hold regardless of any splicing prediction, so BP7's structural prerequisite is not met by this variant class. This determination was made deterministically from the variant's consequence class, prior to LLM adjudication, because no structured VCEP criteria-combination framework governs this case (framework_complete=false); the generic ACMG/AMP 2015 fallback applies.
Rule
deterministic_prefilter: requires_synonymous
Evidence considered
Deterministic consequence classification: missense (prefetch.json normalized protein consequence)
Sources
generic_acmg_combination_rules
Identifiers
Gene
CDKN2A
Transcript
NM_000077.5
Coding (HGVS c.)
NM_000077.5:c.253G>T
Protein (HGVS p.)
NP_000068.1:p.(Ala85Ser)
GRCh38
chr9:21971106 C>A
GRCh37
chr9:21971105 C>A
Location and type
Exon 2 · missense
ClinVar
2563912, 532289
COSMIC
COSV58705440
Gene diagram
Every interpreted variant in CDKN2A, mapped onto the exon structure of NM_000077.5. Select a row to find it on the plot.
CDKN2ANM_000077.5
Fetching transcript structure from UCSC…
Exons
—
Transcript span
—
Strand
—
Variants mapped
—
Source
UCSC ncbiRefSeqCurated
All variants in CDKN2A—Select a row to find it on the plot; use the link to open its page.
AF 3.1e-06 · 5 / 1,604,348 alleles · 0 hom · FAF 1.8e-05
East Asian3 / 44,870
6.7e-05
Remaining individuals1 / 62,342
1.6e-05
African/African American1 / 74,908
1.3e-05
0 alternate alleles recorded in 7 other groups: Admixed American, European (Finnish), Amish, Middle Eastern, South Asian, Ashkenazi Jewish, European (non-Finnish).
gnomAD v2.1
AF 4.3e-06 · 1 / 231,476 alleles · 0 hom
African/African American1 / 14,400
6.9e-05
0 alternate alleles recorded in 7 other groups: Admixed American, Ashkenazi Jewish, East Asian, European (Finnish), European (non-Finnish), Remaining individuals, South Asian.
Statements
gnomAD v4.1
This variant is present in gnomAD v4.1 (AF= 3.11653e-06; MAF= 0.00031%, 5/1604348 alleles, homozygotes = 0) and has highest observed frequency in the East Asian population (AF= 6.68598e-05; MAF= 0.00669%, 3/44870 alleles, homozygotes = 0); grpmax FAF= 1.773e-05.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 4.3201e-06; MAF= 0.00043%, 1/231476 alleles, homozygotes = 0) and has highest observed frequency in the African/African American population (AF= 6.94444e-05; MAF= 0.00694%, 1/14400 alleles, homozygotes = 0).
gnomAD-Canada
Absent from gnomAD-Canada v1.0.
SpliceAI and Pangolin
SpliceAI acceptor lossat -144 bp
0.003
SpliceAI acceptor gainat +0 bp
0.000
SpliceAI donor lossat +122 bp
0.071
SpliceAI donor gainat -135 bp
0.002
Pangolin splice lossmagnitude shown
-0.062
Pangolin splice gain
0.002
Other predictors
REVEL
0.796
BayesDel
0.360
How the predictions were used
PP3SupportingMet (supporting): REVEL score 0.796, above the 0.750 threshold predicting a damaging effect.
Details ▾Hide ▴
Reasoning
Missense variant NM_000077.5:c.253G>T (p.Ala85Ser) in CDKN2A. No ClinGen CSPEC and no gene-specific VCEP materials exist for CDKN2A (cspec found:false; vcep_materials files_found:[]), so the generic ACMG fallback governs. The governing generic in-silico rules (generic_acmg_combination_rules) assign exactly one mechanism per variant type: for missense variants, REVEL only, with REVEL > 0.750 -> PP3 (supporting). The local REVEL v1.3 lookup returns 0.796 for this variant, above the 0.750 threshold, so PP3 is met at supporting strength. For context the score also exceeds the ClinGen SVI-published REVEL PP3_moderate boundary (0.773, Pejaver et al. PMID 36413997), but strength is applied per the governing generic framework. BayesDel score 0.3596 was retrieved but is not usable (no verified published threshold available to the pipeline per the governing rules). The variant's own SpliceAI score (max delta 0.071) is not applied because the splice path is reserved for intronic/synonymous/non-canonical-splice variants and must not clear a missense variant's REVEL-assessed protein-level effect.
Governing framework: no ClinGen CSPEC for CDKN2A and no gene-specific VCEP materials retrieved (case_summary framework_mode = generic_acmg; vcep_materials.json files_found []); generic ACMG/AMP 2015 (PMID 25741868) applies, with the lab's generic in-silico operating thresholds in generic_acmg_classification_rules.md governing PP3/BP4.
generic_acmg_classification_rules.md states: missense variants are evaluated with REVEL only; REVEL > 0.750 -> PP3 (supporting); REVEL < 0.250 -> BP4 (supporting); REVEL in [0.250, 0.750] -> neither met (gray zone); the SpliceAI path applies only to intronic, synonymous, or non-canonical-splice-position variants; a missense variant's own SpliceAI score does not clear its REVEL-assessed protein-level effect.
Local REVEL v1.3 lookup (source_registry key 'revel') returns score 0.796 for NM_000077.5:c.253G>T (chr9:21971106 C>A; p.Ala85Ser), which is > 0.750, meeting the governing generic PP3 (supporting) threshold.
Contextual calibration cross-check: REVEL 0.796 also exceeds the ClinGen SVI PP3_moderate threshold of 0.773 from the published REVEL calibration of Pejaver et al. (PMID 36413997, 'Calibration of computational tools for missense variant classification'); the strength applied here nonetheless follows the governing generic framework assignment of supporting.
BayesDel score 0.3596 was retrieved in prefetch but treated as not usable for PP3/BP4: the governing generic rules state no verified published threshold is available to the pipeline, and no named publication/table threshold can be cited for it.
Variant is a coding missense change (p.Ala85Ser, exon 2, position c.253 is 102 nt from the exon 2 start and not at a canonical splice position), so the missense/REVEL path is the single governing mechanism; SpliceAI (max delta 0.071) was therefore not used to argue no impact.
Still missing
No VCEP/gene-specific PP3/BP4 rule exists for CDKN2A; call rests on the generic framework single-predictor (REVEL) rule.
BayesDel score (0.3596) excluded: no verified published threshold available to the pipeline.
REVEL is the only calibrated missense predictor collected for this case; no independent conservation/other-tool confirmation available.
BP4Not met: REVEL 0.796 is far above the 0.250 no-impact threshold, so the score supports impact rather than no impact.
Details ▾Hide ▴
Reasoning
BP4 requires computational evidence suggesting no impact on the gene/gene product. Under the governing generic framework, the missense in-silico path uses REVEL only, and BP4 requires REVEL < 0.250; the observed REVEL score is 0.796, which instead falls in the PP3 range (> 0.750), so BP4 is not met. The low SpliceAI max delta (0.071) does not apply to this missense variant under the framework's exactly-one-mechanism rule and cannot be used to override the REVEL-assessed protein-level effect. BayesDel (0.3596) is not usable for PP3/BP4 (no verified published threshold). No other computational line suggesting no impact was collected.
Rule
Generic ACMG/AMP (PMID 25741868) BP4 via lab generic in-silico thresholds (generic_acmg_classification_rules.md): BP4 requires REVEL < 0.250 for missense variants. REVEL = 0.796 -> BP4 not met.
Evidence considered
Governing generic framework (generic_acmg_classification_rules.md, based on ACMG/AMP 2015 PMID 25741868) assigns BP4 only when REVEL < 0.250 for a missense variant; REVEL = 0.796 is far above this and above the PP3 threshold of 0.750, so no no-impact signal exists on the governing missense path.
SpliceAI lookup (source_registry key 'spliceai') reports max delta score 0.071 (DS_AG 0.0, DS_AL 0.003, DS_DG 0.002, DS_DL 0.071), i.e. no predicted splice impact, but the variant is a missense substitution and the governing rules reserve the SpliceAI path for intronic/synonymous/non-canonical-splice-position variants and prohibit using it to clear a missense variant's REVEL effect; it was therefore not counted toward BP4.
BayesDel local lookup returns 0.3596 but is not usable for BP4: the governing generic rules state no verified published threshold is available to the pipeline (no named publication/table threshold exists for this pipeline).
ClinVar single-submitter evidence sentences citing 'computational prediction' of deleterious effect (e.g. SCV000912414 internal REVEL threshold, SCV001176765 'predicted to be deleterious by in silico analysis') were noted but are not used as evidence here; no inference is made from another submitter's classification.
Still missing
None for this criterion: REVEL 0.796 unambiguously falls in the damaging-prediction range of the governing rule, excluding BP4.
BP7N/ANot applicable: this is a missense substitution, not a synonymous variant, so the silent-variant criterion does not apply.
Details ▾Hide ▴
Reasoning
BP7 (Richards et al. 2015, PMID:25741868) is defined for a synonymous (silent) variant for which splicing prediction algorithms predict no impact to the splice consensus sequence and the nucleotide is not highly conserved. NM_000077.5:c.253G>T in CDKN2A is a missense substitution predicted to produce NP_000068.1:p.(Ala85Ser). As a result, the encoded protein sequence is altered, so the criterion's silent-variant premise does not hold regardless of any splicing prediction, so BP7's structural prerequisite is not met by this variant class. This determination was made deterministically from the variant's consequence class, prior to LLM adjudication, because no structured VCEP criteria-combination framework governs this case (framework_complete=false); the generic ACMG/AMP 2015 fallback applies.
Rule
deterministic_prefilter: requires_synonymous
Evidence considered
Deterministic consequence classification: missense (prefetch.json normalized protein consequence)
Sources
generic_acmg_combination_rules
Statement
In silico
SpliceAI predicts no significant splice impact for this variant (max delta score = 0.07). REVEL score = 0.796. BayesDel score = 0.359592.
8 paper records listed. 1 cited in the classification. None of the listed paper records explicitly indicates that the variant is named.
Cited in the classification
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology.
Richards et al., ACMG/AMP/CAP 2015 standards and guidelines for sequence variant interpretation - the methodological source defining PS3 (well-established functional studies show a deleterious effect) and BS3 (well-established functional studies show no damaging effect) under the generic framework. It does not mention CDKN2A, c.253G>T or p.Ala85Ser anywhere; only generic illustrative variants are named. Consulted solely as the rule definition for PS3/BS3 adjudication in this case; it contributes no variant- or gene-specific functional evidence.
→ PM2 · moderate→ PP3 · supportingDoes not name this variant
The variant classification approach described in this paper is meant to be applicable to variants in all Mendelian genes whether identified by single gene tests, multi-gene panels, exome sequencing or genome sequencing.
Location Methods (workgroup development section, near statement of scope/applicability)
Context None applicable to the variant; expert-consensus guideline developed by an ACMG/AMP/CAP workgroup via surveys of 100+ North American clinical laboratories (2013-2014) and workgroup testing of the proposed criteria.
Other recorded publications
Outcome of Azacitidine Therapy in Acute Myeloid Leukemia Is not Improved by Concurrent Vorinostat Therapy but Is Predicted by a Diagnostic Molecular Signature.
Craddock CF et al. · 2017
Reports gene-level CDKN2A mutations in 9 of 250 AML/MDS patients treated with azacitidine-based therapy (RAvVA randomized phase II trial), associated with markedly reduced overall survival (median OS 4.5 vs 11.0 months; multivariable HR 10.0). Mutations are described only at gene level as heterozygous predicted loss-of-function events (nonsense in two patients; non-synonymous SNVs in seven, six previously reported); whether p.Ala85Ser falls within that group cannot be determined because no per-variant listing is provided. No functional assay of any specific CDKN2A variant is performed; no PS3/BS3 evidence for p.Ala85Ser.
In our study the CDKN2A mutations were nonsense in two patients and in the other seven were either non-synonymous SNVs that had previously been reported (six patients) or within two codons of a previously reported mutation (one patient).
Discussion, paragraph 2 (final section, p. 15-16); survival statistics in Table 2 'Overall Survival' (Mutations: CDKN2A) and Results section 'Clinical and molecular factors predicting outcome after AZA based therapy'
Management of patients with increased risk for familial pancreatic cancer: updated recommendations from the International Cancer of the Pancreas Screening (CAPS) Consortium.
Homozygous deletions of the p16 tumor-suppressor gene are associated with lymphoid transformation of chronic myeloid leukemia.
Sill H et al. · 1995
This paper does not describe c.253G>T (p.Ala85Ser) or any CDKN2A/p16 point mutation; it is entirely a study of homozygous deletions of the p16 gene (exons 1 and 2) in leukemias, detected by semiquantitative multiplex PCR with Southern blot confirmation in selected cases. Homozygous p16 deletions were found in 5/10 CML lymphoid blast crisis, 5/19 de novo ALL and 1/25 AML patients; deletions were somatically acquired (present only in acute-leukemia DNA in 5 individuals with chronic-phase/remission/constitutional DNA). It provides no functional assay data on p.Ala85Ser and cannot inform PS3/BS3.
Homozygous deletions of the p16 gene were found in 5 of 10 patients with CML in lymphoid BC (50%), 5 of 19 patients with ALL (26%), and 1 of 25 patients with AML (4%). No patient with myeloid BC showed any p16 loss
Results, paragraph 2 (also summarized in the Abstract and Table 1)
The prognostic significance of CDKN2A, CDKN2B and MTAP inactivation in B-lineage acute lymphoblastic leukemia of childhood. Results of the EORTC studies 58881 and 58951.
Mirebeau D et al. · 2006
Abstract-only review (full text not fetched; no paper_extracts entry exists). The paper studies CDKN2A inactivation by deletion and methylation of the 9p21 region in childhood B-lineage ALL (EORTC 58881/58951) and its prognostic significance; it concerns CDKN2A/INK4A-ARF inactivation at the locus level, not any specific CDKN2A sequence variant. The specific variant c.253G>T (p.Ala85Ser) is not mentioned in the abstract and no functional assay of p.Ala85Ser is reported; it was reviewed because the literature triage flagged it as a PS3 candidate, and it does not support PS3 or BS3.
Deletion and methylation of the 9p21 chromosomal region are frequent in childhood acute lymphoblastic leukemia (ALL) but the prognostic significance is controversial.
Abstract (full text not available in case publications directory)
Failure of CDKN2A/B (INK4A/B-ARF)-mediated tumor suppression and resistance to targeted therapy in acute lymphoblastic leukemia induced by BCR-ABL.
Mullighan CG et al. · 2008
Perspective/commentary (not a primary variant study) on deletions of the CDKN2A/B (INK4A/B-ARF) locus in BCR-ABL (Ph+)-driven leukemias: predominantly biallelic whole-cluster deletions occur in ~54% of Ph+ ALL cases by SNP array (rising to 64% by quantitative genomic PCR); the authors propose CDKN2A/B loss removes p16INK4A/p14ARF-imposed tumor suppression based on BCR-ABL+ Arf-null mouse models. Only whole-locus copy-number deletions and Arf-null biology are discussed; the specific missense variant c.253G>T (p.Ala85Ser) is never mentioned and no functional or case data on it are reported. Not usable for PS3/BS3.
In samples taken at the time patients first presented clinically with Ph+ ALL, 84% had already sustained IKFZ1 deletions, whereas >50% exhibited PAX5 or predominantly biallelic losses encompassing the entire CDKN2A (INK4A–ARF) and CDKN2B (INK4B) gene cluster (Table 1).
Section 'Frequent deletions of IKFZ1 (IKAROS), PAX5, and CDKN2A/B (INK4A/B-ARF) in Ph+ ALL', paragraph 1 (text accompanying Table 1)
A practice guideline from the American College of Medical Genetics and Genomics and the National Society of Genetic Counselors: referral indications for cancer predisposition assessment.
Hampel H et al. · 2015
ACMG/NSGC practice guideline on referral indications for cancer predisposition assessment, not a variant-level study. It never mentions c.253G>T (p.Ala85Ser). Gene-level content establishes the inheritance model used in this group's PM3 decision: hereditary melanoma is caused by mutations in the CDKN2A/ARF gene (encoding p16 and p14ARF) and the CDK4 gene; melanoma penetrance in families with CDKN2A mutations is at least 28% (possibly up to ~91% in families with multiple cases); CDKN2A mutations were found in 38% of 466 families with >=3 melanomas. This documents a dominant, incompletely penetrant heterozygous germline susceptibility, i.e., no recessive mechanism for PM3. No functional, case, segregation, or cis/trans data specific to p.Ala85Ser are provided.
Hereditary melanoma is caused by mutations in the CDKN2A/ARF gene, which encodes p16 and p14ARF, and the CDK4 gene. ... The penetrance for melanoma in families with CDKN2A mutations is at least 28%, although it is perhaps as high as 91% in families with multiple cases.83–85 A review of 466 families with at least three cases of melanoma revealed 38% had CDKN2A mutations.86 ... CDKN2A gene mutations seem to be rare in families with pancreatic cancer without any cases of melanoma44 but occur in up to 11% (2/18) of families with both pancreatic cancer and melanoma.87
Hereditary melanoma, also known as familial atypical mole and malignant melanoma (OMIM 155600) section, page 79; also Familial pancreatic cancer section, page 78
Cancer Genetics Risk Assessment and Counseling (PDQ®): Health Professional Version.
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= 3.11653e-06; MAF= 0.00031%, 5/1604348 alleles, homozygotes = 0) and has highest observed frequency in the East Asian population (AF= 6.68598e-05; MAF= 0.00669%, 3/44870 alleles, homozygotes = 0); grpmax FAF= 1.773e-05.
gnomAD v2.1
This variant is present in gnomAD v2.1 (AF= 4.3201e-06; MAF= 0.00043%, 1/231476 alleles, homozygotes = 0) and has highest observed frequency in the African/African American population (AF= 6.94444e-05; MAF= 0.00694%, 1/14400 alleles, homozygotes = 0).
gnomAD-Canada
Absent from gnomAD-Canada v1.0.
ClinVar
This variant has been reported in ClinVar as Uncertain significance (6 clinical laboratories). (ClinVarID = 532289)
In silico
SpliceAI predicts no significant splice impact for this variant (max delta score = 0.07). REVEL score = 0.796. BayesDel score = 0.359592.
Functional
OncoKB did not identify variant-specific reviewed functional evidence for this variant; gene-level curated context is available for reviewer follow-up. CDKN2A, which encodes both a cyclin-dependent kinase inhibitor and an MDM2 inhibitor, is altered by mutation and deletion in various cancers.
Somatic
This variant does not lie in a statistically significant hotspot. This variant has previously been reported in somatic cancers (COSMIC; COSV58705440, n = 1 times).
Cancer Hotspots
This variant does not lie in a statistically significant hotspot.
This variant was interpreted by pipeline 7.2.0.
The current version is 11.0.07.
Re-running queues a fresh interpretation; the result replaces this page when complete.
Re-runs are not free. Blank draws on today's public interpretation allowance; with a code, it uses one of that code's uses.