NTHL1 c.900C>T, p.(Ala300=)

Reasoning

BP7 is in scope: NM_002528.7:c.900C>T is a synonymous substitution (NP_002519.2:p.(Ala300=)) that changes no amino acid and no protein length, so BP7 - not PP3/BP4 - is the computational criterion this variant belongs to. The first prong, no predicted impact on splicing or new splice-site creation, is met unambiguously. SpliceAI returns DS_AG 0.003, DS_AL 0.004, DS_DG 0.002 and DS_DL 0.00, for a maximum delta score of 0.004, with no predicted acceptor or donor gain or loss; that is far below the SpliceAI recommended cutoff of 0.2 for significant splice impact and below the 0.1 no-impact level (Jaganathan et al. 2019, Cell, PMID:30661751). The model's predicted events sit 123-277 nucleotides away from the variant, i.e. nothing local is created or lost. Pangolin outputs recorded alongside (splice gain 0.004, splice loss -0.008) are concordant, although they are not calibrated for this call and are not the basis for it. The second prong, that the nucleotide is not highly conserved, cannot be met with a direct metric: no phyloP, GERP or phastCons value for c.900 is present in this case's evidence. It is nevertheless not contradicted and is supported indirectly - the position tolerates substitution in the population, with gnomAD v2.1 reporting 125/272238 alleles (AF 0.0459%) including 3 homozygotes and gnomAD v4.1 reporting 341/1603718 alleles (AF 0.0213%) including 4 homozygotes, which is not the profile of a highly conserved, constrained nucleotide at a splice-relevant position. No ClinGen CSPEC or gene-specific BP7 rule exists for NTHL1, so generic ACMG/AMP 2015 BP7 was applied. BP7 is met at supporting strength, with the absence of a quantitative conservation score recorded as the remaining gap rather than treated as evidence.

Rule

requires_synonymous_and_no_significant_predicted_splice_impact

Evidence considered

• The variant is synonymous: NM_002528.7:c.900C>T, NP_002519.2:p.(Ala300=), exon 6 of NTHL1; BP7, not PP3/BP4, is the computational criterion in scope for it.
• SpliceAI scores for NM_002528.7:c.900C>T are DS_AG 0.003, DS_AL 0.004, DS_DG 0.002 and DS_DL 0.00, giving a maximum delta score of 0.004, with no predicted acceptor or donor gain or loss; the predicted event positions are DP_AL 123, DP_DL 188, DP_DG 194 and DP_AG 277 nucleotides from the variant.
• The SpliceAI recommended cutoff for significant splice impact is 0.2, with 0.1 as the no-impact level (Jaganathan et al. 2019, Cell, PMID:30661751); the observed maximum delta of 0.004 is far below both, satisfying BP7's no-predicted-splice-impact requirement.
• Pangolin values recorded for the same variant (splice gain 0.004, splice loss -0.008) are concordant with no predicted splice impact but are not calibrated for a benign splice call here, so they are not the basis for the verdict.
• No nucleotide-conservation metric (phyloP/GERP/phastCons) for c.900 is present in the case evidence; the 'not highly conserved' prong is supported indirectly by population tolerance: gnomAD v2.1 AF 0.0459% (125/272238 alleles, 3 homozygotes) and gnomAD v4.1 AF 0.0213% (341/1603718 alleles, 4 homozygotes).
• No ClinGen CSPEC and no NTHL1 gene-specific VCEP/BP7 rule or lookup table exists for this case (cspec found = false; framework_mode = generic_acmg), so the generic ACMG/AMP 2015 BP7 definition was applied.
• ACMG/AMP 2015 (PMID:25741868) defines BP7 for a synonymous variant for which splicing prediction algorithms predict no impact on the splice consensus sequence or creation of a new splice site and where the nucleotide is not highly conserved, and states that predictions from multiple in-silico tools are treated as a single piece of evidence - the reason the single SpliceAI result is scored once, under BP7.

Still missing

• No quantitative nucleotide-conservation metric (phyloP, GERP, phastCons or equivalent) for NTHL1 c.900 is available in this case's evidence, so BP7's 'nucleotide not highly conserved' prong rests on indirect population tolerance of the position rather than on a direct conservation score.
• The variant sits in the last coding exon (exon 6, c.792-*104) 15 nucleotides upstream of the stop codon (c.915); the case carries no RNA/transcript study for this variant, so the negative splice prediction is computational only.

Sources

Reasoning

Governing framework: no NTHL1 ClinGen CSPEC/VCEP specification exists for this case (framework source = generic ACMG/AMP 2015 fallback), so PS4 is assessed under the generic ACMG/AMP 2015 definition (PMID:25741868). PS4 requires that the prevalence of the variant in affected individuals be significantly increased compared with controls. The generic 2015 text states this requirement qualitatively and defines no numeric odds-ratio cutoff, and no NTHL1 VCEP supplies a gene-specific one, so there is no governing numeric threshold and no comparison operator (>=, <=, >, <) applicable to this criterion in this case - no inequality can be written because no value exists to compare. Variant-level review found no case-control enrichment evidence: NM_002528.7:c.900C>T (NP_002519.2:p.Ala300=) is not reported in any affected proband in the retrieved literature. The ACMG/NSGC cancer-predisposition referral practice guideline (PMID:25394175) contains no NTHL1 or c.900C>T content, and the ACMG/AMP 2015 framework paper (PMID:25741868) contains no variant-level data. Instead, the variant is present in unaffected population samples - gnomAD v4.1 AF 0.0002126 (341/1,603,718 alleles) with 4 homozygotes, and gnomAD v2.1 AF 0.0004592 (125/272,238 alleles) with 3 homozygotes - and ClinVar VCV000757993 records only benign-direction laboratory submissions, which is the opposite of case enrichment. PS4 is therefore not met.

Evidence considered

• ClinVar VCV000757993: exact match NM_002528.7(NTHL1):c.900C>T (p.Ala300=); 6 submissions, all single-submitter clinical laboratories or curation; 0 expert-panel submissions; no case cohort or control cohort reported.
• gnomAD v4.1: AF 0.0002126 (341/1,603,718 alleles), 4 homozygotes; popmax South Asian AF 0.0032941 (300/91,072 alleles) - these are population controls, not an affected cohort.
• gnomAD v2.1: AF 0.0004592 (125/272,238 alleles), 3 homozygotes; South Asian AF 0.0038942 (119/30,558 alleles).
• ACMG/NSGC cancer-predisposition referral practice guideline (PMID:25394175) screened for c.900C>T, p.Ala300=, and NTHL1: no variant or gene content, no case or case-control data.
• GeneReviews NTHL1 Tumor Syndrome (PMID:32239880) was flagged as a case-observation source (candidate criteria PS4, PP5) but no full text or abstract was retrievable in this case, so it could not be screened for this variant.
• No NTHL1 VCEP/ClinGen CSPEC specification is available for this case, so no gene-specific PS4 odds-ratio threshold or comparison operator exists; generic ACMG/AMP 2015 (PMID:25741868) states PS4 qualitatively.

Still missing

• No case-control or affected-cohort data for NTHL1 c.900C>T exist in the case materials or retrieved literature; such a comparison (or a VCEP-defined odds-ratio threshold, which no NTHL1 specification supplies) would be required to assess PS4.

Sources

Reasoning

PM2 at supporting strength requires the variant to be absent, or rarer than 0.0001 (0.01%), in a large general population (ClinGen SVI recommendation to apply PM2 at supporting strength; threshold as supplied for this group, attributed to Richards et al. 2015 / ClinGen SVI 2020, PMID:25741868). The variant is neither absent nor rare enough. gnomAD v4.1 all-comers: total AF 0.000212631 (341/1603718 alleles, 4 homozygotes) - about 2.1 times the threshold. gnomAD v2.1 all-comers: total AF 0.000459157 (125/272238 alleles, 3 homozygotes) - about 4.6 times the threshold. Every subset is above the threshold as well, including the rarest sources available: gnomAD v4.1 genomes AF 0.000124723 (19/152338), gnomAD v3.1 non-cancer genomes AF 0.000121659 (18/147954) and gnomAD-Canada 0.000325909 (6/18410). In the patient-relevant South Asian population the frequency is far higher, 0.0032941 (v4.1) and 0.00389423 (v2.1), roughly 33-39 times the 0.0001 threshold, and the variant is observed in the homozygous state in population data. PM2 is therefore not met, and no choice of source - all-comers, non-cancer, or exome-only - would change that. For a recessive gene the ACMG/AMP text also frames PM2 against a race-matched expected carrier frequency; the observed homozygotes and the South Asian frequency comfortably exceed any plausible expectation for this allele.

Rule

generic_acmg_population_frequency: PM2 (supporting) met if AF <= 0.0001 (ClinGen SVI 2020; PMID:25741868)

Evidence considered

• gnomAD v4.1 all-comers (gnomad_v4): total AF 0.000212631 (341/1603718 alleles, 4 homozygotes); exomes 0.000221858 (322/1451380); genomes 0.000124723 (19/152338); highest-population frequency South Asian AF 0.0032941 (300/91072, 4 homozygotes).
• gnomAD v2.1 all-comers (gnomad_v2): total AF 0.000459157 (125/272238 alleles, 3 homozygotes); exomes 0.000510704 (123/240844, 3 homozygotes); genomes 0.000063706 (2/31394); highest-population frequency South Asian AF 0.00389423 (119/30558, 3 homozygotes).
• Non-cancer and exome-only sources, inspected for completeness, are also above the threshold: gnomAD v2.1 non-cancer exomes AF 0.000538666 (123/228342) and gnomAD v3.1 non-cancer genomes AF 0.000121659 (18/147954); no governed specification mandates these sources, and neither would rescue PM2.
• Threshold source: ClinGen SVI recommendation to apply PM2 at supporting strength, with an allele-frequency cutoff of <= 0.0001 (0.01%) - the published value supplied for this group, attributed to Richards et al. 2015 / ClinGen SVI 2020 (PMID:25741868). No NTHL1 ClinGen CSPEC (cspec found=false) or VCEP specification exists to override it.

Sources

Reasoning

PM3 is the allele-placement criterion for recessive disease: the variant must be observed in trans (on the opposite chromosome) with a pathogenic variant in an affected individual. NTHL1 tumor syndrome (familial adenomatous polyposis 3) is an autosomal recessive, biallelic loss-of-function disorder, so PM3's recessive arm is applicable here: the NTHL1 disease mechanism is a homozygous/compound-heterozygous germline NTHL1 lesion (Weren et al. 2015, PMID:25938944, as recorded in this case's gene-level context pvs1_gene_context, whose disease model is stated as 'A germline homozygous mutation in the base-excision repair gene NTHL1 causes adenomatous polyposis and colorectal cancer'). Richards et al. 2015 (PMID:25741868) ties PM3 explicitly to phase, noting that the criterion 'could be upgraded to strong if there were multiple observations of detection of the variant in trans (on opposite chromosomes) with other pathogenic variants'. No such observation exists anywhere in this case's assembled record: there is no affected proband, no trio or family genotype data, no segregation data and no phase-resolved (cis/trans) information of any kind for NM_002528.7:c.900C>T. The only genotype-level observations available are population homozygote counts, and they point the other way - gnomAD v2.1 holds 3 homozygotes (125/272,238 alleles, AF 0.000459; homozygotes concentrated in the South Asian subset, 119/30,558, AF 0.0039) and gnomAD v4.1 holds 4 homozygotes (341/1,603,718 alleles, AF 0.000213; South Asian subset 300/91,072, AF 0.0033), i.e. this allele is carried on both chromosomes in reference-population samples rather than being reported as the second hit of a biallelic NTHL1 genotype. ClinVar (ClinVarID 757993), where the variant is recorded in a germline context under the recessive NTHL1 polyposis phenotype, contains six submissions from clinical laboratories (3 Likely benign, 2 Benign) and none reports a trans configuration with another pathogenic NTHL1 variant; the submissions do not expose per-criterion detail. PM3 is therefore not met on this record; the limitation is the absence of proband/phase data, not a contrary genotype.

Rule

PM3 (Richards et al. 2015, PMID:25741868): for recessive disorders, detected in trans with a pathogenic variant - requires phase-resolved genotype data.

Evidence considered

• NTHL1 tumor syndrome is an autosomal recessive, biallelic disorder: the gene-level disease model retrieved for this case (pvs1_gene_context) gives the disease name as 'A germline homozygous mutation in the base-excision repair gene NTHL1 causes adenomatous polyposis and colorectal cancer' and lists the NTHL1 germline literature, and the corresponding abstract in this case's literature packet (Weren et al. 2015, PMID:25938944) states seven affected individuals from three families carried a homozygous germline NTHL1 nonsense mutation and 'All three families showed recessive inheritance of the adenomatous polyposis phenotype'.
• Richards et al. 2015 (PMID:25741868): 'rule PM3 could be upgraded to strong if there were multiple observations of detection of the variant in trans (on opposite chromosomes) with other pathogenic variants' - confirming PM3 requires a phase-resolved trans observation. The retrieved full text of this PMID contains body text only (Tables 3-4 criterion text absent), so the PM3 definition is applied from the published ACMG/AMP 2015 framework.
• gnomAD v2.1: AF 0.000459 (125/272,238 alleles), 3 homozygotes, population maximum in South Asians (119/30,558 alleles, AF 0.0039, 3 homozygotes). gnomAD v4.1: AF 0.000213 (341/1,603,718 alleles), 4 homozygotes, population maximum in South Asians (300/91,072 alleles, AF 0.0033, 4 homozygotes). Homozygous carriers of this exact allele are present in population reference samples.
• ClinVar ClinVarID 757993 (NM_002528.7(NTHL1):c.900C>T, p.Ala300=): six germline submissions from clinical testing laboratories (GeneDx, Sema4, Ambry Genetics, Sinai Health System, Catalan Institute of Oncology, Labcorp/Invitae) - 3 Likely benign and 2 Benign, review status 'criteria provided, multiple submitters, no conflicts'; condition labels include 'Familial adenomatous polyposis 3' and 'Hereditary cancer-predisposing syndrome'. No submission reports the variant in trans or in cis with another NTHL1 variant, and no per-submission criterion lists are exposed.
• Targeted search of the assembled case evidence (case summary/compact evidence, prefetch, evidence results, ClinVar submission audit, literature pass, paper extracts) for proband, pedigree, segregation, phase (cis/trans) or zygosity evidence on this variant returned nothing: this case is a variant-level classification with no patient genotype attached.
• Not used as a criterion basis: aggregate ClinVar classification labels, which per policy cannot be used to infer PM3 or any other criterion.

Still missing

• No proband/affected-individual genotype, trio, family or segregation data are present in this case, so trans/cis phase for NTHL1 c.900C>T cannot be determined; PM3 must be re-evaluated if a phase-resolved observation of this variant in trans with a pathogenic NTHL1 allele becomes available.
• ClinVar submissions for this variant (ClinVarID 757993) do not expose per-submission criteria or allele-placement detail, so no trans observation can be extracted from them.

Sources

Reasoning

PP5 requires that the exact variant carry a ClinVar expert-panel (or equivalent recognized expert-laboratory) classification of Pathogenic or Likely pathogenic. ClinVar VCV000757993 is an exact variant match (NM_002528.7(NTHL1):c.900C>T, p.Ala300=; exact match confirmed on transcript HGVS), but the record contains zero expert-panel submissions: all six submissions (GeneDx, Sema4, Ambry Genetics, Department of Pathology and Laboratory Medicine/Sinai Health System, Molecular Diagnostics Laboratory/Catalan Institute of Oncology, and Labcorp Genetics/Invitae) are ordinary clinical-laboratory or curation submissions, each with 'criteria provided, single submitter' status, and the aggregate record is 2-star 'criteria provided, multiple submitters, no conflicts'. The aggregate ClinVar label is in any case Benign/Likely benign, not pathogenic. Under the governing rule, ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions must never trigger PP5, and no such assertion can substitute for an expert-panel classification. There is therefore no qualifying classification to cite and PP5 is not met. No numeric threshold applies to PP5.

Evidence considered

• ClinVar VCV000757993 exact match confirmed by transcript HGVS (exact_match true, match_score 16): NM_002528.7(NTHL1):c.900C>T (p.Ala300=).
• Submitter audit: 6 submissions - GeneDx (Likely benign), Sema4 (Likely benign), Ambry Genetics (Likely benign), Sinai Health System (Benign), Catalan Institute of Oncology (Likely benign), Labcorp/Invitae (Benign); is_expert_panel = false for every submission; expert-panel submission count = 0.
• Aggregate ClinVar record: Benign/Likely benign, review status 'criteria provided, multiple submitters, no conflicts' (2 stars), 6 submitters, date last updated 2026-06-06.
• Rule applied: PP5 requires an exact-variant expert-panel Pathogenic/Likely pathogenic classification; ordinary laboratory submissions and aggregate ClinVar labels cannot trigger it, and the aggregate direction here is benign regardless.

Sources

Reasoning

BA1 requires an allele frequency of at least 5% in any general continental population (Richards et al. 2015, PMID:25741868). No dataset for this variant comes close. gnomAD v2.1 all-comers: total AF 0.000459157 (125/272238 alleles, 3 homozygotes), exomes 0.000510704 (123/240844, 3 homozygotes), genomes 0.000063706 (2/31394, 0 homozygotes); the highest-population (popmax) frequency is South Asian AF 0.00389423 (119/30558 alleles, 3 homozygotes) with a grpmax FAF of 0.00332598. gnomAD v4.1 all-comers: total AF 0.000212631 (341/1603718 alleles, 4 homozygotes), exomes 0.000221858 (322/1451380, 4 homozygotes), genomes 0.000124723 (19/152338, 0 homozygotes); popmax South Asian AF 0.0032941 (300/91072 alleles, 4 homozygotes) with a grpmax FAF of 0.00298711. The non-cancer subsets carried alongside these totals are not higher (gnomAD v2.1 non-cancer exomes AF 0.000538666, grpmax FAF 0.00333492; gnomAD v3.1 non-cancer genomes AF 0.000121659, grpmax FAF 0.00224103). The maximum observed population frequency anywhere is therefore ~0.33-0.39% (South Asian), roughly 13-fold below the 5% stand-alone threshold. NTHL1 has no ClinGen CSPEC or VCEP specification (cspec found=false; framework_mode=generic_acmg), so the generic ACMG/AMP 2015 stand-alone threshold of 0.05 was applied, and no NTHL1-specific BA1 threshold exists that could lower the bar. BA1 is not met.

Rule

generic_acmg_population_frequency: BA1 met if AF >= 0.05 in any general continental population (PMID:25741868)

Evidence considered

• gnomAD v2.1 all-comers (gnomad_v2, https://gnomad.broadinstitute.org/variant/16-2089940-G-A?dataset=gnomad_r2_1): total AF 0.000459157 (125/272238 alleles, 3 homozygotes); exomes 0.000510704 (123/240844, 3 homozygotes); genomes 0.000063706 (2/31394, 0 homozygotes); highest-population frequency South Asian AF 0.00389423 (119/30558, 3 homozygotes); grpmax FAF 0.00332598.
• gnomAD v4.1 all-comers (gnomad_v4, https://gnomad.broadinstitute.org/variant/chr16-2039939-G-A?dataset=gnomad_r4): total AF 0.000212631 (341/1603718 alleles, 4 homozygotes); exomes 0.000221858 (322/1451380, 4 homozygotes); genomes 0.000124723 (19/152338, 0 homozygotes); highest-population frequency South Asian AF 0.0032941 (300/91072, 4 homozygotes); grpmax FAF 0.00298711.
• Non-cancer subsets supplied with the case do not raise the frequency: gnomAD v2.1 non-cancer exomes (reported on its exome-only figures) AF 0.000538666 (123/228342, 3 homozygotes; grpmax FAF 0.00333492; South Asian 0.00390471, 119/30476) and gnomAD v3.1 non-cancer genomes AF 0.000121659 (18/147954, 0 homozygotes; grpmax FAF 0.00224103). No NTHL1 VCEP/CSPEC specification exists to mandate a non-cancer or exome-only source, so the all-comers gnomAD v2.1/v4.1 entries remain the default source; the verdict is unchanged under either choice.
• Small-sample artifact noted and excluded: the top subpopulation in the gnomAD v3.1 non-cancer record is HGDP:SINDHI at AF 0.05, but this rests on 2 alternate alleles out of 40 (n=40) and is not a general continental population frequency; the corresponding grpmax FAF is 0.00224103 and was used instead for the BA1 comparison.
• Threshold source: ACMG/AMP 2015 stand-alone benign allele-frequency criterion BA1, >= 0.05 (Richards et al. 2015, PMID:25741868) - the generic published value supplied for this group; no NTHL1-specific BA1 calibration exists.

Sources

Reasoning

BS1 requires an allele frequency greater than that expected for the disorder (Richards et al. 2015, PMID:25741868). Because NTHL1 has no ClinGen CSPEC or VCEP specification and no gene-specific frequency framework is present in this case, the generic published operating threshold for this group was applied: BS1 strong if AF >= 0.01 (1%), the generic SVI default whose per-gene derivation is attributed to Whiffin et al. 2017 (PMID:28518168). Nothing approaches 1%. The maximum frequencies are grpmax FAF 0.00332598 in gnomAD v2.1 (South Asian popmax AF 0.00389423, 119/30558 alleles) and grpmax FAF 0.00298711 in gnomAD v4.1 (South Asian popmax AF 0.0032941, 300/91072 alleles); total all-comers frequencies are 0.000459157 (v2.1) and 0.000212631 (v4.1). The non-cancer subsets are lower or equal: gnomAD v2.1 non-cancer exomes total AF 0.000538666 with grpmax FAF 0.00333492, and gnomAD v3.1 non-cancer genomes total AF 0.000121659 with grpmax FAF 0.00224103. The highest population frequency is thus ~0.33-0.39%, roughly 2.5-fold below the 1% threshold, so BS1 is not met. It should be flagged that a disease-prevalence-derived (Whiffin-style) threshold for an ultra-rare recessive syndrome such as NTHL1 tumor syndrome could plausibly sit nearer the observed South Asian frequency than the generic 1% does; that derivation was not possible here because no retrievable source in this case supplies NTHL1 tumor syndrome prevalence/penetrance (the GeneReviews chapter, PMID:32239880, was not obtained as full text), so the generic 1% default was applied and BS1 is not met on available evidence.

Rule

generic_acmg_population_frequency: BS1 met if AF >= 0.01 (generic SVI default; Whiffin et al. 2017 PMID:28518168)

Evidence considered

• gnomAD v2.1 all-comers: total AF 0.000459157 (125/272238 alleles, 3 homozygotes); highest-population frequency South Asian AF 0.00389423 (119/30558, 3 homozygotes); grpmax FAF 0.00332598.
• gnomAD v4.1 all-comers: total AF 0.000212631 (341/1603718 alleles, 4 homozygotes); highest-population frequency South Asian AF 0.0032941 (300/91072, 4 homozygotes); grpmax FAF 0.00298711.
• gnomAD v2.1 non-cancer exomes: AF 0.000538666 (123/228342, 3 homozygotes), grpmax FAF 0.00333492, South Asian 0.00390471. gnomAD v3.1 non-cancer genomes: AF 0.000121659 (18/147954, 0 homozygotes), grpmax FAF 0.00224103. These non-cancer/exome-only sources were inspected but are not mandated by any NTHL1 specification; they are equal to or lower than the all-comers values and change no verdict.
• Threshold source: generic SVI default BS1 strong if AF >= 0.01, with per-gene derivation preferred (Whiffin et al. 2017, PMID:28518168) - the published value supplied for this group. No NTHL1 ClinGen CSPEC (cspec found=false) or VCEP specification exists to override it, and no gene-specific threshold was derivable (no retrievable prevalence/penetrance source).
• Variant consequence considered: NM_002528.7:c.900C>T is a synonymous substitution, NP_002519.2:p.(Ala300=) (MANE Select transcript), so no protein-level constraint alters the reading of its population frequency.

Still missing

• A gene-specific BS1 threshold derived from NTHL1 tumor syndrome prevalence and penetrance (Whiffin et al. 2017, PMID:28518168 approach) could not be computed: the GeneReviews chapter 'NTHL1 Tumor Syndrome' (PMID:32239880) full text was not retrieved in this case (screenshot only, no machine-readable text) and no other retrieved source provides prevalence or penetrance figures. Because the variant sits at ~0.33-0.39% in South Asians, a lower gene-specific threshold could change this verdict; the generic 1% default was therefore applied and BS1 recorded as not met.

Sources

Reasoning

BS2 (strong benign; Richards et al. 2015, PMID:25741868) requires that the variant be observed in a well-documented healthy adult individual - for a recessive condition, in the homozygous state - and that the disease under investigation be fully penetrant at an early age. The variant is genuinely present in homozygous form in population data: gnomAD v2.1 total 3 homozygotes (all within the South Asian subset, 119/30558 alleles), gnomAD v4.1 total 4 homozygotes (South Asian 300/91072 alleles), and gnomAD v2.1 non-cancer exomes 3 homozygotes; gnomAD v3.1 non-cancer genomes shows 0. Both preconditions nevertheless fail for this gene. First, gnomAD participants are not 'well-documented healthy adults' in the sense BS2 requires: the resource documents only the absence of severe paediatric disease and performs no phenotyping for adult-onset neoplasia, so a homozygote in gnomAD is not documented as free of adenomatous polyposis or colorectal cancer - the ACMG/AMP text itself cautions that subclinical disease is always possible when a general population is used as a control cohort. Second, NTHL1 tumor syndrome (biallelic NTHL1 loss of function; ClinVar condition 'Familial adenomatous polyposis 3' / hereditary cancer-predisposing syndrome) is an adult-onset, incompletely penetrant cancer-predisposition phenotype, not a fully penetrant early-onset disorder, so homozygotes in an adult population database are compatible with the diagnosis rather than evidence against it - the same reasoning that keeps recurrent homozygous MUTYH alleles pathogenic. Because neither BS2 clause is satisfied, the observed homozygotes do not constitute strong benign evidence and BS2 is not met. Human review is recommended: BS2's applicability to recessive adult-onset reduced-penetrance syndromes is a recognised interpretive divergence, and some frameworks would apply BS2 at reduced strength in exactly this situation; no NTHL1-specific specification exists to settle it.

Rule

generic_acmg_BS2: requires a well-documented healthy adult homozygote (for a recessive condition) AND a fully penetrant early-onset disease (PMID:25741868)

Evidence considered

• Homozygote observations (gnomad_v2): gnomAD v2.1 total 3 homozygotes, all in the South Asian subset (AF 0.00389423, 119/30558 alleles, 3 homozygotes); gnomAD v2.1 non-cancer exomes 3 homozygotes (123/228342 alleles).
• Homozygote observations (gnomad_v4): gnomAD v4.1 total 4 homozygotes, all in the South Asian subset (AF 0.0032941, 300/91072 alleles, 4 homozygotes); gnomAD v3.1 non-cancer genomes 0 homozygotes (18/147954 alleles).
• BS2 definition used: ACMG/AMP 2015 (PMID:25741868) states BS2 applies where the disease under investigation is fully penetrant at an early age and the variant is observed in a well-documented healthy adult individual for a recessive (homozygous), dominant (heterozygous), or X-linked (hemizygous) condition.
• Disease context: NTHL1 tumor syndrome is an autosomal recessive, adult-onset (adenomatous polyposis and colorectal/other tumours typically presenting in adulthood) and incompletely penetrant cancer-predisposition phenotype; gnomAD performs no phenotype documentation for adult-onset neoplasia, so its homozygotes do not meet the 'well-documented healthy adult' requirement.
• Variant consequence: NM_002528.7:c.900C>T is synonymous, NP_002519.2:p.(Ala300=); no protein-level effect is expected, but this does not by itself satisfy BS2, whose criteria are population/phenotype based.

Still missing

• BS2's applicability here is a judgment call rather than a data gap: the ACMG/AMP definition requires a fully penetrant early-onset disease plus a well-documented healthy adult, while 3-4 homozygotes are in fact present in population data. Some frameworks apply BS2 at reduced (supporting) strength for recessive adult-onset conditions; no NTHL1-specific specification exists to resolve this, so human review is recommended.
• NTHL1 tumor syndrome penetrance and age-of-onset figures were not retrievable as full text in this case (GeneReviews chapter PMID:32239880 exists only as a screenshot), so the disease-characteristic half of this assessment rests on the established adult-onset, reduced-penetrance nature of the phenotype rather than a citable numeric penetrance estimate.

Sources

Reasoning

BP2 is a phase criterion: it is met when the variant is observed in cis with a pathogenic variant in any inheritance pattern, or observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder (Richards et al. 2015, PMID:25741868). Neither arm can be satisfied here. The trans-with-a-pathogenic-variant arm applies to fully penetrant dominant disease, and NTHL1 tumor syndrome (familial adenomatous polyposis 3) is an autosomal recessive, biallelic loss-of-function disorder - the NTHL1 disease mechanism is a homozygous (biallelic) germline loss-of-function lesion (Weren et al. 2015, PMID:25938944, as recorded in this case's gene context pvs1_gene_context) - so a trans configuration with a pathogenic NTHL1 allele would be the expected configuration of disease rather than benign evidence, and this arm does not apply. The cis arm requires that the variant sit on the same chromosome as a known pathogenic variant, which demands phase-resolved genotype data. This case has no proband, no trio/family or segregation data and no cis/trans information for NM_002528.7:c.900C>T: no source in the assembled record places it on the same allele as any other NTHL1 variant. The available genotype-level data do not substitute for that. The gnomAD homozygote counts (3 homozygotes in v2.1; 4 in v4.1) show the allele borne on both chromosomes in population samples, which is not a cis observation relative to a pathogenic allele and belongs to the BS2/BS1 population argument rather than BP2; and the ClinVar submissions (ClinVarID 757993, six germline submissions, no per-submission criteria exposed) contain no allele-placement detail. BP2 is therefore not met on this record.

Rule

BP2 (Richards et al. 2015, PMID:25741868): 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.

Evidence considered

• Richards et al. 2015 (PMID:25741868) 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.' The retrieved full text of this PMID contains the body discussion only (no Tables 3-4 criterion text), so the definition is applied from the published ACMG/AMP 2015 framework.
• NTHL1 tumor syndrome is an autosomal recessive disorder with a biallelic loss-of-function mechanism (homozygous germline NTHL1 mutation; Weren et al. 2015, PMID:25938944, as recorded in the case's pvs1_gene_context disease model), so BP2's fully penetrant dominant-disorder arm does not apply to this gene/disease.
• ClinVar ClinVarID 757993: six germline submissions from clinical testing laboratories (3 Likely benign, 2 Benign; 'criteria provided, multiple submitters, no conflicts'), none of which reports the variant in cis with a pathogenic variant or exposes per-submission allele-placement detail; condition labels include 'Familial adenomatous polyposis 3'.
• Search of the assembled case record for pedigree, trio, segregation or phase (cis/trans) information on this variant found none; the gnomAD homozygote observations (3 in v2.1, 4 in v4.1) are homozygous carriers in population samples, not a cis configuration with a pathogenic NTHL1 allele.

Still missing

• No phase-resolved (cis/trans) genotype data exist for this variant in this case; BP2 cannot be met or excluded further without a proband genotype showing c.900C>T on the same allele (cis) as a pathogenic NTHL1 variant.
• PMID:28492532 (Sherloc), the reference cited by the Labcorp/Invitae submission for this variant, has no full text or pre-computed extraction in this case, so any Sherloc-specific BP2 sub-usage could not be verified; it would still require the phase data the case lacks.

Sources

Reasoning

BP6 requires that the exact variant carry a ClinVar expert-panel (or equivalent recognized expert-laboratory) classification of Benign or Likely benign. ClinVar VCV000757993 matches the variant exactly (NM_002528.7(NTHL1):c.900C>T, p.Ala300=) and its aggregate label is Benign/Likely benign (2 stars, 'criteria provided, multiple submitters, no conflicts'; three laboratories submit Likely benign and two submit Benign), but the record contains zero expert-panel submissions - every submission is from an ordinary clinical laboratory or curation group with 'criteria provided, single submitter' status. The governing rule explicitly excludes ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions from triggering BP6, so despite the benign aggregate label there is no expert-panel Benign/Likely benign classification on which BP6 could rest. BP6 is therefore not met. No numeric threshold applies to BP6.

Evidence considered

• ClinVar VCV000757993 exact match confirmed by transcript HGVS: NM_002528.7(NTHL1):c.900C>T (p.Ala300=).
• Aggregate record: Benign/Likely benign, review status 'criteria provided, multiple submitters, no conflicts' (2 stars); submissions summarised as Likely benign (3 clinical laboratories) and Benign (2 clinical laboratories).
• Submitter audit: all 6 submissions are single-submitter clinical-laboratory or curation entries with is_expert_panel = false; expert-panel submission count = 0.
• Rule applied: BP6 requires an exact-variant expert-panel Benign/Likely benign classification; aggregate ClinVar labels and ordinary laboratory submissions must never trigger it.

Sources

Reasoning

PS2 (Richards et al. 2015, PMID:25741868) requires a de novo occurrence of the variant, with both parental samples shown by identity testing to be the biological parents, in a patient whose phenotype matches the gene's disease association and whose family history is consistent with de novo inheritance. NM_002528.7:c.900C>T produces NP_002519.2:p.(Ala300=), a synonymous substitution in NTHL1. The consolidated case evidence contains no proband genotype, no parental or trio testing, no family history and no phenotype record: the evidence set comprises variant-level annotations only (ClinVar variation 757993, gnomAD v2.1 and v4.1 including the non-cancer subsets, gnomAD-Canada, SpliceAI, OncoKB, COSMIC, Cancer Hotspots), plus four literature records, none of which reports a de novo occurrence of this variant. Four of the six ClinVar submissions are classified Likely benign and two Benign, but no submission carries proband, pedigree, parental-testing or de novo information, and a submitter's classification cannot be used to infer this criterion. No ClinGen CSPEC/VCEP or local custom framework exists for NTHL1 (final_classification_framework reports generic_acmg_fallback with framework_complete=false), so generic ACMG/AMP 2015 criteria govern. PS2 therefore cannot be resolved as met or not met and is not_assessed; the defining gap is the absence of a phenotyped proband with parental testing.

Evidence considered

• Richards et al. 2015 (PMID:25741868) defines PS2 as strong support: 'A variant observed to have arisen de novo (parental samples test negative), is considered strong support for pathogenicity if the following conditions are met: Both parental samples were shown through identity testing to be the biological parents of the patient. Note: PM6 applies if identity is assumed but not confirmed. The patient has a family history of disease that is consistent with de novo inheritance (e.g. unaffected parents for a dominant disorder)... The phenotype in the patient matches the gene's disease association with reasonable specificity.' The paper contains no NTHL1, c.900C>T or p.(Ala300=) content, so it supplies the criterion definition only, not variant evidence.
• The case evidence carries variant-level annotations only: ClinVar variation 757993 (six germline submissions: GeneDx, Sema4, Ambry Genetics and Catalan Institute of Oncology as Likely benign; Sinai Health System and Labcorp/Invitae as Benign), gnomAD v2.1 and v4.1 plus the non-cancer subsets, gnomAD-Canada, SpliceAI, OncoKB, COSMIC and Cancer Hotspots. None of these records contains a proband, parental genotypes, a pedigree, or a de novo observation.
• Literature screened for this case: PMID:25394175 (ACMG/NSGC cancer-predisposition referral guideline) and PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) mention neither NTHL1-associated data for this variant nor any de novo or family observation of c.900C>T; no other full text relevant to a de novo claim was available.
• No NTHL1-specific ClinGen CSPEC or VCEP framework and no local custom gene framework were found, so the generic ACMG/AMP 2015 criteria and combining rules were applied (final_classification_framework: source generic_acmg_fallback).

Still missing

• No proband-level genotype and no parental (trio) testing for c.900C>T; PS2 requires a de novo occurrence with maternity and paternity confirmed by identity testing.
• No phenotype record and no family history for the individual in whom this variant was found, so phenotype specificity and consistency with de novo inheritance cannot be judged.
• No de novo occurrence of c.900C>T is reported in any source retrieved for this case (ClinVar submissions, gnomAD, COSMIC, OncoKB, Cancer Hotspots, or the four literature records).

Sources

Reasoning

PS3 requires a well-established functional study demonstrating a deleterious effect on gene or protein function (Richards et al. 2015, PMID:25741868). NM_002528.7:c.900C>T in NTHL1 is a synonymous substitution predicted to encode NP_002519.2:p.(Ala300=), so no amino-acid change is produced and the functional question reduces to whether the substitution creates a splicing or other DNA/RNA-level defect. The only functional-adjacent evidence in this case is computational: SpliceAI max delta 0.004 (DS_AG 0.003, DS_AL 0.004, DS_DG 0.002, DS_DL 0.0; Pangolin SG 0.004, SL -0.008) and REVEL/BayesDel/HCI prior not available for this synonymous variant. Computational splice prediction is in-silico evidence (the PP3/BP4 domain) and cannot substitute for a validated assay, so it does not support PS3. No experimental assay of any kind — RNA/minigene splicing assay, NTHL1 DNA glycosylase activity, AP-lyase activity, or cellular oxidative-damage repair readout — is reported for this variant in any source searched for this case: ClinVar VCV000757993 (2-star, 6 submissions) reports no functional assay and cites only framework papers, OncoKB returns Unknown Oncogenic Effect with no variant-specific biological effect, and the ClinVar-linked papers reviewed (PMID:25741868 ACMG/AMP guideline, PMID:25394175 cancer-referral practice guideline, PMID:28492532 Sherloc framework refinement) contain no variant-level functional data; PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) has no retrieved full text in this case. No ClinGen VCEP or gene-specific specification exists for NTHL1 (cspec not found; no gene framework), so generic ACMG/AMP 2015 applies and no gene-specific PS3 strength calibration is available. With no assay results reported in either direction, PS3 is not evaluable and is recorded as not assessed rather than not met.

Evidence considered

• SpliceAI Lookup for NM_002528.7:c.900C>T returned max delta score 0.004 (DS_AG 0.003, DS_AL 0.004, DS_DG 0.002, DS_DL 0.0; Pangolin SG 0.004, SL -0.008; consequence sequence_variant) — computational splice prediction, which is in-silico PP3/BP4 evidence and is not an experimental functional assay.
• ClinVar VCV000757993 (NM_002528.7(NTHL1):c.900C>T p.Ala300=) is Benign/Likely benign, 2-star review status, 6 submissions; no submitter reports a functional assay, and the two criterion-lead submissions cite only framework/guideline PMIDs (25741868 for SCV005920569, 28492532 for SCV001081415).
• OncoKB returned Unknown Oncogenic Effect with no variant-specific biological effect and no citations PMIDs for the NTHL1 A300= query; the gene-level entry only notes NTHL1's base-excision-repair role and germline truncating variants in hereditary colorectal tumors, which is gene-level background and not a variant assay.
• Somatic COSMIC search for NTHL1 c.900C>T returned no result; absence of a somatic record is not functional assay evidence for either PS3 or BS3.
• REVEL, BayesDel and HCI prior were all unavailable for this variant, so no protein-level in-silico functional proxy exists either.
• Reviewed PMID:25741868 (ACMG/AMP 2015 sequence-variant interpretation standards): defines PS3 as well-established functional studies showing a deleterious effect, but names no NTHL1 variant and contains no functional data for this variant.
• Reviewed PMID:25394175 (ACMG/NSGC cancer-predisposition referral practice guideline): no mention of NTHL1 or of c.900C>T/p.(Ala300=) in text, tables or syndrome summaries.
• Reviewed PMID:28492532 (Sherloc ACMG-AMP criteria refinement): a framework/methodology paper reporting aggregate variant classification without variant-level functional data for NTHL1 c.900C>T.
• PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) had no retrievable full text in this case, so no functional-assay content could be extracted.
• Assessment basis: generic ACMG/AMP 2015 (PMID:25741868) fallback, because no ClinGen VCEP/CSPEC or gene-specific framework for NTHL1 exists (cspec found = false; framework_complete = false). PS3 status was not inferred from ClinVar's benign classification or from any other source's final classification.

Still missing

• No functional assay has been performed or reported for NTHL1 c.900C>T (p.Ala300=): no RNA/minigene splice assay, no NTHL1 DNA glycosylase or AP-lyase activity assay, and no cellular oxidative-damage repair readout; PS3 is unevaluable without one.
• No ClinGen VCEP/CSPEC or gene-specific specification for NTHL1 was available (cspec not found; no gene_vcep_dir), so no gene-specific functional-assay framework or PS3 strength calibration exists; generic ACMG/AMP 2015 rules were applied.
• Full text of PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) was not retrieved in this case (screenshot only), so any functional or case-level content in that review could not be inspected for a PS3-relevant assay.

Sources

Reasoning

PM6 (Richards et al. 2015, PMID:25741868) applies when a variant is observed to have arisen de novo (parental samples test negative) but the biological relationship of the parents has not been confirmed by identity testing; it carries moderate weight and still requires an actual de novo observation plus a phenotype that matches the gene's disease association. NM_002528.7:c.900C>T (NP_002519.2:p.(Ala300=)) is a synonymous NTHL1 substitution. The case evidence contains no proband record, no parental testing results and no phenotype description: it consists of variant-level database annotations (ClinVar variation 757993; gnomAD v2.1 and v4.1 and non-cancer subsets; gnomAD-Canada; SpliceAI; OncoKB; COSMIC; Cancer Hotspots) and four literature records, none reporting a de novo occurrence of this variant. ClinVar submissions report classifications only and include no assumed-de-novo evidence, and another source's classification cannot be used to infer this criterion. No NTHL1 CSPEC/VCEP or local custom framework exists, so generic ACMG/AMP 2015 criteria apply. PM6 is therefore not_assessed; without any de novo observation, with or without confirmed parentage, the criterion is unresolvable.

Evidence considered

• Richards et al. 2015 (PMID:25741868) defines the de novo criteria jointly: 'A variant observed to have arisen de novo (parental samples test negative), is considered strong support for pathogenicity if the following conditions are met: Both parental samples were shown through identity testing to be the biological parents of the patient. Note: PM6 applies if identity is assumed but not confirmed.' PM6 is the moderate-weight version of that observation and still presupposes a de novo event.
• No de novo report exists for NM_002528.7:c.900C>T: ClinVar variation 757993 (six germline submissions, four Likely benign and two Benign) carries no proband, parental-testing or de novo data, and the gnomAD v2.1 record (125/272,238 alleles, 3 homozygotes) contains no transmission information.
• PMID:25394175 (ACMG/NSGC cancer referral guideline) does not mention NTHL1 or this variant at any point; PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) was retrieved only as a page image with no searchable text, so no de novo observation could be drawn from either source.
• NTHL1 has no ClinGen CSPEC/VCEP specification and no local custom framework, so the generic ACMG/AMP 2015 framework was applied for this variant.

Still missing

• No reported de novo occurrence of c.900C>T, whether parentage-confirmed or assumed; PM6 requires the variant to be absent from parental samples.
• No parental sample results of any kind, so even the unconfirmed-parentage version of the criterion cannot be evaluated.
• No phenotype for the proband to test against the NTHL1 disease association (biallelic NTHL1 tumor syndrome / adenomatous polyposis, autosomal recessive).

Sources

Reasoning

PP1 (supporting; Richards et al. 2015, PMID:25741868) requires co-segregation of the variant with disease in multiple affected family members, which presupposes a pedigree in which at least two affected relatives carrying the variant have been genotyped. For NM_002528.7:c.900C>T (NP_002519.2:p.(Ala300=)) in NTHL1, the case evidence contains no pedigree, no affected relatives, no genotype data from relatives and no meioses count: the retrieved evidence is entirely variant-level (ClinVar variation 757993; gnomAD v2.1/v4.1 and non-cancer subsets; gnomAD-Canada; SpliceAI; OncoKB; COSMIC; Cancer Hotspots) together with four literature records, none of which reports family members carrying this variant. A submitter's benign classification in ClinVar cannot be used to infer this criterion, and the case contains no segregation table of any kind. No NTHL1 CSPEC/VCEP or local custom framework exists, so generic ACMG/AMP 2015 criteria apply. PP1 is not_assessed: neither co-segregation nor its absence can be demonstrated without family-level genotype data.

Evidence considered

• Richards et al. 2015 (PMID:25741868) weights PP1 as a supporting pathogenic criterion (co-segregation with disease in multiple affected family members) within the pathogenic criteria set, stating that 'Each pathogenic criterion is weighted as very strong (PVS1), strong (PS1-4); moderate (PM1-6), or supporting (PP1-5)'; the PP1 definitional text sits in Table 3 of that paper, whose cell text is not present in the retrieved full text.
• The case evidence contains no pedigree or family genotype data: ClinVar variation 757993 provides six submitter classifications (GeneDx, Sema4, Ambry Genetics, Catalan Institute of Oncology as Likely benign; Sinai Health System and Labcorp/Invitae as Benign) with no relative genotypes, and none of the other retrieved sources (gnomAD v2.1/v4.1, gnomAD non-cancer subsets, gnomAD-Canada, SpliceAI, OncoKB, COSMIC, Cancer Hotspots) reports family data.
• Literature screened for this case provided no NTHL1 c.900C>T family data: PMID:25394175 (ACMG/NSGC cancer referral guideline) does not mention NTHL1 or the variant, and PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) was available only as an image with no text extraction.
• No NTHL1 CSPEC/VCEP specification or local custom framework exists, so the generic ACMG/AMP 2015 framework and combining rules apply.

Still missing

• No pedigree and no genotype data from any relative of the individual carrying c.900C>T; PP1 requires co-segregation in multiple affected family members.
• No documented phenotype for any carrier, so co-segregation of variant with disease cannot be tested.
• No number of informative meioses available in the case evidence.

Sources

Reasoning

PP4 requires that the patient's phenotype be highly specific for a disease with a single genetic etiology. No proband-level clinical information accompanies this case: the case input is the variant alone, with no HPO terms, documented clinical features, age at onset, or family history, and nothing in the case files supplies a patient phenotype to test for specificity. The only phenotype context available is the ClinVar VCV000757993 condition set (Familial adenomatous polyposis 3; Hereditary cancer-predisposing syndrome; not provided), which reflects submitter-selected testing indications rather than a documented patient phenotype, and 'hereditary cancer-predisposing syndrome' is deliberately non-specific and would not satisfy PP4's single-etiology specificity requirement even if it were the proband's own phenotype. The variant itself is synonymous (NP_002519.2:p.Ala300=) with only benign-direction laboratory classifications, giving no phenotype-specificity signal either way. With no patient phenotype to test, PP4 cannot be assessed. No numeric threshold applies to PP4.

Evidence considered

• Case input is the variant alone; no HPO terms, clinical features, age of onset, or family history are present in the case materials (case summary and normalization records), so no patient phenotype could be tested for specificity.
• ClinVar VCV000757993 condition labels: Familial adenomatous polyposis 3; Hereditary cancer-predisposing syndrome; not provided - these are submitter-selected testing indications, and 'hereditary cancer-predisposing syndrome' is non-specific for a single-gene etiology.
• Variant is synonymous NP_002519.2:p.(Ala300=) with only benign-direction laboratory classifications, providing no phenotype-specificity signal.

Still missing

• No proband phenotype is available for this case (no HPO terms, documented clinical features, age at onset, family history, or detailed testing indication), so phenotype specificity for a single-gene disease cannot be evaluated.

Sources

Reasoning

BS3 requires a well-established functional study demonstrating no damaging effect on gene or protein function (Richards et al. 2015, PMID:25741868). NM_002528.7:c.900C>T in NTHL1 is a synonymous substitution predicted to encode NP_002519.2:p.(Ala300=), and no experimental assay — RNA/minigene splicing, NTHL1 DNA glycosylase or AP-lyase activity, or cellular repair readout — has been performed or reported for this variant in any source searched for this case. What is available is computational: SpliceAI max delta 0.004 (Pangolin SG 0.004, SL -0.008), which predicts no splice impact, and REVEL/BayesDel/HCI prior not available. A computational splice prediction, even a strongly negative one, is in-silico evidence belonging to the PP3/BP4 framework and cannot be upgraded into an assay-based BS3. Similarly, the population and database findings in this case are not functional evidence: gnomAD v2.1 MAF 0.0459% (125/272238 alleles, 3 homozygotes; South Asian MAF 0.389%, 3 homozygotes) and gnomAD v4.1 MAF 0.0213% (341/1603718 alleles, 4 homozygotes) bear on BA1/BS1/BS2/PM2, and ClinVar's 2-star Benign/Likely benign classification is explicitly not to be used to infer BS3. No NTHL1 ClinGen VCEP/CSPEC exists (cspec not found; no gene framework), so generic ACMG/AMP 2015 applies with no gene-specific BS3 strength calibration. BS3 is therefore not evaluable and is recorded as not assessed rather than not met.

Evidence considered

• SpliceAI Lookup for NM_002528.7:c.900C>T returned max delta score 0.004 (DS_AG 0.003, DS_AL 0.004, DS_DG 0.002, DS_DL 0.0; Pangolin SG 0.004, SL -0.008) — a computational no-splice-impact prediction, which is in-silico PP3/BP4 evidence and is not an experimental functional assay capable of supporting BS3.
• No experimental functional data for this variant were found in ClinVar VCV000757993 (Benign/Likely benign, 2-star, 6 submissions); submitter criteria cite framework papers only, and ClinVar's own classification was not used to infer BS3.
• OncoKB returned Unknown Oncogenic Effect with no variant-specific biological effect for NTHL1 A300= and no citations, so it provides no assay evidence on the benign side either.
• COSMIC returned no result for NTHL1 c.900C>T; absence of a somatic record is not a functional assay and does not support BS3.
• REVEL, BayesDel and HCI prior were unavailable for this variant; no protein-level in-silico proxy for a non-damaging effect exists.
• Population data in this case (gnomAD v2.1 MAF 0.0459%, 125/272238 alleles, 3 homozygotes, South Asian MAF 0.389% with 3 homozygotes; gnomAD v4.1 MAF 0.0213%, 341/1603718 alleles, 4 homozygotes; absent from gnomAD-Canada) are frequency evidence relevant to BA1/BS1/BS2/PM2, not functional assay evidence, and were not used to support BS3.
• Reviewed PMID:25741868 (ACMG/AMP 2015 standards): defines BS3 as well-established functional studies showing no damaging effect, but contains no NTHL1 variant-level data.
• Reviewed PMID:25394175 (ACMG/NSGC cancer-referral practice guideline): no mention of NTHL1 or of c.900C>T/p.(Ala300=).
• Reviewed PMID:28492532 (Sherloc ACMG-AMP criteria refinement): framework-level refinements only; no variant-level functional data for NTHL1 c.900C>T.
• PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) had no retrievable full text in this case, so no functional content could be extracted.
• Assessment basis: generic ACMG/AMP 2015 (PMID:25741868) fallback, because no ClinGen VCEP/CSPEC or gene-specific framework for NTHL1 exists (cspec found = false; framework_complete = false).

Still missing

• No functional assay has been performed or reported for NTHL1 c.900C>T (p.Ala300=) in either direction, so the absence of a damaging effect cannot be demonstrated experimentally; BS3 is unevaluable without such an assay.
• No ClinGen VCEP/CSPEC or gene-specific specification for NTHL1 was available (cspec not found; no gene_vcep_dir), so no gene-specific BS3 framework or strength calibration exists; generic ACMG/AMP 2015 rules were applied.
• Full text of PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) was not retrieved in this case (screenshot only); it could not be checked for assay or functional data bearing on BS3.

Sources

Reasoning

BS4 (supporting; Richards et al. 2015, PMID:25741868) is the benign counterpart of PP1: it requires demonstrated lack of segregation of the variant with disease among affected members of a family, i.e. a family with at least two affected individuals genotyped for the variant. NM_002528.7:c.900C>T produces NP_002519.2:p.(Ala300=), a synonymous NTHL1 substitution. The case evidence contains no pedigree and no relative genotypes, so no non-segregation can be observed or excluded: the retrieved material is limited to variant-level annotations (ClinVar variation 757993; gnomAD v2.1/v4.1 including non-cancer subsets; gnomAD-Canada; SpliceAI; OncoKB; COSMIC; Cancer Hotspots) and four literature records, none of which reports affected relatives with or without this variant. Population observations such as homozygotes in gnomAD are frequency evidence (a different criterion) and say nothing about segregation in a family. No NTHL1 CSPEC/VCEP or local custom framework exists, so generic ACMG/AMP 2015 criteria apply. BS4 is therefore not_assessed rather than met or not met.

Evidence considered

• Richards et al. 2015 (PMID:25741868) weights BS4 as a supporting benign criterion within the benign set, stating that 'each benign criterion is weighted as stand-alone (BA1), strong (BS1-4) or supporting (BP1-6)'; the BS4 definitional text (lack of segregation in affected members of a family) sits in Table 4 of that paper, whose cell text is not present in the retrieved full text.
• No segregation data are present in the case: ClinVar variation 757993 reports six submitter classifications with conditions listed as Familial adenomatous polyposis 3, Hereditary cancer-predisposing syndrome and 'not provided', but carries no family genotypes; gnomAD v2.1 (125/272,238 alleles, 3 homozygotes, grpmax FAF 0.00333 in South Asians) and the remaining database sources are frequency/predictor data only and address other criteria.
• PMID:25394175 (ACMG/NSGC cancer-predisposition referral guideline) was checked end to end and does not mention NTHL1, this variant, or any family segregation data; PMID:32239880 (NTHL1 Tumor Syndrome, GeneReviews) exists in this case only as a page image with no text available for searching.
• No NTHL1 CSPEC/VCEP specification and no local custom framework were available, so the generic ACMG/AMP 2015 criteria and combining rules were applied.

Still missing

• No family with two or more affected members genotyped for c.900C>T; BS4 requires a demonstrated lack of segregation with disease.
• No non-segregation observation for this variant in any retrieved source, and no pedigree to test it against.
• No phenotype or inheritance pattern documented for the case, so the disease model against which segregation would be assessed is undefined.

Sources

Reasoning

BP5 requires that the variant be observed in an affected individual whose phenotype is explained by an alternate molecular basis (for example a pathogenic variant in a different gene). No case-level observation of this variant exists in the case materials: the case input is the variant alone, with no proband genotype, no co-occurring variant information (for example APC or MUTYH findings in the tested individual), and no clinical note that would establish a second causal explanation. The retrieved literature does not supply such a case either - the ACMG/NSGC cancer-predisposition referral practice guideline (PMID:25394175) contains no NTHL1 or c.900C>T content, and text for the GeneReviews NTHL1 Tumor Syndrome entry (PMID:32239880) could not be retrieved for screening. Because BP5's structural prerequisite (an affected case with an alternate molecular basis for disease) is not documented at all, the criterion cannot be assessed. No numeric threshold applies to BP5.

Evidence considered

• Case materials contain no proband-level genotype, co-occurring variant list, or clinical interpretation; the case input is NM_002528.7:c.900C>T alone, so no alternate molecular basis for disease can be established.
• PMID:25394175 screened for c.900C>T, p.Ala300=, and NTHL1: no variant or gene content, therefore no case with an alternate molecular basis.
• PMID:32239880 (GeneReviews, NTHL1 Tumor Syndrome) full text was not retrievable in this case, so a case observation of this variant could be neither confirmed nor excluded from that source.
• No NTHL1 VCEP/ClinGen CSPEC specification exists for this case; the generic ACMG/AMP 2015 (PMID:25741868) BP5 definition was used.

Still missing

• No proband-level variant list/genotype or interpretation of any co-occurring findings is available, so an alternate molecular basis for disease cannot be established or excluded.

Sources