RAD50 c.1110A>C, p.(Leu370Phe)

Reasoning

PM2 is met: the variant is absent from large population cohorts, which is below the non-VCEP PM2 cutoff (<0.1%). NM_005732.3:c.1110A>C is absent (AF = 0) from gnomAD v2.1 exomes (~125,748 exomes), gnomAD v4.1 exomes (largest dataset, ~1.6M alleles), and gnomAD-Canada v1.0 genomes, with 0 alleles and 0 homozygotes in every dataset. For an autosomal recessive disorder (RAD50 deficiency/NBS-like disorder), PM2 explicitly permits very low or zero frequency, and complete absence across ~2M alleles strongly supports ultra-rarity. There are no threshold artifacts (no subpopulation or FAF exceeds the cutoff because no allele was observed anywhere). Applied at default moderate strength per generic ACMG/AMP 2015 in the absence of a RAD50 VCEP strength calibration.

Rule

Generic ACMG/AMP 2015 PM2 (Richards et al., PMID:25741868; source_registry key generic_acmg_combination_rules): 'Absent in large population cohorts (or at extremely low frequency if recessive).' Applied at the non-VCEP gnomAD cutoff of <0.1%. Met at moderate (default) strength: the variant is completely absent (AF = 0, 0 homozygotes) from gnomAD v2.1, v4.1, and gnomAD-Canada v1.0.

Evidence considered

• gnomad_v2: variant absent (0/125,748 exomes), 0 homozygotes
• gnomad_v4: variant absent (0/~1.6M alleles), 0 homozygotes
• gnomad_canada: variant absent from gnomAD-Canada v1.0 (HostSeq genomes)
• generic_acmg_combination_rules: generic ACMG/AMP 2015 PM2 (PMID:25741868) at default moderate strength; non-VCEP gnomAD cutoff PM2 <0.1% (criteria_task gnomAD cutoff guidance)

Sources

Reasoning

Multiple independent lines of computational evidence suggest no impact on the gene or gene product. Missense predictors: REVEL 0.131 (low, benign range) and BayesDel noAF -0.441836 (negative, benign range) are concordant. Splice-impact predictor: SpliceAI max delta 0.01 with all four delta scores <= 0.01, indicating no predicted effect on splicing. These three independent tools span both the missense in-silico path and the splice-impact path and agree, satisfying the generic ACMG/AMP 2015 BP4 requirement of multiple lines of computational evidence suggesting no impact. Applied at supporting strength per the generic framework; no RAD50 VCEP/CSPEC exists to justify an upgraded strength.

Rule

Generic ACMG/AMP 2015 BP4 (PMID 25741868): multiple lines of computational evidence suggest no impact on gene or gene product (conservation, evolutionary, splicing impact). Met at supporting strength.

Evidence considered

• REVEL score 0.131 - low, benign-range missense prediction
• BayesDel noAF score -0.441836 - negative, benign-range missense prediction
• SpliceAI max delta score 0.01 (DS_AL 0.01; DS_AG/DS_DG/DS_DL 0.0) - no splice impact predicted

Sources

Reasoning

No computational evidence supports a deleterious effect on the gene or gene product. All calibrated computational evidence available for this group is concordant for no effect: SpliceAI max delta score 0.01 (DS_AG 0.0, DS_AL 0.01, DS_DG 0.0, DS_DL 0.0), far below any splice-altering threshold; REVEL 0.131 (low, benign range); BayesDel noAF -0.441836 (negative/benign range). No line of in-silico evidence (splice-impact or missense) predicts pathogenicity, so the generic ACMG/AMP 2015 PP3 criterion (multiple lines of computational evidence supporting deleterious effect) is affirmatively not met.

Rule

Generic ACMG/AMP 2015 PP3 (PMID 25741868): multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact). Not met - every available predictor is concordant benign.

Evidence considered

• SpliceAI max delta score 0.01 (DS_AL 0.01; DS_AG/DS_DG/DS_DL 0.0) - no splice impact predicted
• REVEL score 0.131 - low, benign-range missense prediction
• BayesDel noAF score -0.441836 - negative, benign-range missense prediction

Still missing

• aGVGD scores not available for this variant; not blocking because REVEL, BayesDel, and SpliceAI are concordant benign, so no support for PP3 exists from any available predictor

Sources

Reasoning

BA1 requires an allele frequency in a large, diverse reference population that is incompatible with pathogenicity (generic ACMG/AMP 2015, PMID:25741868, applied at the non-VCEP gnomAD cutoff of >1%; the original 2015 threshold is >5%). NM_005732.3:c.1110A>C is absent from gnomAD v2.1 exomes (gnomad_v2), gnomAD v4.1 exomes (gnomad_v4), and gnomAD-Canada v1.0 genomes (gnomad_canada): allele count 0 in every queried dataset, so the observed allele frequency is 0%, far below both the 1% pipeline cutoff and the 5% ACMG/AMP 2015 cutoff. BA1 is therefore not met under any population metric.

Rule

Generic ACMG/AMP 2015 BA1 (Richards et al., PMID:25741868; source_registry key generic_acmg_combination_rules): allele frequency >1% (non-VCEP gnomAD cutoff; 5% in original 2015 text) is stand-alone benign evidence. Not met: the variant is absent (AF = 0) from gnomAD v2.1, v4.1, and gnomAD-Canada.

Evidence considered

• gnomad_v2: variant absent (search_status 'absent', found=false, no allele count)
• gnomad_v4: variant absent (search_status 'absent', found=false, no allele count)
• gnomad_canada: variant absent from gnomAD-Canada v1.0 (HostSeq genomes)
• generic_acmg_combination_rules: generic ACMG/AMP 2015 BA1 (PMID:25741868) applied at the non-VCEP gnomAD cutoff of >1%

Sources

Reasoning

BS1 requires an allele frequency greater than expected for the disorder (generic ACMG/AMP 2015, PMID:25741868, applied at the non-VCEP gnomAD cutoff of >0.3%). RAD50 deficiency (Nijmegen breakage syndrome-like disorder) is a very rare autosomal recessive disorder, so the expected carrier frequency of a benign RAD50 polymorphism would be orders of magnitude higher than the observed frequency. The variant is absent (AF = 0%) from gnomAD v2.1, gnomAD v4.1, and gnomAD-Canada v1.0, with no subpopulation showing any allele; observed frequency is far below the 0.3% BS1 cutoff. BS1 is therefore not met.

Rule

Generic ACMG/AMP 2015 BS1 (Richards et al., PMID:25741868; source_registry key generic_acmg_combination_rules): allele frequency greater than expected for the disorder. Applied at the non-VCEP gnomAD cutoff of >0.3%. Not met: maximum observed AF is 0% (variant absent from all population databases).

Evidence considered

• gnomad_v2: variant absent (no alleles in any population)
• gnomad_v4: variant absent (no alleles in any population)
• gnomad_canada: variant absent from gnomAD-Canada v1.0
• generic_acmg_combination_rules: generic ACMG/AMP 2015 BS1 (PMID:25741868) applied at the non-VCEP gnomAD cutoff of >0.3%

Sources

Reasoning

BS2 requires observation of the variant in a healthy adult individual in a genotype consistent with the inheritance pattern — for the autosomal recessive RAD50 deficiency/NBS-like disorder, in the homozygous state (or compound heterozygous with a pathogenic allele in a healthy individual). The variant is absent from gnomAD v2.1 and v4.1 (0 alleles, hence 0 homozygotes), absent from gnomAD-Canada v1.0, and has no ClinVar record, so no healthy adult carrier of any genotype is documented in the case evidence. Absence of the variant from population databases is the opposite of the observation BS2 requires; BS2 is therefore not met.

Rule

Generic ACMG/AMP 2015 BS2 (Richards et al., PMID:25741868; source_registry key generic_acmg_combination_rules): 'Observed in a healthy adult individual for a recessive (homozygous), dominant (heterozygous), or X-linked (hemizygous) disorder, with full penetrance expected at an early age.' Not met: the variant is absent from gnomAD v2.1/v4.1/gnomAD-Canada (0 homozygotes, no healthy carriers documented).

Evidence considered

• gnomad_v2: variant absent, 0 alleles and 0 homozygotes
• gnomad_v4: variant absent, 0 alleles and 0 homozygotes
• gnomad_canada: variant absent from gnomAD-Canada v1.0
• clinvar: no ClinVar record for NM_005732.3:c.1110A>C (evidence_sentence 'This variant is absent from ClinVar.'), so no documented healthy-adult observation
• generic_acmg_combination_rules: generic ACMG/AMP 2015 BS2 (PMID:25741868) requires observation in a healthy adult in a genotype consistent with the inheritance pattern

Still missing

• No dedicated healthy-control cohort (beyond gnomAD) was screened for this specific variant, and no case/control literature exists in the workspace; however, complete absence from ~2M gnomAD alleles provides no basis for BS2.

Sources

Reasoning

PS1 (same amino acid change as a previously established pathogenic variant regardless of nucleotide change) cannot be assessed: there is no established pathogenic record of p.Leu370Phe for RAD50 anywhere in the case bundle. ClinVar has no variation record for NM_005732.3:c.1110A>C (search_status 'no_result', variation_ids empty), literature_pass.json found 0 PMIDs, fulltext_map.json is empty, and no publications/ directory exists. The pm5_candidates.json artifact likewise collected 0 same-residue comparator variants (pm5_mode 'unknown', eligible_for_classic_pm5_search false). Without a source documenting that p.Leu370Phe (whether from c.1110A>C or any alternate nucleotide change) is an established pathogenic change, PS1 can be neither met nor ruled out. A human reviewer with ClinVar/LOVD/literature access should verify whether p.Leu370Phe has ever been classified pathogenic at another genomic position.

Rule

PS1 (Richards et al. 2015, PMID:25741868; source_registry key generic_acmg_combination_rules): 'Same amino acid change as a previously established pathogenic variant regardless of nucleotide change.' Requires an established pathogenic record of the identical amino-acid change; no such record exists for RAD50 p.Leu370Phe in this case bundle, so the criterion is not assessable.

Evidence considered

• clinvar: variant absent from ClinVar (no record, no submissions) - no established pathogenic p.Leu370Phe from any nucleotide change is documented
• pm5_candidates: 0 same-residue comparator candidates collected; pm5_mode 'unknown', classic same-residue search skipped
• generic_acmg_combination_rules: PS1 definition per Richards et al. 2015 (PMID:25741868)

Still missing

• No ClinVar record or variant-specific literature reporting p.Leu370Phe (from any nucleotide change) as an established pathogenic variant; ClinVar search for NM_005732.3:c.1110A>C returned no variation

Sources

Reasoning

No de novo occurrence evidence exists for NM_005732.3:c.1110A>C (p.Leu370Phe) in RAD50. The case bundle contains no proband-level clinical data: the variant is absent from ClinVar (no submissions, no case reports), literature_pass.json reports 0 PMIDs, fulltext_map.json is empty, and no publication in the case mentions this variant. There is no information on maternity/paternity-confirmed de novo occurrence, affected status, or family history, so PS2 (de novo with confirmed parentage in a patient with the disease and no family history, Richards et al. 2015) cannot be applied or ruled out.

Rule

PS2 (Richards et al. 2015, PMID:25741868): de novo variant with both maternity and paternity confirmed, in a patient with the disease and no family history. Not applicable to assessment because no de novo/family data were collected.

Still missing

• No de novo occurrence reported for this variant (no proband case report, no parental testing results, no family-history statement) anywhere in the case bundle

Reasoning

PS3 requires well-established in vitro or in vivo functional studies demonstrating a damaging effect on the gene or gene product for this specific variant. No such evidence exists for NM_005732.3:c.1110A>C (p.Leu370Phe): the literature pass returned 0 PMIDs (literature_pass.json note: 'No PMIDs found across evidence sources.'), fulltext_map.json is empty, the publications/ directory contains no papers, OncoKB explicitly reports 'OncoKB did not identify variant-specific reviewed functional evidence for this variant' (Unknown Oncogenic Effect, is_variant_specific=false), and the variant is absent from ClinVar and COSMIC. The only full texts on disk (pvs1_articles, e.g., PMID:32212377) are gene-level RAD50 loss-of-function mechanism papers that do not mention c.1110A>C and therefore cannot serve as variant-specific functional evidence. In silico predictions (REVEL 0.131, BayesDel -0.441836, SpliceAI max delta 0.01) are not functional assays and cannot satisfy PS3; they are handled under PP3/BP4. Evidence insufficient -> not assessed; a validated variant-specific functional assay (e.g., MRN complex formation/activity, DNA double-strand break repair, cellular radiosensitivity/complementation) is required to apply PS3.

Rule

PS3 - Well-established in vitro or in vivo functional studies supportive of a damaging effect on the gene or gene product (Richards et al. 2015, PMID 25741868; generic ACMG/AMP fallback - no RAD50 CSPEC/VCEP exists)

Evidence considered

• OncoKB variant page for RAD50/L370F: no variant-specific reviewed functional evidence (Unknown Oncogenic Effect)
• literature_pass.json: total_pmids=0, criteria_index empty
• fulltext_map.json: empty; publications/ directory: empty
• pvs1_articles full texts grep-checked for c.1110A>C / Leu370 / L370: no mentions (gene-level LoF mechanism only)
• SpliceAI max delta score 0.01 (in silico; not a functional assay)
• REVEL 0.131 and BayesDel -0.441836 (in silico; not functional assays)

Still missing

• No variant-specific in vitro or in vivo functional study of c.1110A>C / p.Leu370Phe was identified (literature_pass total_pmids=0; fulltext_map={}; OncoKB reports no variant-specific reviewed functional evidence; variant absent from ClinVar and COSMIC).
• A validated functional assay demonstrating a damaging effect (e.g., MRN complex assembly/ATPase activity, homologous recombination/DSB repair assays, radiosensitivity or complementation studies) is required before PS3 can be applied.

Sources

Reasoning

PS4 requires case-control or case-series data showing significantly increased prevalence of the variant in affected individuals versus controls. No such data exist for RAD50 c.1110A>C: literature_pass.json returned 0 PMIDs across all evidence sources, fulltext_map.json is empty, no publications directory exists, and ClinVar has no record for the exact variant (no submissions, no referenced case cohorts). Generic ACMG/AMP 2015 (Richards et al., PMID:25741868) governs because no RAD50 CSPEC/VCEP exists (cspec found=false, gene_vcep_dir null).

Rule

generic_acmg_combination_rules (Richards et al. 2015, PMID:25741868) - PS4

Evidence considered

• clinvar (variant absent; no case-control submissions or literature references)

Still missing

• No case-control or case-series enrichment study for RAD50 c.1110A>C identified in any retrieved source; a targeted literature search for this exact variant would be required before PS4 can be assessed.

Sources

Reasoning

PM1 (location in a mutational hotspot and/or critical well-established functional domain without benign variation) cannot be fully assessed. The hotspot route is explicitly negative: the cancer-hotspots lookup returned no statistically significant hotspot for RAD50 L370 (evidence.json hotspots.found = false, record_id 'RAD50 L370', summary 'No hotspot row found'), so PM1 cannot be met via the mutational-hotspot route from collected evidence. The functional-domain route is unevaluable because the case bundle contains no curated domain annotation for RAD50 residue 370 (no UniProt/InterPro domain mapping, no residue-function data in gene_context_RAD50.json or prefetch.json); the variant lies in exon 8 of 25 in the N-terminal half of the 1312-residue protein, but no domain assignment was retrieved. Without domain data, the criterion is left not_assessed rather than not_met.

Rule

PM1 (Richards et al. 2015, PMID:25741868; source_registry key generic_acmg_combination_rules): 'Located in a mutational hot spot and/or critical and well-established functional domain (e.g., active site of an enzyme) without benign variation.' Hotspot evidence is negative for RAD50 L370; no domain annotation is available to evaluate the functional-domain component.

Evidence considered

• evidence bundle (evidence.json hotspots): cancer-hotspots lookup returned no statistically significant hotspot for RAD50 L370 (found=false, 'No hotspot row found')
• generic_acmg_combination_rules: PM1 definition per Richards et al. 2015 (PMID:25741868)

Still missing

• No functional-domain annotation for RAD50 residue 370 (no UniProt/InterPro domain mapping or critical-residue data collected) to evaluate the 'critical and well-established functional domain' route of PM1
• Hotspot route evaluated and negative (cancer-hotspots found no significant hotspot at RAD50 L370), but this alone does not resolve PM1 because the domain route remains unevaluated

Sources

Reasoning

PM3 (Richards et al. 2015, PMID:25741868) requires that, for a recessive disorder, the variant be detected in trans with a pathogenic variant in an affected proband. RAD50 germline deficiency (Nijmegen breakage syndrome-like disorder, 'Human RAD50 deficiency') is a recessive, biallelic-loss-of-function disease context (pvs1_gene_context), so PM3 is nominally applicable to this gene and this variant class (missense). However, the case contains no observational data with which to establish allelic configuration: the variant is absent from ClinVar (no submissions, no variation IDs), the literature pass returned 0 PMIDs (literature_pass.json total_pmids=0, criteria_index empty), and no proband genotypes, pedigree, phase, or second-allele information is present anywhere in the evidence bundle. Without an affected proband observed to carry this variant in trans with a pathogenic RAD50 allele, PM3 cannot be adjudicated. Status set to not_assessed with the gap being absence of proband/phase data, rather than not_met, because the required observation was never available for evaluation.

Rule

generic_acmg_2015: PM3 - for recessive disorders, detected in trans with a pathogenic variant (PMID:25741868)

Evidence considered

• ClinVar search returned no result for RAD50 c.1110A>C / NM_005732.3:c.1110A>C (evidence.json clinvar: search_status=no_result, variation_ids=[], no submissions) - establishes absence of any submitted proband/phase observation
• pvs1_gene_context confirms recessive germline disease mechanism for RAD50 (biallelic LoF, NBS-like phenotype) - establishes PM3 nominal applicability, preventing not_applicable
• literature_pass.json total_pmids=0 - no literature source reporting the variant in any proband

Still missing

• No affected-proband observations of this variant in the case (ClinVar absent, literature_pass total_pmids=0, fulltext_map={})
• No phase data: trans/cis configuration of c.1110A>C relative to any second RAD50 allele was never established
• No evidence of a pathogenic variant in trans (compound heterozygosity) or of a second biallelic hit in any proband

Sources

Reasoning

PM5 (novel missense change at an amino acid residue where a different missense change has been determined pathogenic) cannot be assessed: no alternate missense change at RAD50 residue 370 with an established pathogenic classification was identified in any consulted source. pm5_candidates.json reports pm5_mode 'unknown', eligible_for_classic_pm5_search false, and 0 same-residue candidates (the artifact explicitly instructs: 'Review whether any collected same-residue comparator variants satisfy the governing PM5 rule; do not apply PM5 from this artifact alone'). ClinVar has no record for this variant or searchable comparator at L370, and literature_pass.json found 0 PMIDs. Classic PM5 semantics (a different missense at the same residue classified pathogenic, typically by an expert panel or ClinVar) cannot be confirmed or refuted from the bundle. A human reviewer with ClinVar/literature access should check for any established pathogenic missense at RAD50 p.Leu370 (e.g., p.Leu370Gln/Arg/Val/Pro) to resolve this criterion.

Rule

PM5 (Richards et al. 2015, PMID:25741868; source_registry key generic_acmg_combination_rules): 'Novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before.' Requires an established pathogenic alternate missense at the same residue; no such comparator at RAD50 p.Leu370 exists in the collected evidence, so the criterion is not assessable.

Evidence considered

• pm5_candidates: pm5_mode 'unknown', 0 total candidates, 0 same-residue candidates; recommendation level 'not_applicable' with instruction to review comparators manually
• clinvar: no ClinVar record for NM_005732.3:c.1110A>C and no searchable alternate-missense comparator at residue 370
• generic_acmg_combination_rules: PM5 definition per Richards et al. 2015 (PMID:25741868)

Still missing

• No same-residue alternate missense comparator at RAD50 p.Leu370 with an established pathogenic classification (pm5_candidates.json collected 0 candidates; ClinVar has no record; no variant-specific literature)
• Classic same-residue PM5 semantics could not be confirmed (pm5_mode 'unknown'), so automatic comparator harvesting was skipped

Sources

Reasoning

No assumed-de novo evidence exists for NM_005732.3:c.1110A>C (p.Leu370Phe). The variant has no ClinVar record, no variant-specific literature (literature_pass.json total_pmids = 0), and the bundle contains no proband or parental-testing data from which an unconfirmed-parentage de novo event could be inferred. PM6 (assumed de novo without confirmation of maternity and paternity, Richards et al. 2015) therefore cannot be assessed.

Rule

PM6 (Richards et al. 2015, PMID:25741868): assumed de novo occurrence, but without confirmation of maternity and paternity. Not applicable to assessment because no de novo/family data were collected.

Still missing

• No proband-parent trio or duo data (even with unconfirmed parentage) reporting this variant as de novo

Reasoning

No segregation evidence exists for NM_005732.3:c.1110A>C (p.Leu370Phe). No affected family members carrying the variant are reported anywhere in the bundle: the variant is absent from ClinVar and all literature sources, and no meioses/segregation data were collected. PP1 (co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease, Richards et al. 2015) cannot be applied.

Rule

PP1 (Richards et al. 2015, PMID:25741868): co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease. Not applicable to assessment because no family/segregation data were collected.

Still missing

• No segregation data: no affected relatives genotyped, no count of informative meioses, no segregation analysis for this variant

Reasoning

PP2 (missense variant in a gene that has a low rate of benign missense variation and in which missense variants are a common mechanism of disease) cannot be assessed because neither component can be established from the bundle. No missense constraint data were collected for RAD50: no gnomAD missense z-score or o/e metric, and the HCI prior lookup returned 'gene_not_supported' (hci_prior.found = false). On the mechanism component, the PVS1 gene-context artifact (pvs1_gene_context.json, lof_mechanism_supported = true) characterizes RAD50 germline disease (NBS-like disorder / RAD50 deficiency) as a loss-of-function mechanism, which provides no support for missense being a common disease mechanism and is the only mechanism-level data available. With neither low benign-missense rate nor missense-as-common-mechanism demonstrated, PP2 is left not_assessed.

Rule

PP2 (Richards et al. 2015, PMID:25741868; source_registry key generic_acmg_combination_rules): 'Missense variant in a gene that has a low rate of benign missense variation and in which missense variants are a common mechanism of disease.' Both components require gene-level constraint/mechanism data that are absent for RAD50 in this case bundle.

Evidence considered

• pvs1_gene_context: RAD50 germline disease mechanism characterized as loss-of-function (lof_mechanism_supported=true) - does not support missense as a common disease mechanism
• prefetch.json hci_prior: gene_not_supported - no missense constraint/prior score available
• generic_acmg_combination_rules: PP2 definition per Richards et al. 2015 (PMID:25741868)

Still missing

• No RAD50 missense constraint data (gnomAD missense z-score / o/e metric, or equivalent) to establish a low rate of benign missense variation
• No evidence that missense variants are a common mechanism of RAD50 germline disease; available gene context instead supports loss-of-function as the disease mechanism (pvs1_gene_context.json)

Sources

Reasoning

PP4 requires a phenotype or family history highly specific for a disease with a single genetic etiology. The case bundle contains no proband phenotype or family-history data (case_summary.json has no clinical/phenotype section), so phenotype specificity cannot be evaluated. Gene-level context (pvs1_gene_context) references a distinctive human RAD50-deficiency phenotype (e.g., PMID:32212377), but that literature does not mention the exact variant c.1110A>C and cannot substitute for proband-level phenotype evidence. Human review with access to the requisition/clinical phenotype is needed.

Rule

generic_acmg_combination_rules (Richards et al. 2015, PMID:25741868) - PP4

Evidence considered

• pvs1_gene_context (gene-level RAD50-deficiency phenotype context; exact variant not mentioned)

Still missing

• No proband phenotype or family history information in the case bundle.

Sources

Reasoning

BS3 requires well-established in vitro or in vivo functional studies showing no damaging effect on protein function or splicing for this specific variant. No variant-specific functional study was retrieved (literature_pass total_pmids=0; fulltext_map empty; publications/ empty; OncoKB reports no variant-specific reviewed functional evidence; variant absent from ClinVar and COSMIC). The benign-leaning in silico predictions (REVEL 0.131, BayesDel -0.441836) and SpliceAI (max delta 0.01) are computational predictions, not well-established functional studies, and per generic ACMG/AMP (Richards et al. 2015) they cannot satisfy BS3; they are evaluated under BP4/PP3. Evidence insufficient -> not assessed; a validated functional assay demonstrating normal MRN complex function/DSB repair is required to apply BS3.

Rule

BS3 - Well-established in vitro or in vivo functional studies show no damaging effect on protein function or splicing (Richards et al. 2015, PMID 25741868; generic ACMG/AMP fallback - no RAD50 CSPEC/VCEP exists)

Evidence considered

• OncoKB variant page for RAD50/L370F: no variant-specific reviewed functional evidence (Unknown Oncogenic Effect)
• SpliceAI max delta score 0.01 (in silico; not a functional assay)
• REVEL 0.131 and BayesDel -0.441836 (in silico predictors; not functional assays)
• literature_pass.json: total_pmids=0; fulltext_map empty

Still missing

• No functional study demonstrating no damaging effect of c.1110A>C / p.Leu370Phe on protein function or splicing is available.
• A validated variant-specific functional assay (e.g., MRN complex formation, DSB repair/complementation assays showing wild-type-like activity) is required before BS3 can be applied.

Sources

Reasoning

No non-segregation evidence exists for NM_005732.3:c.1110A>C (p.Leu370Phe). There are no reports of affected family members who do not carry the variant or unaffected carriers (e.g., an unaffected parent/relative with the variant in a dominant context), so BS4 (non-segregation with disease, Richards et al. 2015) cannot be applied. The variant's absence from ClinVar and the empty literature pass leave no source of non-segregation observations.

Rule

BS4 (Richards et al. 2015, PMID:25741868): non-segregation with disease in family members. Not applicable to assessment because no family/segregation data were collected.

Still missing

• No family data demonstrating non-segregation (variant present in unaffected family member(s) or affected family member(s) without the variant)

Reasoning

BP1 (missense variant in a gene for which primarily truncating variants are known to cause disease) cannot be assessed. The bundle establishes that loss-of-function is a supported germline mechanism for RAD50 (pvs1_gene_context.json, lof_mechanism_supported = true, disease 'Human RAD50 deficiency: Confirmation of a distinctive phenotype'), but no curated disease mutation-spectrum data were collected that demonstrate 'primarily truncating variants cause disease' - the criterion requires knowing the full spectrum of disease-causing variants in the gene (including whether missense variants are also causative), not merely that LoF is one mechanism. RAD50-associated phenotypes (NBS-like disorder and cancer-predisposition contexts) cannot be assumed to be truncating-only on the basis of the PVS1 gene-context artifact. With no mutation-spectrum or curated RAD50 disease-variant data in the bundle, BP1 is left not_assessed rather than met or not_met.

Rule

BP1 (Richards et al. 2015, PMID:25741868; source_registry key generic_acmg_combination_rules): 'Missense variant in a gene for which primarily truncating variants are known to cause disease.' Requires disease mutation-spectrum data; only LoF-mechanism support (not a truncating-only spectrum) is available for RAD50 in this case bundle.

Evidence considered

• pvs1_gene_context: loss-of-function supported as a germline mechanism for RAD50 (lof_mechanism_supported=true), but no truncating-only mutation-spectrum claim is made
• generic_acmg_combination_rules: BP1 definition per Richards et al. 2015 (PMID:25741868)

Still missing

• No curated RAD50 disease mutation-spectrum data establishing that primarily truncating variants cause RAD50-associated disease (only LoF-mechanism support exists, which is not equivalent to a truncating-only spectrum)
• No variant-specific evidence that missense variants are non-causative in RAD50 germline disease

Sources

Reasoning

BP2 (Richards et al. 2015, PMID:25741868) applies when the variant is observed in trans with a pathogenic variant for a fully dominant disease/disorder, or is observed in cis with a pathogenic variant in any inheritance pattern. RAD50 is not associated with a fully dominant germline disorder in the case evidence (pvs1_gene_context describes recessive RAD50 deficiency), and no proband, pedigree, or phase data exists to evaluate the cis clause: the variant is absent from ClinVar (no submissions), the literature pass found 0 PMIDs, and no second pathogenic RAD50 allele or phased genotype is reported anywhere in the bundle. Both clauses of BP2 therefore cannot be assessed; status set to not_assessed pending any observational/phase data, which is the sole gap.

Rule

generic_acmg_2015: BP2 - observed in trans with a pathogenic variant for a fully dominant disease/disorder; or observed in cis with a pathogenic variant in any inheritance pattern (PMID:25741868)

Evidence considered

• ClinVar search returned no result for RAD50 c.1110A>C (evidence.json clinvar: search_status=no_result, variation_ids=[]) - no submitted observation of the variant in any phase
• literature_pass.json total_pmids=0 - no literature reporting cis/trans configuration with a pathogenic variant
• pvs1_gene_context lists RAD50 germline disease as recessive deficiency, not a fully dominant disorder - limits applicability of BP2's dominant-trans clause

Still missing

• No proband observations or phased genotypes for this variant in the case (ClinVar absent, literature_pass total_pmids=0)
• No evidence of a pathogenic variant observed in cis with c.1110A>C (any inheritance pattern)
• No fully dominant RAD50 disease context to evaluate the trans clause, and no trans-phase data with a pathogenic variant in a dominant setting

Sources

Reasoning

BP5 applies when a variant is found in a case with an alternative molecular basis for disease. No proband-level data are available in the bundle (no phenotype, no family history, no additional genetic-testing results), so an alternative molecular diagnosis can be neither confirmed nor excluded. Requires human review with access to the full clinical workup.

Rule

generic_acmg_combination_rules (Richards et al. 2015, PMID:25741868) - BP5

Still missing

• No proband clinical/genetic-workup data (e.g., second pathogenic variant in another gene, alternative diagnosis) available in the case bundle.