RAD54L c.1093_1169+15dup, p.?

Reasoning

SpliceAI Lookup (source_registry key 'spliceai') reports a maximal delta score of 0.96 for NM_003579.4:c.1093_1169+15dup, driven by DS_DG = 0.96 (novel donor splice-site gain; DP_DG = 0, i.e., predicted at the variant position) with a secondary DS_DL = 0.41 (donor loss) and DS_AL = 0.02; DS_AG = 0.0. The variant is a 92-bp duplication (c.1093_1169+15) spanning the exon 10/intron 10 boundary, i.e., it recreates the intron 10 5' splice-donor sequence within the duplicated segment, which is mechanistically coherent with the predicted donor gain. Under the SpliceAI published interpretation bands (Jaganathan et al. 2019, PMID 30661751: scores >=0.8 are 'very likely' splice-altering; >=0.5 'likely'; >=0.2 'possible'), 0.96 is a very strong splice-impact prediction, meeting generic ACMG/AMP PP3 ('multiple lines of computational evidence support a deleterious effect on the gene or gene product ... splicing impact', Richards et al. 2015, PMID 25741868). No ClinGen CSPEC/VCEP exists for RAD54L (cspec.found=false; vcep_materials framework found=false, generic_acmg_fallback), so strength is held at the generic supporting level of PP3; no ClinGen SVI-calibrated strength escalation for splice predictors is available in this framework.

Rule

PP3 (Richards et al. 2015, PMID 25741868): Multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.). Applied via generic ACMG fallback because no RAD54L ClinGen CSPEC/VCEP was retrieved.

Evidence considered

• SpliceAI Lookup (source_registry key 'spliceai'): DS_AG=0.0, DS_AL=0.02, DS_DG=0.96, DS_DL=0.41, DP_DG=0, DP_DL=77, max_delta_score=0.96.
• SpliceAI threshold: score >=0.8 = 'very likely' splice-altering, >=0.5 = 'likely', >=0.2 = 'possible' (Jaganathan et al., Cell 2019;176(3):535-548, PMID 30661751). With max delta 0.96 >= 0.8, the prediction supports PP3.
• Generic ACMG/AMP PP3 definition: computational evidence including splicing impact supports a deleterious effect on the gene or gene product (Richards et al. 2015, Genet Med 17(5):405-424, PMID 25741868; framework file output/generic_acmg_classification_rules.md).

Still missing

• No ClinGen SVI-calibrated strength mapping for SpliceAI was available; PP3 strength fixed at 'supporting' per generic ACMG/AMP 2015 even though the raw delta (0.96) falls in the 'very likely' band.
• The splice-altering prediction has not been experimentally validated (no RNA/minigene assay in the case evidence bundle); functional confirmation would be needed to escalate strength.
• REVEL and BayesDel were skipped (not an SNV), so no missense in silico corroboration is available for this indel.

Sources

Reasoning

BS1 requires an allele frequency greater than expected for the disorder (ACMG/AMP 2015, PMID 25741868); this pipeline's non-VCEP gnomAD calibration sets the threshold at AF >0.3%. The variant exceeds this threshold in multiple independent datasets and populations: gnomAD v4.1 East Asian AF 0.734% (326/44,406) and South Asian AF 0.471% (422/89,650), grpmax FAF 0.669%; gnomAD-Canada South Asian AF 0.884% (12/1,358) and East Asian AF 0.754% (10/1,326); gnomAD v2.1 East Asian AF 0.333% (65/19,546). The consistency across gnomAD v2.1, v4.1, and gnomAD-Canada (all GRCh37/GRCh38-normalized to the same allele) argues against a mapping or batch artifact. The presence of 6 homozygotes in gnomAD v4.1 (EAS 4, SAS 1, remaining 1) further indicates a common, non-private allele. BS1 is met at strong strength. Confidence is moderate rather than high because RAD54L germline disease prevalence/penetrance is not established, so 'expected for disorder' rests on the fixed framework threshold.

Rule

BS1: allele frequency greater than expected for the disorder (ACMG/AMP 2015, PMID 25741868); non-VCEP gnomAD calibration threshold >0.3%. Highest population AFs 0.73-0.88% (EAS/SAS) exceed threshold -> met (strong).

Evidence considered

• gnomad_v4: East Asian AF 0.734% (326/44,406) and South Asian AF 0.471% (422/89,650) both exceed 0.3%; grpmax FAF 0.669%; 6 homozygotes (EAS 4, SAS 1, remaining 1)
• gnomad_canada: South Asian AF 0.884% (12/1,358) and East Asian AF 0.754% (10/1,326) exceed 0.3%
• gnomad_v2: East Asian AF 0.333% (65/19,546) just exceeds 0.3%; 1 homozygote
• generic_acmg_combination_rules: BS1 defined as allele frequency greater than expected for disorder (PMID 25741868); non-VCEP gnomAD calibration threshold 0.3%

Still missing

• No established RAD54L germline disease prevalence or penetrance estimate is available; BS1 applied via the fixed 0.3% non-VCEP threshold rather than a disease-derived expected frequency.
• Gene-disease relationship for RAD54L is putative (literature-only, adult-onset cancer susceptibility candidate); if a higher expected allele frequency were justified, BS1 would still hold (observed 0.73-0.88% in EAS/SAS), but review is warranted when disease prevalence becomes established.

Sources

Reasoning

NM_003579.4(RAD54L):c.1093_1169+15dup is a 92-bp tandem duplication spanning the last 77 nt of exon 10 (c.1093_1169, coding) and the first 15 nt of intron 10 (c.1169+1_+15, which includes the GT splice-donor dinucleotide); VariantValidator places it from exon 10 to intron 10 ('10i'). The protein consequence is unknown (NP_003570.2:p.?) and the variant does not fall into any default PVS1 null-variant bucket (nonsense, frameshift, canonical +/-1,2 splice, initiation codon, exonic deletion) under the ClinGen SVI PVS1 framework (PMC6185798); the case pvs1_variant_assessment records consequence_class 'unknown', variant_bucket 'other', and apply_generic_pvs1_framework=false. SpliceAI predicts a strong donor gain (DS_DG=0.96; max delta 0.96) with a weaker donor-loss signal at the original site (DS_DL=0.41): if the novel donor is used, the duplicated intronic+exonic material (92 nt, not a multiple of 3) would be retained as coding sequence, causing a frameshift with premature termination likely subject to NMD (exon 10 is an internal exon followed by intron 10); however, if the original donor is used, the duplicated block lies within intron 10 and is excised, giving a normal transcript. No RNA or functional evidence discriminates between these outcomes, and an in silico splice prediction alone does not establish LOF for PVS1. Gene-level LoF mechanism for RAD54L is supported only by a literature gate (pvs1_gene_context; no official CSPEC/VCEP, mechanism considered putative), and none of the supporting publications mention this variant. PVS1 is therefore not met; RNA/functional data would be required to revisit it.

Rule

PVS1: null variant (nonsense, frameshift, canonical +/-1 or 2 splice, initiation codon, single or multi-exon deletion) in a gene where loss of function is a known mechanism of disease (ACMG/AMP 2015, PMID 25741868; ClinGen SVI PVS1 refinement, Tayoun et al. 2018, PMC6185798)

Evidence considered

• pvs1_variant_assessment: consequence_class 'unknown', variant_bucket 'other', apply_generic_pvs1_framework=false, protein NP_003570.2:p.?
• pvs1_generic_framework (ClinGen SVI PVS1 recommendations, PMC6185798): default PVS1 null-variant buckets do not include this exon/intron-spanning duplication; non-bucket variants require demonstrated LOF
• spliceai: DS_DG=0.96 (donor gain), DS_DL=0.41 (donor loss at original site), max_delta=0.96 - in silico only, no RNA validation
• pvs1_gene_context: RAD54L LoF mechanism supported only via literature gate (supporting PMIDs do not mention this variant; no official VCEP/CSPEC)

Still missing

• RNA or minigene splicing assay to determine whether the novel donor is used and the duplicated 92-nt segment (15 intronic + 77 exonic) is retained in the mature transcript
• Demonstration of a null/LoF transcript or protein for this specific variant (e.g., frameshift-PTC-NMD confirmation)
• A non-putative, gene-specific (VCEP/CSPEC) confirmation that RAD54L LoF is a disease mechanism

Sources

Reasoning

PM2 requires the variant to be absent or at extremely low frequency in controls (ACMG/AMP 2015, PMID 25741868); this pipeline's non-VCEP gnomAD calibration sets the rarity threshold at AF <0.1%. Although the overall gnomAD v4.1 AF (0.0493%) is below 0.1%, the highest population frequencies vastly exceed the threshold: East Asian AF 0.734% (326/44,406, 4 homozygotes) and South Asian AF 0.471% (422/89,650) in gnomAD v4.1, South Asian AF 0.884% in gnomAD-Canada, and East Asian AF 0.333% in gnomAD v2.1. Under standard practice the highest population AF (or grpmax FAF 0.669%) is used for rarity assessment, and all of these are 3-9x above the 0.1% cutoff. The variant is clearly not rare in East Asian/South Asian populations, and 6 homozygotes in gnomAD v4.1 further contradict a private pathogenic allele. PM2 is not met.

Rule

PM2: absent in controls / extremely low frequency (ACMG/AMP 2015, PMID 25741868); non-VCEP gnomAD calibration threshold <0.1%. Highest population AFs 0.33-0.88% (EAS/SAS) exceed threshold -> not met.

Evidence considered

• gnomad_v4: highest population AF East Asian 0.734% (326/44,406) and South Asian 0.471% (422/89,650), grpmax FAF 0.669% - all far exceed the 0.1% PM2 rarity threshold
• gnomad_v2: East Asian AF 0.333% (65/19,546), grpmax FAF 0.270% - exceed 0.1%
• gnomad_canada: South Asian AF 0.884% (12/1,358), East Asian 0.754% (10/1,326) - exceed 0.1%
• generic_acmg_combination_rules: PM2 defined as absent in controls or extremely low frequency (PMID 25741868); non-VCEP gnomAD calibration threshold <0.1%

Still missing

• No RAD54L-specific control cohort or disease prevalence was available; assessment relied on gnomAD population frequencies, which are sufficient to exclude PM2 under the 0.1% threshold.

Sources

Reasoning

PM4 requires a protein length change resulting from an in-frame deletion/insertion in a non-repeat region or from a stop-loss variant. No protein consequence is predicted for this variant (NP_003570.2:p.?, consequence_class 'unknown'), and no in-frame change is established under any plausible outcome: if the original splice donor of intron 10 is used, the duplicated block is spliced out as intron and the transcript is normal (no length change); if the novel donor predicted by SpliceAI (DS_DG=0.96) is used, the retained duplicated segment (15 nt intronic + 77 nt exonic = 92 nt) is not a multiple of three, producing a frameshift with premature termination rather than an in-frame insertion; exon-skipping alternatives would likewise not produce a defined in-frame protein length change. The variant is also not a stop-loss variant. Because the duplication is not a clean in-frame coding indel (it spans the exon 10/intron 10 boundary), PM4 is not met.

Rule

PM4: protein length changes as a result of in-frame deletions/insertions in a non-repeat region or stop-loss variants (ACMG/AMP 2015, PMID 25741868)

Evidence considered

• pvs1_variant_assessment: protein p.? (NP_003570.2), consequence_class 'unknown' - no in-frame length change predicted
• spliceai: predicted donor gain (DS_DG=0.96) implies retention of a 92-nt duplicated segment, which is not a multiple of 3 and would cause a frameshift, not an in-frame insertion
• Variant structure from prefetch.json (VariantValidator): c.1093_1169+15dup spans exon 10 (start_exon '10') into intron 10 (end_exon '10i'), i.e., a complex exon/intron-boundary duplication, not a clean in-frame coding indel

Still missing

• RNA transcript analysis to define the exact coding consequence of the duplication
• Protein-level evidence of any length change (none is predicted in silico)

Sources

Reasoning

PP5 requires a reputable source (per group policy: an exact-variant ClinVar expert-panel classification of Pathogenic or Likely pathogenic) to have reported the variant as pathogenic without evidence available for independent evaluation. The ClinVar submission audit for variation 4281005 (VCV004281005) shows exactly one submission (SCV006560075, CeGaT Center for Human Genetics Tuebingen), classified Likely benign with review status 'criteria provided, single submitter' (1 star); expert_panels is empty, expert_panel_submissions = 0, and there are no validated PMIDs. No expert-panel pathogenic classification exists for this variant — and the only classification present is in the opposite (benign) direction from a non-expert laboratory. Per group policy, ordinary laboratory submissions and aggregate labels must never trigger PP5. Criterion not met.

Rule

PP5: Reputable source recently reports variant as pathogenic, but the evidence is not available to the laboratory to perform an independent evaluation (Richards et al. 2015, PMID 25741868). Group policy: only an exact-variant ClinVar expert-panel Pathogenic/Likely pathogenic classification qualifies; non-expert laboratory submissions never trigger PP5.

Evidence considered

• ClinVar variation 4281005 (source_registry key clinvar): 1 submission (SCV006560075, CeGaT Center for Human Genetics Tuebingen), Likely benign, 'criteria provided, single submitter' (1 star), germline, conditions 'not provided'; expert_panels empty (0 expert-panel submissions); validated_pmids = 0.
• Generic ACMG/AMP 2015 fallback framework (source_registry key generic_acmg_combination_rules, PMID 25741868) defines PP5; group policy restricts PP5 to exact-variant ClinVar expert-panel P/LP classifications, which are absent.

Still missing

• No ClinVar expert-panel (or other reputable expert) classification of this exact variant exists; 0 expert-panel submissions on VCV004281005.

Sources

Reasoning

BA1 requires an allele frequency >5% per ACMG/AMP 2015 (PMID 25741868), and >1% per this pipeline's non-VCEP gnomAD calibration. The maximum allele frequency observed for this variant across all consulted datasets is 0.884% (South Asian, gnomAD-Canada v1.0, 12/1358 alleles); gnomAD v4.1 East Asian AF is 0.734% (326/44,406) with grpmax FAF 0.669%, and gnomAD v2.1 grpmax FAF is 0.270%. All values are below both the 1% and 5% BA1 thresholds. BA1 is therefore not met. The population enrichment in East Asian/South Asian populations (0.33-0.88%) is consistent with a relatively common ancestral allele rather than a private pathogenic variant, but does not reach the BA1 frequency threshold.

Rule

BA1: allele frequency >5% in ExAC/1000 Genomes/ESP (ACMG/AMP 2015, PMID 25741868); non-VCEP gnomAD calibration threshold applied here >1%. Max observed AF 0.884% < 1% -> not met.

Evidence considered

• gnomad_v4: total AF 0.0493% (795/1,612,030 alleles), highest population AF East Asian 0.734% (326/44,406, 4 homozygotes), grpmax FAF 0.669% - all below BA1 threshold
• gnomad_v2: total AF 0.0397% (112/281,762 alleles), East Asian AF 0.333% (65/19,546), grpmax FAF 0.270%
• gnomad_canada: overall AF 0.120% (22/18,406 alleles), South Asian AF 0.884% (12/1,358) - maximum observed AF across all sources
• generic_acmg_combination_rules: BA1 defined as allele frequency >5% in large population datasets (Richards et al. 2015, PMID 25741868); pipeline non-VCEP calibration threshold 1%

Still missing

• Disease prevalence and penetrance for RAD54L germline variants are not established, so an expected-allele-frequency prior could not be independently derived; threshold-based BA1 assessment relied on the fixed 1%/5% cutoffs.

Sources

Reasoning

BS2 requires the variant to be 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 (ACMG/AMP 2015, PMID 25741868). Homozygotes are indeed observed in general-population cohorts (6 in gnomAD v4.1: EAS 4, SAS 1, remaining 1; 1 in gnomAD v2.1 EAS; 0 in gnomAD-Canada), which would be notable for a fully penetrant early-onset recessive disorder. However, RAD54L is only a putative, literature-only germline cancer-susceptibility candidate with adult onset, incomplete penetrance, and no established disease prevalence or phenotype expectations; the full-penetrance-early-onset requirement of BS2 is not satisfied. Homozygosity in a general cohort for a gene without an established fully penetrant phenotype cannot by itself meet BS2. BS2 is therefore not met.

Rule

BS2: 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 (ACMG/AMP 2015, PMID 25741868). Homozygotes present in gnomAD but no established fully penetrant early-onset RAD54L disorder -> not met.

Evidence considered

• gnomad_v4: 6 homozygotes observed (East Asian 4, South Asian 1, remaining 1) - homozygosity present but phenotype/penetrance requirement for BS2 not established
• gnomad_v2: 1 homozygote (East Asian, 1/19,546)
• gnomad_canada: 0 homozygotes (22/18,406 alleles)
• generic_acmg_combination_rules: BS2 requires observation in healthy adults for a disorder with full penetrance expected at early age (PMID 25741868); RAD54L disease context does not satisfy this

Still missing

• No established RAD54L germline disorder with penetrance/age-of-onset data exists; BS2's 'full penetrance at early age' condition cannot be evaluated against a defined phenotype.
• gnomAD cohorts include some affected individuals and are not strictly 'healthy adult' controls; a dedicated control cohort for the relevant phenotype would be needed to revisit BS2.

Sources

Reasoning

BP3 requires an in-frame deletion/insertion located in a repetitive region without known function. None of the preconditions is established. (1) The protein consequence is unknown (NP_003570.2:p.?); the predicted outcomes are a normal transcript (original donor used) or a frameshift with premature termination (novel donor used; 92-nt retention is not a multiple of 3) - neither is an in-frame change. (2) No repeat-masking annotation identifies this locus as a repetitive region: gnomAD lcr and segdup flags are false, and although the duplicated exon-10 sequence is GC-rich with tandem-like internal motifs (e.g., AGGCAGACAGGCAG, CTGCTGCT), coding exon 10 of RAD54L (a SNF2-family chromatin-remodeling helicase) is not a functionally inert repeat region. (3) The duplication spans the canonical exon 10 splice-donor site, a functionally important element, and the affected exon 10 is internal coding sequence. BP3 is therefore not met.

Rule

BP3: in-frame deletions/insertions in a repetitive region without a known function (ACMG/AMP 2015, PMID 25741868)

Evidence considered

• pvs1_variant_assessment: protein p.? (NP_003570.2), consequence_class 'unknown' - in-frame status not established
• spliceai: donor-gain prediction (DS_DG=0.96) supports a frameshift/aberrant-splicing outcome, not an in-frame indel
• gnomad_canada (raw VEP/region flags): lcr=false, segdup=false - no low-complexity or segmental-duplication flag at this locus

Still missing

• RepeatMasker / repeat-region annotation for the c.1093_1169+15 locus (not present in the case evidence bundle)
• RNA evidence establishing any in-frame consequence (currently none predicted)

Sources

Reasoning

BP4 requires multiple lines of computational evidence suggesting no impact on the gene or gene product (including splicing impact). The only applicable computational evidence in this case is SpliceAI, which strongly predicts splice impact (max delta = 0.96, DS_DG = 0.96), the opposite direction of BP4. REVEL and BayesDel were skipped ('not snv or no coords') because the variant is a 92-bp duplication, not a single-nucleotide substitution, so no missense in silico predictions exist for this variant. No benign computational evidence is present. BP4 is therefore not met.

Rule

BP4 (Richards et al. 2015, PMID 25741868): Multiple lines of computational evidence suggest no impact on gene or gene product (conservation, evolutionary, splicing impact, etc.). Not met because available computational evidence predicts impact rather than no impact.

Evidence considered

• SpliceAI Lookup (source_registry key 'spliceai'): max_delta_score=0.96 (DS_DG=0.96), indicating predicted splice impact, contrary to the 'no impact' requirement of BP4.
• REVEL/BayesDel: not run for this variant (prefetch.json and evidence.json: skipped, reason 'not snv or no coords').

Still missing

• No missense in silico predictors (REVEL/BayesDel) could be evaluated because the variant is an indel, not an SNV.

Sources

Reasoning

BP6 requires a reputable source (per group policy: an exact-variant ClinVar expert-panel classification of Benign or Likely benign) to have reported the variant as benign without evidence available for independent evaluation. Although ClinVar variation 4281005 is classified Likely benign, the submission audit shows this is a single ordinary clinical-laboratory submission (SCV006560075, CeGaT Center for Human Genetics Tuebingen; 'criteria provided, single submitter', 1 star) with 0 expert-panel submissions. Per group policy, ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions must never trigger BP6. Criterion not met.

Rule

BP6: Reputable source recently reports variant as benign, but the evidence is not available to the laboratory to perform an independent evaluation (Richards et al. 2015, PMID 25741868). Group policy: only an exact-variant ClinVar expert-panel Benign/Likely benign classification qualifies; non-expert laboratory submissions never trigger BP6.

Evidence considered

• ClinVar variation 4281005 (source_registry key clinvar): 1 submission (SCV006560075, CeGaT Center for Human Genetics Tuebingen), Likely benign, 'criteria provided, single submitter' (1 star); expert_panels empty (0 expert-panel submissions).
• Generic ACMG/AMP 2015 fallback framework (source_registry key generic_acmg_combination_rules, PMID 25741868) defines BP6; group policy restricts BP6 to exact-variant ClinVar expert-panel B/LB classifications, which are absent.

Still missing

• No ClinVar expert-panel benign classification of this exact variant exists; the only classification is a single non-expert clinical laboratory submission.

Sources

Reasoning

PS2 requires a confirmed de novo occurrence of the variant (both biological parents tested, parental identity confirmed) in a proband whose phenotype is consistent with the gene's disease. The case bundle contains no proband, phenotype, pedigree, or parental testing data whatsoever: evidence.json covers only population databases (gnomAD v2.1, v4.1, gnomAD-Canada), ClinVar (single-submitter Likely benign with no criteria or PMIDs), COSMIC (no result), SpliceAI, and OncoKB (failed); literature_pass.json reports 0 PMIDs; fulltext_map.json is empty. Without proband parental testing results, de novo status cannot be established, so PS2 is not assessed.

Rule

PS2: De novo variant (confirmed parental identity) in a proband with the disease and no family history (ACMG/AMP 2015; PMID:25741868)

Still missing

• No proband phenotype or clinical case data in the case bundle
• No parental testing results for the variant (both parents untested or results unavailable)
• No confirmation of parental identity (maternity/paternity testing)

Sources

Reasoning

No well-established functional studies of this exact variant (NM_003579.4:c.1093_1169+15dup, RAD54L) exist in any evidence source for this case. The consolidated bundle contains zero PMIDs (literature_pass.json total_pmids=0), an empty fulltext_map, and an empty publications directory. The only ClinVar submission (SCV006560075, CeGaT, 'Likely benign', criteria provided/single submitter) contains no PMIDs and no functional criteria, and per audit it is not usable for criterion leads; a submitter classification is never functional evidence and cannot substitute for PS3. SpliceAI reports a high max delta score (0.96, DS_DG donor-gain) but this is an in silico splice prediction, not a validated assay, so it cannot support PS3 under the generic ACMG/AMP 2015 rule (PMID 25741868). The variant is a 92-bp duplication spanning the exon 10/intron 10 boundary (donor splice site region; gnomAD-Canada VEP annotation: stop_gained), a class for which RNA/minigene splicing assays or DNA-repair functional assays would be applicable and informative, but none are available. PS3 therefore cannot be met and is recorded as not_assessed pending functional data.

Rule

PS3: Well-established functional studies show a damaging effect on the gene or gene product (Richards et al. 2015, PMID 25741868; generic ACMG/AMP 2015 fallback, source_registry key generic_acmg_combination_rules). Not met due to complete absence of functional assay evidence.

Evidence considered

• Generic ACMG/AMP 2015 PS3 definition (PMID 25741868, source_registry key generic_acmg_combination_rules): PS3 requires well-established functional studies showing a damaging effect on the gene or gene product; no such study is present
• literature_pass.json: total_pmids=0, criteria_index empty - no functional literature identified for this variant
• fulltext_map.json: {} and publications/ directory empty - no full texts available for functional review
• clinvar (VCV004281005 / SCV006560075): single-submitter 'Likely benign', no PMIDs, no criteria provided; unusable for criterion leads and not functional evidence
• spliceai: max delta 0.96 (DS_DG 0.96, DS_DL 0.41) - in silico prediction only, not a well-established functional study; excluded from PS3/BS3 per generic ACMG rule

Still missing

• No functional assay (RNA/minigene splice assay, RT-PCR, or homologous-recombination/DNA-repair functional assay) has been reported for this exact variant
• No PMIDs or full-text publications addressing this variant were retrieved (literature_pass total_pmids=0; fulltext_map empty; publications/ empty)
• ClinVar submission provides no functional evidence, no criteria, and no citations (SCV006560075, rejection reason: no_valid_pmids)

Sources

Reasoning

PS4 requires the prevalence of the variant in affected individuals to be significantly increased compared with controls (case-control enrichment). No case-control or affected-cohort prevalence study exists for NM_003579.4:c.1093_1169+15dup: literature_pass.json reports total_pmids=0 ('No PMIDs found across evidence sources'), the publications/ directory is empty, fulltext_map.json is empty, and no cohort-based enrichment statistics (OR, RR, case/control allele counts) are present anywhere in the evidence bundle. gnomAD v2.1/v4.1 and gnomAD-Canada frequencies (total AF 0.0397%-0.0493%, East Asian AF up to 0.734%, homozygotes observed) document general-population presence but are population-frequency data, not case-control data, and cannot be used to infer enrichment in affected individuals. Evidence is insufficient to assess the criterion; a human reviewer with access to case-level or case-control literature for this exact variant would be needed.

Rule

PS4: The prevalence of the variant in affected individuals is significantly increased compared with controls (Richards et al. 2015, PMID 25741868).

Evidence considered

• gnomAD v2.1 (source_registry key gnomad_v2): total AF 0.000397 (112/281,762 alleles, 1 homozygote), EAS AF 0.00333, grpmax FAF 0.00270 — general-population frequency only, no case/control structure.
• gnomAD v4.1 (source_registry key gnomad_v4): total AF 0.000493 (795/1,612,030 alleles, 6 homozygotes), EAS AF 0.00734, grpmax FAF 0.00669 — general-population frequency only, no case/control structure.
• gnomAD-Canada v1.0 (source_registry key gnomad_canada): AF 0.001195 (22/18,406 alleles) in HostSeq genomes — general-population frequency only, no case/control structure.
• Generic ACMG/AMP 2015 fallback framework (source_registry key generic_acmg_combination_rules, PMID 25741868) defines PS4; no applicable case-control evidence was found to apply it.

Still missing

• No case-control or affected-vs-control enrichment study for NM_003579.4:c.1093_1169+15dup exists in the evidence bundle (literature_pass total_pmids=0; publications/ empty; fulltext_map empty).
• No enrichment statistic (OR/RR with confidence intervals, or case/control allele counts) for this exact variant is available.

Sources

Reasoning

No applicable VCEP/CSPEC specification exists for RAD54L (vcep_materials.json: 'no cspec url available', final_classification_framework.source = generic_acmg_fallback, precedence_rank = 3), so generic ACMG/AMP 2015 (Richards et al., PMID 25741868) governs. PM3 (moderate) requires, for a recessive disorder, observation of the variant in trans with a pathogenic variant, typically in an affected proband with biallelic variants. The case bundle contains no proband or family genotype data, no phase (trans/cis) determination relative to any pathogenic RAD54L variant, no segregation data, and no retrieved literature (literature_pass.json: total_pmids = 0; fulltext_map.json: {}; publications directory empty). The ClinVar record (VCV004281005 / variation 4281005, single germline submission SCV006560075 from CeGaT, classification 'Likely benign', review status 'criteria provided, single submitter', condition 'not provided') contains no genotype, phase, or affected-proband details and no PMIDs; per protocol its final classification is not used to infer any criterion. gnomAD population frequencies (v4.1 total AF 0.0493%, 795/1,612,030 alleles, 6 homozygotes; EAS AF 0.734%) document population presence and homozygosity but do not satisfy PM3's requirement of a pathogenic variant in trans in an affected individual, and homozygosity in population controls is not allelic evidence for PM3. Evidence is insufficient to assess PM3; the criterion is returned as not_assessed.

Rule

Generic ACMG/AMP 2015 (Richards et al. 2015, PMID 25741868), PM3 (moderate, pathogenic): 'For recessive disorders, detected in trans with a pathogenic variant.' Applied because no VCEP/CSPEC specification exists for RAD54L (vcep_materials.json: generic_acmg_fallback, precedence_rank 3).

Evidence considered

• Generic ACMG/AMP 2015 PM3 definition (Richards et al., PMID 25741868): for recessive disorders, detected in trans with a pathogenic variant.
• ClinVar variation 4281005 (VCV004281005): single germline submission (CeGaT, SCV006560075), Likely benign, 'criteria provided, single submitter', condition 'not provided', no PMIDs, no genotype/phase data (prefetch.json submission_audit).
• gnomAD v2.1: total AF 0.03975% (112/281,762 alleles), 1 homozygote, EAS AF 0.33255%.
• gnomAD v4.1: total AF 0.04932% (795/1,612,030 alleles), 6 homozygotes, EAS AF 0.73414%, grpmax FAF 0.00668501.
• gnomAD-Canada v1.0: AF 0.11953% (22/18,406 alleles), 0 homozygotes.
• literature_pass.json: total_pmids = 0; fulltext_map.json: {}; no publication full texts available.

Still missing

• No affected-proband observation of this variant in trans with a pathogenic RAD54L variant (no compound-heterozygote or affected-homozygote case reported).
• No phase (trans/cis) information for this variant relative to any other RAD54L variant in any family.
• Inheritance mode and recessive disease association for RAD54L are not established in this case bundle (no CSPEC/VCEP; ClinVar condition 'not provided').
• No literature or cohort studies retrieved (literature_pass total_pmids = 0; fulltext_map empty) that could report trans observations.

Sources

Reasoning

PM6 applies when a de novo variant is assumed (parental identity not confirmed) in a proband with a consistent phenotype. The case bundle contains no proband or family data: no de novo occurrence is reported for this variant in any source, and no proband phenotype or parental testing information exists. literature_pass.json found 0 PMIDs and the publications directory is empty, so no published de novo case could be consulted. PM6 is not assessed due to absent proband/family evidence.

Rule

PM6: Assumed de novo variant (parental identity not confirmed) in a proband with the disease (ACMG/AMP 2015; PMID:25741868)

Still missing

• No proband de novo occurrence reported in any consulted source
• No parental testing data to support or refute de novo origin
• No proband phenotype information to assess phenotypic consistency

Sources

Reasoning

PP1 requires co-segregation of the variant with disease in affected family members in a gene definitively known to cause the phenotype. No affected relatives, pedigree structure, or segregation data exist in the case bundle, and RAD54L has no established ClinGen VCEP specification defining segregation thresholds for this gene. literature_pass.json reports 0 PMIDs, so no published segregation families could be evaluated. PP1 is not assessed due to absent segregation evidence.

Rule

PP1: Co-segregation with disease in affected family members in a gene definitively known to cause the disease (ACMG/AMP 2015; PMID:25741868)

Still missing

• No family segregation data (affected relatives tested for the variant)
• No meioses/segregation count available for any pedigree
• No gene-specific (VCEP) segregation scoring rule for RAD54L; generic ACMG/AMP 2015 PP1 would require such data if present

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, clinical presentation, or family-history data whatsoever (case_summary.json contains only variant/evidence artifacts; no clinical case description is present). Additionally, RAD54L is at best a putative cancer-predisposition gene (per pvs1_gene_context literature, e.g. candidate ovarian-cancer predisposition gene) and has no established, highly specific single-etiology phenotype that would make PP4 applicable. Without any proband-level phenotypic information, the criterion cannot be assessed from the available evidence.

Rule

PP4: Phenotype or family history highly specific for a disease with a single genetic etiology (Richards et al. 2015, PMID 25741868).

Evidence considered

• Generic ACMG/AMP 2015 fallback framework (source_registry key generic_acmg_combination_rules, PMID 25741868) defines PP4; no proband phenotype or family-history evidence exists in the bundle to apply it.

Still missing

• No proband phenotype, clinical presentation, or family-history data is present anywhere in the case bundle.
• RAD54L lacks an established disease with a single genetic etiology; only generic putative cancer-predisposition context (ovarian cancer candidate gene) is available.

Sources

Reasoning

No functional studies demonstrating a benign (non-damaging) effect of this variant on RAD54L function or splicing were identified. BS3 requires well-established functional studies showing no damaging effect (generic ACMG/AMP 2015, PMID 25741868); none exist in this case's evidence bundle (0 PMIDs, empty fulltext map, empty publications directory). The ClinVar entry is a single-submitter 'Likely benign' classification (SCV006560075) with no criteria and no functional data; per adjudication rules a final classification from another source must never be used to infer a criterion, and population frequency observations (gnomAD homozygotes) belong to the population evidence group (BA1/BS1), not to functional assays. Absence of evidence of damage is not equivalent to evidence of no damage. BS3 is therefore not assessed pending functional assay data.

Rule

BS3: Well-established functional studies show no damaging effect on protein function or splicing (Richards et al. 2015, PMID 25741868; generic ACMG/AMP 2015 fallback, source_registry key generic_acmg_combination_rules). Not assessed due to complete absence of functional assay evidence.

Evidence considered

• Generic ACMG/AMP 2015 BS3 definition (PMID 25741868, source_registry key generic_acmg_combination_rules): BS3 requires well-established functional studies showing no damaging effect; none available
• literature_pass.json: total_pmids=0 - no functional literature supporting a benign effect
• fulltext_map.json: {} and publications/ empty - no full texts available
• clinvar (VCV004281005 / SCV006560075): 'Likely benign' single-submitter entry without criteria or PMIDs; excluded as a BS3 basis per no-inference-from-other-classifications rule
• spliceai: in silico prediction (max delta 0.96) excluded; not a functional study and does not support BS3

Still missing

• No functional assay demonstrating normal/non-damaging effect on splicing or RAD54L DNA-repair function for this exact variant
• No publications, PMIDs, or full texts with functional data were retrieved for this variant
• ClinVar 'Likely benign' (single submitter, no criteria, no PMIDs) cannot serve as functional evidence for BS3

Sources

Reasoning

BS4 requires non-segregation of the variant with disease in affected family members (affected relative(s) who do not carry the variant). No affected family members, pedigree data, or family testing results exist in the case bundle; the ClinVar single-submitter 'Likely benign' classification was deliberately not used as evidence for this or any criterion (per instruction not to infer criteria from another source's final classification). BS4 is not assessed due to absent non-segregation data.

Rule

BS4: Non-segregation with disease in affected family members (ACMG/AMP 2015; PMID:25741868)

Still missing

• No affected family member tested and found negative for the variant
• No pedigree/family testing data of any kind in the case bundle
• ClinVar 'Likely benign' (SCV006560075, single submitter, no criteria provided, no PMIDs) provides no usable non-segregation evidence

Sources

Reasoning

No applicable VCEP/CSPEC specification exists for RAD54L, so generic ACMG/AMP 2015 (Richards et al., PMID 25741868) governs. BP2 (supporting, benign) requires observation of the variant in trans with a pathogenic variant for a fully dominant disease/gene, or in cis with a pathogenic variant in any inheritance pattern. The case bundle contains no such phase observations: there is no proband or family genotype data, no cis/trans determination, no segregation data, no retrieved literature (literature_pass.json total_pmids = 0; fulltext_map.json: {}), and no established fully dominant disease specification for RAD54L (ClinVar condition 'not provided'; no CSPEC). The ClinVar record (VCV004281005, single submitter, 'Likely benign', 'criteria provided, single submitter') provides a final classification without phase data and cannot be used to infer BP2. gnomAD population frequency and homozygote counts (v2.1: 1 homozygote; v4.1: 6 homozygotes; Canada: 0) document population presence only and do not establish cis or trans phase with a pathogenic variant. Evidence is insufficient to assess BP2; the criterion is returned as not_assessed.

Rule

Generic ACMG/AMP 2015 (Richards et al. 2015, PMID 25741868), BP2 (supporting, benign): 'Observed in trans with a pathogenic variant for a fully dominant disease/gene; or observed in cis with a pathogenic variant in any inheritance pattern.' Applied because no VCEP/CSPEC specification exists for RAD54L (vcep_materials.json: generic_acmg_fallback, precedence_rank 3).

Evidence considered

• Generic ACMG/AMP 2015 BP2 definition (Richards et al., PMID 25741868): 'Observed in trans with a pathogenic variant for a fully dominant disease/gene; or observed in cis with a pathogenic variant in any inheritance pattern.'
• ClinVar variation 4281005 (VCV004281005): single germline submission (CeGaT, SCV006560075), Likely benign, 'criteria provided, single submitter', no genotype/phase data, no PMIDs.
• gnomAD v2.1 / v4.1 / gnomAD-Canada v1.0 population records: allele counts and homozygote counts only; no allelic phase information (see PM3 evidence_used for frequencies).
• literature_pass.json: total_pmids = 0; fulltext_map.json: {}; no publication full texts available.

Still missing

• No observation of this variant in cis with a pathogenic RAD54L variant in any individual.
• No observation of this variant in trans with a pathogenic variant in the context of a fully dominant disease/gene; no fully dominant RAD54L disease specification is available in the case bundle.
• No family segregation or phase data for any carrier of this variant.
• No literature or cohort studies retrieved that report allelic configuration of this variant with other RAD54L variants.

Sources

Reasoning

BP5 requires that the variant was found in a case with an alternate molecular basis for disease. No proband-level case data is available in the case bundle (no clinical case description, no segregation/co-occurrence data, no alternate-molecular-diagnosis information). The variant is present in the general population (gnomAD v2.1/v4.1/gnomAD-Canada), which is contextual, but the presence or absence of an alternative molecular diagnosis for any affected individual is unknown. Evidence is insufficient to assess the criterion.

Rule

BP5: Variant found in a case with an alternate molecular basis for disease (Richards et al. 2015, PMID 25741868).

Evidence considered

• Generic ACMG/AMP 2015 fallback framework (source_registry key generic_acmg_combination_rules, PMID 25741868) defines BP5; no proband case or co-occurrence evidence exists in the bundle to apply it.

Still missing

• No proband case data is available; whether the variant was observed alongside an alternate molecular basis for disease is unknown.

Sources