PRPF8 c.2013A>G, p.(Thr671=)

Reasoning

The governing generic ACMG fallback calibration pre-assigns BP4 (supporting) for non-missense variants when SpliceAI max delta < 0.1; observed max delta is 0.004 (single applicable predictor for this synonymous variant per the framework; REVEL/BayesDel/aGVGD not applicable or unavailable). The score sits far below the 0.1 threshold and well outside the 0.1-0.2 gray zone, so BP4 is met at supporting strength with high confidence.

Rule

Generic ACMG fallback PP3/BP4 calibration (output/generic_acmg_classification_rules.md): synonymous/non-missense variants use SpliceAI only; SpliceAI max delta < 0.1 -> BP4 (supporting). Observed max delta 0.004 -> BP4 met (supporting). Source basis: Richards et al. 2015 (PMID 25741868).

Evidence considered

• SpliceAI lookup (spliceai; tool publication Jaganathan et al. 2019, PMID 30661751, Cell 176:535-548) reports max_delta_score 0.004 for NM_006445.3:c.2013A>G.
• Governing PP3/BP4 calibration (output/generic_acmg_classification_rules.md, source basis Richards et al. 2015 PMID 25741868; SpliceAI thresholds as per SVI recommendations reflected in VCEP CSPECs, e.g. ATM, per Walker et al. 2023): for synonymous/non-missense variants SpliceAI max delta < 0.1 -> BP4 (supporting). Observed 0.004 < 0.1, so BP4 is met (supporting).
• Per the governing framework, exactly one mechanism is evaluated per variant and a synonymous variant is scored with SpliceAI only; the missense path (REVEL) is not applicable, and no REVEL/BayesDel/aGVGD scores are available for this variant (REVEL and BayesDel lookups found=false).

Sources

Reasoning

Generic ACMG/AMP 2015 governs (no PRPF8 VCEP/CSPEC; lab non-VCEP PM2 operating threshold AF < 0.1%). To avoid ancestry threshold artifacts, PM2 is evaluated against the highest observed subpopulation frequency: the variant is common in South Asians (SAS AF 0.32009% in gnomAD v2.1; 0.28441% in gnomAD v4.1; 0.14684% in gnomAD-Canada), all far above the 0.1% threshold, and it is observed at 287/1,613,898 alleles overall (v4.1) including 4 homozygotes. Although the overall AF (0.01778-0.03785%) falls below 0.1%, applying PM2 on the overall AF would be a threshold artifact that ignores strong population stratification (the SAS AF is roughly 16-19x the overall AF). The variant is therefore not absent and not at extremely low frequency, and PM2 is not met.

Rule

PM2: Absent in controls or at extremely low frequency in population databases (ACMG/AMP 2015, PMID 25741868); lab non-VCEP operating threshold: AF < 0.1% (evaluated on maximum subpopulation AF to avoid ancestry threshold artifacts)

Evidence considered

• gnomAD v2.1 (key 'gnomad_v2'): SAS subpopulation AF 0.32009% (98/30,616) far exceeds the 0.1% PM2 threshold; overall AF 0.03785%; 2 homozygotes.
• gnomAD v4.1 (key 'gnomad_v4'): SAS subpopulation AF 0.28441% (259/91,066) far exceeds the 0.1% threshold; overall AF 0.01778%; 4 homozygotes. The variant is not rare in South Asians, so PM2 (extremely low frequency) is not met.
• gnomAD-Canada v1.0 (key 'gnomad_canada'): SAS AF 0.14684% (2/1,362) also exceeds the 0.1% threshold; overall AF 0.01086%.
• Evaluation used the maximum subpopulation AF rather than the overall AF to avoid a threshold artifact from population stratification (SAS AF ~16-19x the overall AF); per this consistent approach PM2 is not met.

Sources

Reasoning

Variant NM_006445.3:c.2013A>G is synonymous (NP_006436.3:p.(Thr671=)), so per the governing generic ACMG fallback calibration (no ClinGen CSPEC/VCEP exists for PRPF8) the splice path is the only applicable computational mechanism and SpliceAI is the only applicable predictor. SpliceAI reports max delta score 0.004 (DS_AG 0.004, DS_AL 0.001, DS_DG 0.001, DS_DL 0.001), far below the framework's PP3 threshold of max delta > 0.2 for non-missense variants. No other computational evidence supports pathogenicity: REVEL and BayesDel lookups returned no score for this variant (and REVEL/BayesDel/aGVGD are not applicable to a synonymous variant per the framework), and BayesDel has no verified published threshold available to this pipeline. PP3 is not met.

Rule

Generic ACMG fallback PP3/BP4 calibration (output/generic_acmg_classification_rules.md): non-missense (synonymous) variants use SpliceAI only; SpliceAI max delta > 0.2 -> PP3 (supporting). Observed max delta 0.004 -> PP3 not met. Source basis: Richards et al. 2015 (PMID 25741868).

Evidence considered

• SpliceAI lookup (spliceai; tool publication Jaganathan et al. 2019, PMID 30661751, 'Predicting Splicing from Primary Sequence with Deep Learning', Cell 176:535-548) for NM_006445.3:c.2013A>G (17-1580438-T-C, hg37): DS_AG 0.004, DS_AL 0.001, DS_DG 0.001, DS_DL 0.001, max_delta_score 0.004, variant_consequence 'sequence_variant'.
• Governing PP3/BP4 calibration for this case is the generic ACMG fallback (output/generic_acmg_classification_rules.md, source basis Richards et al. 2015, PMID 25741868): synonymous/non-missense variants are evaluated with SpliceAI only; SpliceAI max delta > 0.2 -> PP3 (supporting). Observed max delta 0.004 is well below 0.2, so no splice-impact computational evidence supports PP3.
• REVEL lookup returned no score (found=false, revel_score=null) and BayesDel lookup returned no score (found=false); per the governing framework a non-missense variant has no REVEL score to use and BayesDel has no verified published threshold available to this pipeline, so neither predictor contributes to PP3 (aGVGD data is likewise absent from the case bundle).

Sources

Reasoning

PP5 under this pipeline's policy applies only to an exact-variant ClinVar expert-panel classification of Pathogenic or Likely pathogenic. The ClinVar submission audit for variation 750629 reports expert_panel_submissions = 0; the sole submission (SCV001073204) is a single clinical laboratory (Labcorp Genetics/Invitae) with review status 'criteria provided, single submitter' (1 star) and an aggregate Benign label. There is no expert-panel assertion of pathogenicity for this variant, so PP5 cannot be met. The variant's aggregate label is Benign, which runs opposite to PP5 in any case.

Rule

ACMG/AMP 2015 PP5 restricted: reputable-source (ClinVar expert panel) Pathogenic/Likely pathogenic for the exact variant only

Evidence considered

• ClinVar variation 750629 submission audit: submissions_total 1, expert_panel_submissions 0, is_expert_panel false, review_status 'criteria provided, single submitter', classification Benign.
• Policy check: ordinary laboratory submissions, aggregate ClinVar labels, and non-expert assertions must never trigger PP5; no qualifying expert-panel classification exists for this exact variant.

Sources

Reasoning

No ClinGen CSPEC/VCEP or local gene-specific framework exists for PRPF8 (vcep_materials.json: cspec found false, gene_vcep_dir null, final_classification_framework source generic_acmg_fallback), so generic ACMG/AMP 2015 (PMID 25741868) governs, with the lab's non-VCEP operating thresholds (BA1 AF > 1%). The highest observed allele frequency for NM_006445.3:c.2013A>G is in the South Asian (SAS) subpopulation: 0.32009% (98/30,616 alleles) in gnomAD v2.1 and 0.28441% (259/91,066 alleles) in gnomAD v4.1; the overall AF is 0.03785% (v2.1, 107/282,696) and 0.01778% (v4.1, 287/1,613,898), and gnomAD-Canada reports 0.01086% (2/18,418). All values, including the population-maximum filtering allele frequency (grpmax FAF 0.26873% v2.1; 0.25596% v4.1), are far below the 1% operating threshold and below the classic 5% BA1 threshold, so BA1 is not met.

Rule

BA1: Allele frequency is too high for a pathogenic variant, standalone benign (ACMG/AMP 2015, PMID 25741868); lab non-VCEP operating threshold: AF > 1% in any gnomAD population

Evidence considered

• gnomAD v2.1 (source_registry key 'gnomad_v2'): total AF 0.000378498 (0.03785%, 107/282,696 alleles, 2 homozygotes); highest subpopulation AF in South Asians 0.00320094 (0.32009%, 98/30,616 alleles, 2 homozygotes); grpmax FAF 0.00268731 (0.26873%). Well below the BA1 operating threshold (>1%) and the classic >5% threshold.
• gnomAD v4.1 (source_registry key 'gnomad_v4'): total AF 0.00017783 (0.01778%, 287/1,613,898 alleles, 4 homozygotes); highest subpopulation AF in South Asians 0.00284409 (0.28441%, 259/91,066 alleles, 4 homozygotes); grpmax FAF 0.00255955 (0.25596%). Well below the BA1 operating threshold.
• gnomAD-Canada v1.0 (source_registry key 'gnomad_canada'): AF 0.00010859 (0.01086%, 2/18,418 alleles, 0 homozygotes); SAS 0.00146843 (0.14684%, 2/1,362); grpmax FAF95 0.00026025 (0.02603%). Below the BA1 threshold.
• No VCEP/CSPEC specification for PRPF8 exists; generic ACMG/AMP 2015 (PMID 25741868) BA1 rule applied with the lab's non-VCEP operating threshold of AF > 1%.

Sources

Reasoning

Generic ACMG/AMP 2015 governs (no PRPF8 VCEP/CSPEC; lab non-VCEP BS1 operating threshold AF > 0.3%). The variant's frequency straddles this threshold depending on dataset and measure: gnomAD v2.1 SAS subpopulation AF is 0.32009% (98/30,616), marginally above 0.3%, but in the larger, current gnomAD v4.1 dataset the SAS AF is 0.28441% (259/91,066), below 0.3%, and the population-maximum filtering allele frequencies (grpmax FAF 0.26873% v2.1; 0.25596% v4.1) are below 0.3% in both datasets; gnomAD-Canada SAS AF is only 0.14684%. Because the only value crossing the 0.3% threshold comes from the older, smaller v2.1 SAS estimate and all robust/maximum-AF measures are below threshold, BS1 is adjudicated not_met, but the borderline SAS frequency (0.28-0.32%, roughly 30-60x the overall AF) warrants human review.

Rule

BS1: Allele frequency is greater than expected for the disorder (ACMG/AMP 2015, PMID 25741868); lab non-VCEP operating threshold: AF > 0.3%

Evidence considered

• gnomAD v2.1 (key 'gnomad_v2'): SAS subpopulation AF 0.00320094 (0.32009%, 98/30,616 alleles) marginally exceeds the 0.3% BS1 operating threshold; overall AF 0.03785% and grpmax FAF 0.26873% are below it.
• gnomAD v4.1 (key 'gnomad_v4'): SAS subpopulation AF 0.00284409 (0.28441%, 259/91,066 alleles) is below the 0.3% threshold; overall AF 0.01778%; grpmax FAF 0.25596%. As the largest and most current dataset (1,613,898 vs 282,696 alleles in v2.1), v4.1 was given precedence, giving not_met.
• gnomAD-Canada v1.0 (key 'gnomad_canada'): overall AF 0.01086%, SAS AF 0.14684%, grpmax FAF95 0.02603% - all well below the 0.3% threshold.
• PRPF8-related disease (autosomal dominant retinitis pigmentosa, RP13) is rare; even the overall AF of 0.018-0.038% is higher than expected for a fully penetrant RP13 allele, which is directionally consistent with a benign allele, but per the lab's deterministic 0.3% operating threshold the criterion is not met on the primary (v4.1) dataset.

Still missing

• The SAS subpopulation AF (0.320% gnomAD v2.1; 0.284% gnomAD v4.1) sits essentially at the 0.3% BS1 operating threshold; whether BS1 should be met depends on which dataset/measure is given precedence (v2.1 SAS AF marginally above vs. v4.1 SAS AF and both grpmax FAFs below). Human review recommended.

Sources

Reasoning

Generic ACMG/AMP 2015 governs (no PRPF8 VCEP/CSPEC). BS2 requires observation of the variant in a healthy adult individual for a dominant disorder with full penetrance expected at an early age. The variant is observed in gnomAD including homozygous individuals (2 homozygotes of 107 alleles in v2.1; 4 homozygotes of 287 alleles in v4.1), which is notable for PRPF8-related autosomal dominant retinitis pigmentosa (RP13), but gnomAD is an unselected general-population cohort whose individuals are not phenotyped as healthy adults, and RP13 has high yet age-dependent penetrance rather than full penetrance expected at an early age. The phenotype requirement of BS2 is therefore not satisfied by the available population data, so BS2 is not met; the homozygous observations are flagged for human review as potential BS2-relevant signal.

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)

Evidence considered

• gnomAD v2.1 (key 'gnomad_v2'): 107/282,696 alleles (AF 0.03785%) including 2 homozygotes in a general-population cohort; observed but without confirmed healthy-adult phenotype.
• gnomAD v4.1 (key 'gnomad_v4'): 287/1,613,898 alleles (AF 0.01778%) including 4 homozygotes; observation in an unselected cohort without RP phenotyping does not meet the BS2 healthy-adult requirement.
• Gene context (gene_context_PRPF8.json): PRPF8 mutations cause autosomal dominant retinitis pigmentosa, an inherited progressive vision loss; penetrance is high but onset/expressivity vary, so 'full penetrance at an early age' is not established for BS2 purposes.

Still missing

• No source provides confirmed unaffected-adult phenotype status for carriers of NM_006445.3:c.2013A>G; gnomAD does not phenotype individuals for retinitis pigmentosa, so the 'healthy adult' requirement of BS2 cannot be verified from the available population data.
• RP13 (PRPF8) penetrance is age-dependent and incomplete at early age, so population-carrier observations alone cannot satisfy BS2's 'full penetrance expected at an early age' condition without explicit unaffected-adult control data.

Sources

Reasoning

BP6 under this pipeline's policy applies only to an exact-variant ClinVar expert-panel classification of Benign or Likely benign. Although the aggregate ClinVar label for variation 750629 is Benign, the submission audit shows expert_panel_submissions = 0: the classification originates from a single ordinary clinical laboratory (Labcorp Genetics/Invitae, SCV001073204) with review status 'criteria provided, single submitter' (1 star). Ordinary laboratory submissions and aggregate ClinVar labels must never trigger BP6, so this criterion is not met. (The laboratory Benign label is not discarded as information, but it cannot be credited as BP6 evidence.)

Rule

ACMG/AMP 2015 BP6 restricted: reputable-source (ClinVar expert panel) Benign/Likely benign for the exact variant only

Evidence considered

• ClinVar variation 750629 submission audit: submissions_total 1, expert_panel_submissions 0, is_expert_panel false, review_status 'criteria provided, single submitter', classification Benign, clinical_laboratories 1.
• Policy check: BP6 requires an exact-variant ClinVar expert-panel Benign/Likely benign classification; the sole 1-star laboratory submission does not qualify.

Sources

Reasoning

BP7 (ACMG/AMP 2015) applies to a synonymous variant with no predicted splice impact whose nucleotide is not highly conserved. The first component is facially satisfied (synonymous p.(Thr671=); SpliceAI max delta 0.004 predicts no splice impact), but BP7 is not applied here because the same single SpliceAI prediction already drives BP4 (met) under the governing framework's pre-assignment for this variant type, and granting BP7 as well would double-count the same prediction across BP4 and BP7, which is prohibited. In addition, no direct conservation score (e.g., PhyloP/GERP) is present in the case bundle to verify the 'nucleotide not highly conserved' clause; only indirect context exists (mid-exon 15 location far from splice consensus; gnomAD v2.1 AF 0.0379% and v4.1 AF 0.0178% with 2-4 homozygotes).

Rule

ACMG/AMP 2015 BP7 (Richards et al., PMID 25741868): synonymous (silent) variant for which splicing prediction algorithms predict no impact to the splice consensus sequence nor the creation of a new splice site AND the nucleotide is not highly conserved. Not applied here: the sole SpliceAI prediction already drives BP4 per the governing framework (double-count guardrail), and no direct conservation score is available.

Evidence considered

• Variant is synonymous: NP_006436.3:p.(Thr671=) / p.(T671=); SpliceAI (spliceai; tool publication Jaganathan et al. 2019, PMID 30661751) predicts no splice impact (max delta 0.004), which facially satisfies the first component of BP7.
• BP7 is withheld to avoid double-counting: the governing generic ACMG fallback calibration (output/generic_acmg_classification_rules.md, source basis Richards et al. 2015 PMID 25741868) assigns this synonymous variant to the SpliceAI-only path, where max delta 0.004 < 0.1 pre-assigns BP4 (supporting); the same prediction cannot also support BP7.
• Indirect context only for the conservation clause: variant lies mid-exon 15 (c.2013; exon 15 spans c.1969-2098), well away from any splice consensus, and is present in gnomAD v2.1 (total AF 0.000378, 107/282696 alleles, 2 homozygotes; gnomad_v2) and gnomAD v4.1 (total AF 0.000178, 287/1613898 alleles, 4 homozygotes; gnomad_v4); no PhyloP/GERP score was retrieved.

Still missing

• Direct nucleotide conservation score (e.g., PhyloP/GERP) for the 'nucleotide is not highly conserved' clause of BP7 is not present in the case bundle.
• The single SpliceAI prediction (max delta 0.004) is already applied to BP4 per the governing framework; applying BP7 additionally would double-count the same prediction across BP4 and BP7, so BP7 is withheld.

Sources

Reasoning

PS2 (de novo, paternity and maternity confirmed) cannot be evaluated: no proband observation with parental testing is present anywhere in the case evidence. The only clinical source, ClinVar VCV000750629, is a single germline clinical-testing submission (SCV001073204, Labcorp) with no criteria codes, no phenotype, and no family/parental testing data. No applicable VCEP/CSpec exists for PRPF8, so the generic ACMG/AMP 2015 definition (PMID 25741868) applies. Absence of any de novo observation leaves the criterion unassessable rather than met or not met.

Rule

PS2: De novo (paternity and maternity confirmed) in a patient with the disease and no family history (strong). Per Richards et al. 2015 (PMID 25741868), generic ACMG/AMP fallback since no PRPF8 VCEP/CSpec exists.

Evidence considered

• ClinVar record VCV000750629 (submission SCV001073204, Labcorp Genetics/Invitae, evaluated 2024-03-09) was reviewed: single germline submission, classification Benign, review status 'criteria provided, single submitter' (1 star), condition 'not provided', no applied criteria codes and no proband, parental, or de novo testing data.
• PMID:28492532 (Sherloc, Nykamp et al. 2017) was the only PMID retrieved; it is a methods/framework refinement paper cited by the ClinVar submission and contains no variant-specific, proband, or de novo evidence.
• Generic ACMG/AMP 2015 framework (PMID 25741868, per source_registry key generic_acmg_combination_rules) applied because no PRPF8 ClinGen CSpec/VCEP specification exists; PS2 rule: de novo (paternity and maternity confirmed) in a patient with the disease and no family history.

Still missing

• No proband family testing data: confirmation of de novo occurrence requires testing of the proband and both biological parents for NM_006445.3:c.2013A>G, with paternity/maternity confirmation.

Sources

Reasoning

No governing VCEP/CSPEC exists for PRPF8 (cspec found: false; gene_vcep_dir: null), so generic ACMG/AMP 2015 applies. PS3 requires well-established functional studies demonstrating a deleterious effect on gene/protein function. The case evidence bundle contains no functional assay data for NM_006445.3:c.2013A>G: literature_pass.json contains a single background-review paper (PMID:28492532, Sherloc criteria-methods paper; variant_mentioned: false, triage_action: background_only), fulltext_map.json is empty, the publications/ directory does not exist, and no minigene splicing, mRNA expression, or protein function assay results are reported for this variant. The only splice-related datum is SpliceAI (max delta score 0.004), an in-silico prediction that does not constitute a well-established functional study and cannot support PS3. Evidence is therefore insufficient to evaluate PS3, and it is returned as not_assessed with the gap identified rather than met or not_met.

Rule

PS3: Well-established in vitro or in vivo functional studies supportive of a damaging effect on the gene or gene product (ACMG/AMP 2015, PMID 25741868)

Evidence considered

• SpliceAI Lookup (source_registry key 'spliceai') was consulted as the only functional-adjacent datum: max delta score 0.004 (DS_AG 0.004, DS_AL 0.001, DS_DG 0.001, DS_DL 0.001), i.e., no predicted splice impact. This is an in-silico prediction and is explicitly insufficient to satisfy PS3, which requires well-established functional studies demonstrating a deleterious effect.
• PMID:28492532 (Sherloc: a comprehensive refinement of the ACMG-AMP variant classification criteria) is the only literature item in the case; it is a criteria-methodology paper (triage theme background_review, variant_mentioned: false, recommended_for_full_review: false) and reports no functional assay for PRPF8 c.2013A>G. It was reviewed to confirm the absence of functional evidence.
• ClinVar record 750629 (single-submitter Benign, 1-star, no criteria-provided signals; evidence_signals.criteria_found empty) was not used as functional evidence, since final classifications from other sources must never be inferred as criterion evidence.
• No VCEP/CSPEC specification for PRPF8 exists; generic ACMG/AMP 2015 (PMID 25741868) PS3 rule was applied.

Still missing

• No well-established functional assay (e.g., minigene splicing assay, mRNA isoform quantification, protein expression/function assay) has been performed on PRPF8 c.2013A>G (p.Thr671=) in any source available to this case; no functional-study publication for this exact variant was retrieved by the literature pass.
• SpliceAI (in-silico, max delta 0.004) is not a functional assay and cannot be used to establish or refute a deleterious effect for PS3.

Sources

Reasoning

PS4 (Richards et al. 2015, PMID:25741868) requires evidence that the prevalence of the variant is significantly increased in affected individuals compared with controls. No case-control or case-series study reporting NM_006445.3:c.2013A>G (p.Thr671=) in affected individuals was identified: the literature pass returned only PMID:28492532 (Sherloc, a criteria-methods paper with no variant data), no phenotype/segregation/case-observation papers were found, and the ClinVar record (variation 750629) carries no condition labels or case counts. gnomAD documents the variant in the general population (v2.1 AF 0.0379%, 107/282,696 alleles, 2 homozygotes; v4.1 AF 0.0178%, 287/1,613,898 alleles, 4 homozygotes), but this is control-side frequency only and cannot be compared against any affected-cohort frequency. The criterion is therefore not_assessed pending any case-control data for this exact variant.

Rule

ACMG/AMP 2015 PS4: prevalence of variant in affected individuals significantly increased vs controls (PMID:25741868)

Evidence considered

• Literature pass identified no case-observation, segregation, or case-control papers for this variant; the only PMID retrieved (28492532, Sherloc criteria refinement) is a methods paper with no variant-level data.
• gnomAD v2.1 (17-1580438-T-C): total AF 0.000378498 (0.0379%, 107/282,696 alleles, 2 homozygotes); highest subpopulation South Asian AF 0.00320094 (98/30,616 alleles, 2 homozygotes); grpmax FAF 0.00268731 - control-side frequency only.
• gnomAD v4.1 (chr17-1677144-T-C): total AF 0.00017783 (0.0178%, 287/1,613,898 alleles, 4 homozygotes); highest subpopulation South Asian AF 0.00284409 (259/91,066 alleles, 4 homozygotes); grpmax FAF 0.00255955 - control-side frequency only.
• No affected-individual allele count or odds-ratio estimate for this variant exists in the evidence bundle, so no PS4 enrichment comparison is possible.

Still missing

• No case-control or case-series study reporting this exact variant (NM_006445.3:c.2013A>G / p.Thr671=) in affected individuals was identified; affected-cohort frequency cannot be compared with the gnomAD population frequency.

Sources

Reasoning

PM6 (assumed de novo without confirmation of paternity and maternity) cannot be evaluated: an assumed de novo event must first be observed in a proband, and no de novo observation or parental testing data exist for NM_006445.3:c.2013A>G in any source in this case. The only sources examined (ClinVar single-submitter record and the Sherloc methods paper) contain no family data. Per the generic ACMG/AMP 2015 fallback, PM6 remains unassessable pending family testing.

Rule

PM6: Assumed de novo, but without confirmation of paternity and maternity (moderate). Per Richards et al. 2015 (PMID 25741868), generic ACMG/AMP fallback since no PRPF8 VCEP/CSpec exists.

Evidence considered

• ClinVar record VCV000750629 reviewed for any assumed-de-novo observation; submission SCV001073204 contains no de novo, parental, or family history data.
• PMID:28492532 (Sherloc) reviewed: background methods paper only, no variant or family observations; nothing supporting an assumed de novo event.
• Generic ACMG/AMP 2015 framework (PMID 25741868) applied: PM6 rule = assumed de novo, but without confirmation of paternity and maternity (moderate).

Still missing

• No de novo observation in a proband: requires testing of proband and both parents; even without paternity/maternity confirmation, an observed de novo event in the proband is a prerequisite that is entirely absent.

Sources

Reasoning

PP1 (co-segregation with disease in multiple affected family members) cannot be evaluated: no segregation analysis, meioses, or affected-family genotype data exist for NM_006445.3:c.2013A>G in any source in this case. The literature pass contains no high-priority papers and no segregation-themed reports; the sole PMID (28492532, Sherloc) has no segregation content; the ClinVar submission carries no family data. PP1 is therefore not_assessed pending segregation data.

Rule

PP1: Co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease (supporting; higher strength may be justified by quantitative segregation analysis). Per Richards et al. 2015 (PMID 25741868), generic ACMG/AMP fallback since no PRPF8 VCEP/CSpec exists.

Evidence considered

• ClinVar record VCV000750629 reviewed for segregation observations; submission SCV001073204 contains no family or segregation data (no criteria codes, condition 'not provided').
• Literature pass (literature_pass.json) contains no high-priority or medium-priority segregation/case-observation papers for this variant; criteria_index is empty; the only paper (PMID:28492532, Sherloc) was reviewed and contains no segregation content.
• Generic ACMG/AMP 2015 framework (PMID 25741868) applied: PP1 rule = co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease (supporting, with strength potentially adjusted by quantitative/Bayesian analysis).

Still missing

• No segregation data: requires genotyping of multiple affected and unaffected family members in a pedigree segregating the relevant phenotype (autosomal dominant PRPF8 disease) to count informative meioses for or against co-segregation.

Sources

Reasoning

PP4 (Richards et al. 2015, PMID:25741868) requires the variant to be found in a case whose phenotype or family history is highly specific for a disease with a single genetic etiology (PRPF8: retinitis pigmentosa 13, autosomal dominant). No clinical case carrying this exact variant was identified: the ClinVar submission (SCV001073204) carries no condition labels, and no variant-specific case report or phenotype study exists in the literature pass. Without any documented patient phenotype attributable to this exact variant, phenotype specificity cannot be evaluated. The criterion is therefore not_assessed pending case-level phenotype data.

Rule

ACMG/AMP 2015 PP4: phenotype or family history highly specific for disease with single genetic etiology (PMID:25741868)

Evidence considered

• ClinVar variation 750629 submission SCV001073204 (Labcorp Genetics/Invitae) was consulted for phenotype attribution and carries empty condition_labels; no phenotype or family-history data are attached.
• Literature pass surfaced no paper reporting this exact variant in a patient with retinitis pigmentosa or any other PRPF8-associated phenotype.
• No phenotype-specificity claim can be grounded for this variant; PP4 is not supported by any source in the evidence bundle.

Still missing

• No case report or cohort describing the phenotype of a patient carrying this exact variant (NM_006445.3:c.2013A>G / p.Thr671=) was identified; the ClinVar submission has no condition labels and the literature pass contains no case-observation paper.

Sources

Reasoning

Generic ACMG/AMP 2015 applies (no PRPF8 VCEP/CSPEC available). BS3 requires well-established functional studies demonstrating no damaging effect on gene/protein function. No such functional study exists for NM_006445.3:c.2013A>G in the case evidence bundle: the literature pass recovered no functional-study paper (only PMID:28492532, a background criteria-methods paper in which the variant is not mentioned), fulltext_map.json is empty, and no wet-lab assay results are present. The SpliceAI max delta score of 0.004 is an in-silico prediction, not a validated functional assay, and cannot substitute for BS3-grade functional evidence. Although the absence of a predicted splice impact and the variant's high population frequency (gnomAD) are consistent with a benign effect, those data belong to other criteria (in-silico BP4/BP7 and population BA1/BS1 groups respectively) and do not constitute BS3 functional evidence. Evidence is insufficient, so BS3 is returned as not_assessed with the gap identified.

Rule

BS3: Well-established in vitro or in vivo functional studies show no damaging effect on protein function or splicing (ACMG/AMP 2015, PMID 25741868)

Evidence considered

• SpliceAI Lookup (source_registry key 'spliceai') was consulted: max delta score 0.004, i.e., no significant splice impact predicted for NM_006445.3:c.2013A>G. This is an in-silico prediction without assay controls and is not a well-established functional study; per the generic ACMG/AMP BS3 definition it cannot be used to apply BS3.
• PMID:28492532 (Sherloc) was reviewed as the sole literature item; it is a criteria-refinement methodology paper with no functional assay data for this variant (variant_mentioned: false, triage background_only), confirming the absence of BS3-grade functional evidence.
• gnomAD v2.1/v4.1 allele frequencies (including SAS subpopulation AF ~0.28-0.32% and homozygous individuals) were noted but are population-frequency evidence for the population criteria group, not functional assay evidence, and were not used for BS3.
• ClinVar 750629 single-submitter Benign classification (1-star, no criteria-provided signals) was not used as functional evidence; other sources' final classifications are not inferred as criterion evidence.

Still missing

• No well-established functional assay (e.g., minigene splicing assay demonstrating normal splicing, or protein function assay demonstrating normal activity) has been performed on PRPF8 c.2013A>G (p.Thr671=); such an assay would be required to apply BS3.
• SpliceAI and other in-silico predictors cannot satisfy BS3, which is reserved for validated functional studies with appropriate assay controls.

Sources

Reasoning

BS4 (non-segregation with disease in affected family members) cannot be evaluated: BS4 requires genotyping of affected family members in pedigrees segregating the disease, and no affected-family non-segregation data exist for NM_006445.3:c.2013A>G in any source in this case. Population allele frequency (gnomAD) and the single-submitter ClinVar Benign label are evaluated under other criteria (BA1/BS1/BP6) and do not constitute BS4 non-segregation evidence. BS4 is therefore not_assessed pending family segregation data.

Rule

BS4: Non-segregation with disease in affected family members (strong). Per Richards et al. 2015 (PMID 25741868), generic ACMG/AMP fallback since no PRPF8 VCEP/CSpec exists.

Evidence considered

• ClinVar record VCV000750629 reviewed for non-segregation observations; none present in submission SCV001073204.
• PMID:28492532 (Sherloc) reviewed: no family, segregation, or non-segregation data for this variant.
• gnomAD frequencies (gnomad_v2: total AF 0.0378%, grpmax FAF 0.27%; gnomad_v4: total AF 0.0178%, grpmax FAF 0.26%; homozygotes present) are population-frequency evidence relevant to BA1/BS1 (other groups), not BS4, which specifically requires non-segregation in affected family members.
• Generic ACMG/AMP 2015 framework (PMID 25741868) applied: BS4 rule = non-segregation with disease in affected family members (strong benign).

Still missing

• No affected-family non-segregation data: requires affected family members (in families with the relevant phenotype) found to lack the variant, i.e. documented absence of segregation in affected relatives.

Sources

Reasoning

BP2 requires observing the variant in trans with a pathogenic variant for a fully dominant disorder/mechanism, or in cis with a pathogenic variant in any inheritance pattern. No such phase observation is documented anywhere in the case bundle: there are no affected-proband or family segregation data, and the only ClinVar record (VCV000750629, Benign, single submitter) reports no co-occurring pathogenic variant and no phase. The evidence needed to evaluate BP2 is entirely absent rather than merely below threshold, so the criterion is not assessed; only proband/family phase information (e.g., a documented pathogenic PRPF8 variant in trans or in cis with c.2013A>G) could support or refute it. Population frequency data (gnomAD v2.1/v4.1 homozygotes, SAS AF ~0.28-0.32%) and ClinVar Benign status address other benign criteria but do not bear on BP2's phase-based precondition.

Rule

Observed in trans with a pathogenic variant for a fully dominant disorder/mechanism; or observed in cis with a pathogenic variant in any inheritance pattern (ACMG/AMP 2015, PMID 25741868)

Evidence considered

• BP2 was evaluated under the generic ACMG/AMP 2015 framework (PMID 25741868) because no ClinGen CSPEC or local gene-specific framework exists for PRPF8 (vcep_materials.json: cspec not found, generic_acmg fallback).
• BP2 requires the variant observed in trans with a pathogenic variant for a fully dominant disorder, or in cis with a pathogenic variant in any inheritance pattern; the case bundle contains no such trans/cis phase observations.
• ClinVar variation 750629 (SCV001073204, Labcorp Genetics/Invitae) classifies the variant Benign with criteria provided by a single submitter, but the submission audit recorded no evidence signals, no co-occurring pathogenic variant, and no phase information.
• PMID:28492532 (Sherloc), the only paper in the case, is a criteria-methodology review; it does not mention c.2013A>G and was not used to support BP2.

Still missing

• No phase (trans/cis) observation of c.2013A>G with any pathogenic PRPF8 variant in an affected proband is documented in the case bundle
• No family segregation or co-occurrence data were provided in the ClinVar submission (SCV001073204) or elsewhere in the case evidence; laboratory proband/family data would be required to assess BP2

Sources

Reasoning

BP5 (Richards et al. 2015, PMID:25741868) requires the variant to be found in a case with a documented alternate molecular basis for disease. No case-level data exist for this variant at all: no case report or cohort describing carriers, no alternate-diagnosis documentation, and no variant-specific literature. Without any reported case, the presence or absence of an alternate molecular cause cannot be evaluated. The criterion is therefore not_assessed pending case-level data.

Rule

ACMG/AMP 2015 BP5: variant found in a case with an alternate molecular basis for disease (PMID:25741868)

Evidence considered

• Evidence bundle contains no case report, cohort, or clinical series involving this exact variant in which an alternate genetic etiology was identified or excluded.
• ClinVar submission audit provides no phenotype or co-occurrence data that could support or refute an alternate molecular basis.

Still missing

• No case-level data documenting the molecular diagnosis of any carrier of this exact variant; an alternate molecular basis for disease has neither been reported nor excluded.