FLCN c.1177-5_1177-3del, p.?

Reasoning

Generic ACMG/AMP (PMID:25741868) applies: no FLCN CSPEC/VCEP exists (vcep_materials.json: 'no cspec url available'), and PVS1 is governed by the ClinGen SVI PVS1 framework (pvs1_generic_framework, PMC6185798). FLCN LoF is an established germline mechanism (pvs1_gene_context: lof_mechanism_supported=true; BHDS, autosomal dominant tumor suppressor). The variant, c.1177-5_1177-3del, is an intronic 3-bp deletion at positions -5 to -3 of the intron 10 acceptor (exon 10i per VariantValidator), so it is not a default-bucket canonical +/-1,2 splice variant and PVS1 requires experimental evidence of a null consequence. That evidence is present: Rossing et al. 2017 (PMID:27734835) mini-gene splicing assay showed the variant converts the wild-type pattern (301 bp, with physiologic partial exon 11 skipping) to a single 177 bp transcript lacking exon 11 entirely. Exon 11 of NM_144997.7 (MANE Select) spans nucleotides 1661-1784 = 124 bp, which is not a multiple of 3; skipping exon 11 is therefore out-of-frame and generates a premature termination codon in exon 12, with NMD predicted (PTC ~c.1301-1432, i.e., >55 nt upstream of the final exon-exon junction at c.1539, not in the last exon). This satisfies the SVI PVS1 path for non-canonical splice variants whose null effect (out-of-frame skipping) is demonstrated by functional RNA assay. No strength downgrade applies: the predicted PTC is not near the C-terminus and NMD is not expected to be escaped. Note the pipeline's pvs1_variant_assessment.json scaffold classifies the variant as 'other/unknown' because it only enumerates default null-variant buckets; that scaffold does not incorporate the functional splicing evidence, which is decisive here.

Rule

PVS1 (very strong): null variant (nonsense, frameshift, canonical +/-1,2 splice, single-/multi-exon deletion) in a gene where LoF is a known mechanism of disease; for splice variants outside the canonical positions, PVS1 requires RNA/functional evidence that the variant produces a null allele (out-of-frame exon skipping / frameshift with predicted NMD) - ClinGen SVI PVS1 framework, PMC6185798; generic ACMG/AMP 2015, PMID:25741868.

Evidence considered

• ClinGen SVI PVS1 recommendations (Tayoun et al. 2018, PMC6185798): non-canonical splice variants qualify for PVS1 when RNA/functional evidence demonstrates a null consequence (out-of-frame exon skipping or frameshifting intron retention); strength is very strong when LOF is a known mechanism and NMD is predicted.
• Rossing et al. 2017 (PMID:27734835), mini-gene splicing assay (pSPL3, FLCN exon 11 with flanking intronic sequence): c.1177-5_1177-3del produced only the 177 bp transcript lacking exon 11, versus wild-type 301 bp (with partial physiologic skipping); authors conclude the variant 'result[s] in aberrant splicing'.
• NM_144997.7 RefSeq annotation (NCBI nuccore, MANE Select): exon 11 = nucleotides 1661-1784 = 124 bp (124 mod 3 = 1); exon 11 skipping is out-of-frame, producing a PTC in exon 12 with NMD predicted (PTC upstream of the final exon-exon junction at c.1539).
• pvs1_gene_context.json: FLCN loss-of-function is a supported germline disease mechanism (BHDS), making the generic PVS1 framework eligible.

Still missing

• No patient-derived (in vivo) RNA/cDNA data are in the bundle; NMD is predicted from the out-of-frame consequence (124-nt exon 11) rather than directly demonstrated.
• PMID:28499369 (Bartram 2017), cited by several ClinVar submissions for functional evidence on this variant, is not available in full text in this case bundle (only background/intro text retrieved); its specific claims could not be independently verified.
• SpliceAI predicts no splice impact (max delta 0.00), discordant with the functional mini-gene assay; RNA-level evidence takes precedence, but the discordance is flagged.

Sources

Reasoning

Two independent, peer-reviewed publications performed validated minigene splicing assays on the exact variant c.1177-5_1177-3del (deletion of the CTC pyrimidines at positions -5 to -3 of intron 10, immediately upstream of the exon 11 acceptor). Bartram et al. 2017 (PMID:28499369): minigene-1 (exons 10-11 context) showed the 3-bp deletion abrogated the native acceptor splice site of exon 11; minigene-2 (exons 10-12 context) showed the mutant produces only the exon-11-skipped transcript (full-length isoform never observed for the mutant), with in-frame fusion of exon 10 to exon 12 generating a frameshift and premature stop codon p.T393Sfs33*; mutant FLCN protein was expressed at significantly reduced levels by Western blot (densitometry n=3, p<0.05). Rossing et al. 2017 (PMID:27734835): pSPL3 minigene (exon 11 and surrounding intronic sequence) in COS-7 cells, WT and mutant in duplicate with Sanger-sequencing verification; WT produced the expected 301 bp transcript plus a weak 177 bp band (baseline partial exon 11 skipping), whereas the mutant produced ONLY the 177 bp exon-11-deleted transcript (complete skipping), and the authors classified the variant as pathogenic (class 5). Both studies include matched WT/mutant controls, sequence-verified constructs, and interrogate a biologically relevant readout: FLCN loss-of-function is a supported germline disease mechanism for Birt-Hogg-Dube syndrome (pvs1_gene_context), and exon 11 skipping yields a premature termination codon expected to lead to a truncated/absent protein. Caveats: exon 11 undergoes partial baseline skipping in WT constructs in both studies (natural alternative splicing), and the assays are minigene-based rather than patient-derived mRNA; however, the mutant consistently eliminates the full-length transcript in two independent laboratories. Per ACMG/AMP 2015 (PMID:25741868) PS3 ('well-established functional studies show a deleterious effect') and the ClinGen SVI functional evidence framework (Brnich et al., Genome Medicine, PMID:31931873) under which validated, replicated assays with appropriate controls support strong evidence, PS3 is met at strong strength. ClinVar submissions (e.g., SCV000271788, SCV004813407, SCV000273659) independently flagged PS3 citing the same primary papers; these were used only as leads and the criterion was adjudicated directly from the primary full texts.

Rule

PS3: Well-established functional studies show a deleterious effect (ACMG/AMP 2015, PMID:25741868); strength calibrated at strong per ClinGen SVI functional evidence framework (Brnich et al., Genome Medicine, PMID:31931873) given two independent validated minigene assays with WT/mutant controls.

Evidence considered

• PMID:28499369 (Bartram 2017): minigene-1 and minigene-2 splicing assays in mouse IMCD cells; the intronic 3-bp deletion abrogates the exon 11 acceptor splice site; mutant yields complete skipping of exon 11 with in-frame exon 10-exon 12 fusion generating frameshift/premature stop p.T393Sfs33*; the full-length wildtype splice isoform was never observed for the mutant; mutant FLCN protein expressed at significantly reduced levels (Western blot, n=3, p<0.05).
• PMID:27734835 (Rossing 2017): pSPL3 mini-gene splicing assay in COS-7 cells (WT and mutant in duplicate, Sanger-verified); c.1177-5_1177-3del results in complete skipping of exon 11 (only the 177 bp exon-11-lacking transcript) versus WT partial skipping; variant classified pathogenic (class 5).
• PMID:25741868 (ACMG/AMP 2015): PS3 defined as 'well-established functional studies show a deleterious effect' (strong).
• PMID:31931873 (ClinGen SVI functional evidence recommendations, Brnich et al., Genome Medicine): validated assays with appropriate controls and replication may support PS3 at strong strength.
• pvs1_gene_context: FLCN loss-of-function is a supported germline disease mechanism, establishing biological relevance of the observed aberrant splicing (exon 11 skipping -> premature termination).

Still missing

• Functional evidence is minigene-based; no patient-derived RNA/cDNA splicing analysis of the endogenous allele is published for this exact variant (Bartram 2017 shows tumor LOH but not mRNA analysis).
• Both minigene studies show baseline partial exon 11 skipping in WT constructs; quantitative isoform ratios (e.g., qRT-PCR) were not reported, so the precise magnitude of the splicing shift is not quantified.

Sources

Reasoning

PM2 (moderate, pathogenic) requires the variant to be absent from - or at extremely low frequency in - large population/control cohorts. The variant is present at extremely low frequency: total AF 0.00121% (3/248,662, gnomAD v2.1) and 0.00043% (7/1,613,730, gnomAD v4.1), maximum subpopulation AF 0.00468% (European Finnish, gnomAD v2.1), 0 homozygotes in both datasets, and absent from gnomAD-Canada v1.0. All observed frequencies are well below this pipeline's documented non-VCEP gnomAD PM2 ceiling (<0.1%) for a rare autosomal dominant disorder. Note that the variant is not literally absent (it is present at ~4-12 per million alleles), and at least one submitting laboratory (Invitae, SCV000632828) treated presence in gnomAD as failing PM2; however, under the pipeline convention for non-VCEP genes, AF <0.1% satisfies PM2, and the ACMG/AMP 2015 definition (variant 'absent from ... a large general population or a control cohort', or below expected carrier frequency) is met in spirit given the observed frequency is orders of magnitude below any plausible BHD carrier frequency. PM2 is met at moderate strength per the governing ACMG/AMP 2015 framework.

Rule

PM2 (moderate, pathogenic): variant absent from large population/control cohorts, or at extremely low frequency; pipeline non-VCEP gnomAD threshold AF <0.1% (ACMG/AMP 2015, PMID 25741868). Met - observed AF 0.00043-0.00121% (max subpopulation 0.00468%), 0 homozygotes, absent from gnomAD-Canada.

Evidence considered

• gnomAD v2.1: total AF 1.20646e-05 (0.00121%, 3/248,662 alleles), max subpopulation AF 4.67683e-05 (Finnish, 0.00468%), 0 homozygotes, grpmax FAF 2.98e-06 (https://gnomad.broadinstitute.org/variant/17-17119819-TGAG-T?dataset=gnomad_r2_1)
• gnomAD v4.1: total AF 4.33778e-06 (0.00043%, 7/1,613,730 alleles), max subpopulation AF 1.5674e-05 (Finnish, 0.00157%), 0 homozygotes, grpmax FAF 1.24e-06 (https://gnomad.broadinstitute.org/variant/chr17-17216505-TGAG-T?dataset=gnomad_r4)
• gnomAD-Canada v1.0 (HostSeq genomes, GRCh38): variant absent (http://127.0.0.1:8001/variant/17-17216505-TGAG-T)
• PM2 criterion definition: variant 'absent from ... a large general population or a control cohort (>1000 individuals)' - moderate evidence for pathogenicity - per ACMG/AMP 2015 (PMID 25741868); pipeline non-VCEP gnomAD PM2 convention is AF <0.1%
• Case-cohort context: variant observed only in affected BHD families (one Danish family with bilateral RCC; previously reported in six BHD families), never reported as a polymorphism (PMID 27734835)

Still missing

• Variant is present (not absent) in gnomAD at 3-7 alleles; PM2 relies on the pipeline's documented non-VCEP AF<0.1% convention rather than strict absence. Race-matched control data specifically for the proband's population were not available.

Sources

Reasoning

PP1 (co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease) is met at supporting strength. FLCN is definitively known to cause Birt-Hogg-Dube syndrome (autosomal dominant; OMIM 135150), satisfying the gene-knowledge requirement. Primary segregation evidence: (1) Rossing 2017 (PMID 27734835, Family 21, Table 1) identified c.1177-5_1177-3del in two affected sisters with bilateral renal cell carcinoma (2 mutation carriers), and states in the Discussion that the variant 'has previously been described in six BHD syndrome families' and 'has been shown to co-segregate with the disease' (citing ref 40 = Kunogi Okura 2013, PMID 24190151); (2) Bartram 2017 (PMID 28499369) reports that family testing of the index patient's kindred 'revealed other members of his family showing features of BHD syndrome to carry the mutation'. Together these establish co-segregation in multiple affected family members across at least two independent families. Strength is limited to supporting because no formal meiosis counts or full pedigree segregation tables are available in the retrievable sources, and the co-segregation claim for prior families rests on a summary statement in Rossing 2017 citing Kunogi Okura 2013 (full text not available in this case bundle for independent verification). Human review recommended to confirm segregation counts from the original family reports before any strength upgrade.

Rule

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

Evidence considered

• Rossing 2017 (PMID 27734835): c.1177-5_1177-3del found in two sisters with bilateral RCC (Family 21, Table 1, 2 mutation carriers); variant previously described in six BHD families and shown to co-segregate with the disease (ref 40: Kunogi Okura 2013, PMID 24190151).
• Bartram 2017 (PMID 28499369): family testing of the index patient's kindred revealed other members with BHD features carrying the mutation.
• ACMG/AMP 2015 (PMID 25741868) PP1 definition consulted as the governing rule.

Still missing

• No countable meioses or segregation LOD scores are reported in the sources available; segregation magnitude cannot be quantified.
• Full text of Kunogi Okura 2013 (PMID 24190151, ref 40 of Rossing 2017), the primary citation for the co-segregation claim in prior families, is not in this case's publication set and was not independently verified.
• Unaffected-carrier and affected-noncarrier statuses within segregating families are not reported, which limits both PP1 confidence and BS4 assessment.

Sources

Reasoning

Carriers of the exact variant consistently present with phenotypes highly specific for Birt-Hogg-Dube syndrome (BHD), a disorder with a single genetic etiology (FLCN). PMID:27734835 (Rossing 2017) states the variant 'has previously been described in six BHD syndrome families with various classical BHD syndrome symptoms (FF, LCs and spontaneous pneumothoraces)' and was found in a Danish family with bilateral RCC. PMID:28499369 (Bartram 2017) reports a family where 'the co-occurrence of chromophobe renal cell carcinoma with familial recurrent pneumothorax made us suspect Birt-Hogg-Dube syndrome', with the variant in the index patient and in family members showing BHD features. Fibrofolliculomas, multiple lung cysts with spontaneous pneumothorax, and chromophobe RCC constitute the highly specific BHD phenotype (single genetic etiology), satisfying generic ACMG/AMP 2015 PP4 at supporting strength. Caveat: the case bundle contains no index-proband clinical phenotype; PP4 rests on documented carrier phenotypes in the literature.

Rule

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

Evidence considered

• PMID:27734835 (Rossing et al. 2017): variant in six prior BHD families with classical symptoms (fibrofolliculomas, lung cysts, spontaneous pneumothorax) plus Danish family with bilateral RCC - phenotype highly specific for BHD
• PMID:28499369 (Bartram et al. 2017): index with chromophobe RCC, familial recurrent pneumothorax, numerous pulmonary cysts; family members with BHD features carry the variant - BHD-specific phenotype
• PMID:19802896 (Lim et al., Hum Mutat 2010): FLCN LSDB; cited in PMID:27734835 as one of the prior reports of this variant in BHD families (variant-level content not directly verifiable here; full text not in this case)
• PMID:25741868 (Richards et al. 2015): rule - PP4 phenotype or family history highly specific for a disease with a single genetic etiology; BHD (OMIM 135150) is caused solely by FLCN mutations
• clinvar (VCV000228691): conditions include Birt-Hogg-Dube syndrome, Pneumothorax - familial, FLCN-related disorder - consistent phenotype spectrum; used as context only

Still missing

• Index case's own clinical phenotype/family history not provided in the case bundle; PP4 derived from literature carrier phenotypes
• Fibrofolliculoma status not documented for all carrier families (e.g., Bartram 2017 index had no classical skin phenotype)

Sources

Reasoning

Generic ACMG/AMP 2015 (PMID 25741868) PS4 requires the prevalence of the variant in affected individuals to be significantly increased compared with controls, i.e. a case-control study with statistical support. For FLCN c.1177-5_1177-3del only case-level observations exist: PMID:27734835 (Rossing 2017) reports the variant in a Danish family (two sisters with bilateral renal cell carcinoma, Table 1, Family 21) and states it 'has previously been described in six BHD syndrome families with various classical BHD syndrome symptoms (FF, LCs and spontaneous pneumothoraces)' (refs 32, 36, 39-41 therein = Furuya 2016, Kunogi 2010, Johannesma 2014, Kunogi Okura 2013, Lim 2010), and PMID:28499369 (Bartram 2017) reports an additional independent family with chromophobe RCC, familial recurrent pneumothorax and lung cysts. The variant is essentially absent from population controls (gnomAD v4.1 AF 4.34e-06, 7/1,613,730 alleles, 0 homozygotes; gnomAD v2.1 AF 1.21e-05, 3/248,662 alleles; gnomAD-Canada absent). This combination (multiple unrelated affected families, near-absence in controls) is strong observational support, but no formal case-control study with odds ratio/statistical significance was performed for this exact variant, so the strong PS4 threshold is not formally demonstrated under the generic framework; a lower-strength PS4 application would require a governing VCEP/specification, which does not exist for FLCN (no CSPEC found).

Rule

ACMG/AMP 2015 (PMID 25741868) PS4 - prevalence of variant in affected individuals significantly increased vs controls (strong); generic framework, no FLCN VCEP/CSPEC

Evidence considered

• PMID:27734835 (Rossing et al., J Hum Genet 2017): variant identified in 1 Danish family, 2 carriers, bilateral RCC; variant previously described in six BHD syndrome families with classical symptoms (FF, LCs, spontaneous pneumothorax); classified class 5; no control cohort screened
• PMID:28499369 (Bartram et al., BMC Cancer 2017): independent family; index with chromophobe RCC + familial recurrent pneumothorax + lung cysts; heterozygous c.1177-5_-3delCTC; other family members with BHD features carry the variant
• gnomad_v4 (v4.1): total AF 4.33778e-06 (7/1,613,730 alleles), 0 homozygotes, grpmax FAF 1.24e-06 - near-absence in general-population controls
• gnomad_v2 (v2.1): total AF 1.20646e-05 (3/248,662 alleles), 0 homozygotes, grpmax FAF 2.98e-06
• clinvar (VCV000228691): 17 submissions (10 Pathogenic, 6 Likely pathogenic, 1 VUS), 14 usable, 8 criterion-lead; used only as submission context, never as PS4 evidence
• PMID:25741868 (Richards et al., Genet Med 2015): governing rule - PS4 requires significantly increased prevalence of the variant in affected individuals vs controls; guideline permits judgment for lower strength (PS4 Note 2) but no specification adopted here

Still missing

• No case-control study (BHD-affected cohort vs matched controls) with statistical significance testing for this exact variant
• Per-cohort proband counts for the six previously described BHD families (Furuya 2016; Kunogi 2010; Johannesma 2014; Kunogi Okura 2013; Lim 2010) could not be verified from full text (full texts not present in this case's publications directory; only attested via PMID:27734835)
• No VCEP/CSPEC PS4 strength calibration exists for FLCN to justify tiered (supporting/moderate) application

Sources

Reasoning

PM4 requires a protein length change resulting from an in-frame insertion/deletion in a non-repeat region or a stop-loss variant (ACMG/AMP 2015, PMID:25741868). At the DNA level the variant is intronic (p.?), and the demonstrated RNA consequence is not an in-frame change: exon 11 skipping removes 124 nt (NM_144997.7 nt 1661-1784), a non-multiple of 3, so the transcript is out-of-frame and the allele behaves as a frameshift/null allele (captured under PVS1), not as an in-frame deletion of protein sequence. PM4 is therefore not met. A single ClinVar assertion (LabCorp SCV004813407) describes the exon 10-to-exon 12 fusion as 'in frame', but this is contradicted by the transcript annotation and by the Rossing mini-gene band arithmetic (301-177 = 124 nt, non-multiple of 3).

Rule

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

Evidence considered

• ACMG/AMP 2015 (PMID:25741868): PM4 applies to in-frame insertions/deletions in non-repeat regions or stop-loss variants; out-of-frame/frameshift consequences are excluded by definition.
• Rossing et al. 2017 (PMID:27734835) + NM_144997.7 annotation: exon 11 skipping removes 124 nt (non-multiple of 3) -> out-of-frame transcript, i.e., a frameshift/null consequence rather than an in-frame protein length change.
• pvs1_variant_assessment.json: protein consequence recorded as p.? (no in-frame protein length change established at the DNA level).

Still missing

• Direct confirmation of the exact protein product (e.g., protein blot or in vivo cDNA sequencing) is not available; however, the out-of-frame determination from the 124-nt exon 11 is unambiguous, so this gap does not change the PM4 call.

Sources

Reasoning

No computational evidence supports a deleterious effect on the gene/gene product. SpliceAI (the only calibrated splice-impact predictor available) reports max delta score = 0.00 (sub-scores pangolin_SG 0.04, pangolin_SL -0.13), far below the published high-recall splice-impact threshold of 0.2 (Jaganathan et al. 2019, PMID 30661751), so no splice impact is predicted for this intronic 3-bp deletion at the exon 11 acceptor (c.1177-5_1177-3del). The missense in-silico path (REVEL/BayesDel/aGVGD) is not applicable: the variant is an intronic deletion, not a missense SNV, and prefetch explicitly skipped these predictors ('not snv or no coords'); no scores were computed. No FLCN VCEP/CSPEC or gene-specific PP3 lookup spreadsheet exists (vcep_materials: 'no cspec url available'), so generic ACMG/AMP 2015 (PMID 25741868) applies. PP3 is therefore not met on either computational sub-path.

Evidence considered

• SpliceAI (registry key: spliceai) max delta score = 0.00 (sub-scores pangolin_SG 0.04, pangolin_SL -0.13); below the published high-recall splice-impact threshold of 0.2 (Jaganathan et al. 2019, PMID 30661751, 'Predicting Splicing from Primary Sequence with Deep Learning', Cell) — no significant splice impact predicted.
• REVEL/BayesDel/aGVGD: not computed — prefetch skipped both ('not snv or no coords'); variant is an intronic 3-bp deletion (c.1177-5_1177-3del), not a missense SNV. BayesDel additionally has no verified published threshold with a nameable publication available to this pipeline and any such score would be treated as insufficiently calibrated for PP3/BP4.
• No FLCN VCEP/CSPEC or gene-specific PP3/BP4 lookup exists (vcep_materials.json: 'no cspec url available'); generic ACMG/AMP 2015 (PMID 25741868) governs.

Sources

Reasoning

BA1 (stand-alone benign) requires an allele frequency greater than expected for the disorder (ACMG/AMP 2015: >5% for stand-alone support; this pipeline's documented non-VCEP gnomAD BA1 convention is >1%). The variant is observed at total AF 0.00121% (3/248,662 alleles, gnomAD v2.1) and 0.00043% (7/1,613,730 alleles, gnomAD v4.1), with a maximum subpopulation AF of 0.00468% (European Finnish, 1/21,382, gnomAD v2.1). Every observed frequency is roughly 200-1000-fold below any BA1 threshold, so no threshold-artifact concern applies in the direction that would trigger BA1. BA1 is not met.

Rule

BA1 (stand-alone benign): AF >5% in a large population database per ACMG/AMP 2015 (PMID 25741868); pipeline non-VCEP gnomAD convention BA1 >1%. Not met - maximum observed AF is 0.00468% (gnomAD v2.1 Finnish), ~200-1000x below threshold.

Evidence considered

• gnomAD v2.1: total AF 1.20646e-05 (0.00121%, 3/248,662 alleles), max subpopulation AF 4.67683e-05 (Finnish, 0.00468%), 0 homozygotes, grpmax FAF 2.98e-06 (https://gnomad.broadinstitute.org/variant/17-17119819-TGAG-T?dataset=gnomad_r2_1)
• gnomAD v4.1: total AF 4.33778e-06 (0.00043%, 7/1,613,730 alleles), max subpopulation AF 1.5674e-05 (Finnish, 0.00157%), 0 homozygotes, grpmax FAF 1.24e-06 (https://gnomad.broadinstitute.org/variant/chr17-17216505-TGAG-T?dataset=gnomad_r4)
• BA1 criterion definition: allele frequency greater than expected for disorder; 'if over 5% then it is considered as stand-alone support (BA1)' per ACMG/AMP 2015 standards and guidelines (PMID 25741868); pipeline non-VCEP gnomAD BA1 convention is >1%

Sources

Reasoning

BS1 (strong benign) requires an allele frequency in a control/general population greater than expected for the disorder. BHD syndrome is a rare autosomal dominant disorder - only a few hundred families have been described worldwide (PMID 35176117; ~200 families as of the Danish study, PMID 27734835) - so the expected disease-allele frequency is far below 0.1%. The observed maximum AF is 0.00468% (gnomAD v2.1 Finnish subpopulation; total AF 0.00043-0.00121%), which is ~60-700x below this pipeline's documented non-VCEP BS1 threshold (>0.3%) and far below any plausible expected frequency for BHD. BS1 is not met.

Rule

BS1 (strong benign): AF greater than expected for the disorder per ACMG/AMP 2015 (PMID 25741868); pipeline non-VCEP gnomAD convention BS1 >0.3%. Not met - max observed AF 0.00468% is far below both the 0.3% convention and any expected BHD allele frequency.

Evidence considered

• gnomAD v2.1: total AF 1.20646e-05 (0.00121%, 3/248,662 alleles), max subpopulation AF 4.67683e-05 (Finnish, 0.00468%), 0 homozygotes (https://gnomad.broadinstitute.org/variant/17-17119819-TGAG-T?dataset=gnomad_r2_1)
• gnomAD v4.1: total AF 4.33778e-06 (0.00043%, 7/1,613,730 alleles), max subpopulation AF 1.5674e-05 (Finnish, 0.00157%), 0 homozygotes (https://gnomad.broadinstitute.org/variant/chr17-17216505-TGAG-T?dataset=gnomad_r4)
• BHD rarity/penetrance context: 'a rare autosomal dominant disorder with variable penetrance... Only a few hundreds of families have so far been described in the literature' (PMID 35176117, abstract); '~ 200 families with BHD syndrome caused by a pathogenic FLCN mutation have been reported worldwide' (PMID 27734835)
• BS1 criterion definition: allele frequency 'greater than expected for the disorder' is strong support for benign interpretation for a rare Mendelian disorder per ACMG/AMP 2015 (PMID 25741868); pipeline non-VCEP gnomAD BS1 convention is >0.3%

Sources

Reasoning

BS2 (strong benign) requires observation of the variant in a well-documented healthy adult with 'full penetrance expected at an early age'. Two conditions fail for this variant. First, BHD syndrome is an adult-onset disorder with variable, age-dependent penetrance (fibrofolliculomas typically appear after age 20; renal cancer and pneumothorax are incompletely penetrant; PMID 27734835, PMID 35176117), so the early-age full-penetrance precondition of BS2 is not satisfied. Second, the variant is present in gnomAD (7 alleles in v4.1, 3 in v2.1, 0 homozygotes in either), but gnomAD is not a healthy cohort and does not provide well-documented unaffected-adult status; the ACMG/AMP 2015 framework explicitly cautions that population databases 'cannot be assumed to include only healthy individuals' (PMID 25741868). BS2 is not met.

Rule

BS2 (strong benign): variant observed in a well-documented healthy adult for a disorder fully penetrant at an early age (ACMG/AMP 2015, PMID 25741868). Not met - BHD is adult-onset with variable penetrance and no documented unaffected adult carriers.

Evidence considered

• gnomAD v2.1: 3/248,662 alleles, 0 homozygotes; gnomAD v4.1: 7/1,613,730 alleles, 0 homozygotes (https://gnomad.broadinstitute.org/variant/17-17119819-TGAG-T?dataset=gnomad_r2_1; https://gnomad.broadinstitute.org/variant/chr17-17216505-TGAG-T?dataset=gnomad_r4)
• BHD is adult-onset with variable penetrance: fibrofolliculomas 'usually do not appear before the age of 20' and phenotype is highly variable (PMID 27734835); 'rare autosomal dominant disorder with variable penetrance' (PMID 35176117)
• BS2 criterion definition: 'if the disease under investigation is fully penetrant at an early age and the variant is observed in a well-documented healthy adult individual' - strong evidence for benign; population databases 'cannot be assumed to include only healthy individuals' per ACMG/AMP 2015 (PMID 25741868)

Still missing

• No clinically documented unaffected adult carriers of c.1177-5_1177-3del are available in this case bundle; gnomAD carrier observations cannot establish healthy-adult status for an adult-onset, variably penetrant disorder.

Sources

Reasoning

BS3 requires well-established functional studies showing no damaging effect on protein function or splicing. No such evidence exists for this variant; to the contrary, two independent validated minigene splicing assays (PMID:28499369, PMID:27734835) demonstrate a clear deleterious effect on splicing (complete skipping of exon 11, loss of the full-length transcript, frameshift/premature stop p.T393Sfs33*). The only 'no impact' signal in the case evidence is SpliceAI (max delta score = 0.00, source key: spliceai), which is an in silico prediction, not a functional assay, and is directly contradicted by the wet-lab splicing data (SpliceAI's model window is insensitive to this non-canonical polypyrimidine-tract deletion). BS3 is therefore not met.

Rule

BS3: Well-established functional studies show no damaging effect on protein function or splicing (ACMG/AMP 2015, PMID:25741868). Not met because functional studies show a damaging effect.

Evidence considered

• PMID:28499369 (Bartram 2017): functional assay demonstrates a deleterious effect on splicing (exon 11 skipping, p.T393Sfs33*, reduced protein expression) - directly contradicts BS3.
• PMID:27734835 (Rossing 2017): functional assay demonstrates complete exon 11 skipping - directly contradicts BS3.
• spliceai: max delta score 0.00 (in silico only; not a functional assay and contradicted by validated minigene assays; insufficient and unreliable for BS3).

Still missing

• No functional assay evidence supports a benign/no-effect interpretation for this variant; the functional data uniformly show a deleterious splicing effect.

Sources

Reasoning

BS4 (lack of segregation in affected family members) is not met: no source in the case bundle reports an affected family member who does not carry c.1177-5_1177-3del, nor any pedigree showing non-segregation. The available family data point in the opposite direction (co-segregation: two affected sisters in Rossing 2017; affected relatives carrying the mutation in Bartram 2017). gnomAD population frequencies (v2.1 AF 1.21e-5; v4.1 AF 4.34e-6, 0 homozygotes) pertain to BS1/BA1 population criteria, not to BS4 segregation evidence, and are not applied here. A definitive BS4 assessment would require systematic testing of multiple affected family members with documented non-carrier affected individuals, which is absent.

Rule

BS4: Lack of segregation in affected family members - ACMG/AMP 2015 (PMID 25741868), strong (BS)

Evidence considered

• Rossing 2017 (PMID 27734835) and Bartram 2017 (PMID 28499369) contain no non-segregation observations; both instead document affected relatives carrying the variant.

Still missing

• No comprehensive family testing dataset (multiple affected members genotyped) is available to detect non-segregation events.
• Penetrance and age-related phenotype onset (BHD features often appear in adulthood) mean that absence of phenotype in young carriers would not constitute BS4 evidence; no such data exist in the bundle.

Sources

Reasoning

BP2 (ACMG/AMP 2015, PMID 25741868) requires a specific observation: the variant seen in trans with a pathogenic variant in a fully dominant disorder, or seen in cis with a pathogenic variant in any inheritance pattern. No such observation exists in this case. The ClinVar record VCV000228691 aggregates 17 submissions (14 usable for classification, 8 with criterion-bearing text); none report the variant in trans or in cis with a second pathogenic FLCN variant, and none report compound heterozygosity. The two publications with the exact variant in full text (Rossing 2017 PMID 27734835; Bartram 2017 PMID 28499369) describe it as the sole germline heterozygous FLCN alteration in autosomal dominant BHD families with no second variant or phase (cis/trans) information. Bartram's tumor-tissue 'homozygous' signal is somatic loss-of-heterozygosity (MLPA-confirmed deletion of the second allele in tumor), which is not a germline cis/trans observation and does not engage BP2. The observation required by BP2 is therefore absent, and the criterion is not met.

Rule

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 (ACMG/AMP 2015, PMID 25741868; generic ACMG fallback — no FLCN CSPEC/VCEP available)

Evidence considered

• ClinVar VCV000228691 (17 submissions, 14 usable for classification, 8 criterion-lead submissions with evidence signals PP1/PS3/PS4/PM2/PP3): no submission reports the variant observed in trans with a pathogenic variant (dominant-disease BP2) or in cis with a pathogenic variant (any-inheritance BP2); no compound-heterozygosity or phase statements.
• PMID:27734835 (Rossing 2017, full text): c.1177-5_1177-3del identified in one family with RCC as the germline FLCN alteration; no second FLCN variant or cis/trans phase data reported.
• PMID:28499369 (Bartram 2017, full text): germline heterozygous c.1177-5_-3delCTC with no second germline FLCN variant; tumor-tissue homozygosity attributed to somatic loss-of-heterozygosity (MLPA-confirmed), which is not a germline cis/trans observation.
• ACMG/AMP 2015 rule text (PMID 25741868): 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'; the required observation is absent in all sources.

Still missing

• No parental or phasing (cis/trans) testing results are available for any reported proband; the BP2 determination rests on the absence of phase observations rather than a documented negative phasing result.
• If a second pathogenic FLCN variant were ever identified and phased in cis or in trans (fully dominant context) with c.1177-5_1177-3del, BP2 would require reassessment.

Sources

Reasoning

BP3 (benign supporting) applies to in-frame indels located in a repetitive region without known function (ACMG/AMP 2015, PMID:25741868). The criterion's premises fail here: (1) the variant is an intronic deletion (c.1177-5_1177-3del, exon 10i) and is not an in-frame coding indel at the protein level; (2) the demonstrated splicing consequence (complete exon 11 skipping, 124 nt) is out-of-frame, not in-frame; and (3) no repeat-region annotation for the deleted genomic interval (GRCh37 chr17:g.17119825_17119827, deleted 'GAG') is available anywhere in the evidence bundle (no repeat-masking source in source_registry.json), so a repeat-context claim could not be substantiated even if the in-frame premise held. BP3 is not met; SpliceAI (max delta 0.00) is insufficient alone to conclude a benign splice effect for BP3.

Rule

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

Evidence considered

• ACMG/AMP 2015 (PMID:25741868): BP3 applies to in-frame indels in repeats without known function; the variant is intronic with an out-of-frame demonstrated consequence, so the in-frame premise is not satisfied.
• SpliceAI Lookup (spliceai): max delta score 0.00 (no significant splice impact predicted) - not a repeat-context observation and superseded for consequence determination by the functional mini-gene assay.
• pvs1_variant_assessment.json: variant bucket 'other', consequence class 'unknown' at DNA level; no repeat-region annotation supplied.

Still missing

• Repeat-masking/RepeatMasker annotation for chr17:g.17119825_17119827 (GRCh37) is not present in the evidence bundle; repeat-region context therefore cannot be confirmed.
• No functional data demonstrating a benign (no-effect) outcome for this allele exist; the available functional data demonstrate aberrant splicing.

Sources

Reasoning

PS2 (de novo with maternity and paternity confirmed) cannot be evaluated: the case bundle contains no proband-level parental testing data for this variant. No ClinVar submission, literature source, or prefetch artifact documents a de novo occurrence of NM_144997.7:c.1177-5_1177-3del in any proband, and none of the reviewed full-text papers (PMID 27734835, 28499369) report parental testing or identity confirmation. The variant is instead recurrent across multiple unrelated BHD families (Danish, Japanese, German, Chinese reports), which is not consistent with a de novo pattern at the population level, but that observation alone neither meets nor refutes PS2 for any specific proband. Governing framework: generic ACMG/AMP 2015 (PMID 25741868), applied because no FLCN CSPEC/VCEP exists (cspec.found=false; framework_mode=generic_acmg).

Rule

PS2: De novo (maternity and paternity confirmed) - ACMG/AMP 2015 (PMID 25741868), strong (PS)

Evidence considered

• ACMG/AMP 2015 joint consensus guideline (PMID 25741868) defines PS2 as a de novo variant with both parental samples tested and biological parentage confirmed by identity testing; no such observation exists for this variant in the bundle.

Still missing

• No proband-parent trio testing data available for this variant; parental genotypes (both parents negative) and identity/maternity-paternity confirmation are required to apply PS2.
• No literature report of a de novo occurrence of c.1177-5_1177-3del was identified in the case bundle.

Sources

Reasoning

PM6 (assumed de novo, without confirmation of maternity and paternity) cannot be evaluated for the same reason as PS2: no proband is described in the case bundle, and no source reports parental samples testing negative for this variant (with or without identity confirmation). PM6 is the fallback de novo criterion and requires at least an assumed-de-novo observation, which is absent. Governing framework: generic ACMG/AMP 2015 (PMID 25741868); no FLCN-specific VCEP/CSPEC specification exists.

Rule

PM6: Assumed de novo, but without confirmation of maternity and paternity - ACMG/AMP 2015 (PMID 25741868), moderate (PM)

Evidence considered

• ACMG/AMP 2015 guideline (PMID 25741868) defines PM6 as assumed de novo without confirmation of maternity and paternity; no de novo observation of c.1177-5_1177-3del is documented in any bundle source.

Still missing

• No parental testing data of any kind (confirmed or assumed de novo) for this variant in any proband.
• Proband phenotype consistency and family history consistent with de novo inheritance cannot be assessed without proband-level clinical data.

Sources

Reasoning

The only available computational line — SpliceAI max delta score = 0.00 (below the published 0.2 threshold, Jaganathan et al. 2019, PMID 30661751) — is consistent with no splice impact, but ACMG/AMP 2015 BP4 requires multiple lines of computational evidence suggesting no impact on gene/gene product (conservation, evolutionary, splicing impact, etc.). No second independent computational line is present in the evidence bundle: REVEL/BayesDel/aGVGD are not applicable to this intronic deletion (not a missense SNV; prefetch skipped them), and no conservation/evolutionary scores (e.g., PhyloP) or additional splice-predictor results (e.g., MaxEntScan/NNSPLICE) for the deleted intronic nucleotides were collected. Per the governing guideline (PMID 25741868), in-silico predictions from different tools are 'considered as a single piece of evidence in sequence interpretation as opposed to independent pieces of evidence', reinforcing that a single no-impact call is insufficient to meet BP4. Evidence is therefore insufficient for a BP4 determination; the criterion is returned as not_assessed rather than met.

Evidence considered

• SpliceAI (registry key: spliceai) max delta score = 0.00 (sub-scores pangolin_SG 0.04, pangolin_SL -0.13); below the published 0.2 splice-impact threshold (Jaganathan et al. 2019, PMID 30661751) — no splice impact predicted; single line only.
• ACMG/AMP 2015 (PMID 25741868): BP4 is defined as multiple lines of computational evidence suggesting no impact on gene/gene product; the guideline states the combination of predictions from different in-silico tools 'are considered as a single piece of evidence in sequence interpretation as opposed to independent pieces of evidence' and that such predictions should not be the sole source of evidence for a clinical assertion.
• No second independent computational line (conservation/evolutionary scores, additional splice predictors such as MaxEntScan/NNSPLICE, or PhyloP for the deleted intronic nucleotides) is present in the evidence bundle.

Still missing

• Only a single computational line is available (SpliceAI no-impact, max delta 0.00); BP4 under ACMG/AMP 2015 requires multiple lines of computational evidence suggesting no impact — no second independent in-silico result (e.g., MaxEntScan/NNSPLICE splice prediction, PhyloP/conservation for the deleted intronic nucleotides) is present in the evidence bundle.
• Missense in-silico predictors (REVEL/BayesDel/aGVGD) are not applicable to this intronic 3-bp deletion and were not computed (prefetch: 'not snv or no coords').

Sources

Reasoning

BP5 (variant found in a case with an alternate molecular basis for disease) requires evidence of a different genetic/etiologic explanation in a carrier. No such evidence exists anywhere in the case evidence bundle: no alternate pathogenic variant, no alternative diagnosis, and no proband-level molecular workup data are reported. BHD is autosomal dominant and the variant co-segregates with disease (attested in PMID:27734835 ref 40), so an alternate molecular basis is not suggested by the data, but absence of proband-level workup means the criterion cannot be affirmatively assessed. Marked not_assessed pending proband-level data.

Rule

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

Evidence considered

• PMID:27734835 (Rossing et al. 2017): variant co-segregates with disease in BHD families (ref 40 therein); no alternate molecular basis reported in the 143-patient Danish cohort
• PMID:28499369 (Bartram et al. 2017): tumor LOH of the FLCN locus in the index patient (Knudson second hit), not an alternate molecular basis
• clinvar (VCV000228691): submitted conditions (Birt-Hogg-Dube syndrome, FLCN-related disorder, Pneumothorax - familial, etc.) show no alternate-etiology assertion
• PMID:25741868 (Richards et al. 2015): rule - BP5 variant found in a case with an alternate molecular basis for disease

Still missing

• No proband-level data on alternate molecular basis (e.g., presence of another pathogenic variant or different genetic etiology) for any carrier of this variant

Sources