MPL c.1513A>T, p.(Ser505Cys)

Reasoning

No gene-specific specification exists for MPL (no CSPEC; vcep_materials.final_classification_framework is the generic ACMG fallback and vcep_materials.domain_tables contains no entry for MPL), so generic ACMG/AMP PM1 was applied and the domain boundaries were derived from the primary literature rather than a VCEP domain table. The variant lies at codon 505 within the single transmembrane domain (TMD) encoded by exon 10 of MPL (VariantValidator assigns c.1513A>T to exon 10), and that domain is explicitly documented as a mutational hotspot for MPL disease: 'The single transmembrane domain (TMD) of the human thrombopoietin receptor (TpoR/... [MPL] protein), encoded by exon 10 of the MPL gene, is a hotspot for somatic mutations associated with myeloproliferative neoplasms (MPNs)' (PMID:31697803). Position 505 is one of only two canonical driver positions in that domain (S505N and W515L/K/R/A), and the same source reports S505N as 'often found as a germline mutation in hereditary thrombocythemia' with a patient annotated 'S505NGer', showing that this residue is hotspot-relevant to germline, gain-of-function MPL disease and not only to somatic neoplasia; consistent with this, OncoKB's gene annotation assigns germline gain-of-function missense variants in this receptor to autosomal dominant thrombocythemia. The criterion is applied at supporting rather than the generic moderate strength because the hotspot documentation derives predominantly from somatic MPN (tumour-context) sequencing, because the functional effect at codon 505 is amino-acid specific (only the Asn substitution is constitutively activating per PMID:31697803, and S505C alone is non-activating per PMID:34845187), and because no benign-variation data for the domain were available to confirm the 'without benign variation' element.

Rule

ACMG/AMP 2015 PM1 (mutational hotspot and/or critical, well-established functional domain without benign variation), applied at supporting strength; generic fallback because no MPL CSPEC/VCEP specification or domain table is available in this case

Evidence considered

• PMID:31697803 (abstract and Introduction): verbatim, 'The single transmembrane domain (TMD) of the human thrombopoietin receptor (TpoR/myeloproliferative leukemia [MPL] protein), encoded by exon 10 of the MPL gene, is a hotspot for somatic mutations associated with myeloproliferative neoplasms (MPNs)' and 'Its transmembrane domain (TMD), encoded by exon 10 of the MPL gene, is a hotspot for mutations associated with myeloproliferative neoplasms (MPNs)'.
• PMID:31697803: 'canonical' MPL exon 10 driver mutations are limited to W515L/K/R/A or S505N, i.e. position 505 is one of the two hotspot residues in the domain; the same paper notes S505N is 'often found as a germline mutation in hereditary thrombocythemia' and reports a patient annotated 'S505NGer' (Figure 2D).
• PMID:31697803: enrichment in the TMD saturation screen was position-restricted, with activating substitutions at S493, L498, V501, L502, L508, S505 and W515 - i.e. the domain contains multiple independent pathogenic positions, supporting a domain-level hotspot rather than an isolated residue.
• PMID:34845187: 'The single pass helical transmembrane domain of the receptor anchors it to the cell surface and the importance of this domain in controlling the activation of the receptor is evident from both biochemical studies and the identification of several important activating mutations in this region', and the variant under assessment is at codon 505 in that domain.
• OncoKB (gene-level annotation): 'Germline gain-of-function mutations are associated with autosomal dominant thrombocythemia 2 ... whereas germline loss-of-function mutations are associated with autosomal recessive congenital amegakaryocytic thrombocytopenia', establishing that the TMD gain-of-function hotspot is a disease-relevant functional domain for germline MPL interpretation.
• VariantValidator mapping of NM_005373.2:c.1513A>T places the variant in exon 10 of MPL (NP_005364.1:p.(Ser505Cys)), consistent with the domain assignment.
• The CancerHotspots lookup for this case queried 'MPL S505' and returned no statistically significant hotspot row; this database result was noted but not treated as decisive against the literature-documented domain hotspot.

Still missing

• No VCEP/CSPEC specification and no entry in vcep_materials.domain_tables exist for MPL, so no authoritative critical-domain list could be checked against residue 505; the domain call rests on the primary literature.
• The 'without benign variation' element of PM1 could not be verified: no gnomAD missense-constraint metric or catalogue of benign variation within the MPL TMD was available in this bundle.
• The observed CancerHotspots lookup for MPL S505 returned no statistically significant hotspot row, which is discordant with the residue being a canonical MPN driver position; the hotspot determination here therefore relies on the published literature and not on the somatic hotspot database check.
• Case phenotype/indication is not recorded in the bundle; this PM1 application presumes the phenotype under assessment is MPL gain-of-function (thrombocythemia / MPN predisposition). Under a loss-of-function congenital amegakaryocytic thrombocytopenia framing the TMD hotspot would not be the relevant domain.

Sources

Reasoning

PM2 (variant absent or at extremely low frequency in population databases) is met at supporting strength. NM_005373.2:c.1513A>T is absent from gnomAD v2.1 (GRCh37, 1-43814978-A-T) and absent from gnomAD v4.1 (GRCh38, chr1-43349307-A-T) - observed allele frequency 0 in two independent, large, well-powered releases - and it has no ClinVar record. Zero frequency is below the 0.1% threshold applied to non-VCEP genes for PM2 in this review, and it satisfies the generic ACMG/AMP 2015 rarity criterion (Richards et al. 2015, PMID:25741868). Supporting strength rather than moderate is used because PM2 is routinely applied at supporting level and because the gene-disease relationship for MPL includes a rare recessive germline phenotype (congenital amegakaryocytic thrombocytopenia spectrum) as well as MPN predisposition; there is no evidence that this specific missense allele is fully penetrant, and rarity alone cannot be escalated. The caveat that MPL exon 10 is a somatic-mutation hotspot (a reason gnomAD hits there may be somatic rather than germline) is irrelevant to a confirmed absence and does not weaken PM2; conversely, absence of a variant from a hotspot region of adequate sequencing depth is genuinely informative.

Rule

generic_acmg_2015: PM2 = absent from population databases (supporting); non-VCEP gnomAD convention PM2 <0.1%

Evidence considered

• gnomAD v2.1 (exome, GRCh37): variant 1-43814978-A-T reported absent, with no allele count or frequency data.
• gnomAD v4.1 (GRCh38): variant chr1-43349307-A-T reported absent, with no allele count or frequency data; two independent releases agreeing on absence supports rarity rather than a single-database artifact.
• ClinVar: no exact match for the variant under any of the searched strings (c.1513A>T, NM_005373.2:c.1513A>T, g.43814978A>T), so there is no conflicting frequency or carrier submission.
• Gene-disease context artifact records a supported MPL germline disease mechanism, i.e. the rarity being assessed is relative to a real germline disease spectrum rather than an unaffected gene; this justifies applying PM2 for this gene.
• PM2 rarity criterion and its supporting-strength application from the generic ACMG/AMP 2015 framework (Richards et al. 2015, PMID:25741868); the non-VCEP gnomAD convention applied in this review places PM2 at allele frequency <0.1%, which a frequency of 0 satisfies.
• Interpretive caveat noted but not applied: published work on MPL/CALR apparent germline variants in population reference datasets raises the possibility that some population-database MPL exon 10 observations are somatic or undiagnosed-MPN clones; because this variant has no population-database record at all, the caveat has no effect on PM2 here.
• gnomAD-Canada v1.0 returned no data due to a network timeout and was therefore not usable as a supporting or refuting dataset.

Still missing

• gnomAD-Canada v1.0 (HostSeq genomes) query failed with a URL timeout, so this secondary population dataset is unverified; it is not expected to change a two-release zero-frequency result but should be re-run if available.
• No per-ancestry allele counts, genotype quality, or site-level coverage metrics were retrieved for the variant position; absence is taken from the gnomAD variant pages returning no record in either release.
• No homozygote-frequency information is available (none exists), so PM2 was applied on absence alone rather than on a homozygote-based argument.
• No MPL VCEP/CSPEC specification was available, so no gene-specific PM2 frequency threshold or PM2-in-hotspot caveat could be applied; generic ACMG/AMP plus the non-VCEP 0.1% convention was used.

Sources

Reasoning

PM5 is satisfied when a different missense change at the same residue as the variant under assessment is established as pathogenic. Codon 505 of MPL has exactly that: p.(Ser505Asn) (S505N). S505N is a 'canonical' MPL exon 10 driver mutation that, together with W515L/K/R/A, 'generate[s] constitutively active receptors and consequent loss of Tpo dependence', accounting for roughly 6% of JAK2 V617F-negative essential thrombocythemia and 14% of primary myelofibrosis cases, and it is 'often found as a germline mutation in hereditary thrombocythemia' (PMID:31697803, including a patient annotated 'S505NGer' in Figure 2D). PMID:34845187 independently states that 'Mutations at the juxtamembrane residue W515 and the transmembrane variant S505N constitute the most common MPL mutations identified in MPN and have been functionally characterized as drivers of pathological activation.' The variant under assessment is a different amino acid change at the same codon (Cys instead of Asn at residue 505) and is not itself an established pathogenic allele (see PS1), which is precisely the situation PM5 addresses. It is applied at supporting strength, not the generic moderate default, because the residue-level functional consequence is amino-acid specific: in the S505N-background deep mutational scan 'substitution of any residue other than asparagine at position 505 resulted in loss of constitutive activity' (PMID:31697803), and S505C alone did not produce constitutive STAT5 signalling or factor-independent growth (PMID:34845187), so the inference from the pathogenic Asn change to the observed Cys change is weaker than a same-residue, same-effect inference. The comparator was established from the primary literature because the automated same-residue harvest returned no candidates and ClinVar holds no record for this variant.

Rule

ACMG/AMP 2015 PM5 (different missense change at an amino acid residue where a different missense change has been established as pathogenic), applied at supporting strength

Evidence considered

• PMID:31697803 (abstract/Introduction): 'canonical' MPL exon 10 driver mutations are 'W515L/K/R/A or S505N, which generate constitutively active receptors and consequent loss of Tpo dependence'; S505N 'activates TpoR by driving strong dimerization'.
• PMID:31697803 (Results and discussion): 'we performed a second DMS screen on the S505N background (Figure 2A), often found as a germline mutation in hereditary thrombocythemia' - directly establishing S505N as a pathogenic change at the same codon as the variant under assessment, including a germline occurrence.
• PMID:31697803 (Figure 2D patient table): patients annotated 'S505NGer', 'S505NSom+V501ASom' and 'S505NGer+Q516RSom', i.e. codon 505 is a recurrent pathogenic position in patient cohorts.
• PMID:31697803: 'as expected, substitution of any residue other than asparagine at position 505 resulted in loss of constitutive activity' - the functional basis for applying PM5 at supporting rather than moderate strength for a different amino acid change at this codon.
• PMID:34845187: 'Mutations at the juxtamembrane residue W515 and the transmembrane variant S505N constitute the most common MPL mutations identified in MPN and have been functionally characterized as drivers of pathological activation', and 'MPL S505C has previously been identified together with W515L in JAK2 V617F-negative ET, although the functional consequences of this double mutation have not been established' - confirming S505N is an established driver while S505C is not, which is the PM5 (not PS1) configuration.
• ClinVar: no record for the variant under assessment, so the pathogenic comparator at the same residue came from the literature and not from a curated submission; the variant's absence from ClinVar is not evidence against PM5.
• pm5_candidates artifact for this case records pm5_mode 'unknown', eligible_for_classic_pm5_search false and zero harvested same-residue candidates; its stated instruction was to review manually whether any same-residue comparator satisfies the governing PM5 rule, which is what was done here.

Still missing

• The automated PM5 comparator harvest returned zero candidates (pm5_candidates: total_candidates_seen = 0, recommendation level 'not_applicable'), and ClinVar has no entry for this variant; S505N's established-pathogenic status therefore rests on the primary literature rather than a curated database assertion, and no ClinVar review status could be attached to it.
• The comparator S505N is pathogenic in a gain-of-function / hereditary thrombocythemia (thrombocytosis) context; if the phenotype under assessment in this case is MPL loss-of-function (congenital amegakaryocytic thrombocytopenia), the residue-based inference would not transfer and PM5 would need to be withdrawn.
• PM5 strength was set at supporting rather than the generic moderate because no residue-level equivalence data exist between the Cys and Asn changes at codon 505, and the published functional data show the opposite direction of effect (non-activating S505C versus strongly activating S505N).

Sources

Reasoning

BP2 (Richards et al. 2015, PMID:25741868) includes the observation of the variant in cis with a pathogenic variant in any inheritance pattern. That configuration is documented for NM_005373.2:c.1513A>T (NP_005364.1:p.(Ser505Cys)): in an essential thrombocythemia patient who later progressed to post-ET myelofibrosis, S505C and W515R - a canonical, well-established pathogenic MPL exon 10 driver - were detected in cis on the same allele, verified by Sanger sequencing, at a 0.37 variant allele ratio (PMID:34845187). Independent cohort data show the same pattern with another canonical driver, W515L, in a JAK2 V617F-negative ET patient (PMID:31697803). Consistent with a non-causal cis passenger, S505C alone produced no constitutive STAT5 signalling and did not support factor-independent growth in Ba/F3 cells; its measured effect was enhancement of signalling driven by the pathogenic W515 variants (PMID:34845187). Applied at supporting strength, the default weight of BP2. Human review is flagged because the documented cis phase comes from a somatic bone-marrow clone rather than a constitutional sample, so the applicability of that phase observation to a germline classification is a judgement call.

Rule

Generic ACMG/AMP BP2 (Richards et al. 2015, PMID:25741868): observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder, or observed in cis with a pathogenic variant in any inheritance pattern; applied at supporting strength.

Evidence considered

• BP2's second clause - observed in cis with a pathogenic variant in any inheritance pattern - is literally satisfied by a phase-resolved patient observation (Richards et al. 2015, PMID:25741868).
• PMID:34845187: 'MPL S505C and W515R mutations were detected in cis (Supplementary Fig. 1), verified by Sanger sequencing, with a 0.37 variant allele ratio'; W515R is a canonical MPL exon 10 driver mutation repeatedly classified as activating/pathogenic in MPN.
• PMID:34845187 also quotes prior work and this patient's context: 'MPL S505C has previously been identified together with W515L in JAK2 V617F-negative ET [8]', i.e. S505C recurs as a second-site allele alongside canonical pathogenic drivers.
• PMID:31697803 reports a further patient with the double combination: 'W515LSom+S505C' (ET, female, age 57; variant allele frequencies listed as 25 and 50), and states that 'Additional published reports of patients with double mutations include V501A+W515R, V501A+W515L, and S505C+W515L, and we identified a second instance of each of these combinations in our patient cohorts.'
• PMID:34845187 functional data consistent with a variant that is not itself the pathogenic allele: 'We found that the S505C did not drive constitutive STAT5 signaling'; 'The MPL S505C variant was not sufficient for factor-independent growth'; the double mutants S505C/W515R and S505C/W515L did signal and proliferate autonomously.
• Caveat recorded for review: the cis partner and the S505C allele were characterised in bone-marrow DNA from a patient with no family history of MPN, so the observation reflects a somatic MPN clone rather than a constitutional genotype.

Still missing

• No constitutional (blood/lymphocyte) phase determination for S505C is available; the cis observation derives from bone-marrow DNA in a somatic MPN clone.
• In PMID:31697803 the S505C/W515L combination had undetermined phase (S505C not annotated as germline or somatic).
• No germline family data are available to confirm or exclude a heritable S505C allele.

Sources

Reasoning

PS1 requires that the same amino acid change, p.(Ser505Cys), is itself a previously established pathogenic variant (regardless of nucleotide change). It is not. First, the variant is absent from ClinVar (no submissions, no review status, no expert-panel assertion), and no other curated database assertion of pathogenicity for p.(Ser505Cys) was collected. Second, the two full-text sources in this case describe S505C as a non-activating secondary variant, not as a pathogenic allele: S505C alone 'did not drive constitutive STAT5 signaling' and 'was not sufficient for factor-independent growth' in Ba/F3 cells (PMID:34845187), and in the saturation mutagenesis study substitution of any residue other than asparagine at position 505 abolished constitutive activity, with S505C appearing in MPN patients only as a co-occurring second-site change in cis with W515L ('W515LSom+S505C') (PMID:31697803). The single COSMIC observation is a somatic count of 1, and the OncoKB 'Likely Oncogenic' label reflects the somatic MPN context rather than a germline pathogenic assertion for this amino acid change. Because no source establishes p.(Ser505Cys) as pathogenic, PS1 is not met.

Rule

ACMG/AMP 2015 PS1 - same amino acid change as a previously established pathogenic variant; not satisfied (no established pathogenic assertion for p.(Ser505Cys))

Evidence considered

• ClinVar: no record for this variant; the evidence collection returned 'This variant is absent from ClinVar' with no variation ID, classification, review status, citations or criterion leads.
• PMID:34845187 (case report/functional study): MPL S505C and W515R were detected in cis in an 83-year-old JAK2 V617F-negative ET patient who progressed to post-ET MF; the authors state 'We found that the S505C did not drive constitutive STAT5 signaling' and 'The MPL S505C variant was not sufficient for factor-independent growth', concluding that S505C 'induces receptor dimerization, but that this is not sufficient for signaling to drive autonomous proliferation'.
• PMID:31697803 (deep mutational scanning): S505C is not among the enriched constitutively activating variants; in the S505N-background screen 'substitution of any residue other than asparagine at position 505 resulted in loss of constitutive activity', and S505C appears only as a second-site variant co-occurring with W515L in patient data.
• OncoKB: variant-level curation is 'Likely Oncogenic' with 'Likely Gain-of-function', derived from somatic MPN literature (PMIDs 31697803 and 34845187); this is not an assertion that the germline p.(Ser505Cys) amino acid change is established pathogenic, so it cannot satisfy PS1.
• No other collected source (population databases, in silico predictors) asserts pathogenicity of p.(Ser505Cys), and absence of the variant from gnomAD v2.1/v4.1 is not evidence of pathogenicity.

Still missing

• No ClinVar or other curated assertion exists for p.(Ser505Cys) (or for an alternative nucleotide change encoding the same amino acid change), so the PS1 determination rests on absence of evidence plus the primary literature characterisation.
• No germline clinical indication/phenotype is recorded in this case bundle; an established pathogenic assertion for p.(Ser505Cys) under a disease context not searched here cannot be formally excluded.

Sources

Reasoning

No well-established functional study demonstrates a damaging effect of MPL c.1513A>T (p.Ser505Cys) on the protein by itself. In PMID:34845187 the substitution was introduced alone into HA-tagged MPL expression constructs (Quikchange mutagenesis, pMX-IRES-GFP): it did not drive constitutive STAT5 transcriptional activity in HEK293T dual-luciferase assays and was not sufficient for factor-independent growth of IL-3-dependent Ba/F3 cells, whereas the positive controls W515R and W515L were autonomously activating in the same experiments. The independent, domain-wide saturation mutagenesis screen in PMID:31697803 (all single amino-acid substitutions across TpoR/MPL transmembrane residues 488-516 in Ba/F3 cells) found only 24 of 580 variants enriched under cytokine withdrawal, all of them either the known drivers (S505N, W515 substitutions) or a small set of other positions, and stated that any substitution other than asparagine at position 505 loses constitutive activity; S505C was therefore not a stand-alone activator in that assay. The only functional consequence demonstrated for S505C is potentiation of signalling that is driven in cis by a canonical W515R/W515L driver, which is a context-dependent modifier effect that cannot be attributed to the variant alone. Assay controls were adequate (wild-type MPL, empty vector and activating W515 controls in the same experiments; DMS enrichment scores correlated with independent growth assays in PMID:31697803), so the negative result reflects a genuine absence of a stand-alone damaging effect rather than assay insensitivity.

Rule

ACMG/AMP 2015 PS3 - well-established functional studies show a damaging effect on gene/protein function. Applied under the generic ACMG/AMP framework; no MPL-specific CSPEC/VCEP specification is available.

Evidence considered

• PMID:34845187 (functional study read in full): S505C, W515L and W515R were introduced individually and as compound S505C/W515L and S505C/W515R constructs into HA-MPL. In HEK293T dual-luciferase STAT5 reporter assays S505C alone did not drive constitutive STAT5 signalling, while the compound S505C/W515R construct produced autonomous STAT5 activity and S505C enhanced W515L-induced autonomous STAT5 activity.
• PMID:34845187: in the IL-3-dependent Ba/F3 factor-free proliferation assay (stable lines sorted for equivalent GFP/surface-HA expression), cells expressing S505C alone did not proliferate without cytokine; only W515R/W515L lines became factor independent, and adding S505C increased the proliferation rate of those W515 lines.
• PMID:34845187: o-phenylene dimaleimide cross-linking of truncated HA-MPL in CHO cells showed that S505C formed more stable dimers (approximately 140 kDa) in the absence of TPO, but the authors concluded these dimers are not in a productive orientation, so dimerisation alone was 'not sufficient for signaling to drive autonomous proliferation'.
• PMID:31697803 (deep mutational scanning, Blood 2020): saturation mutagenesis of the whole human TpoR/MPL transmembrane domain (residues 488-516) expressed in Ba/F3 cells scored cytokine-independent growth over 48 h across 6 replicate selections; only 24 of 580 single substitutions were enriched relative to wild type, comprising the known drivers (S505N, W515 substitutions) plus a few other positions - S505C was not among the constitutively activating variants.
• PMID:31697803: in the second screen on the S505N background the authors state that substitution of any residue other than asparagine at position 505 resulted in loss of constitutive activity, and they report S505C only in combination with a canonical driver (published S505C+W515L patients, plus one additional cohort patient with W515L(somatic)+S505C at VAF 25%/50%).
• oncokb: OncoKB lists MPL S505C as 'Likely Oncogenic' with a biological-effect label of 'Likely Gain-of-function' and cites only PMID:31697803 and PMID:34845187. Because those two papers show that S505C is not activating on its own and only augments a co-occurring W515 driver, the OncoKB gain-of-function label reflects the in-cis modifier effect rather than a stand-alone damaging effect, and OncoKB is a curated knowledge base rather than a primary functional assay. It was therefore not counted as PS3 evidence.
• Search context: no MPL-specific ClinGen CSPEC/VCEP specification or local gene framework exists for this gene (cspec lookup returned no result, gene_vcep_dir null), so PS3 was evaluated under generic ACMG/AMP 2015 with the ClinGen SVI principle that the assay readout must reflect the variant's own effect on the disease-relevant mechanism.

Still missing

• No published assay tested loss-of-function endpoints for p.Ser505Cys (TPO-stimulated JAK2/STAT5 signalling, receptor cell-surface expression, or megakaryocytic differentiation), which is the mechanism that would be relevant if a germline MPL deficiency phenotype is being considered.
• The only positive functional effect of S505C (enhancement of W515R/W515L-driven autonomous STAT5 activity and proliferation) was demonstrated in a single study (PMID:34845187) and only in cis with a canonical driver; whether S505C alone can produce a damaging effect in a germline context is untested.
• No functionally characterised patient carries S505C as the sole MPL exon 10 lesion: the reported case was compound S505C+W515R, and the previously published S505C patients were also S505C+W515L double mutants.
• Both variants of the disease mechanism (gain-of-function thrombocytosis/MPN versus loss-of-function marrow failure) remain untested for S505C alone, so a damaging effect cannot be excluded on the currently available assays.

Sources

Reasoning

PS4 requires the prevalence of NM_005373.2:c.1513A>T (NP_005364.1:p.(Ser505Cys), MPL exon 10) in affected individuals to be significantly increased compared with controls, and no such enrichment exists in any source reviewed. ClinVar holds no record for this variant (exact_match=false, zero submissions, no expert panel), so there are no aggregated case counts or case-control submissions to draw on. The curated full-text literature contains only two variant-carrying reports and neither is an enrichment or case-control study: PMID:34845187 describes a single 83-year-old patient with essential thrombocythemia (negative for JAK2 V617F, CALR frameshifts and MPL W515K/L) in whom S505C was found in cis with a canonical W515R driver at a 0.37 variant allele ratio, with functional assays showing S505C alone did not drive constitutive STAT5 signalling or factor-independent Ba/F3 growth; PMID:31697803 reports S505C only in combination with the canonical W515L driver (published S505C+W515L patients plus one cohort patient with W515L 25% / S505C 50%). No cohort, no control denominator and no frequency comparison was identified, so there is no statistically increased prevalence in affected individuals. Repositories consulted for this group (COSMIC somatic count 1; OncoKB curated label 'Likely Oncogenic') contribute variant-level functional/oncogenicity curation, not case-control case enrichment. PS4 is reported as not_met rather than not_assessed because the criterion's requirement is affirmatively absent rather than untested: the only documented carriers are two co-mutation cases in which a canonical W515 driver provides the principal molecular explanation for the phenotype.

Rule

generic_acmg_2015: PS4 = the prevalence of the variant in affected individuals is significantly increased compared with the prevalence in controls (Richards et al. 2015, PMID:25741868). No MPL CSPEC/VCEP or local gene-specific PS4 rule was available.

Evidence considered

• Criterion definition applied: PS4 = the prevalence of the variant in affected individuals is significantly increased compared with the prevalence in controls (Richards et al. 2015, PMID:25741868), taken from the generic ACMG/AMP framework because no MPL CSPEC/VCEP or local gene-specific specification is available (case_summary.json cspec.found=false; vcep_materials.json final_classification_framework.source=generic_acmg_fallback, official_clingen=false, precedence_rank 3).
• ClinVar search for NM_005373.2:c.1513A>T returned no result (search_status='no_result', exact_match=false, variation_ids=[], evidence_sentence 'This variant is absent from ClinVar.'; searches run as MPL[GENE]+c.1513A>T[VARNAME], NM_005373.2:c.1513A>T[VARNAME], NM_005373.2:c.1513A>T and NC_000001.10:g.43814978A>T[VARNAME]). There is therefore no aggregate case count, no case-control submission and no expert-panel case assessment to inherit for PS4.
• PMID:34845187 (full text, extracted): case-level report of one individual, an 83-year-old with essential thrombocythemia negative for JAK2 V617F, CALR frameshift mutations and MPL W515K/L, in whom 'MPL S505C and W515R mutations were detected in cis (Supplementary Fig. 1), verified by Sanger sequencing, with a 0.37 variant allele ratio' and in whom 'the S505C did not drive constitutive STAT5 signaling' and 'the MPL S505C variant was not sufficient for factor-independent growth'. This is a single affected carrier co-carrying a canonical driver, i.e. no frequency information and no attribution of the phenotype to S505C, so it provides no enrichment signal.
• PMID:31697803 (full text, extracted): 'Additional published reports of patients with double mutations include V501A1W515R, V501A1W515L, and S505C1W515L, and we identified a second instance of each of these combinations in our patient cohorts' with one cohort patient carrying W515L and S505C (VAFs 25% and 50%; Figure 2D). All occurrences are double mutations in affected myeloproliferative-neoplasm patients; no case series with affected-versus-control frequencies was found.
• COSMIC records this exact variant once as a somatic event (COSV65247010, somatic count 1). A somatic single-sample count is not affected-individual prevalence and is not compared against a control set, so it does not satisfy PS4.
• OncoKB (variant-specific 'Likely Oncogenic', biological effect 'Likely Gain-of-function') is a cancer-variant curation resource, not a case-control prevalence dataset; it was consulted but cannot establish the statistically increased prevalence in affected individuals that PS4 requires.
• Population data were checked for completeness and do not help PS4 in either direction: the variant is absent from gnomAD v2.1 and gnomAD v4.1 (no allele counts) and the gnomAD-Canada v1.0 query failed with a URL timeout. Zero population frequency cannot substitute for case enrichment, and PS4 was not inferred from any other source's final classification.

Still missing

• No case-control, cohort-enrichment or case-series frequency study for MPL c.1513A>T exists in the retrieved evidence; the only literature occurrences (n=2 reported cases, both with a canonical W515 driver in cis) cannot support a prevalence comparison.
• No proband-level genotype/phenotype data (including whether this case carries JAK2 V617F, CALR or MPL W515 drivers) is present anywhere in the case bundle, so case attribution of any observed phenotype to S505C alone cannot be tested.
• The variant is a somatic-mutation hotspot residue (MPL exon 10) and is recorded once as a somatic event, so any future case counts would need germline confirmation (e.g. VAF/tissue source) before being used as constitutional case evidence for PS4.
• No MPL-specific ClinGen CSPEC/VCEP or local gene framework exists (cspec.found=false; final_classification_framework.source=generic_acmg_fallback), so no gene-specific PS4 case threshold or case-counting rules could be applied.

Sources

Reasoning

PM3 (Richards et al. 2015, PMID:25741868) requires a recessive disorder and detection of the variant in trans with a pathogenic variant in that gene. No trans observation exists anywhere in this case. The only phase-resolved observation of NM_005373.2:c.1513A>T (NP_005364.1:p.(Ser505Cys)) in the evidence is cis, not trans: in a JAK2 V617F-negative, CALR-negative, MPL W515K/L-negative essential thrombocythemia patient (83 years, no family history of MPN, later post-ET myelofibrosis), S505C and the canonical pathogenic MPL driver W515R were detected in cis, Sanger-confirmed, with a 0.37 variant allele ratio (PMID:34845187). The second cohort paper also shows S505C only in combination with a canonical pathogenic exon 10 driver, W515L, and leaves that phase undetermined (PMID:31697803). In addition, the applicable germline disease mode for this variant class does not favour a recessive setup: MPL germline gain-of-function alleles cause autosomal dominant thrombocythemia 2, whereas the autosomal recessive mode (congenital amegakaryocytic thrombocytopenia) is driven by loss-of-function alleles, and S505C is a transmembrane missense that neither truncates the protein nor shows autonomous activating behaviour on its own (PMID:34845187). With no trans configuration and no proband/second-allele data, PM3 is not met.

Rule

Generic ACMG/AMP PM3 (Richards et al. 2015, PMID:25741868): for recessive disorders, the variant is detected in trans with a pathogenic variant; a documented cis configuration, or absence of any trans observation, does not satisfy PM3.

Evidence considered

• Generic ACMG/AMP 2015 PM3 requires a recessive disorder with the variant detected in trans with a pathogenic variant; a cis configuration is the structural opposite of what PM3 requires (Richards et al. 2015, PMID:25741868).
• PMID:34845187 reports exactly one phase-resolved observation for this variant: 'MPL S505C and W515R mutations were detected in cis (Supplementary Fig. 1), verified by Sanger sequencing, with a 0.37 variant allele ratio' in a JAK2 V617F-negative, CALR-negative, MPL W515K/L-negative ET patient (83 years, 'no family history of MPN') who progressed to post-ET MF.
• PMID:31697803 lists an ET patient with 'W515LSom+S505C' (variant allele frequencies listed as 25 and 50), i.e. co-occurrence with the canonical pathogenic driver W515L, but the table's footnote states that germline versus somatic assignment is only made where VAFs, familial disease or lymphocyte DNA allow it, and S505C carries no such assignment there; no trans configuration is reported.
• The OncoKB gene-level annotation retrieved for MPL states that germline gain-of-function mutations cause autosomal dominant thrombocythemia 2 while germline loss-of-function mutations cause autosomal recessive congenital amegakaryocytic thrombocytopenia, so the recessive mode relevant to PM3 applies to loss-of-function alleles rather than to this transmembrane missense.
• PMID:34845187 shows S505C has no autonomous activity ('We found that the S505C did not drive constitutive STAT5 signaling'; 'The MPL S505C variant was not sufficient for factor-independent growth'), which argues against treating this allele as a straightforward loss-of-function recessive allele.
• The case bundle contains no proband report, pedigree, parental genotypes, or second-variant information for this variant; ClinVar has no record and the variant is absent from gnomAD v2.1 and v4.1 (gnomAD-Canada lookup unavailable).

Still missing

• No proband-level constitutional genotype data (affected proband with a second MPL allele) are available in this case.
• No family/pedigree or parental phase data are available; whether any second MPL variant exists in trans with S505C in a clinically affected individual is unknown.
• In PMID:31697803 the S505C/W515L combination patient had an undetermined configuration and the variant was not annotated as germline or somatic (VAFs listed as 25% and 50%).

Sources

Reasoning

NM_005373.2:c.1513A>T (MPL) is a missense substitution, NP_005364.1:p.(Ser505Cys), exon 10, not a canonical splice-consensus change, so this group's missense protein-impact path governs. No ClinGen CSPEC/VCEP and no local gene-specific MPL framework is available (vcep_materials.final_classification_framework = generic_acmg_fallback), so the generic in-silico calibration applies: REVEL only, PP3 at REVEL > 0.750. The local REVEL v1.3 lookup returned 0.422, which is inside the [0.250, 0.750] gray zone, so the PP3 threshold is not reached. BayesDel (-0.048) is not usable because no verified published threshold is available to this pipeline. The SpliceAI max delta of 0.019 was not considered here: it is a splice-path score and this variant is judged on its protein-impact score, so no splice prediction can support or refute PP3 for a missense variant. No computational protein-impact evidence therefore supports a damaging call.

Rule

Generic ACMG/AMP in-silico calibration for missense variants (applied because MPL has no CSPEC/VCEP or local gene-specific PP3/BP4 rule): REVEL > 0.750 -> PP3 supporting; REVEL 0.250-0.750 -> neither met; REVEL < 0.250 -> BP4 supporting (thresholds from Pejaver et al., PMID 36413997).

Evidence considered

• Variant identity and consequence: NM_005373.2:c.1513A>T -> NP_005364.1:p.(Ser505Cys) (p.S505C), MPL exon 10, consequence_class = missense, canonical splice consensus = false (prefetch.json normalize step; pvs1_variant_assessment.json).
• Missense path selected, not the splice path: a missense substitution is adjudicated on its protein-impact score, so REVEL is the governing predictor for this variant.
• REVEL v1.3 local lookup: score 0.422 (source_registry 'revel'; prefetch.json revel step; evidence.json results.revel). Threshold source: Pejaver et al., PMID 36413997 (ClinGen SVI calibration of REVEL), which places REVEL > 0.750 as PP3 supporting and REVEL < 0.250 as BP4 supporting, with 0.250-0.750 uninformative; 0.422 therefore meets neither.
• BayesDel (no-AF) local score = -0.0484473 (source_registry 'bayesdel'). No named publication and table for a BayesDel PP3/BP4 threshold is available to this pipeline, so this score was treated as not usable and contributed nothing to either criterion.
• SpliceAI lookup for this variant returned max delta score 0.019 (DS_AG 0.011, DS_AL 0.011, DS_DG 0.003, DS_DL 0.019; Pangolin SG 0.01, SL -0.003) (source_registry 'spliceai'; evidence.json results.spliceai). This splice-path score was deliberately not used: the variant is a missense substitution, and 'no predicted splice impact' is not evidence about the substituted amino acid, so it cannot support PP3 or BP4 here.
• Splice-path thresholds (SpliceAI > 0.2 PP3, < 0.1 BP4; Walker et al. 2023, SVI-recommended) were noted but are inapplicable to this missense variant and were not applied.
• No CSPEC and no local MPL framework: case_summary.json cspec.found = false; vcep_materials.json final_classification_framework.source = generic_acmg_fallback, official_clingen = false; therefore the generic in-silico thresholds govern (framework precedence rank 3).
• ClinVar has no record for this variant (evidence.json results.clinvar search_status = no_result), so no external computational-evidence assertion could be inherited.

Still missing

• REVEL 0.422 is in the gray zone, so computational protein-impact evidence is explicitly uninformative for this variant; PP3 cannot be met from in-silico evidence.
• No gene-specific MPL PP3/BP4 lookup table or VCEP specification exists in this case (cspec not found, vcep_materials empty), so a VCEP-preassigned PP3 code cannot be checked or applied.
• BayesDel is uncalibrated in this pipeline (no verified publication/table), so a second protein-impact predictor cannot be used to resolve the REVEL gray zone.

Sources

Reasoning

PP4 requires the patient's phenotype or family history to be highly specific for a disease with a single genetic etiology, and that standard is not met on the material available. First, no proband phenotype, HPO term or family history is recorded anywhere in this case bundle (no clinical fields in case_summary.json, no expert-panel submission, no patient description), so phenotype specificity cannot be demonstrated for this case. Second, the only phenotype associated with this variant in the curated literature is a myeloproliferative neoplasm - essential thrombocythemia in PMID:34845187 and MPN patients in PMID:31697803 - and that phenotype is genetically heterogeneous rather than highly specific for one etiology: the same adult MPN/ET presentation is produced by JAK2 V617F, CALR exon 9 and MPL W515 mutations as well as by triple-negative cases, and MPL itself has a split germline spectrum (gain-of-function thrombocythemia 2 versus loss-of-function congenital amegakaryocytic thrombocytopenia) so an MPN phenotype does not point to a single genetic mechanism. Third, in every documented S505C carrier the phenotype is at least equally explained by a co-occurring canonical activating MPL W515 driver detected in cis (W515R at a 0.37 allele ratio in PMID:34845187; W515L in PMID:31697803), and functional data show S505C alone is not sufficient for autonomous signalling - the opposite of a phenotype that is specific to this variant. PP4 is therefore not met, but human review is flagged because the missing proband phenotype is the single piece of information that could change this determination.

Rule

generic_acmg_2015: PP4 = patient's phenotype or family history is highly specific for a disease with a single genetic etiology (Richards et al. 2015, PMID:25741868). No MPL CSPEC/VCEP or local gene-specific PP4 rule was available.

Evidence considered

• Criterion definition applied: PP4 = patient's phenotype or family history is highly specific for a disease with a single genetic etiology (Richards et al. 2015, PMID:25741868), under generic ACMG/AMP because no MPL CSPEC/VCEP or local gene-specific specification exists (case_summary.json cspec.found=false; vcep_materials.json final_classification_framework.source=generic_acmg_fallback).
• Absence of proband phenotype data: case_summary.json contains no clinical/phenotype section, ClinVar has no submission for this variant (submitter_count=0, expert_panels=[]), and no other artifact in the bundle records HPO terms, onset, platelet indices or family history for this case; PP4's evidential prerequisite is therefore not documented.
• Phenotype heterogeneity of the MPL/MPN disease axis (source_registry 'oncokb' gene and variant summaries): 'MPL, a transmembrane protein receptor, is frequently mutated in myeloproliferative neoplasms including essential thrombocytosis and myelofibrosis', with germline gain-of-function causing autosomal dominant thrombocythemia 2 and germline loss-of-function causing autosomal recessive congenital amegakaryocytic thrombocytopenia; the same clinical phenotype is also caused by JAK2 V617F and CALR mutations, so an MPN/ET phenotype is not highly specific to a single genetic etiology.
• PMID:34845187: the phenotype reported for the S505C carrier is essential thrombocythemia in an 83-year-old with 'MPL S505C and W515R mutations ... detected in cis ... with a 0.37 variant allele ratio', i.e. a genetically heterogeneous adult MPN with a co-occurring canonical MPL W515R driver present - the phenotype does not single out S505C, and the variant alone was 'not sufficient for factor-independent growth'.
• PMID:31697803: 'S505C1W515L' is reported as a double-mutation combination in affected MPN patients (including one cohort patient with VAFs 25%/50%), so again the phenotype in every documented carrier is accompanied by a canonical W515 driver rather than being attributable to S505C alone.
• Gene-level germline context (source_registry 'pvs1_gene_context'): germline_disease_context_found=true with official_source_type='literature_only', lof_mechanism_supported=true, but inheritance=[] and no phenotype/onset terms, so it supports the existence of a germline MPL disease spectrum without providing any specificity claim usable for PP4.
• This criterion was not inferred from any other source's final classification: the OncoKB 'Likely Oncogenic' label and COSMIC somatic record describe biological effect, not phenotype specificity, and were not converted into PP4.

Still missing

• No proband phenotype, HPO terms, age at onset, platelet count, bone-marrow findings or family history are present in the case bundle, so the 'highly specific' prerequisite of PP4 could not be positively tested; if such data exist and are highly specific to the MPL germline spectrum (e.g. early-onset amegakaryocytic thrombocytopenia, or familial thrombocythemia with MPN), PP4 should be re-examined.
• No MPL-specific phenotype checklist, CSPEC/VCEP phenotype-specificity table or local gene framework exists (cspec.found=false; vcep_materials empty; generic_acmg_fallback), so specificity was judged against generic ACMG/AMP expectations only.
• Reported carriers are adults with MPN and co-occurring canonical MPL W515 drivers, so there is no clean S505C-only phenotype description in the literature against which specificity could be measured (n=2 reported cases).
• The germline disease context artifact records a supported MPL germline disease mechanism but lists no inheritance pattern, onset or phenotype terms (inheritance=[]; disease_names unusable), so gene-level phenotype specificity could not be established from it.

Sources

Reasoning

PP5 requires a reputable source to report this exact variant as pathogenic, and for this group that means an exact-variant ClinGen expert-panel Pathogenic/Likely pathogenic classification. No such assertion exists. NM_005373.2:c.1513A>T has no ClinVar record at all: repeated searches (MPL[GENE] AND c.1513A>T[VARNAME], NM_005373.2:c.1513A>T[VARNAME], NM_005373.2:c.1513A>T, NC_000001.10:g.43814978A>T[VARNAME], NC_000001.10:g.43814978A>T and MPL[GENE] AND NM_005373.2:c.1513A>T) returned no result with exact_match=false, variation_ids=[], zero submissions and no expert panel. The only external curated label found, OncoKB's variant-specific 'Likely Oncogenic' (with biological effect 'Likely Gain-of-function'), is a somatic-cancer oncogenicity curation and is explicitly not a ClinVar expert-panel germline classification, so under this group's rule it can never trigger PP5 and was not used as support. No aggregate ClinVar label, ordinary laboratory submission or non-expert assertion was available either, and none would have been sufficient. PP5 is therefore not met: there is no exact-variant, expert-panel pathogenic assertion to inherit.

Rule

generic_acmg_2015: PP5 = reputable source reports the variant as pathogenic (Richards et al. 2015, PMID:25741868), applied only for an exact-variant ClinGen expert-panel Pathogenic/Likely pathogenic classification (supporting strength); non-expert submissions, aggregate ClinVar labels and other curated databases excluded. No MPL CSPEC/VCEP available.

Evidence considered

• Criterion definition applied: PP5 = reputable source reports the variant as pathogenic (Richards et al. 2015, PMID:25741868), restricted for this group to an exact-variant ClinGen expert-panel Pathogenic/Likely pathogenic classification at supporting strength; aggregate ClinVar labels, ordinary laboratory submissions, curated cancer databases and other non-expert assertions cannot trigger it.
• ClinVar lookup (source_registry 'clinvar'): search_status='no_result', found=false, exact_match=false, variation_ids=[], chosen_variation_id=null, variation_name='', with the recorded sentence 'This variant is absent from ClinVar.'; searches covered c.1513A>T, NM_005373.2:c.1513A>T and the GRCh37 genomic description g.43814978A>T. No submission, no review status and no expert-panel classification could be retrieved because no record exists.
• OncoKB (source_registry 'oncokb') returned a variant-specific curated label 'Likely Oncogenic' with biological effect 'Likely Gain-of-function' and PMIDs 31697803 and 34845187. OncoKB is an oncology knowledge base, not a ClinVar expert panel, and its label applies to the somatic oncogenic role; it was therefore explicitly excluded from PP5 and contributed no evidence in either direction.
• No aggregate ClinVar label or non-expert laboratory classification was used: ClinVar has no submissions at all (submitter_count=0, expert_panels=[]), so even the ordinarily insufficient routes to PP5 were unavailable.
• Framework check: no MPL CSPEC/VCEP was retrievable (case_summary.json cspec.found=false; vcep_materials.json note 'no cspec url available', files_found=[]), and the governing framework is the generic ACMG/AMP 2015 fallback (final_classification_framework.source=generic_acmg_fallback), so PP5's generic definition applies unchanged.
• No other source's final classification was used to infer this criterion; the determination rests solely on the absence of an exact-variant expert-panel pathogenic assertion.

Still missing

• No ClinVar record of any kind exists for this variant, so no expert-panel (3-star) classification is available; PP5 cannot be satisfied unless such a submission appears in future.
• No ClinGen CSPEC/VCEP exists for MPL (cspec.found=false, vcep_materials empty), so no gene-specific expert-panel assertion list or PP5 specification could be consulted.
• The variant's apparent absence from the clinical record may reflect that it is predominantly reported as a somatic MPN finding rather than a germline variant; if a germline expert panel later classifies this exact variant, PP5 must be re-evaluated at supporting strength.

Sources

Reasoning

NM_005373.2:c.1513A>T (NP_005364.1:p.(Ser505Cys), MPL exon 10) has an allele frequency of zero in the two large gnomAD releases queried: it is absent from gnomAD v2.1 (GRCh37 locus 1-43814978-A-T) and absent from gnomAD v4.1 (GRCh38 locus chr1-43349307-A-T). BA1 requires a stand-alone benign allele frequency (Richards et al. 2015, PMID:25741868: allele frequency >5%), and even the most permissive non-VCEP gnomAD cut-off used for this review (BA1 >1%) is not remotely approached by a frequency of zero. There is therefore no population-frequency basis for a stand-alone benign classification. No MPL-specific ClinGen CSPEC/VCEP specification exists (cspec.found=false; final_classification_framework found=false, source=generic_acmg_fallback), so the generic ACMG/AMP 2015 thresholds govern. Note that MPL exon 10 is a somatic-mutation hotspot in myeloproliferative neoplasms, so a population-database hit here would itself have been hard to interpret; this does not apply because no hit exists.

Rule

generic_acmg_2015: BA1 = allele frequency >5% (stand-alone benign)

Evidence considered

• gnomAD v2.1 variant page (GRCh37, 1-43814978-A-T) returned 'absent' with no source_data, i.e. no allele count and no allele frequency for this variant.
• gnomAD v4.1 variant page (GRCh38, chr1-43349307-A-T) returned 'absent' with no source_data, independently confirming zero frequency in the newer, larger release.
• The gnomAD-Canada v1.0 (HostSeq genomes) query errored with '<urlopen error timed out>', so it contributed no frequency information in either direction; the recorded status 'Not available in gnomAD-Canada v1.0' reflects an unavailable service, not a confirmed absence.
• BA1 threshold applied from the generic ACMG/AMP 2015 framework (Richards et al. 2015, PMID:25741868); the original framework uses an allele frequency above ~5%, so a frequency of zero is orders of magnitude below the stand-alone benign bar.
• Framework selection: gnomAD is the authoritative population-frequency source for this case; no MPL VCEP specification was retrievable, so generic ACMG/AMP thresholds were applied.

Sources

Reasoning

BS1 requires an allele frequency greater than expected for the disorder. The MPL germline disease spectrum (congenital amegakaryocytic thrombocytopenia and related inherited thrombocytopenia / germline myeloproliferative-neoplasm predisposition) is rare, so the frequency expected for a pathogenic allele is far below 1%. The variant is absent from gnomAD v2.1 and gnomAD v4.1 (observed allele frequency 0), which is below, not above, any plausible disease-expected frequency. Because the observed frequency is zero, it cannot be 'greater than expected for the disorder' under any threshold, generic (Richards et al. 2015, PMID:25741868) or the permissive non-VCEP cut-off used here (BS1 >0.3%). BS1 is therefore not met. The only counter-consideration is that gnomAD exon-10 variation in MPL and CALR can be enriched for somatic/undiagnosed-MPN signal rather than true germline polymorphism, but that concern applies to interpreting frequency hits, and none exists here.

Rule

generic_acmg_2015: BS1 = allele frequency greater than expected for the disorder

Evidence considered

• Observed allele frequency is 0: gnomAD v2.1 and gnomAD v4.1 both report the variant as absent at the matched GRCh37 and GRCh38 loci.
• Disease-expected frequency context: the gene-level germline context artifact records germline_disease_context_found=true with inherited MPL loss-of-function disease mechanism supported from germline-focused literature, i.e. a rare Mendelian phenotype rather than a common polymorphism; BS1 therefore requires a much higher frequency than the observed zero.
• BS1 threshold applied from the generic ACMG/AMP 2015 framework (Richards et al. 2015, PMID:25741868: frequency greater than expected for the disorder); the permissive non-VCEP cut-off used for this review (BS1 >0.3%) is also not approached.
• gnomAD-Canada returned no data due to a network timeout, so it neither supports nor refutes a benign frequency.

Still missing

• gnomAD-Canada v1.0 (HostSeq genomes) lookup failed with a URL timeout, leaving one secondary genome dataset unverified; this cannot convert a zero-frequency result into a BS1-level frequency.

Sources

Reasoning

BS2 requires the variant to be observed in a healthy adult individual, with full penetrance expected at an early age (homozygous for a recessive disorder, heterozygous for a dominant disorder, or hemizygous for an X-linked disorder). No such observation exists. The variant is absent from gnomAD v2.1, gnomAD v4.1 and ClinVar, so there is no population carrier record and no reported homozygote. The only carriers documented in the curated full-text literature are affected individuals: PMID:34845187 reports S505C in cis with W515R (variant allele ratio 0.37) in an 83-year-old patient with JAK2 V617F-, CALR- and W515-negative essential thrombocythemia; PMID:31697803 reports S505C only in combination with the canonical W515L driver in myeloproliferative neoplasm patients. An affected elderly MPN patient is not an unaffected healthy control for BS2, and for a late-onset, incompletely penetrant MPN-predisposition phenotype BS2 would in any case require documented unaffected, elderly, ideally non-clonal carriers, which do not exist in any source reviewed. Somatic origin is also plausible for this residue (MPL exon 10 is a somatic hotspot; COSMIC records this exact variant once as a somatic event), which further undermines any attempt to read the published carriers as constitutional healthy-adult observations.

Rule

generic_acmg_2015: BS2 = observed in a healthy adult individual (full penetrance expected at an early age)

Evidence considered

• gnomAD v2.1 (1-43814978-A-T) and gnomAD v4.1 (chr1-43349307-A-T) both return 'absent'; no carriers, and in particular no homozygotes, are recorded in ~0.8M+ aggregated individuals.
• ClinVar has no record for NM_005373.2:c.1513A>T, so no submitted healthy-carrier or control observation exists there either.
• PMID:34845187 detected MPL S505C and W515R in cis with a 0.37 variant allele ratio in an 83-year-old woman with essential thrombocythemia who was negative for JAK2 V617F, CALR and W515K/L; the authors showed S505C alone did not produce constitutive STAT5 signaling or factor-independent Ba/F3 growth and was not sufficient for autonomous proliferation, indicating an affected carrier rather than a healthy-control observation.
• PMID:31697803 reports S505C in the literature only in combination with W515L (patients with the double mutation; one cohort patient with variant allele frequencies 25% and 50%), i.e. again an affected-patient context, and no unaffected carriers.
• COSMIC records this variant once as a somatic event (COSV65247010, somatic count 1), consistent with the residue being within the MPL exon 10 somatic hotspot, so any apparent carrier must be interpreted with respect to possible somatic/clonal origin rather than assumed constitutional.
• BS2 prerequisites applied from the generic ACMG/AMP 2015 framework (Richards et al. 2015, PMID:25741868); no MPL VCEP specification exists to modify this.

Still missing

• No population-cohort or case-control dataset of unaffected elderly individuals was available for MPL exon 10; the absence of healthy carriers rests on gnomAD v2.1/v4.1 absence plus the affected-only literature, not on an explicit healthy-control genotype study.
• gnomAD-Canada v1.0 lookup failed with a URL timeout, so a secondary genome dataset could not be screened for carriers.

Sources

Reasoning

BS3 is not met because the available assays establish only the absence of autonomous (cytokine-independent) receptor activation, not the absence of a damaging effect on protein function. PMID:34845187 shows S505C alone does not drive constitutive STAT5 signalling and does not support factor-free Ba/F3 growth, but the same experiments show that S505C forms more stable MPL dimers in the absence of TPO and augments autonomous STAT5 activity and proliferation when combined in cis with W515R or W515L - i.e. the variant is measurably not functionally neutral. PMID:31697803 likewise reports S505C in patients only together with a canonical W515 driver, and its second-site screen shows that MPL exon 10 variants of this class act predominantly as modifiers that exacerbate an independent driver mutation. Neither study tested a loss-of-function endpoint (TPO-stimulated JAK2/STAT5 signalling, cell-surface expression, or megakaryocytic differentiation), so no assay addresses the alternative MPL disease mechanism, and the 'no damaging effect' claim rests on non-activation data from one small single-patient functional report plus a domain-wide screen that did not report an individual S505C score. The directionally negative activation data are therefore captured in the PS3 assessment as absence of a stand-alone damaging effect, not as affirmative BS3 evidence.

Rule

ACMG/AMP 2015 BS3 - well-established functional studies show no damaging effect on protein function. Applied under the generic ACMG/AMP framework; no MPL-specific CSPEC/VCEP specification is available.

Evidence considered

• PMID:34845187: the assays that could support a benign/neutral reading are the negative ones - S505C alone produced no constitutive STAT5 reporter activity in HEK293T cells and no cytokine-independent proliferation in Ba/F3 cells, in experiments where W515R and W515L behaved as clear positive controls.
• PMID:34845187: the same study also shows a positive, non-neutral functional consequence of S505C - more stable dimers in the absence of TPO by oPDM cross-linking, and enhanced (not abolished) autonomous STAT5 activity and proliferation when S505C is compounded in cis with W515R or W515L, i.e. 'S505C augments pathological signaling'. This prevents a clean 'no damaging effect' conclusion.
• PMID:31697803: in the domain-wide saturation mutagenesis screen S505C was not constitutively activating as a single substitution, which is directionally consistent with a benign reading, but the study reports the variant in patients only as S505C+W515L (double mutants) and concludes that non-canonical MPL exon 10 variants should be examined as potential drivers or modifiers of oncogenic transformation rather than dismissed as functionally neutral.
• oncokb: OncoKB curates MPL S505C as 'Likely Oncogenic' (biological effect 'Likely Gain-of-function'), a database-level assertion that is not compatible with a 'no damaging effect' claim; its cited basis is the same two papers analysed here.
• Search context: no MPL-specific CSPEC/VCEP specification or local framework exists (cspec lookup returned no result), so BS3 was assessed under generic ACMG/AMP 2015. Because the readout available is activation-specific and the variant demonstrably retains a pro-signalling activity in cis, the evidence was judged insufficient to assert absence of a damaging effect.

Still missing

• No loss-of-function assay (TPO-stimulated signalling, receptor surface expression, megakaryocyte differentiation/maturation) has been performed for p.Ser505Cys, so a functionally damaging effect cannot be formally excluded for a loss-of-function MPL mechanism.
• The pro-signalling modification by S505C was tested only in the presence of a cis W515 driver in a single study (PMID:34845187); no assay of S505C alone against a full dose-response of TPO-stimulated signalling has been reported.
• No independent replication of the dimerisation/cross-linking result or of the Ba/F3 factor-free result exists outside the two papers cited, and both derive from the same small set of MPL exon 10 patient/cohort material.
• Whether the observed modifier effect is sufficient to be considered a 'damaging effect' for BS3 purposes is a judgement call that requires expert review of MPL-specific functional-evidence standards.

Sources

Reasoning

BP1 applies only to a missense variant in a gene for which missense variation is not an established mechanism of disease and disease is caused primarily by truncating variants. That premise fails for MPL. Activating missense substitutions in the exon 10 transmembrane domain are the canonical MPL disease alleles - 'canonical' MPL exon 10 driver mutations W515L/K/R/A and S505N 'generate constitutively active receptors and consequent loss of Tpo dependence' and are found in roughly 6% of JAK2 V617F-negative essential thrombocythemia and 14% of primary myelofibrosis cases, with S505N 'often found as a germline mutation in hereditary thrombocythemia' (PMID:31697803); PMID:34845187 likewise describes S505N as a common, functionally characterised driver of pathological activation. OncoKB's gene annotation makes the same point for the germline setting, assigning germline gain-of-function (missense) variants to autosomal dominant thrombocythemia while assigning germline loss-of-function (largely truncating) variants to recessive congenital amegakaryocytic thrombocytopenia. Although the recessive MPL disorder is indeed dominated by truncating variants, MPL is not a gene in which truncating variants are the primary or only disease-causing class, so BP1 cannot be applied; benignity is also not supported by the population data, as the variant is absent from gnomAD v2.1 and v4.1. Note this is a benign-evidence criterion being declined, not pathogenic evidence being claimed.

Rule

ACMG/AMP 2015 BP1 (missense variant in a gene for which primarily truncating variants are known to cause disease) - not satisfied: missense activation is a well-established MPL disease mechanism

Evidence considered

• PMID:31697803: missense activation is a canonical MPL disease mechanism - 'W515L/K/R/A or S505N ... generate constitutively active receptors', and S505N is 'often found as a germline mutation in hereditary thrombocythemia', which directly contradicts the BP1 premise that MPL disease is caused primarily by truncating variants.
• PMID:34845187: 'Mutations at the juxtamembrane residue W515 and the transmembrane variant S505N constitute the most common MPL mutations identified in MPN and have been functionally characterized as drivers of pathological activation' - missense variants are among the most common MPL disease alleles.
• OncoKB gene-level annotation: germline gain-of-function (missense) variants cause autosomal dominant thrombocythemia 2, while germline loss-of-function (predominantly truncating) variants cause recessive congenital amegakaryocytic thrombocytopenia - i.e. MPL has two distinct mechanisms, so BP1's truncating-predominance premise does not hold.
• Population evidence collected for this case (variant absent from gnomAD v2.1 and v4.1) provides no support for benignity and does not rescue BP1.

Still missing

• The weighting of BP1 is phenotype dependent: under a pure congenital amegakaryocytic thrombocytopenia (loss-of-function) framing the truncating-predominance argument would be stronger, but because MPL also has a well-established activating-missense mechanism (including at the very residue being assessed) BP1 remains inapplicable; the case's clinical indication is not recorded in the bundle to settle this.
• No quantitative breakdown of germline MPL variant classes (truncating versus missense) per disease was available; the determination rests on the mechanism evidence in the two full-text papers plus the OncoKB gene annotation.

Sources

Reasoning

Same single-mechanism determination as PP3, evaluated once for this missense variant: generic calibration uses REVEL only, with BP4 at REVEL < 0.250. The local REVEL score is 0.422, well above that cutoff and inside the 0.250-0.750 gray zone, so benign computational evidence for the protein-level effect is not supported. BayesDel is uncalibrated and unusable, and the low SpliceAI delta (0.019) cannot be repurposed as BP4 for a missense variant, since a non-splice prediction is not evidence about the substituted amino acid. The gray zone is a genuine result rather than missing data, so BP4 is not_met rather than not_assessed.

Rule

Generic ACMG/AMP in-silico calibration for missense variants (no MPL CSPEC/VCEP or local rule available): REVEL < 0.250 -> BP4 supporting; REVEL 0.250-0.750 -> neither PP3 nor BP4 met (thresholds from Pejaver et al., PMID 36413997). BayesDel excluded as uncalibrated.

Evidence considered

• Missense consequence of NM_005373.2:c.1513A>T (p.Ser505Cys) fixed the protein-impact (REVEL) path and excluded the splice path for this criterion.
• REVEL v1.3 local score = 0.422 (source_registry 'revel'; prefetch.json revel step; evidence.json results.revel). BP4 supporting requires REVEL < 0.250 per Pejaver et al., PMID 36413997 (ClinGen SVI REVEL calibration); 0.422 is in the 0.250-0.750 gray zone, so BP4 is not met.
• BayesDel (no-AF) = -0.0484473 was consulted (source_registry 'bayesdel') but rejected as uncalibrated for PP3/BP4 in this pipeline, so it neither supports nor blocks BP4.
• SpliceAI max delta 0.019 (source_registry 'spliceai'; evidence.json results.spliceai) was not used as benign computational evidence: for a missense variant the splice predictor is the wrong mechanism, and using it would double-count a non-coding prediction against the protein-level effect.
• No applicable gene-specific framework: cspec.found = false (case_summary.json) and final_classification_framework.source = generic_acmg_fallback (vcep_materials.json), so the generic REVEL-only rule governs.
• No conflicting computational benign assertion was found externally: this variant has no ClinVar record (evidence.json results.clinvar = no_result) and is absent from gnomAD v2.1 and v4.1, so no population-based benign argument was substituted for BP4 here.

Still missing

• REVEL 0.422 falls in the uninformative gray zone, leaving no calibrated predictor directionality (neither damaging nor benign) for the protein-level effect.
• BayesDel cannot break the tie: no verified published threshold for PP3/BP4 is available to this pipeline.
• No VCEP/CSPEC or local MPL PP3/BP4 lookup spreadsheet exists in this case, so no gene-specific preassigned BP4 could be applied.

Sources

Reasoning

BP5 requires the variant to be found in a case with an alternate molecular basis for disease. For the case under review there is no documented alternate molecular basis: the bundle contains no proband genotype or clinical data at all, so nothing establishes that this individual's phenotype is explained by a different variant. The literature does show every reported S505C carrier also carrying a canonical activating MPL W515 driver in cis (PMID:34845187: S505C+W515R, 0.37 allele ratio; PMID:31697803: S505C+W515L, VAFs 25%/50%), but those are other patients, the co-mutations are somatic clonal events in a hotspot disease, and neither paper presents S505C as a variant excluded from disease causation - on the contrary, both show S505C alone is insufficient for autonomous signalling while still enhancing W515-driven STAT5 activity and proliferation. A variant that demonstrably modifies the co-driver's signalling is exactly the situation in which BP5 should not be used to demote it, and applying BP5 on the basis of co-mutation in unrelated reported patients would misattribute another individual's genotype to this proband. BP5 is therefore not met, with human review flagged because the proband's own molecular workup (e.g. a documented JAK2 V617F, CALR or MPL W515 driver, or another established cause of the phenotype) is the information that would make BP5 applicable.

Rule

generic_acmg_2015: BP5 = variant found in a case with an alternate molecular basis for disease (Richards et al. 2015, PMID:25741868). No MPL CSPEC/VCEP or local gene-specific BP5 rule was available.

Evidence considered

• Criterion definition applied: BP5 = variant found in a case with an alternate molecular basis for disease (Richards et al. 2015, PMID:25741868), under generic ACMG/AMP because no MPL CSPEC/VCEP or local gene-specific specification exists (case_summary.json cspec.found=false; vcep_materials.json final_classification_framework.source=generic_acmg_fallback).
• Case-level alternate basis: the case bundle records only the variant NM_005373.2:c.1513A>T with no second variant, no co-occurring driver result, no phenotype and no family history (case_summary.json compact_evidence and literature_packet; ClinVar submitter_count=0), so no alternate molecular basis is documented for this proband.
• PMID:34845187 (full text, extracted) describes the pattern that would be relevant if it applied to this proband - 'MPL S505C and W515R mutations were detected in cis (Supplementary Fig. 1), verified by Sanger sequencing, with a 0.37 variant allele ratio' - but the patient is a separate 83-year-old individual and the co-mutation is a somatic clonal event; the same paper states 'We found that the S505C did not drive constitutive STAT5 signaling', 'The MPL S505C variant was not sufficient for factor-independent growth' and that S505C 'induces receptor dimerization, but ... is not sufficient for signaling to drive autonomous proliferation', i.e. S505C contributes to the phenotype rather than being excluded from it.
• PMID:31697803 (full text, extracted) reports S505C only in double-mutation carriers: 'Additional published reports of patients with double mutations include V501A1W515R, V501A1W515L, and S505C1W515L, and we identified a second instance of each of these combinations in our patient cohorts' (Figure 2D: W515L 25% / S505C 50%). This documents co-occurrence with a canonical driver in affected patients but provides no proband-level alternate basis for this case.
• Interpretation guard applied: because both sources indicate S505C modifies W515-driven signalling rather than being an inert bystander, BP5 was not awarded from literature co-mutation data; the criterion was left not met pending proband-specific data.
• No other source's final classification was used to infer this criterion; the OncoKB somatic oncogenicity label and the COSMIC somatic record describe biological effect and were not converted into a benign-leaning BP5 call.

Still missing

• No proband-level variant data (co-occurring JAK2 V617F, CALR exon 9, MPL W515 or other established pathogenic findings) are present in the bundle, so the central BP5 question - whether this case has an alternate molecular basis for disease - is unanswered for this individual.
• No proband phenotype is documented, so 'disease' cannot even be specified for the BP5 assessment in this case.
• The co-occurring canonical W515 drivers described in the literature belong to other reported patients and are somatic/clonal findings; they cannot be transferred to this proband, and no case series indicates that all S505C carriers inevitably carry a W515 driver.
• Functional evidence (PMID:34845187, PMID:31697803) shows S505C enhances W515-driven signalling, so S505C is not an inert bystander of the co-driver; whether BP5 may legitimately be applied to a modifier second hit remains a judgement call requiring clinical review.

Sources

Reasoning

BP6 requires a reputable source to report this exact variant as benign, and for this group that means an exact-variant ClinGen expert-panel Benign/Likely benign classification. No such assertion exists: NM_005373.2:c.1513A>T has no ClinVar record at all (multiple search strings, exact_match=false, zero submissions, no expert panels), so there is neither a benign expert-panel classification nor a benign aggregate label to draw on. The only external curated assertion found is OncoKB's variant-specific 'Likely Oncogenic' / 'Likely Gain-of-function' label, which is an oncology curation pointing in the opposite direction and is not a ClinVar expert-panel germline classification; under this group's rule it cannot trigger BP6, and it was not used to argue against the criterion either. Evidence that might be mistaken for BP6 was explicitly not repurposed: the variant's low BayesDel score is uncalibrated in this pipeline and cannot be used, the SpliceAI max delta of 0.019 concerns splicing rather than the missense effect and belongs to a different group, and gnomAD absence is a rarity finding rather than a benign assertion. BP6 is therefore not met.

Rule

generic_acmg_2015: BP6 = reputable source reports the variant as benign (Richards et al. 2015, PMID:25741868), applied only for an exact-variant ClinGen expert-panel Benign/Likely benign classification (supporting strength); non-expert submissions, aggregate ClinVar labels and other curated databases excluded. No MPL CSPEC/VCEP available.

Evidence considered

• Criterion definition applied: BP6 = reputable source reports the variant as benign (Richards et al. 2015, PMID:25741868), restricted for this group to an exact-variant ClinGen expert-panel Benign/Likely benign classification at supporting strength; non-expert assertions, aggregate ClinVar labels and other curated databases cannot trigger it.
• ClinVar lookup (source_registry 'clinvar'): found=false, search_status='no_result', exact_match=false, variation_ids=[], submitter_count=0, expert_panels=[], recorded sentence 'This variant is absent from ClinVar.' (searches included c.1513A>T, NM_005373.2:c.1513A>T and g.43814978A>T). There is therefore no benign assertion of any tier for BP6 to inherit.
• OncoKB (source_registry 'oncokb') returned 'Likely Oncogenic' with biological effect 'Likely Gain-of-function' and PMIDs 31697803/34845187 - an oncology curation, not a ClinVar expert panel, and directionally the opposite of benign; it was excluded from BP6 rather than inverted into evidence.
• Evidence deliberately not converted into BP6: REVEL 0.422 sits in the uninformative 0.250-0.750 interval (Pejaver et al., PMID 36413997) and BayesDel (-0.0484473) has no verified published threshold available to this pipeline, so neither gives a benign computational signal; SpliceAI max delta 0.019 describes splice impact and is handled by the in-silico/splicing group, not repurposed as a benign assertion here.
• Population data do not support BP6 either: the variant is absent from gnomAD v2.1 and gnomAD v4.1 (zero frequency) and the gnomAD-Canada v1.0 query timed out, so no frequency-based benign argument exists.
• Framework check: no MPL CSPEC/VCEP was retrievable (cspec.found=false; vcep_materials.json files_found=[]), and the governing framework is the generic ACMG/AMP 2015 fallback (final_classification_framework.source=generic_acmg_fallback), so BP6's generic definition applies unchanged.
• No other source's final classification was used to infer this criterion.

Still missing

• No ClinVar record (benign or pathogenic) exists for this variant, so no expert-panel benign assertion is available; BP6 cannot be satisfied unless such a submission appears in future.
• No ClinGen CSPEC/VCEP exists for MPL (cspec.found=false; vcep_materials empty), so no gene-specific expert-panel benign assertion list could be consulted.
• No calibrated computational benign argument exists that could inform BP6 here: REVEL is 0.422 (gray zone) and BayesDel lacks a verified published threshold in this pipeline, so BP6 is not merely unmet but unsupported from any direction.

Sources

Reasoning

PS2 requires a confirmed de novo occurrence, i.e. the variant present in the affected proband and demonstrably absent in both biological parents (parentage confirmed). No proband genotype record, no parental testing, no parentage confirmation and no de novo observation for NM_005373.2:c.1513A>T (NP_005364.1:p.(Ser505Cys)) exists in the retrieved evidence: ClinVar has no record for the variant, and neither variant-specific publication reports a proband with parental studies. The one patient-level report describes the variant in bone marrow DNA in cis with W515R at a 0.37 variant allele ratio in an 83-year-old with essential thrombocythemia, and the other reports S505C as a second MPL exon 10 mutation in MPN patients; both are consistent with a somatically acquired lesion and neither establishes a de novo germline origin. PS2 is therefore not_assessed for absence of the required proband/parental data, not because a de novo occurrence was excluded.

Rule

Generic ACMG/AMP 2015 (PMID 25741868): PS2 (pathogenic, strong) - de novo (both maternity and paternity confirmed) in a patient with the disease and no family history. Applied because no MPL CSPEC/VCEP or local gene-specific framework exists. Not met as assessed: no proband/parental testing data and no confirmed de novo report for this variant.

Evidence considered

• PS2 (generic ACMG/AMP 2015, PMID 25741868) requires the variant to be confirmed de novo in a patient with a confirmed parental relationship; no VCEP/CSPEC or local MPL gene-specific specification exists (case_summary.json cspec.found = false; vcep_materials.json final_classification_framework.source = generic_acmg_fallback, official_clingen = false), so the generic definition governs.
• Variant identity checked in all sources: NM_005373.2:c.1513A>T = NP_005364.1:p.(Ser505Cys) = p.S505C, MPL exon 10, GRCh37 1:43814978A>T, GRCh38 1:43349307A>T.
• PMID:34845187 reports one ET patient with MPL S505C and W515R in cis, detected by high-sensitivity targeted sequencing of bone marrow DNA and confirmed by Sanger sequencing, with a 0.37 variant allele ratio; there is no parental testing, no family testing and no de novo statement. The report states 'With no family history of MPN, she presented at 83 years of age'.
• PMID:31697803 Figure 2D lists a 57-year-old female ET patient with W515LSom+S505C (VAF 25%, 50%) and cites earlier reports of S505C+W515L double mutations; S505C carries no 'Ger' (germline) superscript in that annotation scheme, whose legend states that germline vs somatic is only marked 'Where VAF values, knowledge of familial disease, and/or availability of lymphocyte DNA sequences allow differentiation'. No de novo or trio data are given.
• ClinVar has no record for this variant (evidence.json results.clinvar search_status = no_result; case_summary.json clinvar.evidence_sentence = 'This variant is absent from ClinVar'), so no external de novo assertion could be reviewed.
• No case-level source in this evidence set reports a proband or a de novo event for this variant, so PS2 cannot be evaluated; status is not_assessed rather than not_met because the decisive data (parental genotypes) were never collected, not because they were negative.

Still missing

• No proband-level genotype documentation for NM_005373.2:c.1513A>T is available in this case (no clinical case record, no trio data).
• No parental testing for c.1513A>T is reported in any source, and parentage/biological-relationship confirmation is absent, so presence-in-proband/absence-in-both-parents cannot be established.
• No de novo occurrence of this variant is reported in either variant-specific publication; the reported observations are in a somatic myeloproliferative-neoplasm context (bone marrow DNA, VAF 0.37 in cis with W515R; unmarked germline/somatic status at VAF 50% in one ET patient).
• Germline origin itself has not been demonstrated for this variant in any available source (no lymphocyte/non-haematopoietic tissue confirmation), which is the prerequisite for a de novo germline PS2 assessment.

Sources

Reasoning

PM6 requires an assumed de novo occurrence without confirmation of maternity and paternity. Even at that lower evidentiary bar, no source reports NM_005373.2:c.1513A>T as a de novo event: there is no proband record in this case, no parental genotypes, and the variant-specific publications describe the change in a somatic myeloproliferative-neoplasm context (MPL S505C in cis with W515R at VAF 0.37 in bone marrow DNA; S505C as a second exon 10 mutation in MPN patients) without any claim of de novo origin. PM6 is therefore not_assessed for absence of any de novo report, not because a de novo event was excluded.

Rule

Generic ACMG/AMP 2015 (PMID 25741868): PM6 (pathogenic, moderate) - assumed de novo, but without confirmation of maternity and paternity. Applied because no MPL CSPEC/VCEP or local gene-specific framework exists. Not assessed: no de novo report, no proband or parental data for this variant.

Evidence considered

• PM6 (generic ACMG/AMP 2015, PMID 25741868) applies to an assumed de novo occurrence without confirmation of maternity and paternity; no MPL CSPEC/VCEP or local gene-specific specification exists (cspec.found = false; final_classification_framework.source = generic_acmg_fallback), so the generic definition governs.
• PMID:34845187 describes a single 83-year-old ET patient with S505C and W515R in cis (VAF ratio 0.37) detected in bone marrow DNA, and explicitly states no family history of MPN; the variant is reported as a patient-derived double mutation studied functionally, with no statement about de novo occurrence and no parental testing.
• PMID:31697803 reports S505C only as a second-site MPL exon 10 mutation co-occurring with canonical W515 variants (patient W515LSom+S505C, VAF 25% and 50%; published S505C+W515L combinations), i.e. a somatic-driver context, and provides no de novo or family data.
• ClinVar has no record for this variant (evidence.json results.clinvar search_status = no_result), so no submitted de novo assertion exists to review.
• Per the case instruction that PS2/PM6 may only be applied if a specific de novo report for this variant is read, and no such report exists in the retrieved evidence, PM6 cannot be met or denied on evidence and is returned as not_assessed.

Still missing

• No de novo report for c.1513A>T exists in any retrieved source; the required proband observation (affected individual carrying the variant with unaffected parents untested) is entirely absent.
• No parental genotypes or parentage information are available, so neither confirmed (PS2) nor assumed (PM6) de novo status can be described.
• The available patient-level data describe the variant as a somatically acquired MPN lesion (bone marrow DNA; VAF 0.37 in cis with W515R; VAF 50% with an unmarked germline/somatic annotation in one ET patient), and germline status has not been established in any source.
• A de novo assessment would require a clinical case record (proband phenotype, age at onset, parents' status); no such record is present in this case bundle.

Sources

Reasoning

PP1 requires co-segregation of the variant with disease in multiple affected family members (meioses counted in a pedigree). No pedigree, family-member genotyping, or meiosis information exists for NM_005373.2:c.1513A>T in any retrieved source: the variant is reported in individual myeloproliferative-neoplasm patients (one 83-year-old with S505C in cis with W515R, one 57-year-old with W515L+S505C) with no relatives tested, and the single case explicitly documents no family history of MPN. PP1 is therefore not_assessed for absence of segregation data; the absence of a reported family history in a single case is not itself segregation evidence.

Rule

Generic ACMG/AMP 2015 (PMID 25741868): PP1 (pathogenic, supporting) - co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease; strength may be increased with more meioses. Applied because no MPL CSPEC/VCEP or local gene-specific framework exists. Not assessed: no pedigree, no genotyped relatives, no meiosis data for this variant.

Evidence considered

• PP1 (generic ACMG/AMP 2015, PMID 25741868) requires co-segregation with disease in multiple affected family members; no MPL CSPEC/VCEP or local gene-specific specification exists (cspec.found = false; final_classification_framework.source = generic_acmg_fallback), so the generic definition governs.
• PMID:34845187 reports a single ET patient carrying MPL S505C (in cis with W515R, VAF ratio 0.37) and states 'With no family history of MPN'; no relatives were genotyped and no pedigree or meiosis data are presented, so this source neither supports nor refutes co-segregation.
• PMID:31697803 reports S505C in individual MPN patients with multiple MPL exon 10 mutations (patient 57 F ET, W515LSom+S505C, VAF 25% and 50%) and surveys published double-mutation reports; it contains no family study, pedigree or segregation analysis.
• The germline MPL segregation literature referenced in the same sources concerns S505N (hereditary thrombocythemia founder pedigrees, e.g. the eight Italian pedigrees cited in PMID:31697803), a different amino-acid substitution, and cannot be transferred to S505C.
• No case-level family or pedigree data are present anywhere in the evidence bundle, so PP1 is not_assessed rather than not_met: the required observation (genotyped affected relatives) was never made.

Still missing

• No pedigree, no genotyped relatives and no meiosis count are available for c.1513A>T, so co-segregation with disease cannot be counted at any strength.
• Sources report individual patients only (PMID:34845187: one ET patient, no family history of MPN; PMID:31697803: one ET patient with W515LSom+S505C and cited double-mutation reports), none with family studies.
• The only MPL germline family/segregation literature identified concerns the distinct S505N substitution in hereditary thrombocythemia pedigrees; that evidence is variant-non-specific for S505C and cannot be used for PP1.
• A PP1 assessment would require the proband's family structure plus genotype/phenotype data from affected and unaffected relatives; none of this exists in the case bundle.

Sources

Reasoning

PP2 has two required elements: the gene must have a relatively low rate of benign missense variation, and missense variants must be a common mechanism of disease. Only the second element can be evaluated with the evidence available for this case, and it is satisfied for MPL: activating missense substitutions in the exon 10 transmembrane domain are the canonical MPL drivers (S505N, W515L/K/R/A; PMID:31697803; PMID:34845187), and the OncoKB gene annotation assigns germline gain-of-function missense variants to autosomal dominant thrombocythemia. The first element cannot be evaluated: no gene-level missense constraint metric (for example a gnomAD missense Z score or observed/expected ratio) and no rate of benign missense variation for MPL were collected in this bundle, and the absence of this single variant from gnomAD v2.1/v4.1 says nothing about the gene's overall tolerance of missense change. Because a required component of the criterion is unverifiable rather than negative, PP2 is returned as not_assessed instead of met or not_met.

Rule

ACMG/AMP 2015 PP2 (missense variant in a gene with a low rate of benign missense variation and in which missense variants are a common mechanism of disease) - not scored: gene-level missense constraint evidence unavailable

Evidence considered

• PMID:31697803: the disease mechanism of MPL in the transmembrane domain is missense-driven - 'canonical' MPL exon 10 mutations W515L/K/R/A and S505N 'generate constitutively active receptors and consequent loss of Tpo dependence', i.e. missense variants are a common mechanism of MPL disease (second PP2 element satisfied).
• OncoKB gene-level annotation: 'Germline gain-of-function mutations are associated with autosomal dominant thrombocythemia 2 (elevated platelet count and an increased risk for myeloproliferative neoplasm), whereas germline loss-of-function mutations are associated with autosomal recessive congenital amegakaryocytic thrombocytopenia' - again missense (gain-of-function) is a disease mechanism, but this annotation also shows MPL legitimately carries a truncating loss-of-function mechanism, which bears on the missense-benign-rate assumption.
• Population data collected for this case (gnomAD v2.1 and v4.1: variant absent) concern only the single variant and provide no gene-level measure of benign missense variation.
• No missense constraint metric, in silico gene-tolerance score or benign-missense catalogue for MPL was found in the case evidence bundle, so the 'low rate of benign missense variation' element was treated as not_available rather than negative.

Still missing

• No MPL gene-level missense constraint metric (e.g. gnomAD missense Z / observed-to-expected ratio) or benign missense variation rate is present in this case bundle; PP2 cannot be scored to completion without it.
• The bundle also carries no HCI prior or custom PP2 score for MPL (hci_prior found = false, reason 'gene_not_supported'), so no substitute gene-level missense-tolerance quantification was available.
• Consistency caveat: PP2 is gene-level and would apply identically to any MPL missense variant; it should not be used to differentiate pathogenic from non-pathogenic changes at a single residue.

Sources

Reasoning

BS4 requires lack of segregation in affected family members - genotyping of multiple affected relatives with the variant absent in at least one affected relative. No family members were tested for NM_005373.2:c.1513A>T in any retrieved source, so non-segregation was never evaluated. The statement in PMID:34845187 that the patient had 'no family history of MPN' describes a single case with no affected relative genotyped and cannot be scored as non-segregation. BS4 is therefore not_assessed for absence of family genotype data.

Rule

Generic ACMG/AMP 2015 (PMID 25741868): BS4 (benign, strong) - lack of segregation in affected members of a family. Applied because no MPL CSPEC/VCEP or local gene-specific framework exists. Not assessed: no affected relative was genotyped, so non-segregation was never evaluated.

Evidence considered

• BS4 (generic ACMG/AMP 2015, PMID 25741868) requires lack of segregation in affected family members; no MPL CSPEC/VCEP or local gene-specific specification exists (cspec.found = false; final_classification_framework.source = generic_acmg_fallback), so the generic definition governs.
• PMID:34845187: the patient's DNA source was bone marrow cells (high-sensitivity targeted sequencing, VAF ratio 0.37 for S505C/W515R), and the report states she presented with no family history of MPN. No relatives were tested, so the absence of a family history is not a non-segregation observation.
• PMID:31697803: S505C is reported in individual MPN patients (W515LSom+S505C, VAF 25% and 50%) and in cited double-mutation reports; no family members were genotyped and no segregation analysis was performed, so this source provides no BS4-relevant data.
• No case-level family data exist in the evidence bundle (no pedigree, no relative genotypes), so BS4 is not_assessed rather than not_met: lack of segregation was never tested, and a single sporadic patient report cannot substitute for it.

Still missing

• No affected family members were genotyped for c.1513A>T, so absence of the variant in an affected relative (the observation BS4 requires) is not documented anywhere.
• The single reported case (PMID:34845187) states no family history of MPN but tested only bone marrow DNA from the patient; a negative family history without relative genotyping is not non-segregation evidence.
• No pedigree or family structure is available in the case bundle, so the number of affected relatives available for a non-segregation test is unknown.
• BS4 should not be applied from a single sporadic case report; a formal assessment needs genotyping of at least two affected family members (or equivalent family data) that fails to co-segregate.

Sources