Protein Dossier — GPNMB (Transmembrane glycoprotein NMB)
MR feasibility tier: A — Published pQTL-MR estimates exist for this protein (retrieved below - not computed here).
1. Published MR estimates (retrieved, not computed)
| Outcome |
beta |
se |
p |
method |
nSNP |
cis/trans |
coloc |
| Diagnoses - main ICD10: J33 Nasal polyp |
-0.00148 |
0.000544 |
0.00659 |
Inverse variance weighted |
2 |
cis |
NA |
| Diagnoses - main ICD10: J33 Nasal polyp |
-0.00148 |
0.000544 |
0.00659 |
Inverse variance weighted |
2 |
cis |
NA |
| Diagnoses - main ICD10: C50 Malignant neoplasm of breast |
0.00194 |
0.000759 |
0.0106 |
Inverse variance weighted |
2 |
cis |
NA |
| Diagnoses - main ICD10: C50 Malignant neoplasm of breast |
0.00194 |
0.000759 |
0.0106 |
Inverse variance weighted |
2 |
cis |
NA |
| Diastolic blood pressure automated reading |
0.0145 |
0.00609 |
0.0176 |
Inverse variance weighted |
2 |
cis |
NA |
| Diastolic blood pressure automated reading |
0.0145 |
0.00609 |
0.0176 |
Inverse variance weighted |
2 |
cis |
NA |
| Schizophrenia |
0.0521 |
0.0224 |
0.0203 |
Inverse variance weighted |
2 |
cis |
NA |
| Schizophrenia |
0.0521 |
0.0224 |
0.0203 |
Inverse variance weighted |
2 |
cis |
NA |
| Fractured bone site(s): Arm |
-0.0014 |
0.000604 |
0.0205 |
Inverse variance weighted |
2 |
cis |
NA |
| Fractured bone site(s): Arm |
-0.0014 |
0.000604 |
0.0205 |
Inverse variance weighted |
2 |
cis |
NA |
| Systolic blood pressure automated reading |
0.014 |
0.00608 |
0.0212 |
Inverse variance weighted |
2 |
cis |
NA |
| Systolic blood pressure automated reading |
0.014 |
0.00608 |
0.0212 |
Inverse variance weighted |
2 |
cis |
NA |
| …and 154 more outcomes (see JSON) |
|
|
|
|
|
|
|
2. pQTL instrument availability (Tier-B probe)
| Dataset |
Trait |
Author |
Year |
prot-c-5080_131_3 |
GPNMB |
Suhre K |
2019 |
3. GWAS Catalog results — traits with signal at this locus
50 association rows across 21 traits (48 genome-wide significant rows). Associations are loci, not causal claims; the mapped gene at a locus is not necessarily the effector gene.
| Trait |
best p |
lead SNP |
n assoc |
study |
MR status |
| Circulating GPNMB levels (id: OID05139_OID20173) |
1e-666 |
rs75801644 |
5 |
GCST90860688 |
no MR -> candidate analysis |
| Circulating GPNMB levels (id: OID00749_OID20173) |
3e-603 |
rs75801644 |
5 |
GCST90860087 |
no MR -> candidate analysis |
| Transmembrane glycoprotein NMB:Extracellular domain (analyte |
8e-299 |
rs858275 |
1 |
GCST90427313 |
no MR -> candidate analysis |
| Transmembrane glycoprotein NMB:Extracellular domain (analyte |
3e-274 |
rs858275 |
1 |
GCST90426228 |
no MR -> candidate analysis |
| GPNMB protein levels |
4e-242 |
rs191297708 |
3 |
GCST90469393 |
no MR -> candidate analysis |
| Cerebrospinal fluid protein GPNMB levels |
3e-240 |
rs858275 |
1 |
GCST90944341 |
no MR -> candidate analysis |
| Transmembrane glycoprotein NMB:Extracellular domain (analyte |
2e-215 |
rs858275 |
1 |
GCST90427285 |
no MR -> candidate analysis |
| Transmembrane glycoprotein NMB levels |
4e-105 |
rs1881203 |
7 |
GCST90249856 |
no MR -> candidate analysis |
| Transmembrane glycoprotein NMB:Cytoplasmic domain levels |
1e-103 |
rs858275 |
1 |
GCST90427427 |
no MR -> candidate analysis |
| Height |
2e-65 |
rs13240597 |
1 |
GCST90245848 |
MR: beta=-0.00794, p=0.34 (cis) |
| Serum levels of protein GPNMB |
7e-37 |
rs1881203 |
4 |
GCST90090106 |
no MR -> candidate analysis |
| Transmembrane glycoprotein NMB levels (GPNMB.8606.39.3) |
1e-28 |
rs2268748 |
4 |
GCST90243078 |
no MR -> candidate analysis |
| …and 9 more traits (see JSON) |
|
|
|
|
|
4. Phenome map — where this gene is a genetic locus, vs. where MR exists
Top diseases by Open Targets association (of 988 total). Associations are loci, not causal claims. The causal-status column is a four-state triage per pair: established (curated) = a curated clinical assertion exists (ClinGen/G2P/GEL/Orphanet/ClinVar — any validity level, MR adds little); exploratory rare-variant signal = ExWAS burden evidence without curation — a candidate NEW gene-disease relationship; common-variant locus = GWAS signal, classic pQTL-MR territory; multi-layer = burden+GWAS together, an allelic-series candidate (the strongest causal setup). Burden estimand is carrier-vs-noncarrier, not per-SD MR.
| Disease |
genetic assoc. |
burden (ExWAS) |
causal status |
MR status |
| familial primary localized cutaneous amyloidosis |
0.837 |
— |
established (curated) |
no MR -> candidate analysis |
| Parkinson disease |
0.782 |
— |
common-variant locus |
no MR -> candidate analysis |
| amyloidosis cutis dyschromia |
0.608 |
— |
established (curated) |
no MR -> candidate analysis |
| endometriosis |
0.59 |
— |
common-variant locus |
no MR -> candidate analysis |
| temporomandibular joint disorder |
0.51 |
— |
common-variant locus |
no MR -> candidate analysis |
| type 2 diabetes mellitus |
0.43 |
— |
common-variant locus |
no MR -> candidate analysis |
| response to antihypertensive drug |
0.384 |
— |
common-variant locus |
no MR -> candidate analysis |
| hereditary disease |
0.317 |
— |
established (curated) |
no MR -> candidate analysis |
| liver disorder |
0.246 |
— |
common-variant locus |
no MR -> candidate analysis |
Of the 9 rows above, 9 have no MR estimate in this resource. Across all retrieved diseases for this gene: 0 exploratory rare-variant signal(s), 0 multi-layer (allelic-series candidate) pair(s). Final triage still belongs to a statistical geneticist.
5. Downstream annotation (druggability & safety preview)
| Layer |
Result |
| ChEMBL druggability |
2 known modulators (Transmembrane glycoprotein NMB) |
| gnomAD constraint |
pLI=2.9e-29, LOEUF=1.51 — LoF-tolerant |
| GWAS Catalog |
44 unique SNPs / 87 rows |
| ClinVar |
174 records; 4 pathogenic in sample of 30 |
| PharmGKB/ClinPGx |
no annotations |
phenome — Top 30 of 988 associated diseases by overall score. genetic_association aggregates GWAS common-variant AND rare-variant evidence. These are ASSOCIATIONS (loci), not causal claims.chembl — ChEMBL target matched by text search on ‘GPNMB’ and resolved to ‘Transmembrane glycoprotein NMB’ — confirm this is the intended target.clinvar — Pathogenic count is over the 30 record(s) retrieved, NOT over all 174 ClinVar records for this gene; it is a sample, not a rate.pharmgkb — No PharmGKB/ClinPGx clinical annotations (gene may not be a pharmacogene).gwas_traits — Top 20 of 21 traits by best p-value, aggregated from 50 association rows. These are GWAS ASSOCIATIONS (loci), not causal claims; mapped genes at a locus are not necessarily the effector gene.
Sources
uniprot: https://www.uniprot.org/uniprotkb/Q14956 — UniProt release 2026_02 (10-June-2026)mr_outcomes: https://epigraphdb.org/pqtl/ — EpiGraphDB pQTL MR (Zheng et al., Nat Genet 2020) — pre-computed two-sample MR; retrieved, not computed by this agent; EpiGraphDB build 1.0, pQTL dataset v3.0phenome: https://platform.opentargets.org/target/ENSG00000136235/associations — Open Targets data release 26.06chembl: https://www.ebi.ac.uk/chembl/target_report_card/CHEMBL3712919/ — ChEMBL_37 (released 2026-05-01)gnomad: https://gnomad.broadinstitute.org/gene/GPNMB — gnomAD constraint via GraphQL API (reference genome GRCh38)gwas: https://www.ebi.ac.uk/gwas/genes/GPNMB — GWAS Catalog REST (live; release not exposed by this endpoint)clinvar: https://www.ncbi.nlm.nih.gov/clinvar/?term=GPNMB%5Bgene%5D — ClinVar build Build260809-1055.1gwas_traits: https://www.ebi.ac.uk/gwas/genes/GPNMB — GWAS Catalog search API (live; release not exposed)
Provenance
- Generated: 2026-08-14T02:53:32 · Tier: A
- Fully mechanical: every cell above is rendered from tool return values. No language model wrote any part of this dossier.
- MR estimates, where present, are retrieved from published work (EpiGraphDB pQTL, Zheng et al. Nat Genet 2020); nothing is computed here.
- Tool errors this run: none