GOLGA2 (GM130) - a Golgi matrix protein with 249 partners in the one yeast screen that reported it binding AFF4, which had five
also called GOLGA2 · GM130 · Golgin subfamily A member 2 · 130 kDa cis-Golgi matrix protein · Golgin-95 · Q08379
GOLGA2, better known as GM130, is a scaffolding protein that holds the Golgi apparatus together. It was reported as a partner of the nuclear protein AFF4 in one large yeast screen, and in that same screen it bound 249 different proteins.
The same thing, for a scientist
GOLGA2 (UniProt **Q08379, 1,002 aa**) is the cis-Golgi matrix protein GM130, a long coiled-coil golgin that maintains Golgi structure and tethers vesicles. It appears in this graph only because the 2026-09-21 interactome recount listed it among the seventeen unwritten bait-supported partners of the six core proteins, on the strength of **three IntAct rows that are three assay annotations of a single yeast two-hybrid screen** [E1]. **The entry is `speculative` and the number that decides it is this: within that same screen GOLGA2 has 249 distinct partners and AFF4 has 5** [E2]. GOLGA2 is also the highest-degree protein in HuRI at 246 partners, and while it appears there in quantity the AFF4 pair does not appear at all [E2]. UniProt places GOLGA2 at the **cis-Golgi network membrane, the ER-Golgi intermediate compartment and the spindle pole** - and nowhere in the nucleus - while AFF4 is nuclear and chromatin-associated [E3]. `(GOLGA2[tiab] OR GM130[tiab]) AND (MLL[tiab] OR KMT2A[tiab])` returns **0**, and the two leukaemia records title-check to an RPS6KA1 paper and a 1982 adenosine deaminase paper [E4]. **There is no biology here and this entry does not propose any.** It exists because the recount named GOLGA2 as a gap and a reader who finds the name in that list is better served by a record that says why it is not a partner than by no record at all.
This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.
I have never read one of these pictures before
A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.
Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.
Anything in yellow sticks is a bound small molecule, which is usually the drug.
Everything the evidence is
AFF4 as DB-fusion bait and GOLGA2 as AD-fusion prey in yeast, scored in a pooled primary screen, retested on an array, retested pairwise. Three MITAB rows, one publication, one screen, MI-score 0.56, host S. cerevisiae.
No mammalian cell. No endogenous protein. No mapped region, affinity, structure or reciprocal assay. Not recovered in the successor screen. The bait is wild-type AFF4, not AFF1 and not a fusion.
And the prey binds 249 things in the same experiment. A pair in which one member is a 249-degree prey and the other a 5-degree bait tells you about the prey.
Why an entry at all
Two reasons, both about the graph rather than about GOLGA2.
First, the gap list is now public inside this project. docs/coverage.md names GOLGA2 as one of thirteen partners "not mentioned anywhere in the corpus". Leaving it unmentioned means the next person to read that list runs the same searches and reaches the same place. This entry converts a gap into a recorded rejection, which is the distinction brief section Write down what you read and did NOT use asks for.
Second, GM130 is a name a reader will recognise and mis-weight. It is a heavily studied protein with a real literature, and "GM130 is an AFF4 interactor" would read as substantive to anyone who did not check where the row came from. The entry is the check, written down once.
If a future experiment places GM130 near AFF4 or AFF1 in a human cell, this entry should be rewritten, not deleted - the degree argument is a prior, not a proof.
| uniprot | Q08379 |
|---|---|
| gene symbol | GOLGA2 |
| length AA | 1002 |
| alphafold ID | AF-Q08379-F1 |
| protein class | golgin; long coiled-coil cis-Golgi matrix protein |
| uniprot resolution | gene_exact:GOLGA2 AND organism_id:9606 AND reviewed:true returned ONE reviewed row on 2026-09-21; primary gene name read back as GOLGA2; entry GOGA2_HUMAN; 1,002 aa; no gene synonyms. |
| localization | Golgi apparatus, cis-Golgi network membrane; ER-Golgi intermediate compartment membrane; cytoplasm, cytoskeleton, spindle pole (UniProt Q08379, retrieved 2026-09-21). No nuclear annotation. |
| localization evidence code | ECO:0000269 and ECO:0000305 on the cis-Golgi call; ECO:0000269 on the ERGIC and spindle-pole calls |
| domains note | No domain boundaries asserted; GM130 is predominantly coiled-coil and the segment ranges were not read back from UniProt with residue numbers this cycle. |
| shown for fusion | Nobody has looked in any fusion. The assay is wild-type AFF4 against GOLGA2 in yeast; there is no GOLGA2 experiment in AFF1, in any KMT2A fusion, or in any human cell. |
| verdict | Written as a recorded rejection rather than as a partner. The row is real, the interaction is not usable, and the reason is a prey degree of 249 in the experiment that reported it. |
domains
[]
interaction with the core complex
{
"partner": "AFF4",
"partner_accession": "Q9UHB7",
"our_protein_role": "prey (AD fusion) in two rows; neutral component in the validation row",
"methods": [
"psi-mi:\"MI:1112\"(two hybrid prey pooling approach)",
"psi-mi:\"MI:0397\"(two hybrid array)",
"psi-mi:\"MI:1356\"(validated two hybrid)"
],
"host_system": "Saccharomyces cerevisiae (taxid:4932)",
"source_publications": [
"25416956"
],
"distinct_experiments": 1,
"intact_acs": [
"intact:EBI-10542896",
"intact:EBI-10321217",
"intact:EBI-10473953"
],
"intact_miscore": 0.56,
"replicated_in_huri_luck_2020": false,
"region_mapped": false,
"affinity_measured": false,
"structures": [],
"compartment_mismatch": "GOLGA2 cis-Golgi/ERGIC/spindle pole, AFF4 nuclear and chromatin-associated"
}promiscuity controls 2026 09 21
{
"distinct_partners_in_rolland_2014": 249,
"distinct_partners_in_huri_luck_2020": 246,
"distinct_partners_all_of_intact": 652,
"distinct_publications_all_of_intact": 46,
"comparator_aff4_in_rolland_2014": 5,
"share_of_the_2014_map": "roughly 1.8% of ~14,000 reported interactions involve GOLGA2"
}counted absences
{
"kmt2a": {
"query": "(GOLGA2[tiab] OR GM130[tiab]) AND (MLL[tiab] OR KMT2A[tiab])",
"records": 0
},
"leukaemia": {
"query": "(GOLGA2[tiab] OR GM130[tiab]) AND (leukemia[tiab] OR leukaemia[tiab])",
"records_raw": 2,
"records_on_topic": 0,
"pmids": [
"39333927",
"6981287"
]
},
"aliases_searched": [
"GOLGA2",
"GM130"
],
"aliases_not_searched": [
"golgin-95",
"golgin subfamily A member 2"
],
"date": "2026-09-21",
"bound": "zero in titles and abstracts over two aliases; supporting figures not searched"
}- Nothing here touches an intervention. The hazard is editorial: GM130 is a recognisable protein and 'GM130 binds AFF4' would propagate into a summary far more readily than the sentence explaining where the row came from.
- GOLGA2 is annotated at the spindle pole and this graph carries mechanism-mll1-bookmarks-mitotic-chromatin-and-carries-menin-there. Is there any mitotic window in which a Golgi matrix protein and a chromatin-bookmarking complex are in the same compartment? Almost certainly not functionally, but it is the only route by which this pair could be real and nobody has looked.
- What fraction of Rolland 2014's proteins exceed 249 partners? Until that denominator exists, 'GOLGA2 is a hub' is a comparison against AFF4 rather than against the screen.
- E1weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.database
IntAct records AFF4 and GOLGA2 as interacting in three rows, all from PMID 25416956 / IMEx IM-23318, in host Saccharomyces cerevisiae, annotated two hybrid prey pooling approach, two hybrid array and validated two hybrid, MI-score 0.56, with no other source publication.
IntAct molecular interaction database, PSICQUIC MITAB 2.7, query id:Q08379 AND id:Q9UHB7, 3 rows retrieved 2026-09-21 · IntAct (EMBL-EBI) 2026What this evidence cannot show
Three rows, one publication, one screen, one assay class. Not recovered in HuRI (Luck 2020), although IntAct cannot distinguish 'tested and failed' from 'not in the search space'. No region, affinity, structure or mammalian-cell evidence, and the bait is wild-type AFF4.
Evidence that argues against this (1)
- Rolland et al. argue their systematic map is at least as accurate as the literature-curated interactome and deliberately unbiased, so the row is not low-grade by construction. The discount here is prey degree and compartment, both specific to this pair. Rolland T et al. 2014
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
GOLGA2 has 249 distinct partners within PMID 25416956 and 246 within HuRI (PMID 32296183), against AFF4's 5 in the former and 0 rows in the latter; GOLGA2 alone accounts for roughly 1.8% of the approximately 14,000 interactions in the 2014 map.
The four AFF4 two-hybrid partners are three rows of one screen, not three experiments · Oncofusion knowledge graph, ana-the-aff4-two-hybrid-quartet-is-one-screen-and-three-of-the-four-preys-are-hubs 2026What this evidence cannot show
Degrees computed from IntAct's import of each publication on 2026-09-21, deduplicated to distinct accessions. The 1.8% figure divides a deduplicated partner count by the paper's abstract figure for total interactions and is an order-of-magnitude statement, not an exact proportion. No base rate was computed for the screen's degree distribution, so the load-bearing comparison is 249 against AFF4's 5 in the same assay.
Evidence that argues against this (1)
- GOLGA2 is a genuine multivalent tether with a large real interactome - golgins exist to bind many partners - so a high degree is partly biology. The argument this entry makes is that the pair is uninformative about AFF4, which is weaker than saying the pair is false. Huttlin EL et al. 2021
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt annotates GOLGA2 at the Golgi apparatus cis-Golgi network membrane, the endoplasmic reticulum-Golgi intermediate compartment membrane and the cytoplasmic spindle pole, with no nuclear annotation, and describes it as a peripheral cis-Golgi membrane skeleton and vesicle tether; AFF4 is annotated Nucleus and Chromosome.
UniProtKB entry Q08379 (GOGA2_HUMAN, Golgin subfamily A member 2), subcellular location and function comments with evidence codes · UniProtKB/Swiss-Prot 2026What this evidence cannot show
Retrieved 2026-09-21 with evidence codes read, per brief rule 91: the cis-Golgi call carries ECO:0000269 and ECO:0000305, the ERGIC and spindle-pole calls ECO:0000269, and part of the function comment is ECO:0000250 (by similarity) and ECO:0000305 (curator inference). Absence of a nuclear annotation is not a measured absence of nuclear GOLGA2. Yeast two-hybrid forces both constructs into the yeast nucleus by design and cannot detect a compartment mismatch.
Evidence that argues against this (1)
- GOLGA2 is annotated at the spindle pole, so it is not strictly confined to the secretory pathway, and a protein with a mitotic pool is exposed to chromatin during mitosis when the nuclear envelope is absent. That is a real route by which a Golgi protein can meet a chromatin factor, and it is unexplored rather than excluded. Rolland T et al. 2014
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
No literature connects GOLGA2 to KMT2A or MLL: (GOLGA2[tiab] OR GM130[tiab]) AND (MLL[tiab] OR KMT2A[tiab]) returns 0 records, and the two leukaemia records returned by the same alias pair title-check off-beat.
PubMed counted-absence search for GOLGA2/GM130 against MLL, KMT2A and leukaemia, run 2026-09-21 · PubMed (NCBI) 2026What this evidence cannot show
Operand control: the same alias pair against leukemia/leukaemia returns 2, so the query shape returns non-zero and the instrument works. Both hits are off-beat - PMID 39333927 (RPS6KA1 in AML) and PMID 6981287 (adenosine deaminase in lymphoblasts, 1982, where 'GM130' is almost certainly a cell-line or compound code rather than the protein). Aliases searched: GOLGA2, GM130. Not searched: golgin-95, 'golgin subfamily A member 2', GOLGA2P-family pseudogene spellings. Zero in titles and abstracts; supporting figures not searched.
Evidence that argues against this (1)
- GM130 is a four-character alphanumeric token of exactly the shape brief rule 70 warns about - it matches cell-line, medium and compound codes - so this count is exposed to contamination in the direction of inflation, not deflation. The 1982 hit is the demonstration. Huttlin EL et al. 2017
yeast two-hybrid hit in Rolland 2014; prey degree 249 in the same screen, cis-Golgi protein, not usable
the highest-degree prey of the four