RBBP7 - the silent paralogue that is not silent, and the reason this project's paralogue rule finally produced a positive answer
also called RBBP7 · RBAP46 · RbAp46 · histone acetyltransferase type B subunit 2 · retinoblastoma-binding protein 7 · NURF55
RBBP7 is the near-identical twin of RBBP4 - same length, same fold, same job of holding histones - and leukaemia cells make more of it than of RBBP4. Despite that, deleting it does essentially nothing while deleting RBBP4 is lethal, which tells us the two are not interchangeable here even though they look it.
The same thing, for a scientist
RBBP7 (UniProt Q16576, **425 aa, exactly the same length as RBBP4**) is the second WD40 histone-binding subunit shared across PRC2, NuRD, Sin3, CoREST and HAT1 (it is the type-B histone acetyltransferase subunit 2) [E1][E4]. This entry exists for a methodological reason as much as a biological one. This project holds a standing rule that **a paralogue split is only real if the silent paralogue is expressed**, because a gene that is not transcribed scores as non-essential in a CRISPR screen whatever its biology, and the rule had never previously returned a positive. Here it does: across the seven KMT2A-rearranged lines Human Protein Atlas carries, RBBP7 is detected in **7/7** at a median of **139.0 nTPM - higher than RBBP4's 77.5 in every single line** - while its DepMap gene effect in the 11-line index arm is **-0.013** against RBBP4's **-1.486** [E2]. **The most abundant of the two paralogues is the dispensable one.** That is a functional non-redundancy in leukaemia cells and it is not an expression artefact. The developmental literature shows the opposite behaviour in a different window - double knockdown of *Rbbp4* and *Rbbp7*, **but neither alone**, is lethal at the morula-to-blastocyst transition, which the authors describe explicitly as compensation [E3] - so redundancy between this pair is a property of the cell state rather than of the proteins. Neither paralogue has been studied in KMT2A-rearranged leukaemia: `(RBBP4 OR RBBP7) AND (KMT2A OR MLL) AND leukemia` returns **2** records and title-checking finds **0** on topic [E5].
The identifier hazard is worse here than in any case this project has logged
Rule 65 was written after gene_exact:HDAC7 returned HDAC9 first, with a protein name reading Histone deacetylase 7B, and it prescribes three defences: request gene_primary, keep only the primary-name match, and assert accession and length so verify-accessions.py can catch a swap that reading could not.
The third defence does not work for RBBP4/RBBP7.
| RBBP4 | RBBP7 | |
|---|---|---|
| accession | Q09028 | Q16576 |
| primary gene name | RBBP4 | RBBP7 |
| length | 425 aa | 425 aa |
| protein name | Histone-binding protein RBBP4 ... (CAF-I p48) (RBAP48) (Retinoblastoma-binding protein p48) | Histone-binding protein RBBP7 ... (Histone acetyltransferase type B subunit 2) (RBAP46) (Retinoblastoma-binding protein p46) |
The lengths are identical, so a swapped accession passes the length check silently. The protein names differ by a digit. HDAC7/HDAC9 at least gave 952 against 1011. For this pair the accession is the only discriminator, and no automated check in this repository can catch a swap. That is recorded here so a future beat does not believe the length assertion is protecting it 4.
A second retrieval hazard applies: both proteins publish under RBAP46/RBAP48, RbAp46/RbAp48 and NURF55, and the counted absence in 5 was not run over those aliases. By this graph's own retired-and-alias-symbol rule that count is an under-estimate of unknown size.
What 'dispensable' does and does not mean
A DepMap gene effect of -0.013 says that in these lines, over the duration of the screen, knocking out RBBP7 does not impair fitness. It does not say:
- that RBBP7 protein is absent from PRC2 in these cells (RNA is not protein and abundance is not occupancy);
- that RBBP7 has no function (a function with no fitness cost is invisible to the assay);
- that RBBP7 could not substitute under a different selection, such as after chronic RBBP4 inhibition - which is precisely the resistance mechanism a drug developer would need to exclude.
The last point is the live one. If RBBP7 is present, inert, and capable of substituting under pressure, then it is a pre-loaded resistance mechanism for any RBBP4-directed agent, and the screen's null result is the worst possible evidence about it: a redundant paralogue is invisible in exactly the assay used to declare it irrelevant. The preimplantation result 3 shows the pair can compensate when the cell state permits it.
| symbol | RBBP7 |
|---|---|
| uniprot | Q16576 |
| length AA | 425 |
| uniprot primary gene name | RBBP7 |
| uniprot resolution note | gene_exact:RBBP7 AND organism_id:9606 AND reviewed:true returned exactly one row; primary gene name RBBP7 matched; length 425 asserted. WARNING: the paralogue RBBP4 (Q09028) is ALSO 425 aa, so the Rule 65 length check is inert for this pair and the accession is the only discriminator. Resolved 2026-09-20. |
| paralogue | gene-rbbp4 |
| complex | protein-prc2-complex |
| expression hpa nTPM median kmt2a r panel | 139 |
| paralogue split verdict | REAL. RBBP7 is expressed above RBBP4 in 7/7 panel lines and is still dispensable, so the RBBP4 dependency is not an expression artefact. This is the first positive result this project's paralogue-expression control has returned. |
other complexes
[ "NuRD", "Sin3", "CoREST", "HAT1 type-B histone acetyltransferase (subunit 2)", "NURF" ]
essentiality
{
"depmap_kmt2a_r_mean": -0.0131,
"depmap_other_leukaemia_mean": -0.0159,
"depmap_difference": 0.0028,
"depmap_rank_of_17728": 9520,
"depmap_pan_essentiality_all_lines": -0.1547,
"depmap_index_myeloid_mean": -0.0093,
"depmap_index_lymphoid_mean": -0.0198,
"source": "analyses/prc2-dependency-lineage-split/results.json",
"interpretation": "dispensable in every line tested, despite being the more abundant paralogue"
}expression per line nTPM — the numbers behind the chart
{
"SEM": 226.4,
"RS4;11": 120.5,
"MV4-11": 137.3,
"THP-1": 139,
"NOMO-1": 433.3,
"MOLM-13": 153.4,
"OCI-AML-2": 115.9
}depmap kmt2a selectivity
{
"chronos_mean_kmt2a_rearranged": -0.0131,
"chronos_mean_other_leukaemia": -0.0159,
"difference": 0.0028,
"pan_essentiality_all_lines": -0.1547,
"selectivity_rank": 9520,
"rank_denominator": 17728,
"provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
"how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}- If RBBP7 can substitute for RBBP4 under selection, it is a pre-loaded resistance mechanism for any RBBP4-directed agent, and the DepMap null that makes RBBP7 look irrelevant is precisely the assay that cannot detect it.
- RBBP7 is a subunit of the HAT1 type-B histone acetyltransferase as well as of PRC2, so an agent that hit both paralogues would disable replication-coupled histone acetylation alongside Polycomb - a different and broader toxicity than a PRC2-directed drug.
- Is RBBP7 protein incorporated into PRC2 in KMT2A-rearranged blasts? If PRC2 in these cells is RBBP4-only, the dependency asymmetry is explained by complex partitioning and RBBP7 is irrelevant to Polycomb here; if PRC2 contains both, the asymmetry needs a different explanation and RBBP7 becomes a resistance liability.
- Would RBBP7 substitute under chronic RBBP4 inhibition? A single-gene knockout screen is the assay least able to see this, and the preimplantation data show the pair CAN compensate when the cell state permits. A dose-titrated RBBP4 degrader with RBBP7 knockout as a second arm would answer it.
- Does the alias-expanded literature count change the absence? RBAP46, RbAp46 and NURF55 were not searched, and this graph has measured a tenfold under-count from exactly this omission for the septins.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review
RBBP4 and RBBP7 are a pair of histone chaperones that govern the molecular behaviour of histones H3 and H4 and participate in several protein complexes including Polycomb repressive complex 2 and the nucleosome remodeling and deacetylase (NuRD) complex, thereby regulating the cell cycle, histone modifications, DNA damage and cell fate.
Interpreting the molecular mechanisms of RBBP4/7 and their roles in human diseases (Review). · Zhan Y, Yin A, Su X, Tang N, Zhang Z, Chen Y, et al. · Int J Mol Med 2024What this evidence cannot show
A review, cited as a map of complex membership rather than for any primary result. It treats RBBP4 and RBBP7 largely as a unit ('RBBP4/7'), which is exactly the elision this entry exists to break; no statement about relative or differential function should be taken from it.
Evidence that argues against this (1)
- Treating the pair as one unit is contradicted by the data in this entry: in leukaemia cell lines the two have opposite dependency profiles (-1.486 against -0.013) despite RBBP7 being the more abundant transcript, so 'RBBP4/7' is not a safe abbreviation in a haematopoietic context. Q41-prc2-arm 2026
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
RBBP7 is transcribed in 7 of 7 KMT2A-rearranged Human Protein Atlas panel lines at a median of 139.0 nTPM (SEM 226.4, RS4;11 120.5, MV4-11 137.3, THP-1 139.0, NOMO-1 433.3, MOLM-13 153.4, OCI-AML-2 115.9), higher than RBBP4 in every line, while its DepMap gene effect in 11 KMT2A-rearranged leukaemia lines is -0.013 (comparator -0.016, difference +0.003, rank 9,520 of 17,728, pan-essentiality -0.155) against RBBP4's -1.486 - so the paralogue asymmetry passes the expression control and is functional rather than an artefact of a silent transcript.
analyses/prc2-dependency-lineage-split - DepMap (Chronos, via Open Targets) PRC2 panel with lineage split, plus a Human Protein Atlas cell-line RNA expression control · Q41-prc2-arm · Oncofusion knowledge graph, internal analysis 2026What this evidence cannot show
nTPM is bulk RNA from immortalised lines: it is not protein, not complex occupancy, and cannot show whether RBBP7 protein is incorporated into PRC2 in these cells. Human Protein Atlas covers 7 of the 11 index lines (SHI-1 and HB11;19 are absent; SEMK2 and MOLM-14 are the same patients as SEM and MOLM-13). A CRISPR screen measures fitness over a fixed number of doublings and is insensitive to functions with no fitness cost. Raw effects, no z, no permutation null. DepMap release version is not stated by the Open Targets API.
Evidence that argues against this (1)
- The same pair is explicitly compensatory in mouse preimplantation embryos, where double but not single knockdown is lethal. So the leukaemia-line result establishes non-redundancy in that cell state and cannot be generalised to normal haematopoiesis, where an RBBP4-directed agent might be buffered by RBBP7 in a way the leukaemia is not - which would be a therapeutic window rather than a problem. Xiao L et al. 2022
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Double knockdown of Rbbp4 and Rbbp7, but not knockdown of either individually, causes embryonic lethality during the mouse morula-to-blastocyst transition, with cell-cycle block, disrupted lineage specification, disrupted chromatin structure and a dramatic increase in H3.3 and H3K27ac abundance - which the authors describe as a compensatory role between the two paralogues.
Overlapping functions of RBBP4 and RBBP7 in regulating cell proliferation and histone H3.3 deposition during mouse preimplantation development. · Xiao L, Dang Y, Hu B, Luo L, Zhao P, Wang S, et al. · Epigenetics 2022What this evidence cannot show
Abstract read from the eutils record; full text not read. Mouse preimplantation embryos, knockdown not knockout, so 'not individually' means a partial depletion of one paralogue was tolerated rather than that a null was. The complexes named in the abstract are SIN3A, NuRD and CoREST, not PRC2. Whether the compensation is at the level of PRC2 specifically is not established.
Evidence that argues against this (1)
- A separate mouse study using Rbbp4 mutant alleles rather than knockdown finds that loss of RBBP4 alone IS lethal, with blastocysts failing to hatch or arresting, severe apoptosis and histone hyperacetylation. Single-paralogue tolerance may therefore be an artefact of incomplete knockdown rather than genuine compensation, and this entry does not adjudicate between the two reports. Miao X et al. 2020
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
RBBP7 resolves to a single reviewed human UniProt record, Q16576 (RBBP7_HUMAN), primary gene name RBBP7, no listed gene synonyms in the returned row, 425 amino acids, protein name Histone-binding protein RBBP7 (Histone acetyltransferase type B subunit 2) (RBAP46) (Retinoblastoma-binding protein p46) - the same length as its paralogue RBBP4 (Q09028, 425 aa).
UniProtKB reviewed human record Q16576 (RBBP7_HUMAN), resolved with gene_primary and length requested in fields · UniProt 2026What this evidence cannot show
Resolved 2026-09-20. The identical 425-residue length means the length assertion prescribed by Rule 65 is inert for this pair and cannot detect a swap. UniProt length is for the canonical isoform.
Evidence that argues against this (1)
- The aliases are also the retrieval hazard: RBAP46, RbAp46 and NURF55 are in active use and a literature count keyed only to 'RBBP7' will under-retrieve, in the same way that SEPT9 returned twenty records where SEPTIN9 returned one in this project's own measurement. UniProt 2026
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
PubMed `(RBBP4 OR RBBP7) AND (KMT2A OR MLL) AND leukemia` returns 2 records and neither is on topic after a title check, so neither paralogue has a published study in KMT2A-rearranged leukaemia under these symbols.
PubMed esearch count and title check for (RBBP4 OR RBBP7) AND (KMT2A OR MLL) AND leukemia, run 2026-09-20 · NCBI PubMed 2026What this evidence cannot show
The aliases RBAP46, RBAP48, RbAp46, RbAp48 and NURF55 were NOT searched, and this graph's own rule is that a count keyed to current symbols can under-count by an order of magnitude. Per Rule 64 the complex form was run and is larger ('Polycomb repressive complex 2' AND (KMT2A OR MLL) AND leukemia returns 47 against PRC2's 24) but title-checking it recovers no RBBP-specific leukaemia study. This is an absence of published study under these symbols, not evidence of biological irrelevance.
Evidence that argues against this (1)
- The complex-name form is the documented way a subunit-keyed absence goes wrong in this project: PMID 37626123 says 'FACT' and no subunit, and this graph declared the FACT literature empty in one entry while citing that paper in four others. The RBBP absence was therefore cross-checked against the repository as well as against PubMed, and this project's histone-chaperone beat had already scored RBBP4 independently. Q37-histone-chaperones 2026
- part of →PRC2 - the complex, not the subunit, is the unit every experiment in this disease actually resolves
paralogous histone-binding subunit; abundantly expressed and dispensable in leukaemia lines
- co-occurs with →RBBP4 - the shared histone-binding subunit that every leukaemia line needs, and whose identical-length paralogue cannot replace it
paralogue pair, both exactly 425 aa - the accession is the only discriminator
the paralogue-expression control passes: RBBP7 is the more abundant transcript and is dispensable
H3R2me2s EXCLUDES RBBP7 from the H3 tail (Sin3a/NuRD/PRC2 subunit)
RBBP7 is a co-subunit - and the protein PRMT5 H3R2me2s excludes from the H3 tail
EXCLUDES RBBP7, a shared Sin3a/NuRD/PRC2 subunit - the unrevised half of the 2012 model, never re-tested and never reasoned about in this graph
shared histone chaperone