ASF1B - the paralogue that actually does the work in blood: it occupies >70% of H3.3 nucleosomes in fetal liver while HIRA is dispensable there
also called ASF1B · anti-silencing function 1B · histone chaperone ASF1B · Q9NVP2
ASF1B hands new histones to the machines that put them on DNA. In the blood-forming tissue of the fetus - the same tissue this project believes infant leukaemia starts in - ASF1B, not the textbook chaperone HIRA, is the one that decides where the replacement histone goes.
The same thing, for a scientist
ASF1B (Q9NVP2, 202 aa) is one of two human ASF1 paralogues, the upstream H3-H4 donor that feeds both the replication-coupled (CAF-1) and replication-independent (HIRA) deposition routes. **This beat was dispatched expecting an unresolved paralogue question and found a published, blood-specific answer.** In mouse fetal liver, ASF1B is abundant together with H3.3; **HIRA is dispensable in fetal haematopoiesis**; ASF1B predominantly regulates H3.3-encoding genes and erythroid genes with **ASF1A serving a compensatory function**; ASF1B occupies **more than 70% of H3.3 nucleosomes** and determines H3.3 enrichment at erythroid promoters and enhancers, recruiting BRG1 and accumulating H3K27ac [E1]. **Fetal liver is the compartment this graph's `cell-type-fetal-liver-hsc` entry exists for.** If the H3.3 route in fetal haematopoiesis runs through ASF1B rather than HIRA, then the beat's flagship gene is the wrong one for the index disease. ASF1B also has a genuine AML literature. IL-1 beta upregulates ASF1B in AML progenitors compared with healthy progenitors; ASF1s and the TLKs that regulate them are overexpressed across AML genetic subtypes; depleting ASF1s reduces leukaemic growth in vitro and in vivo; **overexpressing ASF1B accelerates leukaemia progression while Asf1b or Tlk2 deletion delays it, and both are dispensable for normal haematopoiesis** [E2]. And DepMap sees none of it: ASF1B difference **+0.065**, pan-essentiality **-0.292**; ASF1A difference **+0.108** at an index mean of **+0.009** [E3]. Neither paralogue is required in any leukaemia line, which is what mutual buffering predicts.
The paralogue rule, applied
This graph's rule is that a paralogue split is only real if the silent paralogue is actually expressed, and analyses/expression-control-for-paralogue-splits/ is the worked example that has already overturned one such claim here.
The ASF1 case is unusual because the literature answers the question directly rather than leaving it to an expression table: the fetal-liver study states that ASF1A serves a compensatory function 1, which is a functional demonstration of the thing an expression table can only make plausible. So:
- ASF1A is expressed and functional in the same cells - that is the paper's own claim, not an inference.
- The split is a split of dominance, not of presence. ASF1B does most of it; ASF1A can cover.
- This is exactly the configuration in which a single-gene CRISPR screen returns nothing, and DepMap does return nothing for both 3.
The HPA cell-line expression control this graph uses for paralogue claims has not been run for ASF1A/ASF1B in KMT2A-rearranged lines, and that is the remaining gap - the functional compensation is shown in mouse fetal liver, not in a human leukaemic blast.
A therapeutic window claimed twice, in two systems, for two different chaperones
The TLK-ASF1 paper states that Asf1b or Tlk2 deletion delays leukaemia progression while these proteins are dispensable for normal haematopoiesis 2. The CHAF1B paper states that halving Chaf1b suppresses leukaemogenesis without impairing normal haematopoiesis (gene-chaf1b E1).
Two independent groups, two different chaperones, the same shape of claim, both in mouse myeloid leukaemia, both against a DepMap picture that says nothing selective is happening. That pattern is worth stating plainly because it cuts both ways:
- It is evidence that chaperone dependence in leukaemia is a dose- and context-dependent phenomenon that pooled CRISPR screens cannot see - the same argument as the CHAF1B dose table.
- It is also the shape of a field-wide reporting bias: 'dispensable for normal haematopoiesis' is the claim every chaperone paper needs in order to be interesting, it is assessed over weeks in a young mouse, and no negative version of it would be published.
This graph should hold both readings. The discriminating experiment is a competitive transplant with long follow-up, which neither paper describes in its abstract.
| uniprot | Q9NVP2 |
|---|---|
| gene symbol | ASF1B |
| length AA | 202 |
| paralogue | gene-asf1a |
| regulated by | Tousled-like kinases TLK1/TLK2 |
| blood role | occupies >70% of H3.3 nucleosomes in mouse fetal liver; determines H3.3 enrichment at erythroid promoters and enhancers; HIRA dispensable in that compartment |
| kmt2a literature records | 0 |
feeds
[ "protein-huca-complex", "CAF-1 (gene-chaf1a/gene-chaf1b)" ]
depmap — the numbers behind the chart
{
"index_mean": -0.1834,
"other_leukaemia_mean": -0.2484,
"difference": 0.0649,
"pan_essentiality": -0.2919,
"n_index": 11
}depmap kmt2a selectivity
{
"chronos_mean_kmt2a_rearranged": -0.1834,
"chronos_mean_other_leukaemia": -0.2484,
"difference": 0.0649,
"pan_essentiality_all_lines": -0.2919,
"selectivity_rank": 15744,
"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."
}- ASF1A compensates for ASF1B, so a selective ASF1B inhibitor has a built-in resistance mechanism requiring no mutation - the paralogue-buffering failure mode this graph has documented four times.
- ASF1 depletion causes rapid induction of alternative lengthening of telomeres, so inhibiting this chaperone could hand a telomere-maintenance mechanism to the surviving cells.
- The TLK-ASF1 axis is driven by IL-1 beta from the microenvironment; anything that reduces marrow inflammation - including the leukaemia responding to treatment - would reduce the dependency and make the target disappear as the disease improves.
- ASF1 feeds both deposition routes, so blocking it slows replication-coupled and replication-independent histone supply together, which by this graph's clearance-clock argument would prolong H3K79 methylation rather than clear it.
- Run this graph's HPA cell-line expression control on ASF1A and ASF1B in KMT2A-rearranged lines. The functional compensation is demonstrated in mouse fetal liver; the expression control in a human leukaemic blast has not been done and the graph's own paralogue rule requires it.
- Is ASF1B the H3.3 chaperone in a fetal-liver-derived infant KMT2A::AFF1 cell? If it is, every argument in this beat that runs through HIRA should run through ASF1B instead.
- Does the IL-1 beta-ASF1B axis operate in infant B-ALL? The AML result is microenvironment-driven and this graph has a marrow-niche beat that has never met this pathway.
- Is 'dispensable for normal haematopoiesis' true on a transplant timescale? Two chaperone papers make the same claim in mouse myeloid leukaemia and neither abstract describes a competitive transplant with long follow-up.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In mouse fetal liver, ASF1B is abundant together with histone variant H3.3; HIRA, the H3.3-specific chaperone, is dispensable in fetal haematopoiesis; ASF1B predominantly regulates H3.3-encoding genes and erythroid genes while ASF1A serves a compensatory function; ASF1B occupies more than 70% of H3.3 nucleosomes and determines H3.3 enrichment at erythroid gene promoters and enhancers; loss of ASF1B de-represses embryonic and fetal globin genes by altering H3.3 and erythroid transcription factor enrichment and chromatin accessibility, with the pathway involving recruitment of BRG1 and accumulation of H3K27ac.
ASF1B promotes erythropoiesis by regulating the establishment and enrichment of H3.3 nucleosomes. · Liu J, Song X, Zhou L, et al. · Nucleic Acids Research 2026What this evidence cannot show
Abstract read from efetch 2026-09-20; full text not read (PMC13153709 exists). **Mouse fetal liver, erythroid lineage, normal development - not leukaemia and not human.** 'HIRA was dispensable' is asserted without the assay being named in the abstract, and an absence of testing is not a negative result; how hard that negative was pushed is unverified here. The >70% occupancy figure is a genome-wide overlap whose denominator is not stated. Graded strong rather than established because the load-bearing transfer - fetal liver to infant leukaemia - is an analogy this entry is making, not a claim the paper makes.
Evidence that argues against this (1)
- In mouse ES and neural precursor cells HIRA IS required for H3.3 at active and repressed genes, so a blanket reading that ASF1B replaces HIRA would contradict the founding genome-wide study of H3.3 deposition. The reconciliation is cell-type-specific redundancy, which means neither result transfers to a leukaemic blast without being redone there. Goldberg AD et al. 2010
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
IL-1 beta upregulates ASF1B in AML progenitors relative to healthy progenitors; ASF1A, ASF1B and the Tousled-like kinases TLK1/TLK2 that regulate them are overexpressed across multiple genetic subtypes of AML; depletion of ASF1s significantly reduces leukaemic cell growth in vitro and in vivo in human models; overexpression of ASF1B accelerates leukaemia progression in a murine model while Asf1b or Tlk2 deletion delays progression, and these proteins are dispensable for normal haematopoiesis; proteomics and phosphoproteomics implicate cell-cycle and DNA-damage pathways.
The TLK-ASF1 histone chaperone pathway plays a critical role in IL-1beta-mediated AML progression. · Lin HY, Mohammadhosseini M, McClatchy J, et al. · Blood 2024What this evidence cannot show
Abstract read from efetch 2026-09-20; full text not read (PMC11340594 exists). **AML across genetic subtypes; no KMT2A-rearranged arm is named in the abstract and the index disease here is infant B-ALL.** 'Dispensable for normal haematopoiesis' is a mouse statement over the duration of the experiments described, which the abstract does not give. The Greek letter in the original title is rendered as 'beta' here because the source markup uses an italic-safe encoding this project has documented as lossy; the PMID and DOI were copied from the same fetched record.
Evidence that argues against this (1)
- DepMap finds neither ASF1A nor ASF1B required in any of 1,208 cell lines, with ASF1A's index mean actually positive (+0.009). A pathway whose depletion 'significantly reduces leukaemic cell growth' in dedicated experiments and is invisible in a genome-wide screen is either buffered by its paralogue or dependent on a microenvironmental signal - IL-1 beta - that pooled in-vitro screens do not supply. Oncofusion knowledge graph (internal analysis) 2026
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
In DepMap CRISPR gene effect, ASF1B's mean effect is -0.183 in 11 KMT2A-rearranged leukaemia lines against -0.248 in 43 other leukaemia lines (difference +0.065) at a pan-essentiality of -0.292; ASF1A's index mean is +0.009 against -0.099 (difference +0.108) at a pan-essentiality of -0.187. Neither paralogue is required in any arm and both are slightly less required in the index arm.
Is any histone chaperone or variant-exchange gene selectively required in KMT2A-rearranged leukaemia lines, and does the answer survive a lineage split? · Oncofusion knowledge graph (internal analysis) 2026What this evidence cannot show
Run 2026-09-20. **This is a negative result and should be read only as one.** A single-gene knockout screen cannot detect a dependency covered by a paralogue, which the fetal-liver study says ASF1A provides. The correct experiment is a double knockout and it has not been run in any leukaemia line.
Evidence that argues against this (1)
- Two independent published studies report growth and progression effects of ASF1 depletion in leukaemia - the TLK-ASF1 AML work and an ASF1A study in CML blast crisis. A genome-wide screen returning nothing for both paralogues is a statement about the screen's design, not about the biology. Yin X et al. 2022
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
ASF1B is UniProt Q9NVP2 (ASF1B_HUMAN), histone chaperone ASF1B / anti-silencing function protein 1 homolog B, 202 amino acids, with no alias gene symbols in the reviewed record; its paralogue ASF1A is a separate reviewed record Q9Y294 of 204 amino acids.
UniProtKB Q9NVP2 (ASF1B_HUMAN), Histone chaperone ASF1B, 202 aa · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026What this evidence cannot show
Both records fetched 2026-09-20 with `gene_primary` requested and read back per rule 62; each query returned exactly one reviewed human record whose primary symbol matched, and neither lists the other as an alias. **The lengths differ by two residues (202 against 204), so the length check that catches a paralogue swap is nearly useless here** - this is the worst case for rule 63 and the accession must be checked directly.
Evidence that argues against this (1)
- Because ASF1A and ASF1B are almost the same length, an accession swap would pass every automated check this project runs. The only defence is that both accessions are asserted explicitly and that no claim in this entry rests on a length. UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
As of 2026-09-20, '(ASF1A OR ASF1B) AND (MLL OR KMT2A)' returns 0 records; '(ASF1A OR ASF1B) AND leukemia' returns 8, of which the on-topic ones are the 2024 Blood TLK-ASF1 AML study, a 2022 ASF1A CML blast-crisis study and a 2026 ASF1A review.
NCBI E-utilities esearch over PubMed for the ASF1 paralogues against KMT2A, MLL and leukaemia, with title checking of all returned records, run 2026-09-20 · US National Library of Medicine 2026What this evidence cannot show
Query translations captured; no [UID] terms. All 8 titles were fetched in full. **The KMT2A zero is bounded by the symbols ASF1A and ASF1B**; a paper framing its result as 'the ASF1 pathway' or 'the TLK-ASF1 axis' without either symbol in title or abstract would not be caught, and the 2024 Blood paper is itself an example of a study that would be missed by a query requiring 'ASF1A' alone.
Evidence that argues against this (1)
- The generic-phrase route found a histone-chaperone/MLL paper that every symbol query in this beat missed, so a symbol-bounded zero for ASF1 is weaker than it looks and should be re-run against 'ASF1 pathway' and 'Tousled-like kinase'. Shimoyama T et al. 2005
- recruits →Histone H3.3 - one protein, two genes, and the reason this graph should not file it as gene-h3f3a
occupies >70% of H3.3 nucleosomes in fetal liver; sets H3.3 enrichment at erythroid regulatory elements
dominant H3.3 chaperone in fetal liver haematopoiesis, where HIRA is dispensable
paralogue pair; ASF1A compensates for ASF1B in fetal liver haematopoiesis
TLK-ASF1 overexpressed across AML subtypes; IL-1beta-driven, no KMT2A arm reported