PAF1c is needed by the normal stem cell and by the leukaemia, and the genes it regulates in each are different
also called PAF1 complex cell context · CDC73 normal versus leukaemic
The PAF1 complex is required both by healthy blood stem cells and by leukaemia cells, so removing it is not selective. But the genes it controls in the two settings are not the same, which is where a selective idea would have to come from.
The same thing, for a scientist
PAF1 itself is pan-essential — depleted below -0.5 in 93.4% of 1,258 screened lines with a KMT2A-rearranged-minus-all-lines difference of +0.012 [E1] — so DepMap carries no cell-type information for it. The informative evidence is at complex level and splits the two contexts: the PAF1c subunit CDC73 is required for mouse haematopoietic stem cell MAINTENANCE while displaying LEUKAEMIA-SPECIFIC gene regulation [E2], and a KDM4A-PAF1 epigenomic network is essential for AML cell self-renewal and survival [E3]. The selectivity, if there is any, is in the target gene set rather than in the requirement.
This is a rule-64 case and it is worth naming
A PAF1-symbol-only enumeration finds two cell-state records in leukaemia 4. The CDC73 result 2 — the only experiment in the literature that separates PAF1c's normal-cell requirement from its leukaemic one — is unreachable from the symbol PAF1, because it is framed at the level of the complex and a different subunit.
Brief rule 64 exists for exactly this: before writing a counted absence on a protein that belongs to a named complex, run the complex form. Here the complex form does not merely widen the count, it supplies the single most load-bearing result.
Identifier trap, verified
PAF1 returns two reviewed human records: Q8N7H5 (PAF1, 531 aa, RNA polymerase II-associated factor 1 homolog) and P28328 (PEX2, 305 aa), which carries PAF1 only as a synonym. HGNC fetched 2026-09-20 gives PAF1 → ENSG00000006712 → Q8N7H5, and that is the protein meant throughout this entry. The length difference of 226 residues means verify-accessions.py would catch a swap here, unlike the RAS family.
| gene | gene-paf1 |
|---|---|
| uniprot | Q8N7H5 |
| length AA | 531 |
| evidence class | 3 |
identifier resolution
{
"symbol": "PAF1",
"hgnc_ensembl": "ENSG00000006712",
"uniprot": "Q8N7H5",
"collision": "P28328 (PEX2, 305 aa) carries PAF1 as a synonym; length difference 226 aa so the accession verifier discriminates"
}cell type
{
"required_in": "essentially all screened lines (floor-compressed); murine HSC via CDC73",
"context": "both normal and leukaemic"
}cell state
{
"maintains": "cell-state-stem-like-self-renewal",
"selectivity_lives_in": "the regulated gene set, not the requirement"
}depmap
{
"pct_all_lines_depleted": 93.4,
"mean_all_lines": -0.83,
"mean_kmt2ar": -0.818,
"floor_compressed": true
}- PAF1c is required by the normal stem cell as well as the leukaemic one, so a complex-level inhibitor has no therapeutic index; only a target-gene-set-selective or interface-selective approach could have one.
- Pan-essentiality at 93.4% of lines means any apparent KMT2A-r selectivity for PAF1 in a small screen is noise at the floor.
- Attributing a CDC73 phenotype to PAF1 would be a rule-64 error in the other direction, and the paper's own leukaemia-specific finding warns that the subunits are not interchangeable.
- Which genes does PAF1c regulate in the KMT2A-rearranged blast that it does not regulate in the normal HSC? The CDC73 work implies the list exists and nobody has made it for PAF1 itself.
- Does the KDM4A-PAF1 network operate in B-lineage KMT2A::AFF1, or only in the AML setting where it was found?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
In DepMap fetched 2026-09-20, PAF1 is depleted below -0.5 in 93.4% of 1,258 screened lines (mean -0.830) and the KMT2A-rearranged arm differs from all lines by +0.012.
Open Targets Platform - DepMap CRISPR gene essentiality (depMapEssentiality), queried by Ensembl gene id 2026-09-20 · Open Targets Platform GraphQL API 2026What this evidence cannot show
Floor compression; the near-zero difference is uninterpretable rather than null. Cancer lines in culture.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT: a fusion-specific PAF1 requirement is published and a +0.012 pan-essential difference cannot represent it. Muntean et al. showed that MLL interacts directly with the PAF complex through sequences flanking the CxxC domain, that PAFc augments MLL- and MLL-AF9-mediated transcriptional activation of Hoxa9, that PAFc knockdown disrupts MLL fusion protein-mediated transcriptional activation AND MLL recruitment to target loci, and - the decisive result - that MLL deletions which abolish the PAFc interaction ELIMINATE MLL-AF9 mediated immortalisation. A requirement that is visible as loss of immortalisation in a colony-replating assay is invisible to a proliferation screen on established lines, which is the assay this evidence item reports. `gene-paf1` in `interactome2/` already states the general form - 'pan-essential subunit of a complex the fusion needs, where the window is in the interface and not in the protein' - and this entry reported the +0.012 without it. Muntean AG et al. 2010
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The PAF1c subunit CDC73 is required for mouse haematopoietic stem cell maintenance but displays leukaemia-specific gene regulation.
The PAF1c Subunit CDC73 Is Required for Mouse Hematopoietic Stem Cell Maintenance but Displays Leukemia-Specific Gene Regulation. · Saha N, Ropa J, et al. · Stem cell reports 2019What this evidence cannot show
Murine (brief rule 73). The result is about CDC73, a different subunit from PAF1, and transferring it to PAF1 assumes the complex acts as a unit, which the paper's own leukaemia-specific finding partly undercuts. Abstract read in this beat.
Evidence that argues against this (1)
- Transferring the CDC73 result to PAF1 assumes PAF1c acts as a unit, and the paper's own central finding - leukaemia-SPECIFIC gene regulation by one subunit - argues that the subunits are not interchangeable. A PAF1-specific normal-versus-leukaemic separation has not been shown. Saha N et al. 2019
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A KDM4A-PAF1-mediated epigenomic network is essential for acute myeloid leukaemia cell self-renewal and survival.
A KDM4A-PAF1-mediated epigenomic network is essential for acute myeloid leukemia cell self-renewal and survival. · Massett ME, Monaghan L, et al. · Cell death & disease 2021What this evidence cannot show
Human AML cells; AML rather than the index B-lineage disease (brief rule 73). Abstract read in this beat. The network framing makes it hard to attribute the requirement to PAF1 alone rather than to KDM4A.
Evidence that argues against this (1)
- PUBLISHED ALTERNATIVE ATTRIBUTION for the self-renewal phenotype. PAF1 does not sit only in a KDM4A network: Sobhian et al. purified a single stable elongation complex containing core active P-TEFb, the PAF1 complex, and the MLL fusion partners AF9, AFF4, AFF1, ENL and ELL together, and showed that the MLL-fusion partners and PAF1 are jointly required for transactivation at the same promoter. An AML self-renewal phenotype on PAF1 depletion is therefore attributable to that assembly at least as readily as to a KDM4A-PAF1 epigenomic network, and the caveat's worry that the network framing makes attribution to PAF1 hard is the smaller half - there are at least three published assemblies PAF1 could be acting in. Sobhian B et al. 2010
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
A bounded PubMed query pairing PAF1 and its complex-name aliases with an eighteen-term cell-state clause and a leukaemia clause returns two records, both fully on topic; the gene clause alone returns 536 and the state clause alone 96,743.
NCBI PubMed E-utilities esearch, db=pubmed, queried 2026-09-20 · NCBI Entrez Programming Utilities 2026What this evidence cannot show
A small count that is entirely signal, which is the opposite failure mode from FLT3's 617 records of residual-disease monitoring. Bounded by titles and abstracts.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT about what 'two records, both fully on topic' bounds. The founding result for PAF1c in this disease is titled 'The PAF complex synergizes with MLL fusion proteins at HOX loci to promote leukemogenesis' and contains none of the eighteen cell-state terms in its title, while reporting the most state-relevant PAFc finding there is: MLL deletions abolishing PAFc interaction eliminate MLL-AF9 immortalisation, i.e. self-renewal capacity. A small count that is entirely signal is not the same as a complete count, and this one misses the paper that established the subject. The caveat frames the small count as the opposite failure mode from FLT3's 617 noisy records; it is the same failure mode - the vocabulary, not the count, is what bounds the result. Muntean AG et al. 2010
the cell-context dossier for this gene
floor-compressed at 93.4% of lines, so DepMap carries no cell-type information for PAF1
MLL deletions that abolish the PAFc interaction ELIMINATE MLL-AF9 immortalisation, and PAFc knockdown disrupts MLL-fusion transcriptional activation and MLL recruitment to target loci (PMID 20541477) - a fusion-specific requirement that a +0.012 pan-essential difference cannot represent, because loss of immortalisation is invisible to a proliferation screen on established lines