start / graph / ana-q113-geo-holds-human-suv39h1-occupancy-data-that-the-encode-only-census-could-not-see

AnalysisstrongQ113-bounded-absence-audit

The GEO sweep that q110 named as free and did not run: SUV39H1 chromatin occupancy HAS been measured in two human cell types, six samples, hg38 bigWigs already deposited - so the writer half of the H3K9me3 absence is measured-not-analysed, not never-measured

also called SUV39H1 GEO sweep · Q113 SUV39H1 overturn

A published entry here said nobody could even ask an archive where the main H3K9 silencing enzyme sits, because the big public atlas holds no experiment for it. The atlas really does hold none, but a different archive holds six samples from two human cell types, already processed into files anyone can read.

The same thing, for a scientist

`absence-no-h3k9-writer-or-reader-localised-at-the-kmt2a-locus` bounded the SUV39H1 archive half with an ENCODE portal census (`type=Experiment, assay_title=TF ChIP-seq, status=released`, per target label) and recorded `SUV39H1: n=0`. It then named the unrun step in its own text: *the SUV39H1 half: a GEO sweep first (free)*. This beat ran it. `esearch db=gds` for `SUV39H1[Title] AND "Homo sapiens"[Organism] AND gsm[ETYP]` returns 75 GSMs; reading all 75 titles and then the SOFT records of the candidates gives **six chromatin-occupancy samples in two independent human systems**: four SUV39H1 ChIP-Seq samples in MCF-7 with and without estradiol (GSM4133857-60, GSE139196, Nature 2021, PMID 34194047) and two SUV39H1 CUT&RUN samples in patient-derived diffuse midline glioma lines SU-DIPG13 and SU-DIPG36 (GSM7037531-32, GSE225028, Neuro-Oncology 2024, PMID 38011799), the latter deposited as **hg38 bigWigs** alongside H3K9me3 in the same cells. ChIP-Atlas independently indexes the MCF-7 set as SRX7030945-48 under antigen SUV39H1, with the CTCF positive control returning HTTP 200 in the same batch. The ENCODE half of the original claim is CORRECT and survives - SUV39H1 has no ENCODE target record under any label, tagged or untagged - and the literature half survives too, including under the alias `KMT1A`, which was never tried and which the MCF-7 samples are literally titled with. What falls is the sentence *for SUV39H1 no archive can be asked either*.

Go deeper

What was run, exactly

esearch db=gds, term SUV39H1[Title] AND "Homo sapiens"[Organism] AND gsm[ETYP], querytranslation read back unchanged, count 75, all 75 esummary titles printed and read one by one. Broadening [Title] to [All Fields] gives 283 and was not title-checked; 75 is the conservative denominator. Candidates whose title suggested an occupancy assay were then opened as SOFT records (acc.cgi?targ=self&form=text&view=brief) and the !Sample_library_strategy and !Sample_characteristics_ch1 = chip antibody: lines read directly, rather than inferred from the title 12.

The remaining 69 are RNA-seq, ATAC-seq or editing-outcome libraries from SUV39H1 knockdown, knockout, CRISPRi and CRISPRoff arms in BJ fibroblasts, U937, CAR T cells, RKO, iPSC and primary T cells, plus one siRNA-screen sample in IMR90. None is occupancy. The 69/6 split is the reason a title count alone would have been wrong in either direction.

The two deposits

GSE225028 - Cut&Run for H3K9me3 and Suv39H1 in SU-DIPG13 and SU-DIPG36 cells (the deposit's own !Series_overall_design). Four samples; GSM7037531 and GSM7037532 carry chip antibody: SUV39H1(CST, #8729S). Processing is recorded per sample: Bowtie2 v2.3.4.1 to hg38, MACS2 v2.1.2 narrow and broad, supplementary files bigWig, Illumina NovaSeq 6000. Published as PMID 38011799, Neuro-Oncology 2024 1. The clone identifier is recorded because rule 74 requires it for an antibody readout: Cell Signaling Technology #8729.

GSE139196 - the ChIP-Seq sub-series of GSE139199, Shape of promoter antisense RNAs regulates ligand-induced transcription activation, Nature 2021, PMID 34194047 2. Four SUV39H1 ChIP-Seq samples in MCF-7, two minus and two plus 40 min estradiol, titled ChIP on SUV39H1 (KMT1A); supplementary files are per-sample bedGraphs. The same four are indexed by ChIP-Atlas as SRX7030945-48 3.

What survives of the original entry, checked rather than assumed

The ENCODE statement survives. /targets/SUV39H1-human/ returns HTTP 404 and type=Target&searchTerm=SUV39H1 returns 404 - no target record exists under any label. This is not the BARD1 failure: BARD1 and RAD50 both return HTTP 200 on the same endpoint in the same batch, so the 404 discriminates 4. The only ENCODE objects mentioning SUV39H1 are 6 Flow-FISH CRISPR screens in K562 where it is a perturbed gene, 96 files, 24 biosamples and 3 gene records - no occupancy. SUV39H2 is subtler and worth recording: it has no bare target record either, but it DOES have one as eGFP-SUV39H2, with zero released experiments attached. A census keyed on the bare symbol would miss a tagged-target label that had experiments; here it does not, but CBX5 likewise carries a 3xFLAG-CBX5 label alongside CBX5 and the general hazard is live.

The literature statement survives, including under the untried alias. KMT1A[tiab] returns 22 records; KMT1A[tiab] AND KMT2A[tiab] and KMT1A[tiab] AND MLL[tiab] both return 0 with non-empty translations and No items found. SUV39H[tiab] AND MLL[tiab] returns 1, PMID 19460867, telomeric small RNAs in mouse ES cells - off topic 5. So nobody has reported SUV39H1 at the KMT2A locus under any spelling.

What falls is the archive claim. The correct class for SUV39H1 is now measured-not-analysed: the data exist, are public, are aligned to hg38, and the analysis has never been run. That is a compute question, not a wet-work question, and it is a different order of cost from what the entry currently says.

The honest limit

Neither system is haematopoietic. MCF-7 is a breast carcinoma line; SU-DIPG13 and SU-DIPG36 are patient-derived H3K27M diffuse midline glioma lines. The graph's question is about a blood cell, so no SUV39H1 occupancy has been measured in a haematopoietic cell remains true and is the claim that should be carried forward. What changed is that a genome-wide SUV39H1 track in a human cell now exists to look at, which it did not an hour ago as far as this corpus was concerned.

Facts and figures

script	no script; E-utilities esearch/esummary and GEO acc.cgi SOFT reads, run inline 2026-09-22
status	run
run date	2026-09-22
run by	Q113-bounded-absence-audit
the question this answers	Does GEO hold a SUV39H1 ChIP-seq experiment in any human cell? - posed verbatim as an open question by absence-no-h3k9-writer-or-reader-localised-at-the-kmt2a-locus, which said 'This beat did not run it.'
result summary	GEO holds 6 human SUV39H1 occupancy samples in 2 cell types; ENCODE holds 0 and that zero is honest; ChIP-Atlas holds 4 of the 6. The archive half of the SUV39H1 absence is overturned, the literature half and the ENCODE statement both survive.

geo sweep

{
  "instrument": "esearch db=gds, querytranslation read back",
  "query": "SUV39H1[Title] AND \"Homo sapiens\"[Organism] AND gsm[ETYP]",
  "querytranslation": "SUV39H1[Title] AND \"Homo sapiens\"[Organism] AND gsm[ETYP]",
  "count": 75,
  "all_titles_read": true,
  "broader_control": {
    "query": "SUV39H1[All Fields] AND \"Homo sapiens\"[Organism] AND gsm[ETYP]",
    "count": 283,
    "note": "not title-checked; the conservative 75 is used"
  },
  "instrument_control": {
    "query": "CTCF[Title] AND \"Homo sapiens\"[Organism] AND gsm[ETYP]",
    "count": 3956
  },
  "occupancy_samples_found": 6,
  "non_occupancy_samples": 69,
  "closes": "6 occupancy + 69 non-occupancy == 75 total"
}

occupancy samples

[
  {
    "gsm": "GSM7037531",
    "series": "GSE225028",
    "cell": "SU-DIPG13",
    "assay": "CUT&RUN",
    "antibody": "SUV39H1 (CST #8729S)",
    "file": "bigWig",
    "build": "hg38",
    "pmid": "38011799"
  },
  {
    "gsm": "GSM7037532",
    "series": "GSE225028",
    "cell": "SU-DIPG36",
    "assay": "CUT&RUN",
    "antibody": "SUV39H1 (CST #8729S)",
    "file": "bigWig",
    "build": "hg38",
    "pmid": "38011799"
  },
  {
    "gsm": "GSM4133857",
    "series": "GSE139196",
    "cell": "MCF-7",
    "assay": "ChIP-Seq",
    "antibody": "SUV39H1 (KMT1A)",
    "condition": "no estradiol",
    "file": "bedGraph",
    "pmid": "34194047"
  },
  {
    "gsm": "GSM4133858",
    "series": "GSE139196",
    "cell": "MCF-7",
    "assay": "ChIP-Seq",
    "antibody": "SUV39H1 (KMT1A)",
    "condition": "no estradiol",
    "file": "bedGraph",
    "pmid": "34194047"
  },
  {
    "gsm": "GSM4133859",
    "series": "GSE139196",
    "cell": "MCF-7",
    "assay": "ChIP-Seq",
    "antibody": "SUV39H1 (KMT1A)",
    "condition": "estradiol 40 min",
    "file": "bedGraph",
    "pmid": "34194047"
  },
  {
    "gsm": "GSM4133860",
    "series": "GSE139196",
    "cell": "MCF-7",
    "assay": "ChIP-Seq",
    "antibody": "SUV39H1 (KMT1A)",
    "condition": "estradiol 40 min",
    "file": "bedGraph",
    "pmid": "34194047"
  }
]

encode recheck 2026 09 22

{
  "instrument": "ENCODE portal, /targets/<SYMBOL>-human/ AND /search/?type=Target&searchTerm=<SYMBOL>, no assay filter",
  "BARD1": {
    "target_record": 200,
    "experiments_all_assays": 2,
    "experiments_TF_ChIP_seq": "HTTP 404 (zero results)"
  },
  "RAD50": {
    "target_record": 200,
    "experiments_all_assays": "HTTP 404 (zero results)"
  },
  "SUV39H1": {
    "target_record": 404,
    "target_search": 404,
    "sitewide_searchTerm_hits": 130,
    "of_which_occupancy": 0,
    "what_they_are": "6 Flow-FISH CRISPR screens in K562 where SUV39H1 is a perturbed gene, 96 Files, 24 Biosamples, 3 Gene records, 1 Publication"
  },
  "SUV39H2": {
    "bare_target_record": 404,
    "tagged_label_found": "eGFP-SUV39H2",
    "experiments_under_tagged_label": 0
  },
  "CBX5": {
    "labels": [
      "CBX5",
      "3xFLAG-CBX5"
    ],
    "experiments_under_CBX5": 4,
    "experiments_under_3xFLAG-CBX5": 0
  },
  "EXO1": {
    "target_record": 404,
    "target_search": 404
  },
  "DNA2": {
    "target_record": 404,
    "target_search": 404,
    "note": "but ChIP-Atlas holds one, SRX3520279 in HCT-116 - the ENCODE zero is not an archive-wide zero"
  }
}

chip atlas 2026 09 22

{
  "instrument": "https://chip-atlas.dbcls.jp/data/hg38/target/<ANTIGEN>.5.tsv, header row parsed for SRX|cell",
  "positive_control_CTCF": "HTTP 200",
  "SUV39H1": {
    "experiments": 4,
    "srx": [
      "SRX7030945",
      "SRX7030946",
      "SRX7030947",
      "SRX7030948"
    ],
    "cells": [
      "MCF-7"
    ]
  },
  "SUV39H2": "HTTP 404 (zero)",
  "SETDB1": {
    "experiments": 14,
    "cells": [
      "293",
      "Hep_G2",
      "K-562",
      "U2OS",
      "WM262",
      "WM451",
      "WM853"
    ]
  },
  "CBX5": {
    "experiments": 29
  },
  "DNA2": {
    "experiments": 1,
    "srx": [
      "SRX3520279"
    ],
    "cells": [
      "HCT_116"
    ]
  },
  "BARD1": "HTTP 404 (zero)",
  "EXO1": "HTTP 404 (zero)",
  "threshold_token_trap": "the URL segment before .tsv is the threshold and must be 5, not 05. CTCF.05.tsv returns HTTP 404 exactly like a genuinely absent antigen, so a sweep written with 05 returns a clean, well-formed, entirely false zero for every protein queried."
}

class change recommended

{
  "entry": "absence-no-h3k9-writer-or-reader-localised-at-the-kmt2a-locus",
  "field": "data.absence_class_per_protein.SUV39H1",
  "currently": "never-measured - no literature report and no ENCODE experiment in any cell type; the GEO individual-deposition route was NOT swept, so a stricter reading of the archive half is measured-not-searched",
  "should_read": "literature half never-measured (no report at the KMT2A locus under SUV39H1, SUV39H, KMT1A or any spelling tried); archive half MEASURED-NOT-ANALYSED - six human occupancy samples in GEO across two cell types (GSE225028 CUT&RUN hg38 bigWigs in SU-DIPG13/SU-DIPG36; GSE139196 ChIP-Seq bedGraphs in MCF-7), none haematopoietic and none read at the KMT2A locus",
  "also_wrong_in_the_title": "'for SUV39H1 no archive can be asked either' - an archive can be asked and has now been asked",
  "reported_not_edited": "this beat does not own that file; the correction is reported to the coordinator with the accessions and PMIDs above"
}

How this could bite us back

Chaetocin, the SUV39H1 inhibitor in the DIPG paper, is a thiodioxopiperazine with well-documented redox and thiol-reactive off-target activity, so any SUV39H1-directed therapeutic argument built on that paper inherits a selectivity problem before it starts.
Correcting an absence from never-measured to measured-not-analysed moves work from the wet lab to the queue, and this project's own ledger shows that a stale 'no work needed here' number suppressed a whole line of enquiry for seven cycles. A measured-not-analysed entry that nobody analyses is not obviously better than a never-measured one.

Still open

Is SUV39H1 present over the 90 kb KMT2A gene body in either deposited system? The GSE225028 bigWigs are hg38 and can be read directly. This beat deliberately did not run it, because a concurrent beat owns signal-level work at the KMT2A exon-11 anchor and the two must not collide. ANSWERED IN CYCLE 66 for the GSE225028 system, by ana-q114-suv39h1-is-not-at-the-kmt2a-locus-in-either-patient-derived-glioma-line, which is the concurrent beat this question deferred to. NO: 0.9381 and 1.1109 over the 90 kb gene body, i.e. 0.94x and 1.11x the random-tile mean and the 39.7th / 49.7th percentile of width-matched CTCF anchors. The MCF-7 system is reported separately because its pre-registered instrument control failed.
GSE225028 deposits H3K9me3 CUT&RUN in the same two lines. Does SUV39H1 occupancy coincide with H3K9me3 there, and if it does not, what is the enzyme doing at the sites where it sits without the mark? ANSWERED IN CYCLE 66 by ana-q114-the-h3k9me3-writer-and-its-own-mark-barely-coincide-in-the-same-nuclei. They barely coincide: Spearman rho 0.148 and 0.078 over 59,821 CTCF-anchor windows in the same nuclei, no different from random tiles, and a quarter to two-fifths of the enzyme's top 1% of sites carry no more mark than the median anchor. What the enzyme is doing at those sites is NOT answered and cannot be from this deposit, which has no IgG and no knockout control.
Does any SUV39H1 occupancy dataset exist in a haematopoietic cell under a name this sweep could not reach - a CUT&RUN deposit whose sample title says only the cell line, for instance? The GEO title index cannot see those, and `CUT&RUN` cannot be searched as a literal string in Entrez because the ampersand parses as boolean AND.

Every source behind this entry (5)

E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
SUV39H1 genomic occupancy was measured by CUT&RUN in two patient-derived diffuse midline glioma lines and deposited as hg38 bigWigs alongside H3K9me3 in the same cells.
Chaetocin-mediated SUV39H1 inhibition targets stemness and oncogenic networks of diffuse midline gliomas and synergizes with ONC201 · Xin D, Liao Y, et al. · Neuro-Oncology 2024
PMID 38011799
What this evidence cannot show
Read from the GEO deposit's own SOFT records, not from the paper's figures. The antibody is CST #8729S and no CUT&RUN validation of it is asserted here; a SUV39H1 knockout control track is not in the deposit.
Evidence that argues against this (1)
- SU-DIPG13 and SU-DIPG36 are H3K27M-mutant lines with a globally redistributed H3K27me3 landscape and an unusual heterochromatin state, so a SUV39H1 distribution measured there is a poor proxy for a normal or a leukaemic nucleus. Bender S et al. 2013
E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Four SUV39H1 ChIP-Seq samples in MCF-7, with and without estradiol, are deposited in GEO and the sample records name the ChIP antibody explicitly.
Shape of promoter antisense RNAs regulates ligand-induced transcription activation · Yang F, et al. · Nature 2021
PMID 34194047
What this evidence cannot show
The deposited supplementary files are bedGraphs produced by that lab's pipeline; the genome build is not stated in the per-sample data_processing lines the way GSE225028's is, so a build check against the chrom table is required before any coordinate is read off them.
Evidence that argues against this (1)
- These are estrogen-receptor-signalling experiments in a breast line. SUV39H1 occupancy there is conditioned on an ER-driven programme that does not exist in a blood cell, so the track answers 'where can SUV39H1 be found' and not 'where is SUV39H1 in the cell this project cares about'. NCBI Gene Expression Omnibus 2021
E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
ChIP-Atlas independently indexes four human SUV39H1 ChIP experiments, and returns a clean zero for BARD1 and EXO1 in the same batch with CTCF passing as a positive control.
ChIP-Atlas hg38 target table for antigen SUV39H1 (SRX7030945, SRX7030946, SRX7030947, SRX7030948; MCF-7) · ChIP-Atlas / DBCLS 2026
What this evidence cannot show
The threshold token in that URL is load-bearing and silently wrong if mistyped: `SUV39H1.05.tsv` returns HTTP 404, and so does `CTCF.05.tsv`, which is how this beat learned the route was broken rather than empty. Every count here was taken with the CTCF control passing in the same batch.
Evidence that argues against this (1)
- ChIP-Atlas indexes from SRA metadata, so it inherits every antigen-naming error upstream and it does not contain the GSE225028 CUT&RUN samples at all - meaning ChIP-Atlas alone would have found 4 of the 6 samples and a beat that stopped there would have under-counted. Oki S et al. 2018
E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
SUV39H1's ENCODE zero is honest - no target record under any label - and it is distinguishable from the BARD1 false 404 because BARD1 and RAD50 return HTTP 200 on the same endpoint.
ENCODE portal target records: /targets/BARD1-human/ 200, /targets/RAD50-human/ 200, /targets/SUV39H1-human/ 404, /targets/EXO1-human/ 404, /targets/DNA2-human/ 404 · ENCODE Project 2026
What this evidence cannot show
Fetched 2026-09-22. A 404 from the ENCODE search endpoint means zero results for that query, not a missing service; the target-record endpoint is the one that discriminates and it is the one used here.
Evidence that argues against this (1)
- The bare-symbol target query is itself incomplete. SUV39H2 has no `SUV39H2` target record but does have `eGFP-SUV39H2`, and CBX5 has both `CBX5` and `3xFLAG-CBX5`. A census keyed on the bare symbol can therefore still under-count, and this one only escapes because both tagged labels carry zero experiments. ENCODE Project 2026
E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
The untried alias KMT1A retrieves 22 records and none of them pairs SUV39H1 with KMT2A or MLL, so the literature half of the original absence survives the spelling test.
NCBI E-utilities esearch db=pubmed: KMT1A[tiab] = 22; KMT1A[tiab] AND KMT2A[tiab] = 0; KMT1A[tiab] AND MLL[tiab] = 0; SUV39H[tiab] AND MLL[tiab] = 1 (PMID 19460867) · NCBI PubMed 2026
What this evidence cannot show
Run through scripts/pubmed_safe.py, which raises rather than returning a zero with an empty querytranslation. Both zeros came back with non-empty translations and warninglist 'No items found.'. This bounds titles and abstracts only - rule 80 - and no full text was searched.
Evidence that argues against this (1)
- PMID 19460867 does pair a Suv39h null with an Mll null in the same mouse ES-cell experiment. It is about telomeric small RNAs, not about occupancy at the Kmt2a locus, so it is off topic - but it is the nearest thing to a counter-example and a reader should see it rather than take the zero on trust. Cao F et al. 2009

Connected to

refutes →No H3K9me3 writer and no H3K9me3 reader has ever been localised at the KMT2A locus IN THE PUBLISHED LITERATURE, in any cell type: 20 queries, 5 alias spellings, 0 on-topic records - and CORRECTED 2026-09-22: the SUV39H1 half is no longer unanswerable from archives. ENCODE still holds zero SUV39H1 experiments (verified by three routes with controls), but GEO holds two deposits, and Q114 read both at this locus and found the writer ABSENT
runs the GEO sweep that entry named as free and unrun; its archive half falls - six human SUV39H1 occupancy samples in two cell types - while its ENCODE statement and its literature half both survive
related to →SUV39H1 (O43463, KMT1A) - the H3K9me3 writer: a menin effector whose binding region OVERLAPS the drug pocket at its N-terminus, a tumour suppressor in MLL-AF9 AML, recruited to double-strand breaks by CTCF, and never localised at the KMT2A locus in any cell type
first chromatin-occupancy datasets for this protein recorded in this graph: GSE225028 (CUT&RUN, hg38 bigWig, two DIPG lines) and GSE139196 (ChIP-Seq, MCF-7)
supports →All 39 bounded absences in this graph, classified by WHAT BOUNDS THEM: 23 rest on an abstract-level literature search, 7 on a threshold or a query filter, and only 2 on reading the primary source
the one overturn out of seven re-runs, and the row that fixes the census's mode-2 exposure count
related to →DNA2 (P51530, 1,060 aa) - EXO1's redundant partner in long-range resection, dual mitochondrial and nuclear, and the gene whose DepMap row is measured in 80 screens where every other cascade member has 1,258
incidental to the ChIP-Atlas sweep: DNA2 has one chromatin-occupancy experiment, SRX3520279 in HCT-116, so its ENCODE 404 is not an archive-wide zero
← supportsGSE225028: SUV39H1 and H3K9me3 CUT&RUN in SU-DIPG13 and SU-DIPG36 - the only published human SUV39H1 occupancy map deposited as a readable hg38 bigWig, and the enzyme and its own mark were profiled in the same cells
the deposit that carries the overturn: hg38 bigWigs, readable at any locus with compute alone
← supportsGSE139196: SUV39H1 ChIP-Seq in MCF-7 with and without 40 minutes of estradiol - a ligand-inducible SUV39H1 occupancy experiment, filed under the retired symbol KMT1A in its own sample titles
the second of the two deposits, and the only SUV39H1 experiment anywhere with two conditions of the same cell
← proposed byAn ENCODE-only archive bound systematically undercounts chromatin-modifying enzymes, because ENCODE's target catalogue is transcription-factor-shaped - measured 3-for-3 on the H3K9 machinery, with ratios of infinity, 2.8x and 7.3x
the SUV39H1 overturn is the observation this hypothesis generalises, and the hypothesis is what makes that observation testable rather than anecdotal
← supportsThree archives, one answer: SUV39H1 has no ENCODE target record under any label (three routes, all empty), ChIP-Atlas's entire hg38 SUV39H1 holding IS the four GSE139196 MCF-7 samples re-indexed, and KMT2A is not among the 26 genes ChIP-Atlas calls SUV39H1 targets
confirms the ENCODE zero by three controlled routes and narrows ChIP-Atlas from a third route to a re-index of one deposit
← supportsRead at last, against a pre-registered class-matched null: in SU-DIPG13 and SU-DIPG36 the SUV39H1 CUT&RUN mean at the KMT2A exon-11 CTCF anchor is 0.0000 and 1.6660 - the first an EXPLICIT zero record, not missing data - and no one of the seven KMT2A intervals in either line exceeds 1.21x the genome-wide random-tile mean. The pre-registered NEGATIVE prior held.
answers its open question on the 90 kb gene body, for the DIPG system