The quiet genome, audited - the famous 1.3 mutations is a count of point mutations in the dominant clone, and all three of those words do work
also called 1.3 non-silent mutations per case · silent mutational landscape of infant MLL-AF4 ALL · low mutation burden claim, measured
Infant KMT2A leukaemia is famous for having almost no mutations - an average of about 1.3 per case. That number is real, but it counts only spelling changes in the DNA, only in the cells that make up the bulk of the leukaemia, and only down to what the machine of the day could see. Checking each of those limits does not overturn the claim; it changes what the claim licenses.
The same thing, for a scientist
`comut-low-mutation-burden-infant-kmt2a-r` records the claim. This entry audits how it was measured, because the number is quoted throughout this graph and in the field as though it were a total mutation burden, and it is not. **The headline.** *'Infant MLL-R ALL has one of the lowest frequencies of somatic mutations of any sequenced cancer, with the predominant leukemic clone carrying a mean of 1.3 non-silent mutations'* [E1]. Two independent series agree on the direction: three infant MLL-AF4 cases by WGS showed 'few somatic changes' in copy number, LOH and SNVs [E2]; and 124 uniformly treated infant B-ALL cases showed 'the most silent mutational landscape reported so far for any sequenced pediatric cancer' [E3]. **Three qualifications, none of which the field usually carries.** 1. **'Non-silent mutations'** means SNVs and indels. It does not count copy-number alterations, of which this disease averages **1.0 per case**, and whose commonest recurrent target is PAX5 in 5 of 22 cases [E1]. On lesion count rather than mutation count the disease is roughly twice as loud, and it has a recurrent second target. 2. **'In the predominant leukemic clone'** is a conditioning clause, not decoration. In the same cohort **40% of all somatic SNVs sat below 30% MAF at 92% mean purity** [E1]. The 1.3 is what is left after the subclonal fraction is set aside, and the subclonal fraction was the larger and more interesting one. 3. **The floor was the instrument's.** WGS at 39x mean haploid coverage, with deep capture at 437x on five cases [E1]. Targeted duplex sequencing of the same disease later found **62% of cancer-associated mutations below standard NGS resolution** [E4]. **Does the claim survive? Yes, comparatively, and that is the form that matters.** The duplex study that lowered the floor by two orders of magnitude found that **infants still had fewer cancer-associated mutations than the older children** [E4], and the index study's own internal comparison stands on its own terms: 1.3 in infants against 6.5 in non-infant MLL-R leukaemia, P = 7.15 x 10^-5 [E1]. Depth moved the floor and preserved the contrast. **One finding does qualify the usual explanation.** The quietness is routinely attributed to the prenatal origin and short latency of the infant disease. But adult KMT2A-rearranged BCP-ALL also shows a **low mutational burden, explicitly described as 'reminiscent of infant BCP-ALL'** [E5]. If adults with the same fusion are quiet too, quietness is at least partly a property of the fusion rather than of infancy - and the developmental explanation this graph leans on is not the only one available.
What was measured, cohort by cohort
| study | n | assay | stated burden | qualifier |
|---|---|---|---|---|
| St. Jude PCGP 1 | 22 WGS infant MLL-R; 47 infant MLL-R total | WGS 39x; 232-gene panel; SNP array; 437x capture on 5 | 1.3 non-silent mutations | in the predominant leukemic clone; 40% of all SNVs below 30% MAF |
| St. Jude, non-infant arm 1 | 20 MLL-R older children | same | 6.5 per case (P = 7.15e-5 vs infants) | and 45% with epigenetic-regulator mutations |
| Dobbins 2 | 3 infant MLL-AF4 pro-B ALL | WGS, tumour and matched germline | 'few somatic changes' | n = 3; no number given in the abstract |
| Agraz-Doblas 3 | 124 infant B-ALL, Interfant-99/06 | genome-wide multi-layer | 'the most silent mutational landscape reported so far for any sequenced pediatric cancer' | recurrent mutation only in K-RAS and N-RAS |
| Pilheden 4 | 20 infants, 7 children KMT2A-r ALL | targeted duplex, floor ~0.01-0.1% | 87 cancer-implicated mutations; 62% below standard NGS | infants had fewer than the children |
| GRAALL 5 | 141 adults, KMT2A-r BCP-ALL | panel profiling | low mutational burden, 'reminiscent of infant BCP-ALL' | adults, same fusion class |
Is it an artefact of small cohorts? No. 22 whole genomes is small, but 124 cases genome-wide 3 and 47 infant MLL-R cases on a panel 1 are not, and three independent groups on three continents report the same direction. The small-cohort worry applies to individual gene frequencies within the disease - where 4/22 and 5/22 counts are genuinely fragile - and not to the aggregate burden claim.
Is it an artefact of shallow sequencing? Partly, in the sense that the specific number 1.3 is a 39x-conditioned, dominant-clone-conditioned figure. Not in the sense that would matter: the one study that sequenced deeply enough to answer it found the infant-versus-child contrast intact 4.
What the number is allowed to be used for
Allowed. That infant KMT2A-rearranged ALL has an exceptionally low burden of clonal point mutations relative to other paediatric cancers and to older children with the same fusion. That no second recurrent point-mutation driver has emerged. That the fusion is therefore close to sufficient.
Not allowed. That the disease has no recurrent second lesion - PAX5 focal loss is present in 23% of whole-genome-sequenced cases 1. That it has almost no mutations in total - 62% of the cancer-associated mutations in this disease sit below the depth at which the number was derived 4. That the quietness is caused by infancy - adults with the same fusion are quiet too 5.
And one usage that needs particular care, because this graph does it: the number is used to argue that there is no cooperating lesion to co-target therapeutically. That argument is sound on the point-mutation evidence and it is stronger when the copy-number evidence is added, not weaker, because PAX5 and CDKN2A losses are not druggable either. The correction is to the wording, not to the conclusion.
| audited claim | comut-low-mutation-burden-infant-kmt2a-r |
|---|---|
| verdict | the claim survives, with three qualifications to its wording |
| survives because | the one study that lowered the detection floor by roughly two orders of magnitude still found infants carrying fewer cancer-associated mutations than older children with the same fusion |
| unexpected qualification | adult KMT2A-rearranged BCP-ALL also has a low mutational burden described as reminiscent of infant BCP-ALL, so quietness is not uniquely a property of the infant developmental window |
| not an artefact of | small cohorts - 124 cases genome-wide and 47 infant KMT2A-r cases on a panel, from three independent groups, agree on direction |
qualifications
[ "'non-silent mutations' counts SNVs and indels only; the disease also averages 1.0 copy-number alteration per case, whose commonest recurrent target is PAX5 in 23% of whole-genome-sequenced cases", "'in the predominant leukemic clone' is a conditioning clause: 40% of all somatic SNVs in the same cohort sat below 30% MAF at 92% purity, and were excluded from the 1.3", "the floor was 39x WGS; duplex sequencing later found 62% of cancer-associated mutations in this disease below standard NGS resolution" ]
- The quiet-genome claim is the foundation of the single-target therapeutic argument in this disease. Every qualification in this entry narrows what that argument may say, and the copy-number qualification in particular means 'no second lesion' is false as written even though 'no second druggable lesion' remains true.
- A mutation burden derived at one depth is a property of the assay. This graph should not quote 1.3 without the two conditioning clauses attached to it, and several existing entries do.
- If the burden is low because latency is short and the cell is fetal, then treatments that extend survival without eradicating disease are creating, for the first time, a version of this leukaemia with time to accumulate mutations - which is exactly what the early-relapse TP53 and IKZF1 alterations look like.
- A quiet genome is also a genome with few neoantigens. The same property that makes the fusion look sufficient makes immune-based approaches harder, and nothing in this graph currently states that trade-off.
- What is the genome-wide mutation burden of infant KMT2A-rearranged ALL at duplex depth? The 62% figure comes from a targeted panel; nobody has run a duplex-depth genome-wide measurement in this disease.
- Is adult KMT2A-rearranged BCP-ALL quiet for the same reason infant disease is, or for a different one? If the fusion enforces quietness independent of age, the prenatal-origin explanation this graph relies on is at most partial.
- Why is adult KMT2A-rearranged AML not quiet when adult KMT2A-rearranged BCP-ALL is? Lineage appears to matter as much as age and no entry in this graph addresses it.
- What is the total LESION burden - SNVs, indels, CNAs and structural variants together - per case in this disease? Every published figure counts one class, and the number the field quotes counts the smaller one.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Whole-genome, exome, RNA and targeted DNA sequencing of 65 infants (47 MLL-rearranged) and 20 older children with MLL-rearranged leukaemia found that infant MLL-rearranged ALL has one of the lowest somatic mutation frequencies of any sequenced cancer, with the predominant leukaemic clone carrying a mean of 1.3 non-silent mutations against 6.5 per case in older children (P = 7.15e-5); 40% of somatic SNVs were detected below 30% mutant allele frequency despite 92% mean tumour purity; the average was 1.0 copy-number alteration per infant MLL-R case, with PAX5 the most frequently affected gene among structural events not caused by the rearrangement; and older children had frequent (45%) mutations in epigenetic regulators, rarely mutated in infants.
The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias. · Andersson AK, Ma J, Wang J, et al. · Nature Genetics 2015What this evidence cannot show
The headline 1.3 is explicitly conditioned on the dominant clone and is a count of non-silent SNVs and indels only. WGS was at 39x mean haploid coverage on 22 cases; the deep-capture arm that establishes the subclonal structure covers five cases at an average of 437x. The infant-versus-older-child comparison is 22 against 20 cases and the older-child arm mixes ALL and AML. Full text read via efetch db=pmc (PMC4553269) after Europe PMC fullTextXML returned HTTP 500.
Evidence that argues against this (1)
- Targeted duplex sequencing of the same disease found 62% of cancer-associated mutations below standard NGS resolution, so the 1.3 figure is a floor-conditioned measurement and the true number of cancer-associated variants per case is larger. Pilheden M et al. 2022
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Whole-genome sequencing of diagnostic leukaemic blasts and matched germline samples from three infants with MLL-AF4 pro-B ALL revealed few somatic changes across copy-number abnormalities, loss of heterozygosity and single-nucleotide variants, demonstrating that only a very small number of mutations are necessary to generate infant MLL-rearranged leukaemia.
The silent mutational landscape of infant MLL-AF4 pro-B acute lymphoblastic leukemia. · Dobbins SE, Sherborne AL, Ma YP, et al. · Genes, Chromosomes and Cancer 2013What this evidence cannot show
Three patients. No mutation count is given in the abstract, and no sequencing depth. Its value is as an independent replication of direction by a different group, not as a quantitative estimate. The conclusion drawn - that few mutations are necessary - is an inference about sufficiency from an observation about count, and three genomes cannot establish it. Abstract read.
Evidence that argues against this (1)
- Three cases cannot distinguish a genuinely quiet disease from a sampling of its quietest members, and the deeper sequencing available since shows a substantial mutational reservoir below the depth used here. Pilheden M et al. 2022
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Multi-layered genome-wide analysis of 124 de novo infant B-ALL cases uniformly diagnosed and treated on Interfant-99/06 found the most silent mutational landscape reported so far for any sequenced paediatric cancer, with recurrent mutations exclusively in K-RAS and N-RAS, subclonal and frequently lost at relapse.
Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis. · Agraz-Doblas A, Bueno C, Bashford-Rogers R, et al. · Haematologica 2019What this evidence cannot show
'Most silent reported so far' is a superlative relative to what had been sequenced by 2019 and is not a measurement. Sequencing depth is not given in the abstract. The cohort is infant B-ALL including non-KMT2A-rearranged cases; the abstract does not separate the burden figure by fusion status. Abstract read; full text not retrieved for this entry.
Evidence that argues against this (1)
- The same abstract notes a larger number of non-recurrent, non-silent mutations alongside the recurrent RAS ones, so 'silent' describes the recurrence structure more than the raw count. Agraz-Doblas A et al. 2019
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Targeted duplex sequencing combined with standard NGS on 20 infants and 7 children with KMT2A-rearranged ALL identified 87 nonsynonymous mutations previously implicated in cancer, of which 62% were below the resolution of standard NGS and only 33 exceeded 2% cellular prevalence, with infants having fewer such mutations than the children.
Duplex Sequencing Uncovers Recurrent Low-frequency Cancer-associated Mutations in Infant and Childhood KMT2A-rearranged Acute Leukemia. · Pilheden M, Ahlgren L, Hyrenius-Wittsten A, et al. · HemaSphere 2022What this evidence cannot show
A targeted panel, so the 87 mutations are bounded by panel design and cannot be converted into a genome-wide burden - this study qualifies the 1.3 figure's floor without replacing the figure. 27 patients. A variant at 0.1% cellular prevalence cannot be assigned to a leukaemic rather than a normal cell by sequencing alone. Abstract read; full text not retrieved for this entry.
Evidence that argues against this (1)
- Because the assay is a targeted panel of genes already known to be altered in KMT2A-rearranged leukaemia, it is enriched by design for finding exactly what it found, and the 62%-below-NGS figure is a statement about those genes rather than about the genome. Pilheden M et al. 2022
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Molecular profiling of 141 adults with KMT2A-rearranged B-cell precursor ALL, drawn from 1,091 Philadelphia-negative BCP-ALL patients in three consecutive GRAALL trials, showed a low mutational burden in this subtype, described by the authors as reminiscent of infant BCP-ALL.
Genetic alterations and MRD refine risk assessment for KMT2A-rearranged B-cell precursor ALL in adults: a GRAALL study. · Kim R, Bergugnat H, Pastoret C, et al. · Blood 2023What this evidence cannot show
The burden statement is qualitative in the abstract - no mutation count per case is given - and the profiling is panel-based rather than genome-wide, so it is not directly comparable to the 1.3 figure derived from WGS. It is B-lineage ALL only; adult KMT2A-rearranged AML by contrast carries DNMT3A, TP53 and KRAS mutations at prognostically meaningful rates. Abstract read; full text not retrieved.
Evidence that argues against this (1)
- Adult KMT2A-rearranged AML is NOT quiet in the same way - DNMT3A, KRAS and TP53 all carry prognostic weight there - so the 'quietness follows the fusion' reading holds for the B-lineage disease and fails for the myeloid one, which means lineage, not just fusion, is part of the explanation. Hernandez-Sanchez A et al. 2024
audits how the 1.3 figure was measured: point mutations only, dominant clone only, 39x floor
- related to →Subclonal cooperating lesions lost at relapse - what was actually measured, and why 'lost' is the wrong word
the same instrument audit from the burden side rather than the clonality side
the copy-number qualification: the commonest recurrent second lesion is not a point mutation
the therapeutic negative rests on the audited burden claim, not the unqualified one
an ITD in a non-hotspot region: another lesion class the mutation-count framing under-reports