CoChat
Sign Up
Cited by 15Neuroinflammation and Neurodegeneration MechanismsAlzheimer's disease research and treatmentsSingle-cell and spatial transcriptomics

Somatic cancer variants enriched in Alzheimer’s disease microglia-like cells drive inflammatory and proliferative states

Open access
Huang, August Yue, Zhou, Zinan, Talukdar, Maya, Enyenihi, Liz, Miller, Michael B., Chhouk, Brian, Rosen, Ila, Zheng, Mengyue +27 more·2026·Cell
DOI 10.1016/j.cell.2026.03.040PMC PMC13107304
Open this paper in CoChat — free account

Abstract

Alzheimer's disease (AD) is a neurodegenerative condition characterized by microglia-mediated neuroinflammation. Deep (>1,000×) panel sequencing of 311 brain samples revealed enrichment of somatic single-nucleotide variants (sSNVs) in cancer driver genes in AD brains, especially in genes associated with clonal hematopoiesis (CH). These sSNVs were associated with clonal expansion and carried by both microglia-like brain macrophages (MLBMs) in multiple brain regions as well as paired blood, suggesting a likely hematopoietic origin. Single-nucleus RNA sequencing data from 62 additional AD and control brains revealed increased somatic copy number variants (sCNVs) associated with CH in AD MLBMs, whereas single-cell multi-omic analyses demonstrated that sSNV- and sCNV-carrying MLBMs exhibited inflammatory and proliferative transcriptional signatures characteristic of disease-associated microglia. These signatures were recapitulated in induced pluripotent stem cell-derived microglia-like cells with TET2, ASXL1, and DNMT3A variants. These findings suggest that clonal somatic driver variants in MLBMs are enriched in AD, potentially promoting neuroinflammation and neurodegeneration.

AI snapshot

Written from the paper’s full text.

TL;DR

Deep panel sequencing found more somatic cancer-driver variants, especially clonal-hematopoiesis genes, in Alzheimer’s disease (AD) brains than controls, with variants enriched in microglia-like brain macrophages (MLBMs). Single-cell and iPSC models link these variants to inflammatory and proliferative disease-associated microglial states.

Key findings

  • In ROSMAP prefrontal cortex (190 AD, 121 controls), AD brains had more somatic single-nucleotide variants (sSNVs) in 149 cancer driver genes (p = 0.001, OR = 1.3; stringent list p = 0.008, OR = 1.6), remaining significant after covariates (p = 0.03); age was independently associated with sSNV burden (p = 0.002).
  • The top ten most commonly mutated genes were affected in 39% of AD patients versus 20% of controls; TET2, ASXL1, KMT2D, ATRX, and CBL had nominally more variants in AD (p < 0.05, one-tailed), though not significant after correction across 149 genes. AD variants showed higher VAFs and stronger positive selection, with >350% increase in positively selected cells.
  • All 10 AD variants and 7 of 8 control tumor-suppressor variants tested were enriched 2- to 438-fold in CSF1R+ MLBMs versus neurons; AD variants were present across multiple cortex regions and cerebellum, absent in neurons, and detected in matched blood, suggesting a hematopoietic origin.
  • In SEA-AD snRNAseq (31 AD, 31 controls), AD microglia-CAMs had nominally more mosaic chromosomal alterations (4 vs 1) and 8-fold more mCA-carrying cells (p = 0.06); mutant microglia-CAMs upregulated immune-activation genes and overlapped DAM genes (p = 0.04). iPSC-derived microglia-like cells with DNMT3A, TET2, or ASXL1 variants were enriched in DAM and proliferation states and showed pro-inflammatory/proliferation pathway upregulation.

Methods

  • UMI-based hybrid capture panel targeting 149 cancer driver genes, sequenced at >1000X after UMI collapsing, on ROSMAP prefrontal cortex DNA from 190 AD and 121 controls; variants validated by amplicon sequencing.
  • Fluorescence-activated nuclei sorting for CSF1R+ MLBMs and NeuN+ neurons, snRNAseq, GoT-ChA simultaneous snATACseq/genotyping, and GO-TEN genotyping/transcriptomics; SEA-AD snRNAseq was used for mCA calling with CONICSmat.
  • Isogenic iPSC lines with DNMT3A R882H, TET2 DelE3-E11, or ASXL1 E635fs*15 were differentiated to iMGLs and analyzed by scRNAseq; analyses included burden tests, regression, dN/dS selection, permutation, differential expression, GO, and GSEA.

Limitations

  • The paper states SEA-AD sample size was too small for mCA enrichment in AD to reach statistical significance; individual gene enrichments were not significant after multiple-testing correction; the MCI group was not included in downstream analyses due to limited sample size.
  • The authors state definitive lineage origin of mutant MLBMs requires lineage-tracing studies, and an unidentified shared microglia-macrophage progenitor cannot be excluded; GO-TEN had modest genotyped cell numbers.
  • Evident from the text: several key findings are nominal or trend-level, and ethnicity information was not available for the cohorts.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.cell.2026.03.040

Cites

111 of 111 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

26 of 35 cited sources checked · 13 supported · 2 flagged · accuracy not yet measured

  • Premisecitation

    The somatic BRAF p.V600E variant in the microglial lineage has been causally implicated in degeneration of neurons secondary to clonal expansion and activation of mutant microglia in both mouse models and humans.

    “Moreover, the somatic BRAF p.V600E variant, a common cancer driver variant, in the microglial lineage has been causally implicated not in cancer, but instead in degeneration of neurons secondary to clonal expansion and activation of mutant microglia in both mouse models and humans29.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [29] Mass E, Jacome-Galarza CE, Blank T, Lazarov T, Durham BH, Ozkaya N, Pastore A, Schwabenland M, Chung YR, Rosenblum MK, (2017). A somatic mutation in erythro-myeloid progenitors causes neurodegenerative disease. Nature 549, 389– 393. 10.1038/nature23672. 28854169 PMC6047345 source

      The paper’s Abstract reports that BRAFV600E-mutant microglia expand and drive neuronal loss in mice, with BRAF inhibition reducing the pathology. It also reports activated BRAFV600E-positive microglia at sites of neuronal loss in human histiocytosis patients, but describes the human findings as shared features rather than a demonstration of causation in humans. Page numbers are not available in our copy. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
  • Premisecitation

    Two recent studies investigating CHIP variants primarily using whole-exome sequencing data from blood cells found no effect or a protective effect on Alzheimer's disease risk.

    “Two recent studies investigating CHIP variants primarily using whole-exome sequencing (WES) data from blood cells found no effect24 or a protective effect25 on AD risk.”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [24] Kessler MD, Damask A, O’Keeffe S, Banerjee N, Li D, Watanabe K, Marketta A, Van Meter M, Semrau S, Horowitz J, (2022). Common and rare variant associations with clonal haematopoiesis phenotypes. Nature 612, 301– 309. 10.1038/s41586-022-05448-9. 36450978 PMC9713173 source

      The paper describes identifying CHIP carriers using exome sequencing data, but the passages from its Phenotypic associations and Longitudinal disease risk sections do not report an Alzheimer’s disease result. Searches for Alzheimer’s disease, AD risk, and dementia found no passage establishing the attributed finding of no effect on Alzheimer’s risk. (LLM, conf=0.72)

      Experimental — accuracy not yet measured
    • cites [25] Bouzid H, Belk JA, Jan M, Qi Y, Sarnowski C, Wirth S, Ma L, Chrostek MR, Ahmad H, Nachun D, (2023). Clonal hematopoiesis is associated with protection from Alzheimer’s disease. Nat Med. 10.1038/s41591-023-02397-2. PMC10353941 37322115 source

      The Results (Association between CHIP and AD dementia; page not available) report a protective association between CHIP and Alzheimer’s dementia, including in the blood-derived WES replication cohort. But the discovery cohorts identified CHIP from blood-derived whole-genome sequencing, so describing this study as primarily using WES is not supported by the cited passages. The separate “no effect” finding is attributed to citation 24, not this paper. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • Findingown data

    The enrichment of somatic variants in Alzheimer's disease was confirmed when considering only CHIP variants previously identified from blood whole-genome sequencing datasets.

    “We further confirmed the enrichment of somatic variants in AD when considering only CHIP variants previously identified from blood whole-genome sequencing datasets25,37 (Figure S3A–C).”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [25] Bouzid H, Belk JA, Jan M, Qi Y, Sarnowski C, Wirth S, Ma L, Chrostek MR, Ahmad H, Nachun D, (2023). Clonal hematopoiesis is associated with protection from Alzheimer’s disease. Nat Med. 10.1038/s41591-023-02397-2. PMC10353941 37322115 source
    • cites [37] Bick AG, Weinstock JS, Nandakumar SK, Fulco CP, Bao EL, Zekavat SM, Szeto MD, Liao X, Leventhal MJ, Nasser J, (2020). Inherited causes of clonal haematopoiesis in 97,691 whole genomes. Nature 586, 763– 768. 10.1038/s41586-020-2819-2. 33057201 PMC7944936 source
  • Findingown data

    The somatic single-nucleotide variants were carried by microglia-like brain macrophages in multiple brain regions and in paired blood.

    “These sSNVs were associated with clonal expansion and carried by both microglia-like brain macrophages (MLBMs) in multiple brain regions as well as paired blood, suggesting a likely hematopoietic origin.”
    Quote in the paperSubstantial claim
  • Findingown data

    The somatic single-nucleotide variants enriched in Alzheimer's disease brains were associated with clonal expansion.

    “These sSNVs were associated with clonal expansion and carried by both microglia-like brain macrophages (MLBMs) in multiple brain regions as well as paired blood, suggesting a likely hematopoietic origin.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 34
  • May combine two claims. 32

  • Classified as probably not a substantial claim.

    Premisecitation

    During aging or neurodegeneration, bone marrow-derived blood monocytes can infiltrate the brain through a disrupted blood-brain barrier and differentiate into microglia-like cells.

    “Moreover, recent studies have reported that during aging or neurodegeneration, bone marrow-derived blood monocytes can infiltrate the brain through a disrupted blood-brain barrier (BBB) and differentiate into microglia-like cells, exhibiting transcriptional and functional similarity to brain-resident microglia13–15.”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites [13] Lund H, Pieber M, Parsa R, Han J, Grommisch D, Ewing E, Kular L, Needhamsen M, Espinosa A, Nilsson E, (2018). Competitive repopulation of an empty microglial niche yields functionally distinct subsets of microglia-like cells. Nat Commun 9, 4845. 10.1038/s41467-018-07295-7. 30451869 PMC6242869 source

      The Results and Discussion report that bone marrow-derived monocytes entered the brain and became microglia-like cells after experimental microglial depletion, but explicitly state that engraftment occurred without blood-brain barrier breakdown. The paper does not establish this process during aging or neurodegeneration. (LLM, conf=0.88)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured
    • cites [14] Wilk CM, Cathomas F, Torok O, Le Berichel J, Park MD, Bigenwald C, Heaton GR, Hamon P, Troncoso L, Scull BP, (2023). Circulating senescent myeloid cells infiltrate the brain and cause neurodegeneration in histiocytic disorders. Immunity 56, 2790– 2802 e2796. 10.1016/j.immuni.2023.11.011. 38091952 PMC11587932 source

      The Discussion reports that circulating BRAFV600E+ myeloid cells entered the brain alongside BBB breakdown and differentiated into senescent CD11a+ macrophages in a model of LCH-associated neurodegeneration. It does not establish that this occurs generally during aging or neurodegeneration, or that the resulting cells are transcriptionally and functionally similar to brain-resident microglia; its extension to age-associated conditions is explicitly speculative. The indexed copy provides section labels but no page numbers. (LLM, conf=0.87)

      Experimental — accuracy not yet measured
    • cites [15] Silvin A, Uderhardt S, Piot C, Da Mesquita S, Yang K, Geirsdottir L, Mulder K, Eyal D, Liu Z, Bridlance C, (2022). Dual ontogeny of disease-associated microglia and disease inflammatory macrophages in aging and neurodegeneration. Immunity 55, 1448– 1465 e1446. 10.1016/j.immuni.2022.07.004. 35931085 source

  • May combine two claims.

    Study design and conductcitation

    UMI-based panel sequencing detects somatic variants with variant allele fractions as low as 0.1% with much higher sensitivity and precision than previous methods not employing consensus error correction.

    “By generating a consensus sequence from multiple reads derived from the same original DNA molecule, UMI-based panel sequencing exponentially reduces base-calling errors and detects somatic variants with VAFs as low as 0.1% (Figure S1C–D), with much higher sensitivity and precision than previous methods not employing consensus error correction30.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [30] Ganz J, Maury EA, Becerra B, Bizzotto S, Doan RN, Kenny CJ, Shin T, Kim J, Zhou Z, Ligon KL, (2022). Rates and Patterns of Clonal Oncogenic Mutations in the Normal Human Brain. Cancer Discov 12, 172– 185. 10.1158/2159-8290.CD-21-0245. 34389641 PMC8758513 source

      The Results (Identification of Variants from Deep-Targeted Sequencing; page unavailable) limit the analyzed variants to 0.5%–15% VAF and report sensitivity comparable to other studies, not much higher. The Discussion (page unavailable) says sensitivity at 0.5% VAF was approximately 10% and that detecting still-lower-frequency events remained challenging, contrary to the claimed 0.1% detection capability. (LLM, conf=0.88)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Somatic variants accumulate in all cell types that have been studied, both during normal development and during aging.

    “Somatic variants accumulate in all cell types that have been studied, both during normal development and during aging16–18.”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites [16] Lodato MA, Rodin RE, Bohrson CL, Coulter ME, Barton AR, Kwon M, Sherman MA, Vitzthum CM, Luquette LJ, Yandava CN, (2018). Aging and neurodegeneration are associated with increased mutations in single human neurons. Science 359, 555– 559. 10.1126/science.aao4426. 29217584 PMC5831169 source

      The Main Text reports that somatic single-nucleotide variants accumulate with age in human neurons and notes that variants are already present in fetal progenitor cells. It does not establish accumulation during both development and aging in all studied cell types; its direct analysis is limited to neurons. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [17] Li R, Du Y, Chen Z, Xu D, Lin T, Jin S, Wang G, Liu Z, Lu M, Chen X, (2020). Macroscopic somatic clonal expansion in morphologically normal human urothelium. Science 370, 82– 89. 10.1126/science.aba7300. 33004515 source
    • cites [18] Li R, Di L, Li J, Fan W, Liu Y, Guo W, Liu W, Liu L, Li Q, Chen L, (2021). A body map of somatic mutagenesis in morphologically normal human tissues. Nature 597, 398– 403. 10.1038/s41586-021-03836-1. 34433965 source

  • May combine two claims.

    Premisecitation

    A more recent study utilizing deep panel sequencing revealed an increased burden of CHIP variants in Alzheimer's disease blood samples compared to age-matched controls, primarily driven by low-VAF variants, and this increase was independently confirmed in whole-genome sequencing of ~30,000 cases and controls.

    “A more recent study utilizing deep panel sequencing revealed an increased burden of CHIP variants in AD blood samples compared to age-matched controls, primarily driven by low-VAF variants26, and this AD-related increase of CHIP variants was independently confirmed in whole-genome sequencing (WGS) of ~30,000 cases and controls26.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [26] Choi J, Park KS, Guen YL, Park J-H, Zhou Z, Enyenihi L, Rosen I, Choi Y-H, Walsh CA, Kim J-W, (2025). Clonal Hematopoiesis Mutations Increase Risk of Alzheimer’s Disease with APOE ε3/ε3 Genotype. bioRxiv. 10.1101/2025.05.19.654981. source

  • May combine two claims.

    Premisecitation

    All of the Alzheimer's disease-enriched genes represent critical tumor suppressor genes widely implicated in various cancers and clonal hematopoiesis.

    “All of these AD-enriched genes represent critical TSGs widely implicated in various cancers33 and CH34.”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [33] Tate JG, Bamford S, Jubb HC, Sondka Z, Beare DM, Bindal N, Boutselakis H, Cole CG, Creatore C, Dawson E, (2019). COSMIC: the Catalogue Of Somatic Mutations In Cancer. Nucleic Acids Res 47, D941– D947. 10.1093/nar/gky1015. 30371878 PMC6323903 source

      The Cancer gene census section (page unavailable) describes genes implicated in cancer and explains that some are classified as tumour suppressor genes. It does not identify the AD-enriched genes or establish that all of them are tumour suppressors implicated in various cancers. The separate clonal-hematopoiesis assertion is attached to citation 34, not this source. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
    • cites [34] Zink F, Stacey SN, Norddahl GL, Frigge ML, Magnusson OT, Jonsdottir I, Thorgeirsson TE, Sigurdsson A, Gudjonsson SA, Gudmundsson J, (2017). Clonal hematopoiesis, with and without candidate driver mutations, is common in the elderly. Blood 130, 742– 752. 10.1182/blood-2017-02-769869. 28483762 PMC5553576 source

      The 'Association of driver mutations with clonal hematopoiesis' section reports significant CH associations for several genes, including TET2, DNMT3A, ASXL1, and PPM1D. It does not identify the citing paper’s AD-enriched genes or establish that all of them are implicated in CH. The separate cancer citation (33) is not assessed here. (LLM, conf=0.83)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    The dN/dS ratio test confirmed that somatic variants in Alzheimer's disease brains experienced stronger positive selection than somatic variants in control brains, with greater than a 350% increase in positively selected cells.

    “Direct examination of the signal of positive selection on these variants using the dN/dS ratio test36 confirmed that somatic variants in AD brains experienced stronger positive selection than somatic variants in control brains (Figure 2H-I), with greater than a 350% increase in positively selected cells (Figure 2J).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [36] Martincorena I, Raine KM, Gerstung M, Dawson KJ, Haase K, Van Loo P, Davies H, Stratton MR, and Campbell PJ (2017). Universal Patterns of Selection in Cancer and Somatic Tissues. Cell 171, 1029– 1041 e1021. 10.1016/j.cell.2017.09.042. 29056346 PMC5720395 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Large-scale genetic association studies have identified Alzheimer's disease risk variants in a growing list of microglia-related genes.

    “The importance of microglia in Alzheimer’s disease (AD) pathogenesis has been demonstrated by large-scale genetic association studies, which have identified AD risk variants in a growing list of microglia-related genes1–4.”
    Quote in the paperNot a substantial claim3 of 4 sources checked
    • cites [1] Kunkle BW, Grenier-Boley B, Sims R, Bis JC, Damotte V, Naj AC, Boland A, Vronskaya M, van der Lee SJ, Amlie-Wolf A, (2019). Genetic meta-analysis of diagnosed Alzheimer’s disease identifies new risk loci and implicates Abeta, tau, immunity and lipid processing. Nat Genet 51, 414– 430. 10.1038/s41588-019-0358-2. 30820047 PMC6463297 source
    • cites [2] Jansen IE, Savage JE, Watanabe K, Bryois J, Williams DM, Steinberg S, Sealock J, Karlsson IK, Hagg S, Athanasiu L, (2019). Genome-wide metaanalysis identifies new loci and functional pathways influencing Alzheimer’s disease risk. Nat Genet 51, 404– 413. 10.1038/s41588-018-0311-9. 30617256 PMC6836675 source
    • cites [3] Hardy J, and Escott-Price V (2019). Genes, pathways and risk prediction in Alzheimer’s disease. Hum Mol Genet 28, R235– R240. 10.1093/hmg/ddz163. 31332445 source
    • cites [4] Bellenguez C, Kucukali F, Jansen IE, Kleineidam L, Moreno-Grau S, Amin N, Naj AC, Campos-Martin R, Grenier-Boley B, Andrade V, (2022). New insights into the genetic etiology of Alzheimer’s disease and related dementias. Nat Genet 54, 412– 436. 10.1038/s41588-022-01024-z. 35379992 PMC9005347 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Once abnormally reactive in Alzheimer's disease, microglia can promote synaptic and neuronal loss while exacerbating tau proteinopathy.

    “Once abnormally reactive in AD, microglia can promote synaptic and neuronal loss while exacerbating tau proteinopathy11,12.”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [11] Hickman S, Izzy S, Sen P, Morsett L, and El Khoury J (2018). Microglia in neurodegeneration. Nat Neurosci 21, 1359– 1369. 10.1038/s41593-018-0242-x. 30258234 PMC6817969 source
    • cites [12] Bohlen CJ, Friedman BA, Dejanovic B, and Sheng M (2019). Microglia in Brain Development, Homeostasis, and Neurodegeneration. Annu Rev Genet 53, 263– 288. 10.1146/annurev-genet-112618-043515. 31518519 source

  • May combine two claims.

    Findingown data

    Most Alzheimer's disease somatic variants in ASXL1 were nonsense variants broadly distributed across the encoded protein, including two recurrent alleles observed in AD patients, similar to ASXL1 variants in clonal hematopoiesis events of blood.

    “Most AD somatic variants in ASXL1 were nonsense variants broadly distributed across the encoded protein, including two recurrent alleles observed in AD patients, similar to what is seen in ASXL1 variants in CH events of blood34 (Figure S2D).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [34] Zink F, Stacey SN, Norddahl GL, Frigge ML, Magnusson OT, Jonsdottir I, Thorgeirsson TE, Sigurdsson A, Gudjonsson SA, Gudmundsson J, (2017). Clonal hematopoiesis, with and without candidate driver mutations, is common in the elderly. Blood 130, 742– 752. 10.1182/blood-2017-02-769869. 28483762 PMC5553576 source

      Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    Alzheimer's disease patients showed missense variants in TET2 that clustered in its critical oxygenase domains, a similar mutational pattern to that seen in clonal hematopoiesis, which was not observed in aged controls.

    “AD patients also showed missense variants in TET2 that clustered in its critical oxygenase domains (Figure 2F), a similar mutational pattern to that seen in CH34 (Figure S2D) that was not observed in our aged controls.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [34] Zink F, Stacey SN, Norddahl GL, Frigge ML, Magnusson OT, Jonsdottir I, Thorgeirsson TE, Sigurdsson A, Gudjonsson SA, Gudmundsson J, (2017). Clonal hematopoiesis, with and without candidate driver mutations, is common in the elderly. Blood 130, 742– 752. 10.1182/blood-2017-02-769869. 28483762 PMC5553576 source

      The ‘Association of driver mutations with clonal hematopoiesis’ section reports that TET2 mutations are significantly associated with CH. Our copy does not describe TET2 missense variants clustering in oxygenase domains, so it does not establish the specific mutational-pattern similarity for which this source is cited. The AD patients and aged controls are the citing authors’ own observations, not findings attributed to this source. (LLM, conf=0.81)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    A few small studies of primary brain samples have hinted that cancer-associated variants and clonal overgrowth might occur in brain microglia.

    “A few small studies of primary brain samples have hinted that cancer-associated variants and clonal overgrowth might occur in brain microglia, which proliferate throughout life25,27,28.”
    Quote in the paperSubstantial claim3 of 3 sources checked
    • cites [25] Bouzid H, Belk JA, Jan M, Qi Y, Sarnowski C, Wirth S, Ma L, Chrostek MR, Ahmad H, Nachun D, (2023). Clonal hematopoiesis is associated with protection from Alzheimer’s disease. Nat Med. 10.1038/s41591-023-02397-2. PMC10353941 37322115 source
    • cites [27] Keogh MJ, Wei W, Aryaman J, Walker L, van den Ameele J, Coxhead J, Wilson I, Bashton M, Beck J, West J, (2018). High prevalence of focal and multi-focal somatic genetic variants in the human brain. Nat Commun 9, 4257. 10.1038/s41467-018-06331-w. 30323172 PMC6189186 source

      The Results and Discussion sections on multiple regional mutations and brains with more than one mutation report cancer-associated variants in human brain samples and discuss possible clonal origins or subclone evolution. They do not identify the affected brain cells as microglia or show clonal overgrowth of microglia; some variants are instead attributed to blood-cell lineages. The paper also does not establish that microglia proliferate throughout life. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
    • cites [28] Vicario R, Fragkogianni S, Weber L, Lazarov T, Hu Y, Hayashi SY, Craddock BP, Socci ND, Alberdi A, Baako A, (2024). A microglia clonal inflammatory disorder in Alzheimer’s Disease. eLife. 10.7554/eLife.96519.2. PMC11908784 40085681 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Clonal hematopoiesis increases in prevalence with age and is associated with increased risk of hematologic malignancies and cardiovascular disease, likely through inflammatory effects of mutant cells.

    “Clonal expansion of blood cells, called clonal hematopoiesis (CH), increases in prevalence with age and is associated with increased risk of hematologic malignancies, but also unexpectedly with increased risk of cardiovascular disease20,21, likely through inflammatory effects of mutant cells22.”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites [20] Genovese G, Kahler AK, Handsaker RE, Lindberg J, Rose SA, Bakhoum SF, Chambert K, Mick E, Neale BM, Fromer M, (2014). Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med 371, 2477– 2487. 10.1056/NEJMoa1409405. 25426838 PMC4290021 source

      The Results section on advancing age reports that clonal hematopoiesis was more common in older participants, and the Results section on subsequent hematologic cancer reports an increased cancer risk. Our copy does not report an association with cardiovascular disease; its Discussion mentions cardiovascular disease only as an analogy for preventive biomarkers. The proposed inflammatory mechanism has a separate citation in the citing sentence and is not attributed to this source. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [21] Jaiswal S, Natarajan P, Silver AJ, Gibson CJ, Bick AG, Shvartz E, McConkey M, Gupta N, Gabriel S, Ardissino D, (2017). Clonal Hematopoiesis and Risk of Atherosclerotic Cardiovascular Disease. N Engl J Med 377, 111– 121. 10.1056/NEJMoa1701719. 28636844 PMC6717509 source
    • cites [22] Avagyan S, Henninger JE, Mannherz WP, Mistry M, Yoon J, Yang S, Weber MC, Moore JL, and Zon LI (2021). Resistance to inflammation underlies enhanced fitness in clonal hematopoiesis. Science 374, 768– 772. 10.1126/science.aba9304. 34735227 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The normal iPSC line N-2.12-D-1-1 served as the parental line for all CRISPR/Cas9-edited lines used in this study.

    “The previously described normal induced pluripotent stem cell (iPSC) line N-2.12-D-1-1 served as the parental line of all CRISPR/Cas9-edited lines used in this study106.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [106] Kotini AG, Chang CJ, Boussaad I, Delrow JJ, Dolezal EK, Nagulapally AB, Perna F, Fishbein GA, Klimek VM, Hawkins RD, (2015). Functional analysis of a chromosomal deletion associated with myelodysplastic syndromes using isogenic human induced pluripotent stem cells. Nat Biotechnol 33, 646– 655. 10.1038/nbt.3178. 25798938 PMC4464949 source

  • May combine two claims.

    Findingown data

    Somatic variants in Alzheimer's disease brains showed significantly higher variant allele fractions than variants in control brains, with the effect more pronounced in tumor suppressor genes, clonal hematopoiesis-related genes, or AD-enriched genes.

    “Somatic variants in AD brains also showed significantly higher VAFs than did variants in control brains, with the effect being more pronounced in TSG, CH-related genes, or AD-enriched genes (Figure 2G and S2F).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    More frameshift somatic insertions and deletions were observed in Alzheimer's disease brains (5 in AD versus 2 in control), though this enrichment did not reach significance.

    “Besides sSNVs, we also observed more frameshift sIndels in AD brains (5 in AD versus 2 in control; Table S3), though this enrichment did not reach significance in this small sample size.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Five genes-TET2, ASXL1, KMT2D, ATRX, and CBL-harbored nominally more somatic variants in Alzheimer's disease brains than controls (p < 0.05, one-tailed proportion test), though these individual gene enrichments were not significant after multiple hypothesis testing correction across 149 assayed genes.

    “Five genes—TET2, ASXL1, KMT2D, ATRX, and CBL—harbored nominally more somatic variants in AD brains than controls (Figure 2E; p < 0.05, one-tailed proportion test), though these individual gene enrichments were not significant after multiple hypothesis testing correction across 149 assayed genes.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    A greater somatic single-nucleotide variant burden in Alzheimer's disease was observed for tumor suppressor genes but not for oncogenes.

    “When we divided cancer driver genes into (proto-)oncogenes and tumor suppressor genes (TSGs), we observed a greater sSNV burden in AD for TSGs but not for oncogenes (Figure 2D).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Another 4-9% of the sorted nuclei were classified as CNS-associated macrophages, while the remaining cells represented various other brain cell types.

    “Interestingly, another 4-9% of the nuclei were classified as CNS-associated macrophages (CAMs; Figures 3A and S4B), a recently identified class of brain-resident macrophages with high expression of MS4A7 and MRC139, while the remaining cells represented various other brain cell types.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [39] Masuda T, Sankowski R, Staszewski O, and Prinz M (2020). Microglia Heterogeneity in the Single-Cell Era. Cell Rep 30, 1271– 1281. 10.1016/j.celrep.2020.01.010. 32023447 source

  • May combine two claims.

    Findingown data

    The mutational spectrum of somatic single-nucleotide variants is consistent with the cell division/mitotic clock signature SBS1 (cosine similarity 0.92), suggesting that these variants predominantly occurred during cell division.

    “The mutational spectrum of sSNVs is consistent with the cell division/mitotic clock signature SBS1 (Figure S2A; cosine similarity 0.92), suggesting that these variants predominantly occurred during cell division.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Age was an independent factor positively correlated with somatic single-nucleotide variant burden (p = 0.002) and the proportion of somatic single-nucleotide variant carriers.

    “we also found age was an independent factor positively correlated with sSNV burden (Figure 2C; p = 0.002, linear regression) and the proportion of sSNV carriers (Figure S2C)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    All 10 Alzheimer's disease variants enriched in CSF1R+ cells were also detected in blood, with a trend toward a positive correlation between variant allele fractions in brain CSF1R+ nuclei and blood (p = 0.052).

    “Analysis of matched blood DNA showed that all 10 AD variants enriched in CSF1R+ cells were also detected in blood, with a trend toward a positive correlation between VAFs in brain CSF1R+ nuclei and blood (Figures 3F and S4F; p = 0.052, Pearson correlation).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    UMAP visualization of the integrated object showed no visible clustering by sample ID or individual ID, consistent with successful integration.

    “UMAP visualization of the integrated object showed no visible clustering by sample ID or individual ID, consistent with successful integration (Figure S7).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    All 7 profiled Alzheimer's disease variants were globally present in CSF1R+ cells sorted from four distinct cerebral cortex regions and the cerebellum, but absent in neuronal populations across all brain regions.

    “Regional analysis of these somatic CHIP variants showed that all 7 profiled AD variants were globally present in CSF1R+ cells sorted from four distinct cerebral cortex regions and the cerebellum, but absent in neuronal populations across all brain regions (Figures 3E and S4E).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    75-77% of CSF1R+ sorted nuclei from AD and control brains are transcriptionally microglia, with an additional 4-9% being CNS-associated macrophages (CAM).

    “Our snRNAseq result confirmed that 75-77% of CSF1R+ sorted nuclei from the AD and control brains are transcriptionally microglia, with an additional 4-9% being CNS-associated macrophages (CAM; Figure 3A).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Minimal blood contamination was observed in the sorted CSF1R+ population, with only 1% monocytes and absence of other major blood cell types.

    “We also observed minimal blood contamination in the sorted CSF1R+ population, with only 1% monocytes and the absence of other major blood cell types including red blood cells, T-cells, and B-cells (Figures 3A and S4B).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Clonal somatic driver variants in microglia-like brain macrophages are enriched in Alzheimer's disease and potentially promote neuroinflammation and neurodegeneration.

    “These findings suggest that clonal somatic driver variants in MLBMs are enriched in AD, potentially promoting neuroinflammation and neurodegeneration.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Clonal expansion driven by somatic variants in genes regulating cell proliferation and fitness is considered the major cause of cancer, but has also been reported in various non-cancerous cells, often in the absence of visible pathology.

    “Clonal expansion, driven by somatic variants in genes regulating cell proliferation and fitness, is considered the major cause of cancer, but has also been recently reported in various non-cancerous cells, often in the absence of visible pathology19.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [19] Martincorena I, and Campbell PJ (2015). Somatic mutation in cancer and normal cells. Science 349, 1483– 1489. 10.1126/science.aab4082. 26404825 source

  • May combine two claims.

    Study design and conductown data

    Thirty-five of 39 tested somatic single-nucleotide variants and 8 of 10 somatic insertions and deletions successfully validated in newly extracted DNA samples, confirming high accuracy of somatic variant calling even for variants with variant allele fractions as low as 0.1%.

    “Thirty-five of 39 (90%) tested sSNVs and 8 of 10 (80%) sIndels successfully validated in newly extracted DNA samples from the corresponding PFC samples, confirming the high accuracy of our somatic variant calling strategy even for those with VAFs as low as 0.1% (Figure S1E–G).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    CHIP and mosaic chromosomal alterations represent two primary categories of somatic variants that propel the progression of clonal hematopoiesis.

    “Clonal hematopoiesis of indeterminate potential (CHIP, denoting somatic single-nucleotide variants (sSNVs) and somatic insertions and deletions (sIndels) in cancer driver genes) and mosaic chromosomal alterations (mCAs, denoting chromosome-level somatic copy number variants (sCNVs) enriched in CH) represent two primary categories of somatic variants that propel the progression of CH23.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] Ahmad H, Jahn N, and Jaiswal S (2023). Clonal Hematopoiesis and Its Impact on Human Health. Annu Rev Med 74, 249– 260. 10.1146/annurev-med-042921-112347. 36450282 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microglia originate from erythro-myeloid progenitors in the yolk sac during early embryogenesis.

    “As the primary macrophages in the central nervous system (CNS), microglia originate from erythro-myeloid progenitors in the yolk sac during early embryogenesis5.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [5] Saijo K, and Glass CK (2011). Microglial cell origin and phenotypes in health and disease. Nat Rev Immunol 11, 775– 787. 10.1038/nri3086. 22025055 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Ten trials of 1000 permutations were completed to ensure the robustness of p-values.

    “Ten trials of 1000 permutations were completed to ensure the robustness of p-values.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    As a final step to ensure that scType'd microglia-CAMs were highly similar to their corresponding pre-annotated counterparts, only scType'd microglia-CAMs in clusters where over 50% of cells were pre-annotated microglia-CAMs were retained as high-confidence.

    “As a final step to ensure that scType’d microglia-CAMs were highly similar to their corresponding pre-annotated counterparts, pre-annotated and scType’d microglia-CAMs derived from the same donor were merged into a single Seurat object and processed, normalized, and clustered using the Louvain algorithm. Clusters in which over 50% of cells were pre-annotated microglia-CAMs were identified, and”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    A candidate was considered validated as somatic variant if the read fraction of the mutant allele was more than three times as high as the fractions of the other two error alleles in all three amplicons, or if the read fraction of the mutant allele in the corresponding brain sample was significantly higher than the fraction in an unrelated negative control brain sample for all three amplicons.

    “A candidate was considered validated as somatic variant (Figure S1E–G) if 1) the read fraction of the mutant allele was more than three times as high as the fractions of the other two error alleles in all three amplicons (somatic-I); or 2) the read fraction of the mutant allele in the corresponding brain sample was significantly higher than the fraction in an unrelated negative control brain”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    For per-individual sample analysis, a Benjamini-Hochberg correction was used to control for multiple-hypothesis testing.

    “One-tailed Fisher’s exact tests were used to test for cell type and microglial state enrichment analyses and were performed both with all samples pooled to maximize power as well as per individual sample. For per-individual sample analysis, a Benjamini-Hochberg correction was used to control for multiple-hypothesis testing.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    A mixing experiment was performed to benchmark the performance of the designed panel and variant calling pipeline.

    “A mixing experiment was performed to benchmark the performance of the designed panel and variant calling pipeline.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    For sSNV calling, candidates were required to have a likelihood ≥0.5, ≥100 total reads, and ≥4 mutant-supporting reads.

    “For sSNV, MosaicHunter calculated the likelihood of the presence of a mutant allele, and only the candidates with a 0.5 or higher likelihood, 100 or more total reads, and 4 or more mutant-supporting reads were considered.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    An mCA was considered validated if its macrophage-to-neuron ratio significantly deviated from the window of 0.99-1.01, and such deviation could only be seen in the donors in which the mCA was originally called.

    “An mCA was considered validated if its macrophage-to-neuron ratio significantly deviated from the window of 0.99-1.01, and such deviation could only be seen in the donors in which the mCA was originally called.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The study achieved comparable depth and coverage between AD and control samples, with more than 1000X average depth and more than 80% coverage of targeted regions at >500X for consensus reads.

    “As a result, we achieved comparable depth and coverage between the AD and control samples, with more than 1000X average depth and more than 80% coverage of the targeted regions at >500X for consensus reads (Table S2 and Figure S1A).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The diagnosis of AD was based on the consensus conclusion from all postmortem data generated by neurologists with expertise in dementia and neurodegeneration.

    “The diagnosis of AD was based on the consensus conclusion from all postmortem data generated by neurologists with expertise in dementia and neurodegeneration.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductown data

    Raw mCA calls were filtered to increase specificity using six criteria: Bonferroni adjusted p < 0.05; <25% ambiguous cells; median expression of putative mCA-carrying cells > or < 1.96 SD of putative normal cells; no negative control regions showed larger difference; expression of putative normal cells within 1.96 SD of baseline; and same mCA not called in a different cell-type from the same个体.

    “In brief, a putative mCA was retained if it met the following criteria: 1) Bonferonni adjusted p-value < 0.05; 2) <25% ambiguous cells (cells with a posterior probability > 0.25 and < 0.75); 3) median expression of putative mCA-carrying cells is > or < 1.96 standard deviations of putative normal cells of the same type for amplifications or deletions, respectively; 4) no negative control regions (i”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The genotype of the isolated iPSC clones was verified by Sanger sequencing.

    “Multiple independent clones with each variant were isolated and their genotype were verified by Sanger sequencing.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    6 samples were filtered out due to not meeting at least one of the following sample-specific metrics: microglia-CAM AUC > 0.9, ExN AUC > 0.9, fraction of pre-annotated ExN typed by scType as microglia-CAM < 0.1, and total number of pre-annotated and scType'd microglia-CAMs > 50.

    “In addition to filtering of individual cells, 6 samples were filtered out due to not meeting at least one of the following sample-specific metrics: 1) microglia-CAM AUC > 0.9, 2) ExN AUC > 0.9, 3) fraction of pre-annotated ExN typed by scType as microglia-CAM < 0.1, and 4) total number of pre-annotated and scType’d microglia-CAMs > 50.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Postmortem samples in all studies were collected and de-identified following the protocol of the corresponding Institutional Review Board with informed consent.

    “Postmortem samples in all studies were collected and de-identified following the protocol of the corresponding Institutional Review Board with informed consent.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    Candidates were excluded as germline variants if they had ≥30% variant allele fraction, did not significantly deviate from binomial distribution for heterozygous variants (p ≥ 0.05), or were present in polymorphism databases or had ≥0.01% population allele frequency in gnomAD.

    “We further excluded candidates as germline variants if i) they have a 30% or higher variant allele fraction; 2) the counts of mutant-supporting and total reads do not significantly deviate from the binomial distribution for heterozygous variants (p ≥ 0.05); 3) they are present in the polymorphism databases (dbSNP73, the 1000 Genomes Project74, the Exome Sequencing Project75, and the Exome Aggregat”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    Two pipelines were developed to balance sensitivity and specificity: a stringent pipeline keeping variants detected in one sample and absent in others, and a sensitive pipeline additionally allowing variants exclusively present or enriched in AD or control group (two-sample Z-test p < 0.05).

    “To balance the sensitivity and specificity of our sSNV and sIndel detection, we developed two different pipelines when considering the recurrent presence across multiple individuals. The “stringent” pipeline only kept the variants that were detected in one sample and were completely absent in any other samples, whereas the “sensitive” pipeline additionally allowed the variants that were exclusiv”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    For microglia-CAMs, putative mCAs were additionally filtered if the number of scType'd non-ambiguous cells was ≤1.5 times the number of pre-annotated non-ambiguous cells for both altered and wild-type cells.

    “For microglia-CAMs, putative mCAs were additionally filtered if the number of scType’d non-ambiguous cells (posterior probability < 0.25 or > 0.75) was ≤ 1.5 times the number of pre-annotated non-ambiguous cells for both altered and wild-type cells.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    A subset of AD patients (n=121) with age distribution matched to healthy controls was selected using MatchIt, and the resulting subset closely mirrored the age distribution of the control group.

    “To further control for the potential impact of age on our burden analysis, we selected a subset of AD patients (n = 121) with an age distribution matched to that of the healthy controls. This was done using the R package MatchIt (v4.5.5)90, with the function matchit(Cogdx ~ Age + Sex + PMI, distance = “glm”, method = “optimal”), where Cogdx indicates clinical status (AD or control), and PMI denot”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Only consensus reads supported by two or more reads in both strands were kept.

    “We only kept the consensus reads that were supported by two or more reads in both strands.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    High-quality cells were retained for scRNAseq analysis based on percent mitochondrial gene expression < 20% and 2000 < nFeature_RNA < 7000.

    “High-quality cells were retained based on meeting the following quality-control metrics: 1) percent mitochondrial gene expression < 20% and 2) 2000 < nFeature_RNA < 7000.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    All CRISPR/Cas9-edited iPSC clones used in this study were confirmed to have a normal karyotype.

    “All clones were confirmed to have a normal karyotype.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    The linear regression models adjusted for potential effects of age and sex.

    “In addition, we adjusted for the potential effects of these variables, including age and sex, in the linear regression models described in the following sections.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The purity of microglia-enriched brain macrophage sorting was slightly reduced in some AD samples, with a marginally higher proportion of non-microglia/CAM nuclei.

    “To note, the purity of our microglia-enriched brain macrophage sorting was slightly reduced in some AD samples, exhibiting a marginally higher proportion of non-microglia/CAM nuclei as determined by snRNAseq (Fig. S3A).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Multiple independent iPSC clones with each variant were isolated.

    “Multiple independent clones with each variant were isolated and their genotype were verified by Sanger sequencing.”
    Quote in the paperNot a substantial claim
53 of 53

  • The authors state this as an assumption.

    Non-somatic drivers of Alzheimer’s disease—such as amyloid, tau, and APOE4—contribute to an inflammatory brain environment that exerts lifelong effects on clonal competition within the microglial population.

    “Our data suggest a model (Figure 7) in which non-somatic drivers of AD—such as amyloid, tau, and APOE4—contribute to an inflammatory brain environment that exerts lifelong effects on clonal competition within the microglial population.”

  • The authors state this as an assumption.

    Age-related disruption of the blood-brain barrier promotes clonal evolution by introducing microglia-like brain macrophages into the microglia pool.

    “Age-related disruption of the BBB may further promote clonal evolution by introducing MLBMs into the microglia pool.”

  • The authors state this as an assumption.

    Chronic inflammation induces the proliferation of microglia, in which positive selection favors cells carrying cancer driver variants that provide fitness advantages over those without.

    “Chronic inflammation induces the proliferation of microglia, in which positive selection favors cells carrying cancer driver variants that provide fitness advantages over those without.”

  • The authors state this as an assumption.

    Operation of clonal selection over decades would lead to the high fraction of mutant cells observed.

    “Operation of this clonal selection over decades would be expected to lead to the high fraction of mutant cells that we observe”

  • The authors state this as an assumption.

    As mutant clones accumulate additional variants over time, microglia carrying them become increasingly abnormal in their transcriptomic patterns and their ability to sustain versus damage neighboring neurons.

    “as mutant clones accumulate additional variants over time, microglia carrying them would be expected to become increasingly abnormal in their transcriptomic patterns and their ability to sustain—versus damage—neighboring neurons.”

  • The authors state this as an assumption.

    Almost all sSNVs are heterozygous in carrier cells.

    “given that almost all the sSNVs should be heterozygous in carrier cells”

  • The authors state this as an assumption.

    The number of profiled cells in panel sequencing is consistent for each brain.

    “Assuming a consistent number of profiled cells in panel sequencing for each brain”
7 of 7

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • DNMT3A editing template has two different variant labels

    Minor

    The methods say the editing procedure introduced DNMT3A R882H, but the description of that procedure labels the mutant donor plasmid G12D. The text does not reconcile the two variant identities.

1 of 1

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Authors
Huang, A. Y., Zhou, Z., Talukdar, M., Enyenihi, L., Miller, M. B., Chhouk, B., Rosen, I., Zheng, M., Zhou, M., Yang, A., Stronge, E., Durens, M., Nguyen, M., Choi, J., Zhao, B., Khoshkhoo, S., Kim, J., Andersen, R., An, Z., Cheng, Y., Ganz, J., Mekerishvili, L., Travaglini, K. J., Gabitto, M. I., Hodge, R. D., Kaplan, E. S., Belk, J. A., Landau, D., Lein, E. S., De Jager, P. L., Bennett, D. A., Marro, S. G., Papapetrou, E. P., Lee, E. A., & Walsh, C. A.
Affiliations
  • Broad Institute (US)
  • Boston Children's Hospital (US)
  • Icahn School of Medicine at Mount Sinai (US)
  • Tisch Cancer Institute
  • New York Genome Center (US)
  • Allen Institute for Brain Science (US)
  • Stanford University (US)
  • Rush University Medical Center (US)
Corresponding author
  • Huang, August Yue
  • Marro, Samuele G.
  • Papapetrou, Eirini P.
  • Lee, Eunjung Alice
  • Walsh, Christopher A.
ORCID
Venue
Cell
Publisher
Elsevier BV
Published
2026-06
Volume
189
Issue
12
Pages
3719-3735.e24
ISSN
0092-8674 (ISSN-L)
DOI
10.1016/j.cell.2026.03.040
PMC
PMC13107304
OpenAlex
W7155086195
Cited by
15
References
103
Citations by year
2026: 15
FWCI
35.51
Type
journal article
Language
en
Open access
Licence
Locations
  • Cell (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion)
Funders
  • Boston Children's Hospital The Manton Center for Orphan Disease Research
  • Edward P Evans Foundation
  • Howard Hughes Medical Institute
  • National Institutes of Health
  • Alzheimer's Association
  • Allen Family Philanthropies
  • Harvard Stem Cell Institute
  • BrightFocus Foundation
  • US Department of Defense
  • Doris Duke Charitable Foundation Inc
Topics
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
  • Health Sciences › Medicine › Physiology › Alzheimer's disease research and treatments
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
Keywords
Alzheimer's disease, somatic mutations, clonal hematopoiesis, disease-associated microglia, neuroinflammation, cancer driver genes, TET2, ASXL1, DNMT3A, hematopoietic origin, microglia-like cells, induced pluripotent stem cell-derived microglia, clonal expansion
MeSH terms
Clonal Hematopoiesis, Alzheimer Disease, Animals, Brain, Humans, Inflammation, Macrophages, Microglia, Cell Proliferation, DNA Copy Number Variations

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-19.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-19.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review