CoChat
Sign Up
PreprintCited by 2,225Neuroinflammation and Neurodegeneration MechanismsSingle-cell and spatial transcriptomicsMultiple Sclerosis Research Studies

Complex cell-state changes revealed by single cell RNA sequencing of 76,149 microglia throughout the mouse lifespan and in the injured brain

Open access
Hammond, Timothy R., Dufort, Connor, Dissing-Olesen, Lasse, Giera, Stefanie, Young, Adam, Wysoker, Alec, Walker, Alec J., Segel, Michael +7 more·2018·Immunity
DOI 10.1101/406140PMC PMC6655561
Open this paper in CoChat — free accountOpen PDF

Abstract

Microglia, the resident immune cells of the brain, rapidly change states in response to their environment, but we lack molecular and functional signatures of different microglial populations. In this study, we analyzed the RNA expression patterns of more than 76,000 individual microglia during development, old age and after brain injury. Analysis uncovered at least nine transcriptionally distinct microglial states, which expressed unique sets of genes and were localized in the brain using specific markers. The greatest microglial heterogeneity was found at young ages; however, several states - including chemokine-enriched inflammatory microglia - persisted throughout the lifespan or increased in the aged brain. Multiple reactive microglial subtypes were also found following demyelinating injury in mice, at least one of which was also found in human MS lesions. These unique microglia signatures can be used to better understand microglia function and to identify and manipulate specific subpopulations in health and disease.

AI snapshot

Written from the paper’s full text.

TL;DR

Single-cell RNA sequencing of 76,149 mouse microglia across development, aging, and demyelinating injury identifies at least nine transcriptionally distinct microglial states, with the greatest diversity early in development and state shifts in aging and injury. The study defines state-specific markers, validates several populations by in situ hybridization, and finds a Ccl4+ microglial subset in mouse lesions and human multiple sclerosis lesions.

Key findings

  • Across all ages and conditions, clustering revealed nine unique microglial states, with cluster sizes ranging from 0.2% to 24% of all microglia; canonical microglial genes were not uniformly expressed, as only C1qa, Fcrls, and Trem2 were expressed in all clusters.
  • Microglial diversity was greatest at E14.5 and P5 and lower at P30 and P100; aging and injury redistributed states, including increases in Clusters 8 and 9. In aged P540 versus P100 mice, there were 2–4 times more microglia in aging-enriched Clusters OA2 and OA3, with OA2 expressing inflammatory genes such as Ccl4, Ccl3, and Il1b and OA3 expressing interferon-response genes including Ifitm3, Rtp4, and Oasl2.
  • In LPC-induced demyelination at 7 days post-lesion, injury-responsive Cluster IR2 was almost entirely composed of lesion microglia and subclustered into IR2.1–2.4; IR2.2 upregulated Cxcl10, and IR2.2/IR2.3 upregulated Ccl4. IRM shared a 12-gene core signature with disease-associated microglia (DAM), including Spp1, Lpl, and Apoe, but also had unique interferon-response genes.
  • Sex had almost no effect on microglial diversity or cluster proportions across E14.5, P4/P5, and P100; the only observed difference was female enrichment of the small Cluster Sx6 (~0.5% of microglia, observed only at P4/P5, enriched for Cd74, Ccl24, and Arg1). In human MS lesions, approximately 30% of microglia in the lesion were Ccl4+.

Methods

  • Microglia were FACS-purified from C57BL/6J mice at E14.5, P4/P5, P30, P100, and P540, plus adult P100 mice after focal LPC-induced demyelination at 7 days post-lesion; at least three to four biological replicates per age were collected, totaling 76,149 sequenced microglia from 41 animals.
  • Cells were processed with a droplet-based RNA-seq approach (10× Genomics Chromium), sequenced to ~40,000–60,000 reads/cell, and analyzed using an independent component analysis (ICA)-based clustering pipeline, including subclustering and canonical correlation analysis (CCA) for comparison with published DAM data.
  • Findings were validated with single-molecule fluorescent in situ hybridization (smFISH), chromogenic in situ hybridization, and immunohistochemistry in mouse tissue, and with human MS tissue from the UK MS Tissue Bank; statistical analyses included ANOVA with Tukey’s post hoc tests, t-tests, and gene-marker thresholds of >1.5 fold change and P < 1E-4 with Benjamini–Hochberg correction.

Limitations

  • Stated: further analysis is needed to establish whether the identified states are transient or represent terminal differentiation; lineage tracing is needed to determine the origin of Cluster 6 microglia and whether they give rise to mature microglia; manipulation of aged inflammatory microglia is needed to test whether they contribute to age-related brain inflammation; the small female-enriched Cluster Sx6 requires further validation; and whether other mouse lesion-enriched genes translate to human microglia needs confirmation.
  • Evident: the main dataset is mouse-based, and the injury analysis focuses on a single model (LPC-induced demyelination) at one time point (7 days post-lesion); human MS validation was limited to five control/patient samples and centered on Ccl4/Tmem119 staining.

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

Integrity signals

Registry

Crossref · 10.1101/406140

Cites

84 of 84 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 32 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Cluster 6 microglia also upregulate other Ms4a family members, including Ms4a6c, Ms4a6b, Ms4a6d, Tmem176a, and Tmem176b.

    “In addition to Ms4a7, Cluster 6 microglia also expressed increased transcripts of other Ms4a family members including Ms4a6c, Ms4a6b, Ms4a6d, Tmem176a, and Tmem176b (Fig 2a).”
    Quote in the paperNot a substantial claim
  • Findingown data

    Cluster 6 microglia upregulate Ccr1.

    “Cluster 6 microglia also upregulated chemokine (C-C motif) receptor 1 (Ccr1)”
    Quote in the paperNot a substantial claim
  • Findingown data

    Cluster 3 marker genes Fabp5, Mif, Ldha, and Tpi1 are also upregulated, though less strongly, in P4/P5 Cluster 4 Spp1+ microglia.

    “The top genes, including Fabp5, Mif, Ldha, and Tpi1, in Cluster 3 microglia were also upregulated in the Cluster 4 Spp1+ microglia found at P4/P5, albeit to a lower extent (Fig 4c, Fig S4b).”
    Quote in the paperNot a substantial claim
  • Findingown data

    At E14.5, about 10-15% of all microglia belong to Cluster 3.

    “At E14.5, approximately 1015% of all microglia belonged to Cluster 3 (Fig 4b).”
    Quote in the paperNot a substantial claim
  • Findingown data

    Ms4a6c is upregulated in microglia from injured tissue.

    “Ms4a6c was also upregulated in microglia from the injured tissue (Fig 2c, Fig S4d)”
    Quote in the paperNot a substantial claim
  • May combine two claims. 65
  • Classified as probably not a substantial claim. 31

  • May combine two claims.

    Findingown data

    Ms4a7+ microglia were sparsely distributed throughout the embryonic brain, and Ms4a7+ macrophages were present at brain borders.

    “Ms4a7+ microglia were sparsely distributed throughout the embryonic brain (Fig 2d,f), with a population of Ms4a7+ macrophages at the brain borders (not shown).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Fabp5+Mif+ microglia were distributed throughout the E14.5 brain and often formed small clusters.

    “Fabp5+Mif+ microglia were distributed throughout the E14.5 brain and often formed small clusters (Fig 4d).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Proliferative cluster markers were comparable in magnitude at E14.5 and P4/P5 and highly specific to those clusters.

    “Markers for each of the proliferative populations were comparable in magnitude at E14.5 and P4/P5 and were highly specific to the proliferative clusters (Fig 4i, Fig S4a).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Cluster 2a markers include Mcm6, Pcna, and Rrm2; 2b markers include Ube2c, Birc5, and H2afx; 2c markers include Hist1h2bc, Ccnb2, and Cenpa.

    “The top cluster markers for Cluster 2a were minichromosome maintenance complex component 6 (Mcm6), proliferating nuclear cell antigen (Pcna), and Rrm2. For Cluster 2b, they were Ube2c, baculoviral IAP repeat-containing 5 (Birc5), and H2A histone family, member X (H2afx), and for Cluster 2c they were histone cluster 1, H2bc (Hist1h2bc), cyclin B2 (Ccnb2), and Cenpa.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Cluster 9 consists mainly of microglia from focal white matter injury and shows increased Ifi27l2a expression.

    “Cluster 9 was composed predominantly of microglia from the focal white matter injury and had increased expression of the gene interferon, alpha-inducible protein 27 like protein 2A (Ifi27l2a).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    A major microglial state, Cluster 4, appears at P4/P5 and is barely present at other time points.

    “At P4/P5, we found one major microglial state (Cluster 4) that was barely found at any other time point (Fig 3a,b).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The axon tracts containing ATM later become heavily myelinated, but ATM largely disappear before myelination.

    “The axon tracts where ATM were concentrated will eventually become heavily myelinated, but ATM are largely gone before myelination occurs.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    At E14.5, proliferative clusters 2a-2c together make up about 40% of microglia, and at P5 about 35%.

    “At E14.5, the most populous microglial states were proliferative (Clusters 2a,2b,2c) and comprised a total of approximately 40% of all E14.5 microglia and 35% of P5 microglia (Fig 4g,h).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Some microglial states, including chemokine-enriched inflammatory microglia, persist throughout life or increase in the aged brain.

    “several states – including chemokine-enriched inflammatory microglia – persisted throughout the lifespan or increased in the aged brain.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Cluster 6 microglia are predominantly from E14.5, comprising about 10% of E14.5 microglia and no more than 3% at other ages.

    “Cluster 6 microglia were predominantly from E14.5 versus any other age or condition analyzed, as approximately 10% of all E14.5 microglia (but no more than 3% of microglia at any other age) fall into Cluster 6 (Fig 2b).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Most microglia expressed the canonical genes Fcrls, P2ry12, Cx3cr1, Trem2, and C1qa, but only C1qa, Fcrls, and Trem2 were uniformly expressed across all clusters.

    “Gene expression analysis showed that the canonical microglial genes (Fcrls, P2ry12, Cx3cr1, Trem2, and C1qa) were highly expressed by most of the analyzed cells, but only three (C1qa, Fcrls, Trem2) were uniformly expressed in all clusters (Fig 1e)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Some microglial clusters were dominated by cells from specific ages, while others contained cells from multiple ages.

    “Some clusters were dominated by microglia from specific ages, whereas others contained microglia from several ages”
    Quote in the paperSubstantial claim

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

    Findingown data

    Non-microglial monocytes and macrophages constituted 0-3% of cells and were retained in clustering as a reference.

    “A small percentage (0-3%) of non-microglial monocytes and macrophage populations were also found (Cd45high cells) but were left in the clustering analysis to provide a point of reference for comparison with the various microglial states (Fig 1b,g, Mono/Mac cluster).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Cluster 4 microglia have an amoeboid morphology and upregulate the lysosomal markers Lamp1 and Cd68.

    “These cells, which assume an amoeboid morphology (data not shown), also upregulate the lysosomal markers lysosomal-associated membrane protein 1 (Lamp1) and Cd68 (Table S1, Fig S5e)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    At E14.5, a distinct microglial subpopulation (Cluster 6) uniquely expresses Ms4a7 and transcriptionally resembles brain border macrophages.

    “At E14.5, we identified a distinct subpopulation of microglia that uniquely express Ms4a7 and share a similar transcriptional profile with brain border macrophages (Fig 2, Cluster 6).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The Fabp5+Mif+ microglial population is found almost exclusively in the embryonic brain and is not restricted to a particular brain region.

    “Together, these data show that this unique population of Fabp5+Mif+ microglia was found almost exclusively in the embryonic brain and were not restricted to a particular brain region.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Cluster 8 cells occur at most ages in smaller numbers and peak developmentally at P5.

    “Cluster 8 cells were found at most ages, albeit in smaller numbers, with a developmental peak at P5.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Aging did not add or remove microglial clusters but progressively expanded clusters that are rare in adolescent and adult samples.

    “Aging did not lead to the appearance or disappearance of any clusters, but rather caused a progressive expansion of clusters that typically have few very cells in adolescent and adult samples (P30 and P100, Fig 1d) (Fig 5a–c).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Cluster Sx6 is the smallest microglial cluster (~0.5%), appears only at P4/P5, and is enriched for Cd74, Ccl24, and Arg1.

    “Cluster Sx6 is the smallest identified microglial cluster (~0.5% of microglia), was observed only at P4/P5 (Fig S6c–d, Fig 1d), and was enriched for the genes Cd74, chemokine (C-C motif) ligand 24 (Ccl24), and Arg1 (Fig S6e).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    MC2.1, MC2.4, and MC2.5 represented P2ry12-high, Spp1-high, and Ccl4-high microglia, respectively.

    “These clusters represented P2ry12high (MC2.1, data not shown), Spp1high (MC2.4, Fig S7h), and Ccl4high (MC2.5, data not shown) microglia.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Chromogenic ISH showed a significant increase in Ccl4+Fcrls+ cells in P540 brains, with no significant regional distribution differences.

    “we used chromogenic ISH and found a significant upregulation of doublepositive cells in P540 brains (Fig 5h), however no significant differences in regional distribution were found.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Ccl4+Fcrls+ microglia were distributed among other microglia and were otherwise indistinguishable.

    “Ccl4+Fcrls+ microglia were distributed amongst the other microglia but were otherwise indistinguishable (Fig 5f).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Male microglia uniquely expressed Eif2s3y, and female microglia expressed Xist.

    “microglia from males uniquely expressed the Y chromosome gene Eif2s3y and microglia from females expressed the X inactivation gene Xist (Fig S6f).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Ccl4+ microglia are rare in the healthy developing brain, peak slightly at P5, and increase substantially in P540 brain.

    “There were very few Ccl4+ microglia in the healthy developing brain, with a small developmental peak at P5 but a large increase in the percentage of these cells was found in the P540 brain (Fig 1d).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    A portion of Cluster 3 microglia cluster with proliferative microglia and are enriched for degradation pathways for proliferation mediators.

    “a portion of the Cluster 3 microglia physically cluster with the proliferative microglia (Fig 1b) and were also enriched for degradation pathways for mediators of proliferation (Fig S5d)”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Birc5 is also expressed by proliferative microglia during early brain development.

    “a gene also expressed by proliferative microglia during early brain development (Fig 4g)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Clusters 2 and 3 share some transcriptional overlap, with Mif, Ldha, and Tpi1 also expressed at lower levels in proliferative microglia.

    “The top cluster markers in Clusters 2 and 3 also shared some transcriptional overlap. For example, Mif, Ldha, and Tpi1 were also expressed by proliferative microglia but to a lesser degree (Fig 4c,i).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    At least 9 microglial subpopulations were defined, with unique transcriptional and spatial signatures.

    “We defined at least 9 subpopulations of microglia and provide unique transcriptional and spatial signatures that give us hints about their possible functions.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    These microglia are present in small numbers during development and expand in aging and injury.

    “These microglia are present in small numbers during development and expand in two different contexts: aging and injury.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    A unique microglial state, Cluster 6, was enriched at E14.5 and defined by Ms4a7, Ccr1, and Mrc1.

    “Here, we identified a unique microglial state (Cluster 6) that was enriched at E14.5 and defined by the genes Ms4a7, Ccr1, and Mrc1”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Injury-responsive microglia upregulated other pathways found in developmental subpopulations, including proliferation (Birc5), metabolic activity (Mif), and Ms4a family (Ms4a6c) markers.

    “we also found that injury-responsive microglia upregulated other pathways found in developmental subpopulations including markers of proliferation (Birc5), metabolic activity (Cluster 3 microglia [Mif]), and those belonging to the Ms4a family (Ms4a6c).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Microglia responding to injury re-express developmental markers, but their transcriptional states do not completely overlap.

    “We found that microglia responding to injury do re-express developmental markers, but their transcriptional states do not completely overlap.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Other early developmental microglial states were enriched for metabolism, growth, motility, and proliferation pathways and were scattered throughout the brain.

    “Other key microglial states in early development were enriched for pathways associated with cell metabolism, growth, motility (Cluster 3) and proliferation (Clusters 2a-c) and were scattered throughout the brain.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The only microglia enriched for inflammatory signals are in Cluster 8/Cluster OA2/Cluster IR2.2 and IR2.3 and express Ccl3, Ccl2, Ccl7, Ccl9, Ccl12, Il1b, and Tnf.

    “we found that the only microglia enriched for inflammatory signals are in Cluster 8/Cluster OA2/Cluster IR2.2, IR2.3 and express Ccl3, Ccl2, Ccl7, Ccl9, Ccl12, Il1b and Tnf.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Ccl4 staining was predominantly in Tmem119+ microglia, although a small percentage of Ccl4+/Iba1+ amoeboid immune cells were also observed.

    “Ccl4 staining was found predominantly in Tmem119+ microglia, although a small percentage of Ccl4+/Iba1+ amoeboid immune cells were also observed (data not shown).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Cluster 4 was restricted to unmyelinated axon tracts at P4/P5 and had a highly activated profile.

    “The other microglial subpopulation in the developing brain, Cluster 4, was restricted to unmyelinated axon tracts at P4/P5 and assumed a highly activated profile”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Clusters 3 and 4 have an overlapping transcriptional signature and possible relationship.

    “suggesting an overlapping transcriptional signature and possible relationship between the two states.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Batch and replicate effects were removed from the clustering analysis by excluding independent components that captured them.

    “For both rounds of analysis, independent components that captured batch or replicate effects were removed before clustering analysis (Fig S2d), as previously described (Saunders et al., 2018).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Saunders et al., 2018 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    These data demonstrate that microglia exist in multiple definable states that change over development, aging, and injury.

    “Together, these data demonstrate that microglia exist in multiple definable states that change over the course of development, aging, and injury.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Cluster 6 microglia reside in the embryonic brain parenchyma and bear considerable transcriptional overlap with brain border macrophages.

    “Altogether, these results identify Cluster 6 microglia as a subpopulation that reside in the embryonic brain parenchyma and bears considerable transcriptional overlap with brain border macrophages.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Some microglial states are present across development while others are more transitional.

    “indicating that some microglial states are present across development while others are more transitional (Fig 1c,d).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationown data

    Existing tools and marker definitions may need to be updated.

    “suggesting that existing tools and marker definitions may need to be updated.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    Percoll purification did not alter the microglial subpopulations recovered, with only a slight shift in the relative percentage of one of the nine clusters.

    “To ensure this step did not affect our recovery of microglial subpopulations, we Percoll purified microglia at P5, when we see considerable microglial diversity, and compared these microglia to those isolated without Percoll – finding no change in the subpopulations present and only a slight shift in the relative percentages of microglia in one of the nine clusters (Fig S1).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    These data provide a comprehensive map of microglial state changes and identify genes that define several specific microglial states.

    “these data provide a comprehensive map of these changes and identify genes that define several of these specific microglial states.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The dataset includes at least three to four biological replicates per age, totaling 76,149 microglia from 41 animals.

    “at least three to four biological replicates per age were collected for a total of 76,149 sequenced microglia from 41 total animals (Fig 1b).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    Sequencing depth and median UMI and gene counts were comparable across all ages and conditions.

    “Cells were sequenced to comparable sequencing depths (~40,000–60,000 reads/cell), and had a similar median unique molecular identifier (UMI) count and median gene number in all ages and conditions (Fig S2a).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    smFISH quantification closely matched the sequencing percentages and confirmed the presence of Ms4a7+ microglia in the developing brain.

    “These differences found by smFISH closely mimic the percentages of Ms4a7+ microglia found in our sequencing data and confirm that these cells are present in the developing brain.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    Fcrls was uniformly expressed by all microglia in every cluster and was used as a microglial marker.

    “Microglia were marked by Fcrls, which was uniformly expressed by all microglia in every cluster (Fig 2d).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The widespread distribution of both proliferative and metabolically active microglia suggests that both cell subtypes give rise to mature microglia in the adult brain.

    “The widespread distribution of both proliferative and metabolically active microglia suggests that both cell subtypes give rise to mature microglia in the adult brain, but lineage tracing studies will be needed to track their progression.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Microglial activation is a tailored response involving both generalized and selective transcriptional programs.

    “These results suggest that microglial activation is a tailored response in which microglia activate both generalized and selective transcriptional programs.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S1 (asserted).

    Interpretationreasoning

    This small subpopulation of microglia contributes to age-related brain inflammation.

    “suggests that this small subpopulation of microglia contributes to age-related brain inflammation; however manipulation of these microglia will be essential for testing this hypothesis.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Franceschi et al., 2007 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Human Tmem119 expression is maintained by resident microglia in MS lesions and is not expressed by infiltrating immune cells.

    “since its expression is maintained by resident human microglia in MS lesions and is not expressed by infiltrating immune cells (Zrzavy et al., 2017).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Zrzavy et al., 2017 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    This quantification supports the sequencing data and shows an increase in the number of Cluster OA2 microglia with aging.

    “This quantification supports our sequencing data and shows an increase in the number of Cluster OA2 microglia with aging.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    These effects were in response to demyelination and not a consequence of injection.

    “suggesting that these effects were in response to demyelination and were not a consequence of the injection (Fig 7c).”
    Quote in the paperSubstantial claim

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

    Interpretationreasoning

    Ccl4 staining was perinuclear, most likely in the Golgi.

    “and was perinuclear, most likely in the Golgi, where cytokines are often packaged into vesicles for release.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Disease Associated Microglia were recently identified in the 5xFAD mouse model of Alzheimer's disease.

    “Another major class of activated microglia called Disease Associated Microglia (DAM) were recently identified in a mouse model of Alzheimer’s disease (5xFAD) (Keren-Shaul et al., 2017).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Keren-Shaul et al., 2017 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Demyelinated mouse lesions and human MS tissue trigger diverse microglial activation responses.

    “Diverse microglial activation responses are triggered in demyelinated mouse lesions and human MS tissue”
    Quote in the paperSubstantial claim

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

    Study design and conductown data

    The sex-comparison analysis clustered 49,445 cells across ages and sexes.

    “A total of 49,445 cells were clustered together across all the ages and sexes.”
    Quote in the paperNot a substantial claim

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

    Interpretationown data

    Other cells might also upregulate these pathways.

    “suggesting other cells might also upregulate these pathways.”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    For chromogenic in situ quantification, sample and probe identities were concealed from the researcher, who counted cells under blinded conditions.

    “The images were coded to obscure the identity of each sample as well as the probes used. A blinded researcher used the manual counting feature in ImageJ to count cells labeled with green substrate (Ccl4+) or co-labeled with green and red substrate (Ccl4+Fcrls+).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Microglia assume many distinctive states that change over time.

    “Our results suggest that microglia assume many distinctive states that change over time”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    Re-expression of these markers could be an integral part of the activation process and might be necessary to stimulate a quick response to pathology.

    “Re-expression of these markers could be an integral part of the activation process and might be necessary to stimulate a quick response to pathology.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Blocking TGF-beta signaling has no major effect on microglial cell number but causes a drastic gene-expression shift, including widespread upregulation of genes found in Cluster 6 and brain border macrophages, including Mrc1.

    “blocking TGF-beta signaling (which helps to confer microglia identity, (Butovsky et al., 2014)) has no major effect on microglia cell number, but causes a drastic shift in their gene expression including widespread upregulation of (or failure to downregulate) genes found in Cluster 6 microglia and in brain border macrophages, including Mrc1 (Wong et al., 2017).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Butovsky et al., 2014 source

      Not checked in the free run.

    • cites Wong et al., 2017 source

      Not checked in the free run.

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

    Premisecitation

    Microglia become activated and aberrantly remove synapses in early stages of pathology in mouse models of Alzheimer's disease, well before cognitive decline and plaque formation.

    “microglia become activated and aberrantly remove synapses in the early stages of pathology in mouse models of Alzheimer’s disease, well before cognitive decline and plaque formation (Hong et al., 2016).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Hong et al., 2016 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Identification of robust markers for different microglia states in disease will allow development of tools to more specifically track and visualize microglial activation, inflammation, and disease progression.

    “identification of robust markers for different microglia states in disease will allow for the development of tools to more specifically track and visualize microglial activation, inflammation, and disease progression.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    Markers of these unique and potentially pathogenic microglia could be used as biomarkers or therapeutic targets.

    “which could mean that markers of these unique – and potentially pathogenic – microglia could be used as biomarkers or therapeutic targets.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Evidence of aberrant microglial activation has been found in many neurodevelopmental and neurodegenerative diseases, but the diversity and specificity of these responses need further exploration.

    “evidence of aberrant microglia activation has been found in many forms of neurodevelopmental and neurodegenerative disease, but the diversity of these responses and specificity of the microglia response to each disease needs further exploration (Salter and Stevens, 2017).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Salter and Stevens, 2017

      Not checked in the free run.

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

    Study design and conductown data

    Pathway enrichment analysis used hypergeometric testing with Benjamini-Hochberg correction for multiple testing.

    “Pathway analysis was performed using Hypergeometric testing and a Benjamini Hochberg correction for multiple testing.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Independent components linked to batch or sample replicate were excluded before clustering the remaining cells.

    “ICs corresponding to batch or replicate effects were removed from analysis, and the cells were then clustered based on their values for the remaining ICs.”
    Quote in the paperSubstantial claim

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

    Study design and conductnone given

    Human MS brain tissue was obtained through a UK prospective donor scheme with full ethical approval from the UK MS Tissue Bank.

    “Brain tissue was obtained through a UK prospective donor scheme with full ethical approval from the UK Multiple Sclerosis (MS) Tissue Bank (MREC/02/2/39).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Current PET imaging ligands, including TSPO, are not selective for microglia and are better markers of neuroinflammation.

    “Current PET imaging ligands, including TSPO, are not selective for microglia and are better markers of neuroinflammation (Politis et al., 2012).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Politis et al., 2012 source

      Not checked in the free run.

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

    Study design and conductnone given

    An ACD 3-plex negative control probe was used on one section per slide to confirm signal specificity.

    “Additionally, the ACD 3-plex negative control probe was used on one section per slide to confirm signal specificity.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    Spectral gating was performed on the LSM 710 microscope to narrow emission spectra for Alexa 488 and Atto 550 dyes and minimize autofluorescence.

    “We performed spectral gating on the LSM 710 microscope to narrow the collected emission spectra for the Alexa 488 and Atto 550 dyes and to minimize the autofluorescence.”
    Quote in the paperNot a substantial claim

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

    Interpretationcitation

    These microglia could also arise from ontologically distinct progenitors.

    “These microglia could also arise from ontologically distinct progenitors”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Recent transplantation experiments suggest that Ms4a7 and other Cluster 6 markers are specific to hematopoietic-derived cells rather than yolk-sac progenitors.

    “recent transplantation experiments suggest that Ms4a7 and other markers in Cluster 6 are specific to hematopoietic derived cells versus those derived from yolk-sac progenitors (Bennett et al., 2018).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bennett et al., 2018 source

      Not checked in the free run.

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

    Study design and conductcitation

    The Ccl4 antibody's specificity was verified using secondary-antibody-only controls and preadsorption with recombinant Ccl4 protein, following a previously described method.

    “Ccl4 antibody specificity was tested using secondary antibody only controls as well as with antibody preadsorption using recombinant Ccl4 protein, as previously described (Sekar et al., 2016).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Sekar et al., 2016 source

      Not checked in the free run.

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

    Study design and conductnone given

    All animal experiments were reviewed and overseen by the institutional animal use and care committee at Boston Children's Hospital in accordance with NIH guidelines.

    “All experiments were reviewed and overseen by the institutional animal use and care committee at Boston Children’s Hospital in accordance with all NIH guidelines for the humane treatment of animals.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    These microglial states can be defined by unique markers and localized within the brain.

    “states that can be defined by unique markers and localized within the brain.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Another group showed that these microglia were the predominant source of Igf1 in the early postnatal brain, and Igf1 conditional knockout in microglia affected developing cortical neuron health at the same age.

    “Another group showed that these microglia were the predominant source of Igf1 in the early postnatal brain, and Igf1 conditional knockout in microglia affected the health of developing cortical neurons at the same age (Ueno et al., 2013).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Ueno et al., 2013 source

      Not checked in the free run.

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

    Premisecitation

    Metabolic state is directly manipulated by cytokine and pattern-recognition receptors, and elevated glycolytic activity has been linked to immune cell activation.

    “Metabolic state is directly manipulated by cytokine and pattern-recognition receptors, and elevated glycolytic activity has been tied to immune cell activation (Everts et al., 2014; O’Neill and Pearce, 2016).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Everts et al., 2014 source

      Not checked in the free run.

    • cites O’Neill and Pearce, 2016

      Not checked in the free run.

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

    Premisecitation

    Lipofuscin accumulates in the aged brain and causes autofluorescence, which motivated using chromogenic RNAscope detection for 18-month-old mice.

    “The RNAscope® 2.5 HD Duplex Detection Kit (Chromogenic) was used in experiments involving 18 month-old mice to avoid autofluorescence from lipofuscin accumulation in the aged brain (Xu et al., 2008).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Xu et al., 2008

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Many signals expressed in this subpopulation can be highly damaging to the brain; Il1b and Tnf can cause neurotoxicity, and chemokines can attract infiltrating immune cells that exacerbate pathology.

    “Many of the signals expressed in this subpopulation can be highly damaging to the brain. For example, Il1b and Tnf can both cause neurotoxicity (Takeuchi et al., 2006; Ye et al., 2013), and infiltrating immune cells attracted by the chemokines expressed in the Ccl4+ subpopulation can exacerbate pathology (Gadani et al., 2015).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Takeuchi et al., 2006 source

      Not checked in the free run.

    • cites Ye et al., 2013

      Not checked in the free run.

    • cites Gadani et al., 2015 source

      Not checked in the free run.

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

    Study design and conductnone given

    A negative control probe (ACD 2-plex) was included on one section per slide in the chromogenic in situ experiments.

    “The ACD 2-plex negative control probe was used on one section per slide for control purposes.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    Cluster-defining genes had to exceed a 1.5-fold change and a P value below 1E-4, and were adjusted with Benjamini-Hochberg false discovery rate correction.

    “Genes that define each cluster of microglia exceed a minimum threshold of 1.5 fold change and a P-value less than 1E-4 and were adjusted using Benjamini–Hochberg false discovery rate (FDR) correction.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    This small subpopulation may be a specialized group uniquely primed to produce an inflammatory response.

    “it is possible that they are a specialized group uniquely primed to produce an inflammatory response.”
    Quote in the paperSubstantial claim

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

    Interpretationreasoning

    These parenchymal microglia could be microglia that have not fully differentiated.

    “These microglia, which were present in the brain parenchyma, could be microglia that have yet to fully differentiate.”
    Quote in the paperNot a substantial claim
86 of 86

  • The paper relies on this without support.

    Microglia rapidly change states in response to their environment.

    “Microglia, the resident immune cells of the brain, rapidly change states in response to their environment”

  • The paper relies on this without support.

    There is a lack of molecular and functional signatures of different microglial populations.

    “but we lack molecular and functional signatures of different microglial populations”

  • The paper relies on this without support.

    Microglia have historically been characterized as resting, M1, or M2 based on simple in vitro stimulation methods.

    “microglia have historically been characterized as ‘resting’, ‘M1’ (proinflammatory), or ‘M2’ (anti-inflammatory) - based on simple in vitro stimulation methods.”

  • The paper relies on this without support.

    It is still unclear when microglia diverge from brain border neighbors and how quickly they differentiate.

    “However, it is still unclear when microglia diverge from their brain border neighbors and how quickly they differentiate.”

  • The paper relies on this without support.

    It is not until microglia infiltrate the brain parenchyma and are exposed to brain-derived signals that they achieve their unique identity.

    “It is not until microglia infiltrate the brain parenchyma and are exposed to brain-derived signals that they achieve their unique identity.”

  • The paper relies on this without support.

    The extent of the microglial response to aging has not been comprehensively studied, and it remains unclear how, where, and to what extent microglia change with normal aging.

    “However, the extent of the microglial response to aging has not been comprehensively studied, and it remains unclear how, where (in the brain), and to what extent microglia change with normal aging.”

  • The paper relies on this without support.

    It is unclear whether microglia tailor their responses to specific injury types or whether distinct microglial populations exist in pathological tissue.

    “However, it is still unclear whether or how microglia tailor their response to specific injury types or whether distinct populations of microglia exist in pathological tissue, as we lack both the markers and tools to identify and characterize different activation states, should they exist.”

  • The paper relies on this without support.

    Microglia respond robustly and rapidly to demyelination in the LPC model.

    “third, microglia respond robustly and rapidly to demyelination in this model”

  • The paper relies on this without support.

    Microglial activation is largely confined to the lesion where demyelination occurred.

    “microglial activation is largely confined to the lesion where the demyelination has occurred.”

  • The paper relies on this without support.

    At 7 days post-lesion, microglia are highly responsive and the lesion is transitioning from myelin debris cleanup to initial remyelination.

    “We focused our analysis on 7 days post-lesion (7 dpl), a time point at which microglia are highly responsive and the lesion is undergoing a transition from myelin debris cleanup to the initial phases of remyelination (Fig 6a).”

  • The paper relies on this without support.

    Demyelinated lesions contain a high density of activated microglia, generally in the core, where they can remove myelin debris and release factors into the lesion microenvironment.

    “Demyelinated lesions contain a high density of activated microglia, generally in the core, where they can remove myelin debris and release factors into the lesion microenvironment.”

  • The paper relies on this without support.

    In MS, infiltrating immune cells specifically target myelin and cause white matter lesions.

    “In MS, infiltrating immune cells specifically target myelin and cause white matter lesions.”

  • The paper relies on this without support.

    Microglia migrate, grow, and proliferate in response to injury.

    “Because microglia migrate, grow, and proliferate in response to injury”
13 of 13

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Aging-related cell states exhibited upregulation of genes (e.g., B2m and Trem2 in Micro-3, Xdh in Endo-3, Gfap and C4b in Astro-2) previously shown to be enriched in activated microglia, endothelial cells, and astrocytes.

    “These aging-related cell states exhibit upregulation of genes (e.g., B2m and Trem2 in Micro-3, Xdh in Endo-3, and Gfap and C4b in Astro-2) (Figure 2C) that have been previously shown to be enriched in microglia, endothelial cells, and astrocytes activated by inflammation and/or aging.”

    From Molecular and spatial signatures of mouse brain aging at single-cell resolution (2023) · cites this paper as [29]

    The Results section on inflammatory and interferon-responsive microglia in the aged brain (page unavailable) reports age-enriched microglial states that upregulate inflammatory genes. It does not identify B2m or Trem2 as markers of those states, or report the claimed Xdh enrichment in endothelial cells or Gfap and C4b enrichment in astrocytes. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Previous transcriptomic studies have revealed widespread changes in cell state in specific cell types with age, with many highlighting increased inflammation as a key aspect of brain aging.

    “Previous transcriptomic studies have revealed widespread changes in cell state in specific cell types with age, with many of these studies highlighting a role for increased inflammation as a key aspect of brain aging.17–20,25–30”

    From Molecular and spatial signatures of mouse brain aging at single-cell resolution (2023) · cites this paper as [29]

    The Results section reports an age-related increase in inflammatory and interferon-responsive microglial states, supporting the inflammation aspect of the claim. It also says these states comprise only a small fraction of microglia and that the vast majority are unaltered or only slightly altered by aging, so this study does not support describing its cell-state changes as widespread. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • Microglial and astrocytic activation has also been reported in aged rodent and human brains.

    “Microglial and astrocytic activation has also been reported in aged rodent and human brains”

    From Molecular and spatial signatures of mouse brain aging at single-cell resolution (2023) · cites this paper as [29]

    The Results section, “Small populations of inflammatory and interferon-responsive microglia emerge in the aged brain” (page not specified), reports an increase in these microglial states in aged mice, supporting the rodent microglia portion of the claim. It says these cells remain a small fraction of microglia; the passages do not report astrocytic activation or activation in aged human brains. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • For microglia and astrocytes, genes highly upregulated with age overlapped substantially with genes previously reported as upregulated in activated (reactive) states (e.g., Gfap and C4b for astrocytes; B2m and Lyz2 for microglia).

    “For microglia and astrocytes, we observed that the genes highly upregulated with age overlapped substantially with genes that have been previously reported as being upregulated in the activated (or “reactive”) state of these cell types (e.g., Gfap and C4b for astrocytes; B2m and Lyz2 for microglia)”

    From Molecular and spatial signatures of mouse brain aging at single-cell resolution (2023) · cites this paper as [29]

    The Results section, “Small populations of inflammatory and interferon-responsive microglia emerge in the aged brain” (page unavailable), reports that aging increases small populations of microglia with inflammatory or interferon-response gene expression. It does not establish substantial overlap between age-upregulated and reactive-state genes across microglia and astrocytes, or identify the cited examples B2m, Lyz2, Gfap, and C4b as such genes. (LLM, conf=0.82)

    Experimental — accuracy not yet measured

  • Single-cell deep sequencing methods revealed regional-, age- and sex-dependent differences of the microglial population and various expression profiles defining microglial subpopulations.

    “Recent progress in single-cell deep sequencing methods allowed for more refined differentiation and revealed regional-, age- and sex-dependent differences of the microglial population, and a growing number of studies demonstrate various expression profiles defining microglial subpopulations.”

    From Beyond Activation: Characterizing Microglial Functional Phenotypes (2021) · cites this paper as [102]

    The Results sections on developmental subpopulations and axon tract–associated microglia report age-dependent transcriptional states and a subpopulation localized to particular brain regions. But the Results section on sex says clustering differed almost not at all between males and females, with only one very small female-enriched cluster; it concludes that diversity was largely unaffected by sex. Thus the sex-dependent part overstates this paper's findings. (LLM, conf=0.90)

    Experimental — accuracy not yet measured

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Hammond, T. R., Dufort, C., Dissing-Olesen, L., Giera, S., Young, A., Wysoker, A., Walker, A. J., Segel, M., Nemesh, J., Saunders, A., Macosko, E., Franklin, R. J. M., Piao, X., McCarroll, S., & Stevens, B.
Affiliations
  • Broad Institute (US)
  • Boston Children's Hospital (US)
  • Wellcome/MRC Cambridge Stem Cell Institute (GB)
Corresponding author
Stevens, Beth
ORCID
Venue
Immunity
Publisher
openRxiv
Published
2018-08-31
Volume
50
Issue
1
Pages
253-271.e6
ISSN
1074-7613 (ISSN-L)
DOI
10.1101/406140
PMC
PMC6655561
OpenAlex
W2901838561
Cited by
2,225
References
83
Citations by year
2026: 257 · 2025: 312 · 2024: 304 · 2023: 343 · 2022: 351 · 2021: 308 · 2020: 218 · 2019: 123 · 2018: 4 · 2017: 1 · 2013: 1 · 2012: 1
FWCI
72.90
Type
journal article
Language
en
Open access
Locations
  • Immunity (publishedVersion)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nd)
  • PubMed (publishedVersion)
  • Europe PMC (PubMed Central) (acceptedVersion)
  • PubMed Central (acceptedVersion)
  • Apollo (University of Cambridge) (submittedVersion)
  • Apollo
Funders
  • Howard Hughes Medical Institute
  • Simons Foundation
  • International Rett Syndrome Foundation
  • Simons Foundation Autism Research Initiative
  • Rosetrees Trust
  • Lundbeckfonden (R182-2014-2844)
  • Stanley Center for Psychiatric Research, Broad Institute
  • Medical Research Council (MC_PC_12009)
  • National Institute of Mental Health (P50MH112491)
Topics
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Health Sciences › Medicine › Pathology and Forensic Medicine › Multiple Sclerosis Research Studies
Keywords
microglia, microglial heterogeneity, microglial states, brain aging, brain injury, multiple sclerosis, reactive microglia, microglial development, single-cell RNA sequencing
MeSH terms
Adaptation, Physiological, Aging, Animals, Brain, Brain Injuries, Cell Differentiation, Demyelinating Diseases, Humans, Longevity, Mice, Inbred C57BL, Multiple Sclerosis, Sequence Analysis, RNA, Microglia, Mice, Single-Cell Analysis

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 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

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