CoChat
Sign Up
Cited by 344Single-cell and spatial transcriptomicsNeuroinflammation and Neurodegeneration MechanismsCell Image Analysis Techniques

Single-Nucleus RNA-Seq Is Not Suitable for Detection of Microglial Activation Genes in Humans

Open access
Thrupp, Nicola, Sala Frigerio, Carlo, Wolfs, Leen, Skene, Nathan G., Fattorelli, Nicola, Poovathingal, Suresh, Fourne, Yannick, Matthews, Paul M. +4 more·2020·Cell Reports
DOI 10.1016/j.celrep.2020.108189PMC PMC7527779
Open this paper in CoChat — free accountOpen PDF

Abstract

Single-nucleus RNA sequencing (snRNA-seq) is used as an alternative to single-cell RNA-seq, as it allows transcriptomic profiling of frozen tissue. However, it is unclear whether snRNA-seq is able to detect cellular state in human tissue. Indeed, snRNA-seq analyses of human brain samples have failed to detect a consistent microglial activation signature in Alzheimer's disease. Our comparison of microglia from single cells and single nuclei of four human subjects reveals that, although most genes show similar relative abundances in cells and nuclei, a small population of genes (∼1%) is depleted in nuclei compared to whole cells. This population is enriched for genes previously implicated in microglial activation, including APOE, CST3, SPP1, and CD74, comprising 18% of previously identified microglial-disease-associated genes. Given the low sensitivity of snRNA-seq to detect many activation genes, we conclude that snRNA-seq is not suited for detecting cellular activation in microglia in human disease.

AI snapshot

Written from the paper’s full text.

TL;DR

Comparing single-cell and single-nucleus RNA-seq of human microglia from four subjects, most genes had similar abundance, but a small set (~1%) was depleted in nuclei and enriched for microglial activation genes. The authors conclude snRNA-seq is not suited for detecting cellular activation states in human microglia.

Key findings

  • Most genes showed similar normalized abundance in microglial cells and nuclei: 98.6% fell along the diagonal (Pearson’s correlation coefficient = 0.92, p < 2.2e-16). However, 246 genes (1.1%) were less abundant in nuclei (fold change < −2, padj < 0.05), and 67 genes (0.3%) were more abundant in nuclei (fold change > 2, padj < 0.05).
  • Nuclei-depleted genes were enriched for microglial activation signatures: 45 of 257 orthologous DAM genes (NES = −2.16, padj = 3.6e-05), 28 of 200 orthologous ARM genes (NES = −2.01, padj = 3.6e-05), LPS-upregulated genes (NES = −1.86, padj = 3.6e-05), and Mathys et al. Mic1 markers (NES = −2.14, padj = 3.6e-05). Examples included APOE, CST3, SPP1, and CD74; the abstract states this comprises 18% of previously identified microglial-disease-associated genes.
  • In public datasets, cell-abundant genes were depleted in other single-nucleus microglia relative to single-cell microglia (mean microglial Z score 7.95 for cells-to-nuclei, versus 0.01 for cell-to-cell and 0.81 for nuclei-to-nuclei comparisons). Reanalysis of Gerrits et al. data found the same nuclei-depleted pattern when comparing fresh cells to frozen nuclei, but little difference between fresh cells and fresh nuclei.
  • snRNA-seq recovered major cell types from human tissue: 37,060 nuclei yielded 7 major cell types, including microglia (11.3%), but clustering did not convincingly recover differences between other microglial subpopulations.

Methods

  • Paired scRNA-seq of FACS-sorted microglia and snRNA-seq from the same four human temporal cortex samples (epilepsy surgery; ages 7 male, 20, 24, and 50 female) using the 10X Genomics single-cell gene expression v2 kit. After preprocessing, 3,927 microglial nuclei remained; differential abundance analysis used 3,721 nuclei and 14,435 cells after discarding non-microglial clusters.
  • Differential abundance was performed with Seurat using a two-sided Wilcoxon rank sum test with Bonferroni correction; genes with padj < 0.05 and fold change > |2| were considered significant. Functional analysis used GSEA (fgsea) against ranked log fold changes and gene sets from prior studies.
  • Public dataset comparisons used the EWCE package and a MicroglialDepletion package with bootstrapped Z scores across 8 public single-cell or single-nucleus datasets; Gerrits et al. data were downloaded from GEO (GSE135618) and reanalyzed for fresh versus frozen nuclei.

Limitations

  • Stated in the paper: the study used non-diseased tissue, so the authors cannot dismiss the possibility that mRNA localization distributions change in more activated microglia. They also state they cannot attribute cell-versus-nucleus abundance differences to a biological mechanism versus a technical artifact, and the reason for differences between fresh and frozen nuclei is unclear.
  • Stated in the paper: deeper sequencing or increased sample size could possibly compensate for lack of sensitivity, but the sparse nature of snRNA-seq and high human heterogeneity suggest this may remain a problem.
  • Evident from the text: the primary comparison relies on only four non-AD epilepsy surgery subjects, with one patient sample removed due to poor quality, so the findings about disease-associated activation are inferred from gene-set enrichment rather than direct AD patient snRNA-seq in this study.

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

Integrity signals

Registry

Crossref · 10.1016/j.celrep.2020.108189

Cites

48 of 48 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

21 of 22 cited sources checked · 12 supported · 1 flagged · accuracy not yet measured

  • Premisecitation

    snRNA-seq analyses of human brain samples have failed to detect a consistent microglial activation signature in Alzheimer's disease.

    “snRNA-seq analyses of human brain samples have failed to detect a consistent microglial activation signature in Alzheimer’s disease.”
    • Also stated in: § Introduction · hedged
      “a number of high-profile snRNA-seq studies of microglia in human AD (Del-Aguila et al., 2019; Grubman et al., 2019; Mathys et al., 2019) have been unable to recover an equivalent activation signature or, possibly more revealing, a consistent activation signature.”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites Del-Aguila et al., 2019 source

      The Results section (page unavailable) says the study did not sequence enough microglial cells to replicate the previously reported disease-associated microglia (DAM) marker findings. But it found five markers consistently associated with microglia across all three samples. The claim overgeneralizes this limited finding to a failure to detect any consistent microglial activation signature in human Alzheimer’s disease samples. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites Grubman et al., 2019 source
    • cites Mathys et al., 2019 source

      In “Identification of AD-associated cellular subpopulations” (page not specified), the paper reports an AD-associated human microglial subpopulation, Mic1, with immune-related marker genes that significantly overlap disease-associated microglial signatures. It also reports immunohistochemical confirmation of MHC class II–expressing microglia in AD-pathology samples. This is contrary to attributing a failure to detect a microglial activation signature to this study, though the paper does not establish consistency across all human snRNA-seq studies. (LLM, conf=0.87)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured
  • Findingown data

    Markers of the human microglial response to AD (Mic1) are depleted in nuclei.

    “Markers of this response (referred to by Mathys et al., 2019 as “Mic1”) had a NES score of −2.14 (padj = 3.6e-05), indicating that they were depleted in nuclei (Figure 2C).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Mathys et al., 2019 source
  • Findingown data

    Genes upregulated by LPS stimulation in mice are depleted in nuclei.

    “Genes upregulated by lipopolysaccharide (LPS) stimulation in mice (Gerrits et al., 2019) also showed depletion in nuclei (NES = −1.86, padj = 3.6e-05; Figure 2C; Figure S2F).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Gerrits et al., 2019 source
  • Findingcitation

    Most DAM genes were not identified as activation genes using the Mathys et al. snRNA-seq protocol.

    “The study identified 28 DAM genes as marker genes of the Mic1 response cluster (shown in orange in Figure 2F); however, most DAM genes were not identified as activation genes using their snRNA-seq protocol (purple in Figure 2F).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Mathys et al., 2019 source

      The Identification of AD-associated cellular subpopulations section reports a significant overlap between Mic1 markers and mouse disease-associated microglia genes, while describing the states as distinct. It does not say that most DAM genes were missed or quantify how many were not identified as activation genes by its snRNA-seq analysis. Page numbers were not available in our copy. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
  • Findingown data

    45 of 257 orthologous DAM genes are depleted in nuclei.

    “we found a strong depletion of genes associated with mouse microglial activation: 45 of 257 orthologous DAM genes (Keren-Shaul et al., 2017; NES = −2.16, padj = 3.6e-05; Figures 2C and 2D)”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Keren-Shaul et al., 2017 source
  • Classified as probably not a substantial claim. 28
  • May combine two claims. 12

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

    Findingown data

    Mitochondrial mRNA is enriched among genes more abundant in cells.

    “as was mitochondrial mRNA (NES = −1.69, padj = 2.7e-04, gene set extracted from Ensembl’s BioMart; Zerbino et al., 2018).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Zerbino et al., 2018 source

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

    Findingown data

    Ribosomal protein genes are depleted in nuclei.

    “RNA of genes coding for ribosomal proteins was also depleted in nuclei (NES = −2.37, padj = 3.6e-05), as previously described (Habib et al., 2017).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Habib et al., 2017 source

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

    Findingown data

    Genes with longer coding sequences are enriched in nuclei.

    “whereas longer CDSs were enriched (NES = 1.91, padj = 3.6e-05), as already observed in earlier snRNA-seq studies (Bakken et al., 2018).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Bakken et al., 2018 source

      In Results, under “Comparison of nuclear and whole cell transcriptomes” (page unavailable), Bakken et al. report that nucleus-enriched genes were longer than cell-enriched genes when expression was calculated from intronic and exonic reads. They do not establish that the coding sequences specifically were longer; they caution that long introns may account for the apparent enrichment. The NES and adjusted P-value in the citing sentence are not presented as Bakken et al.’s results. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Study design and conductcitation

    The single-cell protocol performed dissociation at 4°C to mitigate potential activation, making it unlikely that dissociation procedures account for the differences.

    “For our single-cell protocol, we performed dissociation at 4°C to mitigate potential activation (Mancuso et al., 2019), so it is unlikely the dissociation procedures account for the differences.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Mancuso et al., 2019 source

      The Methods section, “Isolation of human and mouse microglia from the mouse brain” (page unavailable), describes tissue dissociation followed by centrifugation and antibody labeling at 4°C. It does not say dissociation itself was performed at 4°C or that cold processing mitigated activation. The conclusion about the citing authors’ own differences is not attributed to Mancuso et al. (LLM, conf=0.82)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    Synaptosome genes and ambient RNA are enriched within nucleus-abundant genes.

    “This is supported by the enrichment of synaptosome genes (NES = 1.82, padj = 3.6e-05, Figure 2A; Hafner et al., 2019) and ambient RNA—mRNA originating not from intact cells/nuclei but from free-floating transcripts (Macosko et al., 2015)—(NES = 1.71, padj = 9.2e-05; Figure 2A; Figure S2D) within the nucleus-abundant genes.”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites Hafner et al., 2019 source
    • cites Macosko et al., 2015 source

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

    Findingown data

    In agreement with Gerrits et al., fresh nuclei versus cells showed very little difference.

    “In agreement with the results of Gerrits et al. (2019), when we plotted the normalized abundance of fresh nuclei versus cells, we saw very little difference between the two populations (Figure S3C).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Gerrits et al., 2019 source

  • May combine two claims.

    Study design and conductcitation

    The lab has previously used the same protocol to study microglial activation in mice, and no activation due to technical artifacts was seen.

    “Our lab has previously used the same protocol to study microglial activation in mice, and no activation due to technical artifacts was seen (Sala Frigerio et al., 2019).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Sala Frigerio et al., 2019 source

      The Results and STAR★METHODS (page not provided) describe studying microglial activation in mice and keeping tissue cold to minimize activation caused by isolation. They do not report that no activation due to technical artifacts was observed; minimizing a risk is not evidence that it was absent. (LLM, conf=0.87)

      Experimental — accuracy not yet measured

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

    Premisecitation

    The improved v3 chemistry yields a higher number of reads and genes per cell in PBMCs.

    “The improved v3 chemistry yields a higher number of reads and genes per cell, as shown in a comparison using peripheral blood mononuclear cells (PBMCs) (Ding et al., 2020).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Ding et al., 2020 source

      The Results section, “Similar relative ranking of method sensitivity across experiments” (page unavailable), reports that 10x Chromium v3 detected the most genes per PBMC cell among the high-throughput methods. It does not report that v3 yielded more reads per cell; its higher UMI count is not a read count. The claim also omits the high-throughput-methods qualifier. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

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

    Findingown data

    5,068 genes were exclusively detected in nuclei.

    “Additionally, 3,248 genes were only detected in cells, and 5,068 genes were exclusively detected in nuclei.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    A second population of 67 genes (0.3%) is more abundant in nuclei than cells.

    “A second population of 67 genes (0.3%) was found to be more abundant in nuclei (fold change > 2, padj < 0.05; red in Figure 1A).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Single-nucleus clustering recovers cytokine clusters and macrophage clusters well.

    “A comparison of single-nucleus and single-cell clustering suggested that cytokine clusters and macrophage clusters recovered well by using single-nucleus methods”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Findingown data

    In the Gerrits frozen nuclei dataset, microglial clusters (clusters 3,5 and 6) were identified using known markers including P2RY12, CSF1R, CX3CR1, resulting in 2,659 microglial nuclei.

    “In the frozen nuclei dataset, microglial clusters (clusters 3,5 and 6 in Figures S3A and S3B) were identified using known markers including P2RY12, CSF1R, CX3CR1 (resulting in 2,659 microglial nuclei), and extracted for downstream analysis.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The nuclei were able to recover a cytokine response cluster (CRM), marked by CCL3, CCL4, and an activation-like cluster, equivalent to the "in vitro microglia" identified in the original study (original markers included APOC1, GPNMB, SPP1, APOE).

    “Our nuclei were able to recover a cytokine response cluster (CRM), marked by CCL3, CCL4, and an activation-like cluster, equivalent to the “in vitro microglia” identified in the original study (original markers included APOC1, GPNMB, SPP1, APOE).”
    Quote in the paperSubstantial claim

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

    Findingown data

    Cluster Nuc4 showed similarities with the "in vitro 2" cluster, suggesting it could be a cluster of activation, however all 5 overlapping genes were mitochondrial genes.

    “Cluster Nuc4 showed similarities with the “in vitro 2” cluster, suggesting it could be a cluster of activation, however all 5 overlapping genes were mitochondrial genes.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    The cell macrophage (CAM) and cell cytokine (CRM) clusters showed the largest overlaps with Nuc1 and Nuc5 (27 and 24 of 40 markers, respectively).

    “The cell macrophage (CAM) and cell cytokine (CRM) clusters showed the largest overlaps with Nuc1 and Nuc5 (27 and 24 of 40 markers, respectively).”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Cluster Nuc3 markers RPS12, TPT1, FTL, RPS18 and EEF1A1 also appeared as markers of "in vitro 2."

    “Cluster Nuc3 markers RPS12, TPT1, FTL, RPS18 and EEF1A1 also appeared as markers of “in vitro 2.””
    Quote in the paperNot a substantial claim

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

    Findingown data

    The dataset contained 7 main cell types with proportions: oligodendrocytes 34.0%, excitatory neurons 27.0%, interneurons 11.3%, oligodendrocyte precursors 9.4%, microglia 11.3%, astrocytes 6.0%, and endothelial cells 1.1%.

    “We identified 7 main cell types in 37,060 nuclei: oligodendrocytes (ODC, 34.0%), excitatory neurons (27.0%), interneurons (11.3%), oligodendrocyte precursors (OPC 9,4%), microglia (11.3%), astrocytes (6.0%), and endothelial cells (1.1%).”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Downsampling of cell counts to match nucleus sequencing depth resulted in 1,304 mean reads per cell, compared with the original count of 3,979 reads per cell.

    “This resulted in 1,304 mean reads per cell, compared with the original count of 3,979 reads per cell.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Other clusters only showed overlaps of between 1 and 5 genes.

    “Other clusters only showed overlaps of between 1 and 5 genes.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    The integrated dataset consisted of 37,060 nuclei, with a mean UMI count of 4,305 counts per nucleus, and 1,791 genes per nucleus.

    “The dataset consisted of 37,060 nuclei, with a mean UMI count of 4,305 counts per nucleus, and 1,791 genes per nucleus.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Cluster Nuc0 showed an overlap of 5 genes with "in vivo HM," and cluster Nuc2 showed an overlap of 2 genes with "in vivo HM."

    “Cluster Nuc0 showed an overlap of 5 genes with “in vivo HM,” and cluster Nuc2 showed an overlap of 2 genes with “in vivo HM.””
    Quote in the paperNot a substantial claim

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

    Findingown data

    Cluster Nuc3 showed similar overlaps between "in vitro 1" and "in vitro 2" (5 genes).

    “Cluster Nuc3 showed similar overlaps between “in vitro 1” and “in vitro 2” (5 genes).”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Microglial nuclei reclustering identified 3,721 microglia (expressing MEF2A, P2RY12, CX3CR1, CSF1R), a macrophage cluster (enriched for CD163 and MRC1, 67 nuclei), a neutrophil cluster (72 nuclei), and a cluster containing microglial as well as astrocytic markers (marked by GFAP, 68 nuclei).

    “We identified 3,721 microglia (expressing MEF2A, P2RY12, CX3CR1, CSF1R), a macrophage cluster (enriched for CD163 and MRC1, 67 nuclei), a neutrophil cluster (72 nuclei), and a cluster containing microglial as well as astrocytic markers (marked by GFAP, 68 nuclei).”
    Quote in the paperNot a substantial claim

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

    Findingown data

    After pre-processing, 3,927 nuclei remained, with a mean UMI count of 1,328 and a mean gene count of 879 genes per nucleus.

    “After pre-processing, 3,927 nuclei remained, with a mean UMI count of 1,328 and a mean gene count of 879 genes per nucleus.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    The CAM (macrophage) cluster (CD163, MRC1) separated out from the bulk of the microglia and was easily recognizable based on its MS40 score.

    “Finally, the CAM (macrophage) cluster (CD163, MRC1), separated out from the bulk of the microglia, and was easily-recognizable based on its MS40 score.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Selecting 40 markers allowed the authors to align cellular and nuclear clusters in an almost one-to-one fashion.

    “Selecting 40 markers allowed us to align cellular and nuclear clusters in an almost one-to-one fashion (see Figure S1F).”
    Quote in the paperNot a substantial claim

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

    Interpretationnone given

    snRNA-seq is viable for identification of cell types in tissue for which cell dissociation is problematic.

    “although snRNA-seq offers a viable alternative to scRNA-seq for the identification of cell types in tissue for which cell dissociation is problematic”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductreasoning

    The study used non-diseased tissue and cannot dismiss the possibility that mRNA localization distributions change in more activated populations of microglia.

    “One caveat of the present study is that we have used non-diseased tissue, and we cannot dismiss the possibility that mRNA localization distributions change in more activated populations of microglia.”
    Quote in the paperSubstantial claim

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

    Study design and conductown data

    To ensure the high number of cells relative to nuclei did not affect the result, the authors randomly sampled 3,927 cells from the count matrix.

    “To ensure the high number of cells relative to nuclei did not affect the result (Figure S2C), we randomly sampled 3,927 cells from the count matrix.”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    In the fresh nuclei and cells, examination of these markers in the clustering confirmed that all cells were microglial.

    “In the fresh nuclei and cells, examination of these markers in the clustering confirmed that all cells were microglial.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    The study confirms its results using publicly available data.

    “We confirm our results using publicly available data.”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    Nuclei from one patient sample (RM101.1) were removed due to poor quality (low read and gene count).

    “Following alignment, nuclei from one patient sample (RM101.1) were removed due to poor quality (low read and gene count).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Figure S1C confirms that only microglial clusters were extracted from the original data.

    “Figure S1C, highlighting known microglial markers (P2RY12, CSF1R, DOCK8 and CX3CR1) as well as markers for other celltypes (MBP for oligodendrocytes, GRIA1 for neurons, AQP4 for astrocytes and CLDN5 for endothelial nuclei), confirms that we extracted only microglial clusters from the original data.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    The neutrophil and ambiguous clusters were discarded, leaving only microglia and brain macrophages for downstream analysis.

    “The neutrophil and ambiguous clusters were discarded, leaving only microglia and brain macrophages for downstream analysis (Figure S1D).”
    Quote in the paperSubstantial claim
34 of 34

  • The paper relies on this without support.

    scRNA-seq of whole cells provides an accurate reference for the true cellular transcriptome against which snRNA-seq can be judged.

    “To compare gene abundance in single microglial cells (14,823 cells) and nuclei (3,940 nuclei), we performed a differential abundance analysis between cells and nuclei from the 4 subjects (Figure 1A).”

  • The paper relies on this without support.

    The limitations of snRNA-seq observed with the 10X Genomics v2 kit and frozen tissue generalize to snRNA-seq generally in human disease.

    “we conclude that snRNA-seq is not suited for detecting cellular activation in microglia in human disease.”

  • The paper relies on this without support.

    The four human subjects and non-diseased tissue are representative enough to support conclusions about human microglial activation in disease.

    “One caveat of the present study is that we have used non-diseased tissue, and we cannot dismiss the possibility that mRNA localization distributions change in more activated populations of microglia.”

  • The paper relies on this without support.

    The nuclei-depleted genes are depleted due to technical limitations of snRNA-seq rather than biological differences in nuclear mRNA localization.

    “However, we are unable to attribute the differences in cellular and nuclear abundance to a biological mechanism (for instance, mRNA trafficking) or a technical artifact of the methodology.”

  • The authors state this as an assumption.

    Nuclei with less than 700 counts from the original unfiltered raw count matrix were empty drops rather than nuclei.

    “under the assumption that these were empty drops rather than nuclei”
5 of 5

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?

  • Nuclear RNA may not fully capture microglial states or localized transcriptomic changes.

    “Our study has several limitations: isotropic fractionation and read depth cut-offs may bias cell recovery based on their nuclear content; nuclear RNA may not fully capture microglial states82 or localized transcriptomic changes; and pathology burden is based on per-sample averages instead of on the spatial context of each cell.”

    From Single-cell multiregion dissection of Alzheimer’s disease (2024) · cites this paper as [82]

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.

  • Interneuron percentage differs between Results and Methods

    Minor

    For the same 37,060-nucleus cell-type breakdown, the Results give interneurons as 11.2%, while the Methods give 11.3%.

  • The 246-gene group is described in opposite directions

    Minor

    The original 246 genes are reported as less abundant in nuclei, but the ambient-RNA comparison calls the original 246 genes less abundant in cells.

  • The 67-gene group is described in opposite directions

    Minor

    The original 67 genes are reported as more abundant in nuclei, but the later comparison calls 67 genes underrepresented in nuclei.

3 of 3

Cites

Flagged first, then by citations

2 of the 50 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Related

Authors
Thrupp, N., Sala Frigerio, C., Wolfs, L., Skene, N. G., Fattorelli, N., Poovathingal, S., Fourne, Y., Matthews, P. M., Theys, T., Mancuso, R., de Strooper, B., & Fiers, M.
Affiliations
  • Vlaams Instituut voor Biotechnologie (BE)
  • UK Dementia Research Institute (GB)
  • KU Leuven (BE)
  • University of London (GB)
Corresponding author
Fiers, Mark
ORCID
Venue
Cell Reports
Publisher
Elsevier BV
Published
2020-09
Volume
32
Issue
13
Pages
108189
ISSN
2211-1247 (ISSN-L)
DOI
10.1016/j.celrep.2020.108189
PMC
PMC7527779
OpenAlex
W3091491962
Cited by
329
References
50
Citations by year
2026: 40 · 2025: 70 · 2024: 48 · 2023: 57 · 2022: 54 · 2021: 57 · 2020: 2 · 2012: 1
FWCI
12.36
Type
journal article
Language
en
Open access
Licence
Locations
  • Cell Reports (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • Lirias (KU Leuven) (submittedVersion, other-oa)
  • PubMed Central (submittedVersion, cc-by-nc-nd)
  • UCL Discovery (University College London) (submittedVersion)
  • Spiral (Imperial College London) (submittedVersion, cc-by-nc-nd)
Funders
  • Fonds Wetenschappelijk Onderzoek
  • Roche
  • KU Leuven
  • Biogen
  • Vlaams Instituut voor Biotechnologie
  • European Resuscitation Council (834682)
  • Alzheimer&apos;s Association
  • Vlaamse regering
  • Alzheimer Society
  • NIHR Imperial Biomedical Research Centre
  • European Commission
  • Novartis
  • UK Dementia Research Institute (FWO 1830717N)
  • GlaxoSmithKline
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Biophysics › Cell Image Analysis Techniques
Keywords
single-nucleus RNA sequencing, microglial activation, human microglia, apolipoprotein E, CST3, osteopontin, CD74, Alzheimer's disease
MeSH terms
Humans, Sequence Analysis, RNA, Microglia, Gene Expression Profiling, 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 2026-10-13.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-13.

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