CoChat
Sign Up
Cited by 30MicroRNA in disease regulationPlant Molecular Biology ResearchCancer-related molecular mechanisms research

A curated human cellular microRNAome based on 196 primary cell types

Open access
Patil, Arun H, Baran, Andrea, Brehm, Zachary P, McCall, Matthew N, Halushka, Marc K·2022·GigaScience
DOI 10.1093/gigascience/giac083PMC PMC9404528

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Background: An incomplete picture of the expression distribution of microRNAs (miRNAs) across human cell types has long hindered our understanding of this important regulatory class of RNA. With the continued increase in available public small RNA sequencing datasets, there is an opportunity to more fully understand the general distribution of miRNAs at the cell level. Results: From the NCBI Sequence Read Archive, we obtained 6,054 human primary cell datasets and processed 4,184 of them through the miRge3.0 small RNA sequencing alignment software. This dataset was curated down, through shared miRNA expression patterns, to 2,077 samples from 196 unique cell types derived from 175 separate studies. Of 2,731 putative miRNAs listed in miRBase (v22.1), 2,452 (89.8%) were detected. Among reasonably expressed miRNAs, 108 were designated as cell specific/near specific, 59 as infrequent, 52 as frequent, 54 as near ubiquitous, and 50 as ubiquitous. The complexity of cellular microRNA expression estimates recapitulates tissue expression patterns and informs on the miRNA composition of plasma. Conclusions: This study represents the most complete reference, to date, of miRNA expression patterns by primary cell type. The data are available through the human cellular microRNAome track at the UCSC Genome Browser (https://genome.ucsc.edu/cgi-bin/hgHubConnect) and an R/Bioconductor package (https://bioconductor.org/packages/microRNAome/).

AI snapshot

Written from the paper’s full text.

TL;DR

This paper curates public small RNA sequencing data into a human cellular microRNAome of 2,077 samples from 196 primary cell types, plasma, and platelets. It provides cell-type-resolved miRNA expression estimates, normalization benchmarks, and resources through UCSC Genome Browser and an R/Bioconductor package.

Key findings

  • From 6,054 candidate SRA runs, 4,184 were processed with miRge3.0 and curated to 2,077 samples from 196 cell types across 175 studies; 2,406 runs included 329 technical replicates. Of 2,731 miRBase v22.1 miRNAs, 2,452 (89.8%) were detected, and 99.5% of MirGeneDB 2.1 miRNAs were detected.
  • Among 323 mature MirGeneDB miRNAs with ≥100 RPM in at least one cell type, 108 were cell specific/near specific, 59 infrequent, 52 frequent, 54 near ubiquitous, and 50 ubiquitous.
  • DESeq2 VST normalization gave the highest cell-type classification accuracy (96.8%) versus raw counts (56.5%), ComBat-Seq (79.3%), RUVg (94%), and RUVr (86.3%); immune cells reached 99% accuracy and neurons 93.6%.
  • Plasma miRNA deconvolution of 85 samples across 30 cell types indicated RBCs and platelets were major contributors (38%); RBC contribution was reported as 27.4% in the stacked composition. Older miRNAs were more frequently expressed than younger ones (average VST 7.4 vs. 3.0, P = 2.7e–13).

Methods

  • SRA query on 22 January 2021 yielded 58,117 runs from 1,872 studies; manual curation excluded paired-end, cancer, exosome, and non-Illumina runs, leaving 6,054 downloaded runs. After adapter/UMI filtering, 4,184 runs were processed with miRge3.0 against miRBase v22.1, with most downstream analyses using MirGeneDB 2.1 miRNAs.
  • UMAP-based outlier removal and exclusion of 608 RNA Atlas runs produced 2,077 final samples from 196 cell types. Normalization methods were compared using leave-one-study-out cross-validation and Euclidean distance to cell-type centroids; DESeq2 VST was selected.
  • Cell specificity was classified for MirGeneDB guide-strand miRNAs with average RPM ≥100 in at least one of 198 cell types. Plasma contributions were estimated with CIBERSORT using 85 plasma samples and a 30-cell-type reference; tissue comparisons used four tissue samples processed through miRge3.0.

Limitations

  • The paper states that cellular microRNAome data come primarily from cultured cells, and ex vivo conditions can alter miRNA expression; neurons in particular are stem-cell derived, may be more primitive, shared miRNAs with stem cells, and lacked some brain-tissue coexpression patterns.
  • The authors state normalization across many resources was challenging, only Illumina TruSeq small RNA or similarly processed samples were included, 608 RNA Atlas runs were removed due to poor clustering, some treated cells were initially included, and remaining biases relate to library preparation, treated cells, RNA Atlas exclusion, and cell culture passaging.
  • The paper notes that cell-type specificity depends on the cell types compared, many cell-specific miRNAs were lowly expressed (~100–500 RPM), tissue analysis conveyed only a general idea of cellular contributions without exact deconvolution, and no novel miRNAs were searched.

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

Integrity signals

Registry

Crossref · 10.1093/gigascience/giac083

Cites

53 of 53 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

2 of 47 cited sources checked · 1 supported · 0 flagged · accuracy not yet measured

  • Findingcitation

    The neurons did not display some of the coexpression patterns of miRNAs described in brain tissue.

    “The neurons also did not display some of the coexpression patterns of miRNAs described in brain tissue [25].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [25] Cherone, 2019, Cotargeting among microRNAs in the brain, Genome Res source
  • Findingown data

    The dataset was curated to 2,077 samples from 196 unique cell types derived from 175 separate studies.

    “This dataset was curated down, through shared miRNA expression patterns, to 2,077 samples from 196 unique cell types derived from 175 separate studies.”
    Quote in the paperSubstantial claim
  • Findingown data

    59 miRNAs were designated as infrequent.

    “Among reasonably expressed miRNAs, 108 were designated as cell specific/near specific, 59 as infrequent, 52 as frequent, 54 as near ubiquitous, and 50 as ubiquitous.”
    Quote in the paperNot a substantial claim
  • Findingown data

    miR-199a-5p is enriched in fibroblast foreskin (n = 14).

    “miR-184 (Mir-184_3p) is highly enriched in conjunctival epithelial cells (n = 8), miR-199a-5p (Mir-199-P1-v1_5p*) in fibroblast foreskin (n = 14), and miR-373-3p (Mir-430-P7_3p) in iPSC (induced pluripotent stem cell) fibroblasts (n = 32).”
    Quote in the paperNot a substantial claim
  • Findingown data

    miR-122-5p is nearly exclusive to hepatocytes.

    “For example, miR-122-5p (Mir-122_5p) is nearly exclusive to hepatocytes, while miR-205-5p (Mir-205-P4_5p) is a more generic epithelial miRNA.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. 30
  • May combine two claims. 31

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

    Premisecitation

    miRNAs demonstrate altered expression levels in disease states that may have biomarker potential.

    “They also demonstrate altered expression levels in disease states that may have biomarker potential [3, 4].”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [3] Mitchell, 2008, Circulating microRNAs as stable blood-based markers for cancer detection, Proc Natl Acad Sci source
    • cites [4] Halushka, 2019, Opportunities for microRNAs in the crowded field of cardiovascular biomarkers, Annu Rev Pathol source

  • May combine two claims.

    Findingown data

    Most plasma miRNAs were RBC derived, followed by platelets, mesenchymal cells, and immune cells.

    “Similar to a recent manuscript on cell-free RNA in plasma [40] and consistent with other miRNA-based studies [41–43], we observed that most plasma miRNAs were RBC derived, followed by platelets, mesenchymal cells, and immune cells.”
    Quote in the paperSubstantial claim1 of 4 sources checked
    • cites [40] Vorperian, 2022, Cell types of origin of the cell-free transcriptome, Nat Biotechnol source
    • cites [41] Kirschner, 2011, Haemolysis during sample preparation alters microRNA content of plasma, PLoS One source
    • cites [42] Cheng, 2013, Plasma processing conditions substantially influence circulating microRNA biomarker levels, PLoS One source
    • cites [43] Pritchard, 2012, Blood cell origin of circulating microRNAs: a cautionary note for cancer biomarker studies, Cancer Prev Res source

      The Results report that blood cells could be a major source of plasma miRNAs and that RBC-associated miRNAs increase substantially in hemolyzed plasma. But the Discussion explicitly says the study did not assess the relative contributions of different blood cell types. It therefore does not support RBC predominance or the stated ranking of platelets, mesenchymal cells, and immune cells. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

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

    Findingown data

    Over 40.7 billion reads were processed.

    “Over 40.7 billion reads were processed.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In total, ∼9 billion of the ∼20 billion trimmed reads of the 2,406 runs mapped to guide and passenger miRNAs covering 89.8% of miRNAs in miRBase and 99.5% of all miRNAs in MirGeneDB2.1.

    “In total, ∼9 billion of the ∼20 billion trimmed reads of the 2,406 runs mapped to guide and passenger miRNAs covering 89.8% of miRNAs in miRBase [15] and 99.5% of all miRNAs in MirGeneDB2.1 (99.8% of mature strands) (Supplementary Table S3) [16].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [15] Kozomara, 2011, miRBase: integrating microRNA annotation and deep-sequencing data, Nucleic Acids Res source
    • cites [16] Fromm, 2022, MirGeneDB 2.1: toward a complete sampling of all major animal phyla, Nucleic Acids Res source

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

    Findingreasoning

    This limited the number of studies that appeared in the final analysis.

    “This limited the number of studies that appeared in the final analysis.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    4,184 of the 6,054 datasets were processed through miRge3.0.

    “From the NCBI Sequence Read Archive, we obtained 6,054 human primary cell datasets and processed 4,184 of them through the miRge3.0 small RNA sequencing alignment software.”
    Quote in the paperNot a substantial claim

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

    Findingnone given

    The microRNAome is easily accessible through the UCSC GenomeBrowser or through an R/Bioconductor experimental data package.

    “It is easily accessible through the UCSC GenomeBrowser or through an R/Bioconductor experimental data package.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Of 640 samples from the RNA-Atlas project, 608 nonimmune cells consistently clustered together irrespective of their class/cell type expression and were removed.

    “Further, of 640 samples from the RNA-Atlas project [17], 608 (the nonimmune cells) consistently clustered together irrespective of their class/cell type expression. These were also removed, resulting in 2,077 final samples from 196 cell types.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [17] Lorenzi, 2021, The RNA Atlas expands the catalog of human non-coding RNAs, Nat Biotechnol source

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

    Findingcitation

    miR-1 is a known myomiR with skeletal and cardiac muscle expression, and was noted in a colon sample here and was seen in a prostate sample previously.

    “For example, miR-1 (Mir-1-P1), a known myomiR with skeletal and cardiac muscle expression, was noted in a colon sample here and was seen in a prostate sample previously [38, 39].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [38] Kumar, 2018, Cell-type specific expression of oncogenic and tumor suppressive microRNAs in the human prostate and prostate cancer, Sci Rep source
    • cites [39] Sweetman, 2008, Specific requirements of MRFs for the expression of muscle specific microRNAs, miR-1, miR-206 and miR-133, Dev Biol source

  • May combine two claims.

    Findingown data

    The microRNAome consists of 2,077 samples from 2,406 runs representing 196 cell types, platelets, and plasma and covering 89.8% of miRNAs in the miRBase reference database.

    “It consists of 2,077 samples from 2,406 runs representing 196 cell types, platelets, and plasma and covering 89.8% of miRNAs in the miRBase reference database.”
    Quote in the paperSubstantial claim

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

    Findingown data

    For 173 of the 2,077 samples, there were 329 technical replicates, resulting in 2,406 total runs.

    “For 173 of the 2,077 samples, we had 329 technical replicates, resulting in 2,406 total runs.”
    Quote in the paperNot a substantial claim

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

    Findingnone given

    The data are available through the human cellular microRNAome track at the UCSC Genome Browser and an R/Bioconductor package.

    “The data are available through the human cellular microRNAome track at the UCSC Genome Browser (https://genome.ucsc.edu/cgi-bin/hgHubConnect) and an R/Bioconductor package (https://bioconductor.org/packages/microRNAome/).”
    Quote in the paperSubstantial claim

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

    Findingnone given

    Not all miRNAs were easily placed into one of the five expression-pattern categories.

    “Not all miRNAs were easily placed in a category.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In colon, miR-192-5p/215-5p is expressed exclusively in epithelial cells, while miR-103a-3p/107 is more ubiquitously expressed.

    “For example, in colon, miR-192-5p/215-5p (Mir-192-P2_5p/Mir-192-P1_5p) is expressed exclusively in epithelial cells, while miR-103a-3p/107 (Mir-103-P4_3p/Mir-103-P1_3p) is more ubiquitously expressed.”
    Quote in the paperNot a substantial claim

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

    Findingreasoning

    Assay type filtering resulted in 6,054 runs that were downloaded.

    “This resulted in 6,054 runs that were downloaded using fasterq-dump/fastq-dump of the NCBI SRA Tookit (version 2.9.2) (https://github.com/ncbi/sra-tools) [46].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [46] Kodama, 2012, The Sequence Read Archive: explosive growth of sequencing data, Nucleic Acids Res source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    The majority (∼46%) of the reads are mature miRNAs (∼9 billion), of which ∼95.8% are guide miRNA reads.

    “The read distribution across various small RNA types is provided in Fig. 1B, where the majority (∼46%) of the reads are mature miRNAs (∼9 billion), of which ∼95.8% are guide miRNA reads.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    The search of the NCBI Sequence Read Archive on 22 January 2021 yielded 58,117 runs corresponding to 1,872 studies.

    “This search was performed on 22 January 2021 and yielded 58,117 runs corresponding to 1,872 studies (Fig. 1A).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    There was a very minor miRNA signal for brain-enriched miRNA, miR-9, and the contributions of neurons and astrocytes to the plasma miRNAs were estimated at 0.43% and 0.89%, respectively.

    “Of note, there was a very minor miRNA signal for brain-enriched miRNA, miR-9 (Mir-9-P1), and the contributions of neurons and astrocytes to the plasma miRNAs were estimated at 0.43% and 0.89%, respectively.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    An average of 462 mature MirGeneDB miRNAs were identified across each of the 13 cell classes with the median range between 200 and 350 miRNAs.

    “An average of 462 mature MirGeneDB miRNAs were identified across each of the 13 cell classes with the median range between 200 and 350 miRNAs (Fig. 1D).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The neurons in this study all had high levels of miR-9, a classic neural maker, and also shared many miRNAs with stem cells and were notable for being adjacent to stem cells in the UMAP.

    “In fact, while the neurons in this study all had high levels of miR-9 (Mir-9-P1), a classic neural maker, they also shared many miRNAs with stem cells and were notable for being adjacent to stem cells in the UMAP of Fig. 2.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    The average number of miRNAs detected across all 2,077 samples was 550.

    “The average number of miRNAs detected across all 2,077 samples was 550.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The highest classification accuracy was for immune cells (99%), while the lowest accuracy was for neurons (93.6%), where ∼6% matched fibroblasts rather than neurons.

    “The highest accuracy was for immune cells (99%), while the lowest accuracy was for neurons (93.6%), where ∼6% matched fibroblasts, rather than neurons (Table 2).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The majority (108) of evaluated miRNAs (323) were considered 'cell specific/near specific,' but many of these were more lowly expressed (∼100-500 RPM).

    “The majority (108) of evaluated miRNAs (323) were considered “cell specific/near specific,” but many of these were more lowly expressed (∼100–500 RPM).”
    Quote in the paperSubstantial claim

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

    Findingown data

    The processed dataset includes 329 duplicate runs of 173 samples, for a total of 2,406 runs.

    “We also included 329 duplicate runs of 173 samples, for a total of 2,406 runs processed.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The top 10 highest expressed miRNAs in each of 4 tissue samples are both from specific cells and generally expressed across numerous cells.

    “As seen in Fig. 5, the top 10 highest expressed miRNAs in each sample are both from specific cells and generally expressed across numerous cells.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The main cell types of the four tissue samples (colon, liver, spleen, lymph node) are present in the dataset.

  • May combine two claims.

    Interpretationown data

    Patterns of cell enrichment and ubiquitousness are similar to those reported earlier with a few new associations based on new-to-this-study cell types.

    “Patterns of cell enrichment and ubiquitousness are similar to those reported earlier with a few new associations based on new-to-this-study cell types.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    The short, 7- to 8-nt seed sequence for target recognition has innumerable potential targets in the genome, resulting in a vast overinterpretation of miRNA regulatory roles.

    “One is the short, 7- to 8-nt seed sequence for target recognition that has innumerable potential targets in the genome, resulting in a vast overinterpretation of miRNA regulatory roles [5].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [5] Pinzon, 2017, microRNA target prediction programs predict many false positives, Genome Res source

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    The authors chose to only include samples that had library preparations using Illumina TruSeq small RNA kits or that appeared similarly processed.

    “We chose to only include samples that had library preparations using Illumina TruSeq small RNA kits or that appeared similarly processed.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    miRNAs are short, ∼22-nt, critical regulatory elements that repress protein translation and promote degradation of messenger RNA.

    “MicroRNAs (miRNAs) are short, ∼22-nt, critical regulatory elements that repress protein translation and promote degradation of messenger RNA (mRNA) [1, 2].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [1] Bartel, 2004, MicroRNAs: genomics, biogenesis, mechanism, and function, Cell source

      Not checked in the free run.

    • cites [2] Mendell, 2012, MicroRNAs in stress signaling and human disease, Cell source

      Not checked in the free run.

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

    Premisenone given

    miRNAs are recognized as functional regulators of development and cellular biology.

    “miRNAs are recognized as functional regulators of development and cellular biology.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Previously, the authors described a cellular microRNAome based upon 46 primary cell types from 161 samples.

    “Previously, we described a cellular microRNAome based upon 46 primary cell types from 161 samples [12].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [12] McCall, 2017, Toward the human cellular microRNAome, Genome Res source

      Not checked in the free run.

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

    Premisecitation

    FANTOM5 reported data from 123 cell types from 304 samples.

    “Separately, FANTOM5 reported data from 123 cell types from 304 samples [11].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [11] de Rie, 2017, An integrated expression atlas of miRNAs and their promoters in human and mouse, Nat Biotechnol source

      Not checked in the free run.

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

    Study design and conductnone given

    Metadata curation excluded any sequencing runs that corresponded to paired-end sequencing, cancer cells, exosomes, and non-Illumina sequencing platforms.

    “This curation excluded any sequencing runs that corresponded to paired-end sequencing, cancer cells, exosomes, and non-Illumina sequencing platforms.”
    Quote in the paperSubstantial claim

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

    Study design and conductcitation

    All runs were evaluated for adapter sequences, and any samples with barcodes, unique molecular identifiers (UMIs), or adapter sequences on both ends were not processed (n = 1,870 runs were removed).

    “All runs were evaluated for adapter sequences, and any samples with barcodes, unique molecular identifiers (UMIs), or adapter sequences on both ends were not processed (n = 1,870 runs were removed) due to the use of miREC in the processing [47].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [47] Zhang, 2021, Aberration-corrected ultrafine analysis of miRNA reads at single-base resolution: a k-mer lattice approach, Nucleic Acids Res source

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

    Study design and conductreasoning

    In the tissue analysis, only a general idea of cellular contributions is conveyed as the exact composition of each tissue with a deconvolution technique was not employed.

    “In our tissue analysis, only a general idea of cellular contributions is conveyed as the exact composition of each tissue with a deconvolution technique was not employed.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Too many 'novel miRNAs' lack clear miRNA features and have only served to further complicate the miRNA field.

    “Too many “novel miRNAs” lack clear miRNA features and have only served to further complicate the miRNA field [17, 35, 36].”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [17] Lorenzi, 2021, The RNA Atlas expands the catalog of human non-coding RNAs, Nat Biotechnol source
    • cites [35] Londin, 2015, Analysis of 13 cell types reveals evidence for the expression of numerous novel primate- and tissue-specific microRNAs, Proc Natl Acad Sci source

      Not checked in the free run.

    • cites [36] Blanco-Dominguez, 2021, A novel circulating microRNA for the detection of acute myocarditis, N Engl J Med source

      Not checked in the free run.

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

    Premisecitation

    It is unclear what short RNA sequences should legitimately be considered bona fide miRNAs versus some other form of non-miRNA.

    “A second is what short RNA sequences should legitimately be considered bona fide miRNAs versus some other form of non-miRNA that has yet to be accurately characterized [6].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [6] Fromm, 2020, Quo vadis microRNAs?, Trends Genet source

  • Classified as probably not a substantial claim.

    Premisenone given

    Some cell types, neurons in particular, are cultured only as a derivation from a stem cell.

    “Additionally, some cell types, neurons in particular, are cultured only as a derivation from a stem cell.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationnone given

    The complexity of cellular microRNA expression estimates recapitulates tissue expression patterns and informs on the miRNA composition of plasma.

    “The complexity of cellular microRNA expression estimates recapitulates tissue expression patterns and informs on the miRNA composition of plasma.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductreasoning

    Because normalization occurred on all samples before leave-one-study-out cross-validation, there was a common bias toward increasing homogeneity in the ComBat-Seq, RUVr, and RUVg approaches, likely inflating overall accuracy but not affecting accuracy rank.

    “As normalization occurred on all samples prior to the leave-one-out approach, there was a common bias toward increasing homogeneity of samples in all of the ComBat-Seq, RUVr, and RUVg approaches, likely inflating overall accuracy but not affecting accuracy rank.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationnone given

    The study presents the largest human cellular microRNAome project generated to date, which largely agrees with and expands upon prior knowledge of cell-type miRNA expression patterns.

    “In conclusion, we present the largest human cellular microRNAome project generated to date, which largely agrees with and expands upon prior knowledge of cell-type miRNA expression patterns.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisereasoning

    The distribution of miRNA expression among cell types is unclear.

    “A third is the unclear distribution of miRNA expression among cell types.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Early publications of general miRNA expression focused on tissues, which comprise numerous cell types, and the localization of miRNAs, whether cell specific or ubiquitous, has not been established.

    “The second is that early publications of general miRNA expression focused on tissues, which comprise numerous cell types, and the localization of miRNAs, whether cell specific or ubiquitous, has not been established [8, 9].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [8] Lagos-Quintana, 2002, Identification of tissue-specific microRNAs from mouse, Curr Biol source
    • cites [9] Landgraf, 2007, A mammalian microRNA expression atlas based on small RNA library sequencing, Cell source

  • May combine two claims.

    Study design and conductreasoning

    The cellular microRNAome has several important biases related to the library preparation method, inclusion of some treated cells, exclusion of most RNA Atlas samples, and cell culture passaging rather than in vivo isolation.

    “Thus, our cellular microRNAome has several important biases. These relate to the library preparation method, inclusion of some treated cells, exclusion of most RNA Atlas samples, and cell culture passaging rather than in vivo isolation.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    MirGeneDB identifies the evolution origin of each miRNA as a node of origin for either the individual miRNA (locus) or the miRNA family (family).

    “MirGeneDB identifies the evolution origin of each miRNA as a node of origin for either the individual miRNA (locus) or the miRNA family (family) [16].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [16] Fromm, 2022, MirGeneDB 2.1: toward a complete sampling of all major animal phyla, Nucleic Acids Res source

  • May combine two claims.

    Interpretationown data

    The study generally replicated miRNA expression estimates from prior studies while significantly adding to the number of samples upon which those estimates are based.

    “It generally replicated miRNA expression estimates from prior studies [11, 12, 23], while significantly adding to the number of samples upon which those estimates are based.”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [11] de Rie, 2017, An integrated expression atlas of miRNAs and their promoters in human and mouse, Nat Biotechnol source

      Not checked in the free run.

    • cites [12] McCall, 2017, Toward the human cellular microRNAome, Genome Res source

      Not checked in the free run.

    • cites [23] Halushka, 2018, Big strides in cellular microRNA expression, Trends Genet source

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    The cell-type specificity of any miRNA is dependent on the sample types to be compared.

    “The cell-type specificity of any miRNA is dependent on the sample types to be compared.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationcitation

    A tissue signal is a composite of a number of different cell types, and it can easily but wrongfully be assumed that one's miRNA of interest is present in a cell type of interest if its expression estimate is only obtained from tissues.

    “As noted herein, a tissue signal is a composite of a number of different cell types, and it can easily but wrongfully be assumed that one's miRNA of interest is present in a cell type of interest if its expression estimate is only obtained from tissues [33].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [33] Kent, 2014, Lessons from miR-143/145: the importance of cell-type localization of miRNAs, Nucleic Acids Res source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    miRGeneDB has clearly established itself as the resource for bona fide miRNAs, while miRBase still includes scores of dubiously identified miRNAs.

    “miRGeneDB has clearly established itself as the resource for bona fide miRNAs, while miRBase still includes scores of dubiously identified miRNAs [6, 14, 16, 31, 32].”
    Quote in the paperSubstantial claim0 of 5 sources checked
    • cites [6] Fromm, 2020, Quo vadis microRNAs?, Trends Genet source
    • cites [14] Fromm, 2020, MirGeneDB 2.0: the metazoan microRNA complement, Nucleic Acids Res source
    • cites [16] Fromm, 2022, MirGeneDB 2.1: toward a complete sampling of all major animal phyla, Nucleic Acids Res source
    • cites [31] Griffiths-Jones, 2006, miRBase: the microRNA sequence database, Methods Mol Biol source
    • cites [32] Kozomara, 2019, miRBase: from microRNA sequences to function, Nucleic Acids Res source

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

    Interpretationreasoning

    An absence of miR-1 in any reasonably expressed cell type from these tissues suggests an alternative cell state or simple gap in cellular coverage of tissues.

    “Skeletal and cardiac myocytes are not believed to be in these tissues, and an absence of miR-1 (Mir-1-P1) in any reasonably expressed cell type from these tissues suggests an alternative cell state or simple gap in our cellular coverage of tissues.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Skeletal and cardiac myocytes are not present in colon or prostate tissues.

  • Classified as probably not a substantial claim.

    Premisecitation

    RUVg requires ubiquitous miRNAs from across the datasets to serve as spike-in controls.

    “RUVg requires ubiquitous miRNAs from across the datasets to serve as spike-in controls.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [50] Risso, 2014, Normalization of RNA-seq data using factor analysis of control genes or samples, Nat Biotechnol source
51 of 51

  • The paper relies on this without support.

    Skeletal and cardiac myocytes are not present in colon or prostate tissues.

    “Skeletal and cardiac myocytes are not believed to be in these tissues”

  • The paper relies on this without support.

    The 30 representative cell types used for CIBERSORT deconvolution are sufficient to estimate the major contributors to plasma miRNA.

    “We deconvoluted 85 plasma samples from 30 representative cell types using CIBERSORT [21] and determined the major contributors to plasma miRNA are RBCs and platelets (38%) (Fig. 5E, F).”

  • The paper relies on this without support.

    The main cell types of the four tissue samples (colon, liver, spleen, lymph node) are present in the dataset.

    “we obtained miRNA expression estimates of 4 tissue samples (colon, liver, spleen, lymph node) for which the main cell types are present in our dataset”
3 of 3

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.

  • Figure 5 describes 139 plasma samples, but the analysis used 85

    Minor

    The Figure 5 caption says its plasma estimates cover 139 samples, while both the results and methods specify deconvolution of 85 plasma samples.

  • The cell-specific category is called a majority despite being smaller than the other categories combined

    Minor

    Only 108 of the 323 evaluated miRNAs were classified as cell specific or near specific. The four other reported categories total 215, so 108 is not a majority.

  • The stated tissue-to-cell ratio cannot trigger the described cap

    Minor

    The methods define the ratio as tissue RPM divided by cell-type RPM. If cell-type expression exceeds tissue expression, that ratio is below 1, so it cannot need capping at 1.25 as the next sentence states.

3 of 3

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Patil, A. H., Baran, A., Brehm, Z. P., McCall, M. N., & Halushka, M. K.
Affiliations
  • Johns Hopkins University (US)
  • University of Rochester Medicine (US)
Corresponding author
Halushka, Marc K
ORCID
Venue
GigaScience
Publisher
Oxford University Press (OUP)
Published
2022
Volume
11
ISSN
2047-217X (ISSN-L)
DOI
10.1093/gigascience/giac083
PMC
PMC9404528
OpenAlex
W4293091797
Semantic Scholar
e161d7e1db51436be012c36ba2ac1c915024296e
Cited by
30
References
53
Citations by year
2026: 7 · 2025: 7 · 2024: 9 · 2023: 6 · 2022: 1
FWCI
2.05
Type
journal article
Language
en
Open access
Licence
Locations
  • GigaScience (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
Funders
  • National Heart, Lung, and Blood Institute (R01HL137811)
  • National Institute of General Medical Sciences (R01GM130564, R01GM139928)
  • University of Rochester
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › MicroRNA in disease regulation
  • Life Sciences › Agricultural and Biological Sciences › Plant Science › Plant Molecular Biology Research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Cancer-related molecular mechanisms research
Keywords
microRNA expression, miRNAome, small RNA sequencing, UCSC Genome Browser, Bioconductor package
MeSH terms
Humans, Software, Sequence Alignment, Genome, Sequence Analysis, RNA, MicroRNAs

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 2026-11-07.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-11-07.

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