Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Analysis
  • Published:

Inference of age-associated transcription factor regulatory activity changes in single cells

Nature Aging volume 2, pages 548–561 (2022)Cite this article

Subjects

Abstract

Transcription factors (TFs) control cell identity and function. How their activity is altered during healthy aging is critical for an improved understanding of aging and disease risk, yet relatively little is known about such changes at cell-type resolution. Here we present and validate a TF activity estimation method for single cells from the hematopoietic system that is based on TF regulons, and apply it to a mouse single-cell RNA-sequencing atlas, to infer age-associated differentiation activity changes in the immune cells of different organs. This revealed an age-associated signature of macrophage dedifferentiation, which is shared across tissue types, and aggravated in tumor-associated macrophages. By extending the analysis to all major cell types, we reveal cell-type and tissue-type-independent age-associated alterations to regulatory factors controlling antigen processing, inflammation, collagen processing and circadian rhythm, that are implicated in age-related diseases. Finally, our study highlights the limitations of using TF expression to infer age-associated changes, underscoring the need to use regulatory activity inference methods.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Derivation and validation of hematopoietic transcription factor regulons.
Fig. 2: Validation of hematopoietic transcription factor regulons in scRNA-seq data.
Fig. 3: Age-associated differentiation activity changes in myeloid cells.
Fig. 4: Age-associated macrophage signature in lung tumor macrophages.
Fig. 5: Consistent age-associated increased or decreased regulatory activity patterns across cell and tissue types.
Fig. 6: Validation of age-associated DOROTHEA results in an independent skin aging 10x scRNA-seq dataset.
Fig. 7: Improved sensitivity of transcription factor activity over transcription factor expression to detect age-related variation.

Similar content being viewed by others

Data availability

Data analyzed in this manuscript is publicly available from the GEO (www.ncbi.nlm.nih.gov/geo/) under accession numbers GSE56046 (DNAm data) and GSE130973 (skin aging scRNA-seq dataset), from EBI ArrayExpress (https://www.ebi.ac.uk/arrayexpress) under accession numbers E-MTAB-6149 and E-MTAB-6653 (lung cancer scRNA-seq dataset). FACS-sorted expression data from purified blood cell subtypes are available from https://haemosphere.org/datasets/show. The lung cancer bulk RNA-seq dataset is available from the TCGA data portal (http://tcgaportal.org/). The TMS scRNA-seq data are available from https://doi.org/10.6084/m9.figshare.8273102.v2. The regulons for the 328 hematopoietic TFs and their cell-specific lineage information are provided in Supplementary Data 1. The DOROTHEA regulons are available from https://saezlab.github.io/dorothea. ChIP–seq data were downloaded from the ChIP–seq Atlas (https://chip-atlas.org/). Processed data are available from the corresponding author upon reasonable request. Source data are provided with this paper.

Code availability

SCIRA functions for estimating TFA are available from the scira R package (http://github.com/aet21/scira/). We also provide an R markdown file and associated data objects from the figshare repository (https://figshare.com/articles/software/R-markdown_file_and_data_objects_for_estimating_TFA_in_TMS_lung-tissue_scRNA-seq_dataset/17167085) that illustrate in a few examples how to estimate TFA and how to correlate it with age in the TMS dataset.

References

  1. Lopez-Otin, C., Blasco, M. A., Partridge, L., Serrano, M. & Kroemer, G. The hallmarks of aging. Cell 153, 1194–1217 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. Donertas, H. M., Fabian, D. K., Valenzuela, M. F., Partridge, L. & Thornton, J. M. Common genetic associations between age-related diseases. Nat Aging 1, 400–412 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  3. Fabian, D. K., Fuentealba, M., Donertas, H. M., Partridge, L. & Thornton, J. M. Functional conservation in genes and pathways linking ageing and immunity. Immun. Ageing 18, 23 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  4. Vijg, J. & Kennedy, B. K. The essence of aging. Gerontology 62, 381–385 (2016).

    Article  PubMed  Google Scholar 

  5. Brunauer, R., Alavez, S. & Kennedy, B. K. Stem cell models: a guide to understand and mitigate aging? Gerontology 63, 84–90 (2017).

    Article  CAS  PubMed  Google Scholar 

  6. Graf, T. & Enver, T. Forcing cells to change lineages. Nature 462, 587–594 (2009).

    Article  CAS  PubMed  Google Scholar 

  7. Yamanaka, S. & Blau, H. M. Nuclear reprogramming to a pluripotent state by three approaches. Nature 465, 704–712 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Zhou, X., Sen, I., Lin, X. X. & Riedel, C. G. Regulation of age-related decline by transcription factors and their crosstalk with the epigenome. Curr. Genomics 19, 464–482 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Palmer, D., Fabris, F., Doherty, A., Freitas, A. A. & de Magalhaes, J. P. Ageing transcriptome meta-analysis reveals similarities and differences between key mammalian tissues. Aging 13, 3313–3341 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Tabula Muris, C. A single-cell transcriptomic atlas characterizes ageing tissues in the mouse. Nature 583, 590–595 (2020).

    Article  CAS  Google Scholar 

  11. Martinez-Jimenez, C. P. et al. Aging increases cell-to-cell transcriptional variability upon immune stimulation. Science 355, 1433–1436 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Levy, O. et al. Age-related loss of gene-to-gene transcriptional coordination among single cells. Nat Metab. 2, 1305–1315 (2020).

    Article  CAS  PubMed  Google Scholar 

  13. Teschendorff, A. E. & Wang, N. Improved detection of tumor suppressor events in single-cell RNA-seq data. NPJ Genom. Med. 5, 43 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Holland, C. H. et al. Robustness and applicability of transcription factor and pathway analysis tools on single-cell RNA-seq data. Genome Biol. 21, 36 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Basso, K. et al. Reverse engineering of regulatory networks in human B cells. Nat. Genet. 37, 382–390 (2005).

    Article  CAS  PubMed  Google Scholar 

  17. Margolin, A. A. et al. ARACNE: an algorithm for the reconstruction of gene regulatory networks in a mammalian cellular context. BMC Bioinformatics 7, S7 (2006).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  18. Chen, Y., Widschwendter, M. & Teschendorff, A. E. Systems-epigenomics inference of transcription factor activity implicates aryl-hydrocarbon-receptor inactivation as a key event in lung cancer development. Genome Biol. 18, 236 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  19. Garcia-Alonso, L., Holland, C. H., Ibrahim, M. M., Turei, D. & Saez-Rodriguez, J. Benchmark and integration of resources for the estimation of human transcription factor activities. Genome Res. 29, 1363–1375 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. de Graaf, C. A. et al. Haemopedia: an expression atlas of murine hematopoietic cells. Stem Cell Reports 7, 571–582 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

  21. Kanamori, M. et al. A genome-wide and nonredundant mouse transcription factor database. Biochem. Biophys. Res. Commun. 322, 787–793 (2004).

    Article  CAS  PubMed  Google Scholar 

  22. Oki, S. et al. ChIP-Atlas: a data-mining suite powered by full integration of public ChIP–seq data. EMBO Rep. 19, e46255 (2018).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  23. Jojic, V. et al. Identification of transcriptional regulators in the mouse immune system. Nat. Immunol. 14, 633–643 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Heng, T. S., Painter, M. W. & Immunological Genome Project, C. The Immunological Genome Project: networks of gene expression in immune cells. Nat. Immunol. 9, 1091–1094 (2008).

    Article  CAS  PubMed  Google Scholar 

  25. Tabula Muris, C. et al. Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris. Nature 562, 367–372 (2018).

    Article  CAS  Google Scholar 

  26. Setty, M. et al. Characterization of cell fate probabilities in single-cell data with Palantir. Nat. Biotechnol. 37, 451–460 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Elyahu, Y. et al. Aging promotes reorganization of the CD4 T cell landscape toward extreme regulatory and effector phenotypes. Sci. Adv. 5, eaaw8330 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Jonkman, T. H. et al. Functional genomics analysis identifies T and NK cell activation as a driver of epigenetic clock progression. Genome Biol. 23, 24 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Moskowitz, D. M. et al. Epigenomics of human CD8 T cell differentiation and aging. Sci. Immunol. 2, eaag0191 (2017).

    Article  Google Scholar 

  30. Hu, B. et al. Distinct age-related epigenetic signatures in CD4 and CD8 T Cells. Front. Immunol. 11, 585168 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Hu, B. et al. Transcription factor networks in aged naive CD4 T cells bias lineage differentiation. Aging Cell 18, e12957 (2019).

    PubMed  PubMed Central  Google Scholar 

  32. Gustafson, C. E., Cavanagh, M. M., Jin, J., Weyand, C. M. & Goronzy, J. J. Functional pathways regulated by microRNA networks in CD8 T cell aging. Aging Cell 18, e12879 (2019).

    Article  PubMed  CAS  Google Scholar 

  33. Kurachi, M. et al. The transcription factor BATF operates as an essential differentiation checkpoint in early effector CD8+ T cells. Nat. Immunol. 15, 373–383 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Quigley, M. et al. Transcriptional analysis of HIV-specific CD8+ T cells shows that PD-1 inhibits T cell function by upregulating BATF. Nat. Med. 16, 1147–1151 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Sokalski, K. M. et al. Deletion of genes encoding PU.1 and Spi-B in B cells impairs differentiation and induces pre-B cell acute lymphoblastic leukemia. Blood 118, 2801–2808 (2011).

    Article  CAS  PubMed  Google Scholar 

  36. Butcher, S., Chahel, H. & Lord, J. M. Review article: ageing and the neutrophil: no appetite for killing? Immunology 100, 411–416 (2000).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Feinberg, M. W. et al. The Kruppel-like factor KLF4 is a critical regulator of monocyte differentiation. EMBO J. 26, 4138–4148 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Alder, J. K. et al. Kruppel-like factor 4 is essential for inflammatory monocyte differentiation in vivo. J. Immunol. 180, 5645–5652 (2008).

    Article  CAS  PubMed  Google Scholar 

  39. Reynolds, L. M. et al. Age-related variations in the methylome associated with gene expression in human monocytes and T cells. Nat. Commun. 5, 5366 (2014).

    Article  PubMed  Google Scholar 

  40. Feinberg, M. W. et al. Kruppel-like factor 4 is a mediator of proinflammatory signaling in macrophages. J. Biol. Chem. 280, 38247–38258 (2005).

    Article  CAS  PubMed  Google Scholar 

  41. Liao, X. et al. Kruppel-like factor 4 regulates macrophage polarization. J. Clin. Invest. 121, 2736–2749 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Luo, X. et al. Kruppel-like factor 4 is a regulator of proinflammatory signaling in fibroblast-like synoviocytes through Increased IL-6 expression. Mediators Inflamm. 2016, 1062586 (2016).

    PubMed  PubMed Central  Google Scholar 

  43. Dann, E., Henderson, N. C., Teichmann, S. A., Morgan, M. D. & Marioni, J. C. Differential abundance testing on single-cell data using k-nearest neighbor graphs. Nat. Biotechnol. 40, 245–253 (2021).

  44. Lambrechts, D. et al. Phenotype molding of stromal cells in the lung tumor microenvironment. Nat. Med. 24, 1277–1289 (2018).

    Article  CAS  PubMed  Google Scholar 

  45. Cancer Genome Atlas Research Network. Comprehensive molecular profiling of lung adenocarcinoma. Nature 511, 543–550 (2014).

    Article  CAS  Google Scholar 

  46. Kriegsman, B. A. et al. Frequent loss of IRF2 in cancers leads to Immune evasion through decreased MHC class I antigen presentation and increased PD-L1 expression. J. Immunol. 203, 1999–2010 (2019).

    Article  CAS  PubMed  Google Scholar 

  47. Alvarez, M. J. et al. Functional characterization of somatic mutations in cancer using network-based inference of protein activity. Nat. Genet. 48, 838–847 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Jo, A. et al. PARIS farnesylation prevents neurodegeneration in models of Parkinson’s disease. Sci. Transl. Med. 13, eaax8891 (2021).

    Article  CAS  PubMed  Google Scholar 

  49. Podolsky, M. J. et al. Age-dependent regulation of cell-mediated collagen turnover. JCI Insight 5, e137519 (2020).

    Article  PubMed Central  Google Scholar 

  50. Xia, J. et al. SIRT1 deacetylates RFX5 and antagonizes repression of collagen type I (COL1A2) transcription in smooth muscle cells. Biochem. Biophys. Res. Commun. 428, 264–270 (2012).

    Article  CAS  PubMed  Google Scholar 

  51. Donato, A. J., Morgan, R. G., Walker, A. E. & Lesniewski, L. A. Cellular and molecular biology of aging endothelial cells. J. Mol. Cellular Cardiol. 89, 122–135 (2015).

    Article  CAS  Google Scholar 

  52. Wang, Y. et al. Global transcriptomic changes occur in aged mouse podocytes. Kidney Int. 98, 1160–1173 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. O’Brown, Z. K., Van Nostrand, E. L., Higgins, J. P. & Kim, S. K. The inflammatory transcription factors NFκB, STAT1 and STAT3 drive age-associated transcriptional changes in the human kidney. PLoS Genet. 11, e1005734 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  54. Thompson, J. C. et al. Gene signature of antigen processing and presentation machinery predicts response to checkpoint blockade in non-small cell lung cancer and melanoma. J. Immunother. Cancer 8, e000974 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  55. Solé-Boldo, L. et al. Single-cell transcriptomes of the human skin reveal age-related loss of fibroblast priming. Commun Biol. 3, 188 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  56. Li, Z. & Rasmussen, L. J. TIP60 in aging and neurodegeneration. Ageing Res. Rev. 64, 101195 (2020).

    Article  CAS  PubMed  Google Scholar 

  57. Schaum, N. et al. Ageing hallmarks exhibit organ-specific temporal signatures. Nature 583, 596–602 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Krausgruber, T. et al. IRF5 promotes inflammatory macrophage polarization and TH1–TH17 responses. Nat. Immunol. 12, 231–238 (2011).

    Article  CAS  PubMed  Google Scholar 

  59. Mantovani, A. & Longo, D. L. Macrophage checkpoint blockade in cancer back to the future. N. Engl. J. Med. 379, 1777–1779 (2018).

    Article  PubMed  Google Scholar 

  60. Galdiero, M. R. et al. Tumor-associated macrophages and neutrophils in cancer. Immunobiology 218, 1402–1410 (2013).

    Article  CAS  PubMed  Google Scholar 

  61. Porta, C. et al. Macrophages in cancer and infectious diseases: the ‘good’ and the ‘bad’. Immunotherapy 3, 1185–1202 (2011).

    Article  CAS  PubMed  Google Scholar 

  62. Chen, H. H. et al. IRF2BP2 reduces macrophage inflammation and susceptibility to atherosclerosis. Circ. Res. 117, 671–683 (2015).

    Article  CAS  PubMed  Google Scholar 

  63. Frieler, R. A. et al. Myeloid-specific deletion of the mineralocorticoid receptor reduces infarct volume and alters inflammation during cerebral ischemia. Stroke 42, 179–185 (2011).

    Article  PubMed  Google Scholar 

  64. Cruz, S. A. et al. Loss of IRF2BP2 in microglia increases inflammation and functional deficits after focal ischemic brain injury. Front. Cell Neurosci. 11, 201 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  65. Xu, L. S. et al. ETV6–RUNX1 interacts with a region in SPIB intron 1 to regulate gene expression in pre-B cell acute lymphoblastic leukemia. Exp. Hematol. 73, 50–63 (2019).

    Article  CAS  PubMed  Google Scholar 

  66. Cartwright, T., Perkins, N. D. & Wilson, C. L. NFKB1: a suppressor of inflammation, ageing and cancer. FEBS J. 283, 1812–1822 (2016).

    Article  CAS  PubMed  Google Scholar 

  67. Concetti, J. & Wilson, C. L. NFKB1 and cancer: friend or foe? Cells 7, 133 (2018).

    Article  CAS  PubMed Central  Google Scholar 

  68. Acosta-Rodriguez, V. A., Rijo-Ferreira, F., Green, C. B. & Takahashi, J. S. Importance of circadian timing for aging and longevity. Nat. Commun. 12, 2862 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  69. Tonsfeldt, K. J. et al. The contribution of the circadian gene Bmal1 to female fertility and the generation of the preovulatory luteinizing hormone surge. J. Endocr. Soc. 3, 716–733 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  70. Rakshit, K. & Matveyenko, A. V. Induction of core circadian clock transcription factor Bmal1 enhances beta cell function and protects against obesity-induced glucose intolerance. Diabetes 70, 143–154 (2021).

    Article  CAS  PubMed  Google Scholar 

  71. Breen, D. P. et al. Sleep and circadian rhythm regulation in early Parkinson disease. JAMA Neurol. 71, 589–595 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  72. Hulme, B. et al. Epigenetic regulation of BMAL1 with sleep disturbances and Alzheimer’s disease. J. Alzheimers Dis. 77, 1783–1792 (2020).

    Article  CAS  PubMed  Google Scholar 

  73. Ehlen, J. C. et al. Bmal1 function in skeletal muscle regulates sleep. eLife 6, e26557 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  74. Oh, G. et al. Circadian oscillations of cytosine modification in humans contribute to epigenetic variability, aging, and complex disease. Genome Biol. 20, 2 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  75. Oh, E. S. & Petronis, A. Origins of human disease: the chrono-epigenetic perspective. Nat. Rev. Genet. 22, 533–546 (2021).

  76. Babagana, M. et al. Hedgehog dysregulation contributes to tissue-specific inflammaging of resident macrophages. Aging 13, 19207–19229 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  77. Horvath, S. DNA methylation age of human tissues and cell types. Genome Biol. 14, R115 (2013).

    Article  PubMed  PubMed Central  Google Scholar 

  78. Smyth, G. K. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3, 3 (2004).

    Article  Google Scholar 

  79. Zhuang, J., Widschwendter, M. & Teschendorff, A. E. A comparison of feature selection and classification methods in DNA methylation studies using the Illumina Infinium platform. BMC Bioinformatics 13, 59 (2012).

    Article  CAS  Google Scholar 

  80. Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies and species. Nat. Biotechnol. 36, 411–420 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  81. Jablonski, K. A. et al. Novel markers to delineate murine M1 and M2 macrophages. PLoS ONE 10, e0145342 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  82. Aryee, M. J. et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics 30, 1363–1369 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  83. Troyanskaya, O. et al. Missing value estimation methods for DNA microarrays. Bioinformatics 17, 520–525 (2001).

    Article  CAS  PubMed  Google Scholar 

  84. Teschendorff, A. E. et al. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450k DNA methylation data. Bioinformatics 29, 189–196 (2013).

    Article  CAS  PubMed  Google Scholar 

  85. Yang, Z., Jones, A., Widschwendter, M. & Teschendorff, A. E. An integrative pan-cancer-wide analysis of epigenetic enzymes reveals universal patterns of epigenomic deregulation in cancer. Genome Biol. 16, 140 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

Download references

Acknowledgements

This work was supported by National Science Foundation of China grants (31771464, 32170652 and 31970632). We thank the TMS consortium for making their data open access and to everyone else who shares data in public repositories.

Author information

Authors and Affiliations

  1. CAS Key Laboratory of Computational Biology, Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China

    Alok K. Maity, Xue Hu, Tianyu Zhu & Andrew E. Teschendorff

  2. UCL Cancer Institute, Paul O’Gorman Building, University College London, London, UK

    Andrew E. Teschendorff

Authors
  1. Alok K. Maity
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Xue Hu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Tianyu Zhu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Andrew E. Teschendorff
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

A.E.T. conceived the study, performed the statistical analyses and wrote the manuscript. A.K.M., X.H. and T.Z. helped with analyses.

Corresponding author

Correspondence to Andrew E. Teschendorff.

Ethics declarations

Competing interests

The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. The authors declare no competing interests.

Peer review

Peer review information

Nature Aging thanks Amir Bashan, Joao Pedro de Magalhaes and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Comparison of validation accuracies between SCIRA and Dorothea and validation of mouse hematopoietic TF-regulons.

a) Barplots compare validation accuracies of estimated regulatory activities for blood-cell-lineage specific TF-regulons, obtained using the TF-regulons from SCIRA or those from the DOROTHEA database, as indicated. Top-panel is for mouse RNA-Seq ImmGen dataset which profiled purified blood cell subtypes from the various lineages as shown. Lower-panel is for the mouse ImmGen2 dataset. B = B-cell-lineage, DC = dendritic cell, Mon/Mac=Monocyte/Macrophage lineage, MPP = multipotent progenitor, Neu=Neutrophil, NK = Natural Killer, RPP = Restricted potency progenitor, T = T-cell lineage, All=accuracy obtained over all lineages. b) Diffusion maps on the left display diffusion components DC1 & DC2, of the transcription factor activity (TFA) matrix defined over the hematopoietic TFs and 4142 human bone marrow cells from Setty et al, with cells colored by cell-type as annotated in Setty et al (HSC = hematopoietic stem cell, Pre=precursor/multipotent progenitor, Mono=monocytes, DC = dendritic cells, Ery=Erythroid cells, top panel), and with cells colored by estimated diffusion pseudotime (DPT, lower panel). TFA was estimated using our SCIRA-derived TF-regulons. Tip-points are marked in red, displaying the starting branch point in the lower left corner, and the two terminal branch points representing the endpoints of the erythroid (green dashed line) and myeloid (red dashed line) branches. The diffusion maps to the right have cells colored by the estimated TFA of TFs that are known to be specific for HSCs (HOXB5), monocytes (CEBPB), erythrocytes (GATA1) and dendritic cells (IRF8) (top-panel). Lower panel displays the TFA of these same TFs as a function of diffusion pseudotime along their respective branches.

Extended Data Fig. 2 Cross-tissue comparison of age-associated TFA patterns lymphoid cells.

a) Heatmap panels display the age-associated patterns of differentiation activity (TFA) for lineage-specific TFs in different tissues and in corresponding cell-types of that lineage. Clockwise, panels represent the patterns for T-cell, B-cell and NK lineages. Each heatmap displays the signed statistical significance level as indicated, where the P-value is derived from the t-statistic of a regression of TFA against age, adjusting for sex. b) Violin plot displaying estimated TFA vs age-group for Lef1 in lung Cd4 + T-cells. Pearson Correlation Coefficient (PCC) and associated P-value are shown. c) Scatterplot of the t-statistics of association of TFA with age for T-cell specific TFs as derived in CD4 + T-cells from lung (x-axis) vs. T-cells from spleen (y-axis). TFs significant in both tissues are colored. Pearson Correlation Coefficient and P-value between the two tissues is given. d) Violin plot displaying estimated TFA vs age-group for Spib in lung B-cells. Pearson Correlation Coefficient (PCC) and associated P-value are shown.

Extended Data Fig. 3 Cross-tissue comparison of age-associated TFA patterns neutrophils.

Heatmap displays age-associated differentiation activity (TFA) changes for neutrophils from lung and spleen, as inferred from the Tabula Muris Senis dataset. Scatterplot displays the corresponding t-statistics of association of TFA with age. TFs colored in red are significant in both tissues. Pearson Correlation Coefficient (PCC) and associated two-sided P-value between lung and spleen is given.

Extended Data Fig. 4 Age-associated DNA methylation at macrophage-specific TFs in purified monocyte samples.

For 4 macrophage-specific TFs that exhibit dysregulation of regulatory activity with age in monocytes/macrophages, as inferred from the TMS 10X scRNA-seq dataset, we display the signed statistical significance (y-axis) of age-association of Illumina 450k DNA methylation probesm as derived from the 1200 purified monocyte sample set from Reynolds et al. In the y-axis we display the sign of the age-associated t-statistic times the negative log10[Q-value], where Q-value is the FDR-estimate. X-axis labels the genomic position of the probes. Probes in red are significantly associated with age (Q < 0.05).

Extended Data Fig. 5 Validation of hematopoetic TF regulons in human datasets and TFA patterns of macrophage-specific TFs in lung adenocarcinoma.

a-b) Validation of hematopoetic TF regulons in human datasets. a) Barplots displaying the validation accuracy of blood cell-type specific SCIRA TF-regulons in 3 independent human FACS mRNA expression datasets (Ebert, Schultze & de Graf) from the Haemopedia resource. Accuracy was estimated as the fraction of cell-type specific TFs whose regulons predicted a significantly higher TFA in the corresponding blood cell types compared to all other cell-types, as assessed using a one-sided t-test (P < 0.05). Only cell-types with at least 5 samples were included in this analysis. Total number of cell-sorted samples per dataset were: n = 211 (Ebert), n = 384 (Schultze), n = 42 (de Graf). b) Left panel: tSNE-diagram of a 10X scRNA-Seq peripheral blood mononuclear cell (PBMCs) dataset from Zheng et al. Right panel: Barplot displaying the validation accuracy of blood cell type specific TF-regulons derived from cell-sorted bulk expression data from Haemopedia in the scRNA-Seq PBMC data from Zheng et al. c-e) Pattern of TFA of macrophage-specific TFs in lung adenocarcinoma TCGA set. c) Color bars displaying the t-statistics (t) and P-values (P) between the TFA of 11 macrophage-specific TFs and normal-cancer status using only paired samples (n = 45 pairs). Thus, cyan indicates lower TFA in tumor vs normal. d) As a), but now adjusted for macrophage marker (LYZ & CD14) expression. e) Kaplan Meier overall survival curves for all primary LUAD samples, with samples stratified into low, middle and high tertiles according to IRF2 TFA activity. Hazard Ratio and chi-square test P-value derive from a Cox-regression of IRF2 TFA (treated as continuous variable).

Extended Data Fig. 6 Regulon size comparison between DOROTHEA and SCIRA and their influence on age-associations.

a) Violin plots compare the regulon sizes of the DOROTHEA regulons vs the hematopoietic TF regulons built with SCIRA. The number of TF-regulons in each group is displayed below. Scatterplots to the right plot the regulon size (x-axis) vs. the fraction of regulon genes where the interaction between TF and gene is positive (fPOS, y-axis). b) Panels display scatterplots of statistical significance (-log10P, y-axis) vs AUC (x-axis) derived from a two-sided Wilcoxon-test comparing the regulon size of DOROTHEA TF-regulons that were significantly associated with age (Linear model P < 0.001) vs the size of TF-regulons that were not associated with age, as assessed in cell-types within the given tissue. Each datapoint represents one distinct cell-type within the tissue. The green dashed-line indicates the line AUC = 0.5. Points to the right (AUC > 0.5) are cell-types for which the regulon sizes of age-associated TF-regulons were higher compared to non age-associated ones, with the y-axis indicating statistical significance level. c) Boxplots display the SCIRA TF regulon size vs a binary variable indicating whether the TFA derived from the TF-regulon was significantly associated with age (SigTF) or not (NonSigTF), as assessed using a P < 0.001 threshold. The number of TF-regulons in each category is given below each boxplot. The P-value comparing the regulon sizes between the significant and non-significant regulons is indicated within each panel and is derived from a two-tailed Wilcoxon rank sum test. The bar within each boxplot represents the median, the box itself the interquartile range (IQR), and whiskers extend to 1.5 times the IQR.

Extended Data Fig. 7 Age-associated changes of TFA for selected TFs across 75 cell-types.

Heatmap displays the signed significance of the age-associated t-statistics of TFA with age for 6 selected TFs across 75 cell-types and 11 tissue types as indicated. The 6 TFs were selected on the basis of displaying highly significant skews towards either increased or decreased regulatory activity with age.

Extended Data Fig. 8 Rfx5 activity changes in heart cells and inverse association with Col1a2 expression.

a) Violin plots display the regulatory activity (TFA) of Rfx5 against age [months] for 4 cell-types in heart, as indicated. The number of cells at each timepoint is indicated above violin plot. P-value shown is from a linear regression of TFA against age, adjusting for sex. b) Heatmap of significance statistics for Rfx5 TFA against age (including Col1a2 in its regulon), Rfx5 TFA against age not including Col1a2 in its regulon (‘no Col1a2’) and for Col1a2 expression against age.

Extended Data Fig. 9 Validation of Nfkb1 increased activity in T-cells using CD4 + T-cell NFKB1 ChIP-Seq targets (5 kb).

Left panels: Genome plots display the signed statistical significance of NFKB1 target expression with age, as assessed in T-cells from the given tissue-type. Gene targets within 5 kb of the NFKB1 ChIP-Seq peak are shown with the color indicating increased (magenta) or decreased (cyan) mRNA expression with age. Alternating grey/black bars indicate successive mouse chromosomes. Right panels: Violin plots of the corresponding Spearman Rank Correlation Coefficients (SCC), with the P-value derived from a one-tailed t-test, testing for the null that the average SCC = 0. The number of ChIP-Seq targets is 126. Similar results are seen when defining NFKB1 ChIP-Seq targets as having a peak within 10 kb and 1 kb of the gene.

Extended Data Fig. 10 Association of antigen processing machinery (APM) and Rfx5 regulatory activity with age.

a) Boxplots display the Spearman rank correlation coefficient (SCC) between the gene-expression of APM genes (n = 16) with age, stratified according to tissue and cell-type. Boxes shown in magenta (cyan) indicate cell-types where Rfx5 activity increased (decreased) significantly with age (P < 0.05). P-values above boxplots derive from a two-tailed t-test, testing if the average SCC values are significantly different from zero. P-values in red are those with P < 0.05. Number of datapoints in each boxplot is 16. The bar within each boxplot represents the median, the box itself the interquartile range (IQR), and whiskers extend to 1.5 times the IQR. b) A scatterplot of the average SCC with age (x-axis) against the t-statistic of association between Rfx5 regulatory activity and age (y-axis), with each datapoint representing one tissue-cell-type pair (as in a)). Datapoints highlighted in red indicate consistent significant associations, those in blue indicate inconsistent associations. Overall Spearman Correlation Coefficient (SCC) and two-sided P-value are given.

Supplementary information

Reporting Summary

Supplementary Tables 1–6

Supplementary Data 1

Regulons for all 328 hematopoietic TFs, with columns of the data matrix labeling the TFs.

Source data

Source Data Fig. 1

Statistical source data (enrichment table for hematopoietic TFs with ChIP–seq data, with columns labeling the TF, number of regulon genes, number of direct targets, AUC and associated one-sided P value)

Source Data Fig. 2

Statistical source data (validation accuracies summarized by cell lineage, tissue and technology type, and method.

Source Data Fig. 3

Statistical source data (significance levels, that is, sign(t-statistic) × −log10(P value) of age associations for macrophage-specific TFs and macrophage/monocyte populations in various tissues).

Source Data Fig. 4

Statistical source data (TFA values for the 11 macrophage-specific TFs and all lung macrophages, with cells labeled by barcode and normal/tumor status.

Source Data Fig. 5

Statistical source data (significance levels, that is, sign(t-statistic) × −log10(P value) of age associations for 168 selected TFs and all cell types considered).

Source Data Fig. 6

Statistical source data (TFA values for DOROTHEA TFs across keratinocyte cells, cells labeled by barcode and age group).

Source Data Fig. 7

Statistical source data (z-statistics of differential TFA and DE for ARNTL).

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Maity, A.K., Hu, X., Zhu, T. et al. Inference of age-associated transcription factor regulatory activity changes in single cells. Nat Aging 2, 548–561 (2022). https://doi.org/10.1038/s43587-022-00233-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s43587-022-00233-9

This article is cited by

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing