CoChat
Sign Up
PreprintCited by 100Single-cell and spatial transcriptomicsGene Regulatory Network AnalysisGene expression and cancer classification

The DOI matches its Crossref record; its reference list is not available.

Biologically informed deep learning to infer gene program activity in single cells

Open access
Lotfollahi, Mohammad, Rybakov, Sergei, Hrovatin, Karin, Hediyeh-zadeh, Soroor, Talavera-López, Carlos, Misharin, Alexander V, Theis, Fabian J.·2022·Nature Cell Biology
DOI 10.1101/2022.02.05.479217PMC PMC9928587

Preprint — published as 10.1038/s41556-022-01072-x · Paper page

Open this paper in CoChat — free account

Abstract

The increasing availability of large-scale single-cell datasets has enabled the detailed description of cell states across multiple biological conditions and perturbations. In parallel, recent advances in unsupervised machine learning, particularly in transfer learning, have enabled fast and scalable mapping of these new single-cell datasets onto reference atlases. The resulting large-scale machine learning models however often have millions of parameters, rendering interpretation of the newly mapped datasets challenging. Here, we propose expiMap, a deep learning model that enables interpretable reference mapping using biologically understandable entities, such as curated sets of genes and gene programs. The key concept is the substitution of the uninterpretable nodes in an autoencoder’s bottleneck by labeled nodes mapping to interpretable lists of genes, such as gene ontologies, biological pathways, or curated gene sets, for which activities are learned as constraints during reconstruction. This is enabled by the incorporation of predefined gene programs into the reference model, and at the same time allowing the model to learn de novo new programs and refine existing programs during reference mapping. We show that the model retains similar integration performance as existing methods while providing a biologically interpretable framework for understanding cellular behavior. We demonstrate the capabilities of expiMap by applying it to 15 datasets encompassing five different tissues and species. The interpretable nature of the mapping revealed unreported associations between interferon signaling via the RIG-I/MDA5 and GPCRs pathways, with differential behavior in CD8 + T cells and CD14 + monocytes in severe COVID-19, as well as the role of annexins in the cellular communications between lymphoid and myeloid compartments for explaining patient response to the applied drugs. Finally, expiMap enabled the direct comparison of a diverse set of pancreatic beta cells from multiple studies where we observed a strong, previously unreported correlation between the unfolded protein response and asparagine N-linked glycosylation. Altogether, expiMap enables the interpretable mapping of single cell transcriptome data sets across cohorts, disease states and other perturbations.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1101/2022.02.05.479217

Cites

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

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

The self-check needs the paper’s full text.

Citations

No citation lists for this paper

No citation lists for this paper yet — checked again daily.

References 123 · Cited by 100

Research works printed but not linked: 130

  1. 2021, Single-cell CUT&Tag profiles histone modifications and transcription factors in complex tissues, Nat. Biotechnol., 39, 825, 10.1038/s41587-021-00869-9 · DOI · Paper page
  2. 10.1038/nmeth.4380 · DOI · Paper page
  3. Stoeckius, M. et al. Cell ‘hashing’ with barcoded antibodies enables multiplexing and doublet detection for single cell genomics. bioRxiv 237693 (2017). · DOI
  4. 8/2019, scGen predicts single-cell perturbation responses, Nat. Methods, 16, 715
  5. Lotfollahi, M. et al. Compositional perturbation autoencoder for single-cell response modeling. bioRxiv 2021.04.14.439903 (2021). · DOI
  6. Gayoso, A. et al. scvi-tools: a library for deep probabilistic analysis of single-cell omics data. bioRxiv 2021.04.28.441833 (2021). · DOI
  7. Hao, Y. et al. Integrated analysis of multimodal single-cell data. http://biorxiv.org/lookup/doi/10.1101/2020.10.12.335331 (2020). · DOI · Paper page
  8. 2021, RNA velocity—current challenges and future perspectives, Mol. Syst. Biol., 17, e10282, 10.15252/msb.202110282 · DOI · Paper page
  9. 10.7554/eLife.27041 · DOI · Paper page
  10. 2020, Cells of the adult human heart, Nature, 588, 466, 10.1038/s41586-020-2797-4 · DOI · Paper page
  11. 10.1038/s41593-020-00794-1 · DOI · Paper page
  12. 10.1038/s41586-020-2922-4 · DOI
  13. 2021, Mapping the evolution of T cell states during response and resistance to adoptive cellular therapy, Cell Rep., 37, 109992, 10.1016/j.celrep.2021.109992 · DOI
  14. Lotfollahi, M. et al. Mapping single-cell data to reference atlases by transfer learning. Nat. Biotechnol. 1–10 (2021). · DOI · Paper page
  15. 10.1186/s13059-020-1926-6 · DOI · Paper page
  16. 2021, Efficient and precise single-cell reference atlas mapping with Symphony, Nat. Commun., 12, 5890, 10.1038/s41467-021-25957-x · DOI
  17. 2019, Fast, sensitive and accurate integration of single-cell data with Harmony, Nat. Methods, 16, 1289, 10.1038/s41592-019-0619-0 · DOI · Paper page
  18. 2021, Differentiable biology: using deep learning for biophysics-based and data-driven modeling of molecular mechanisms, Nat. Methods, 18, 1169, 10.1038/s41592-021-01283-4 · DOI · Paper page
  19. Lotfollahi, M. , Dony, L. , Agarwala, H. & Theis, F. J. Out-of-distribution prediction with disentangled representations for single-cell RNA sequencing data . in Workshop on Computational Biology, ICML (unknown, 2020). · DOI
  20. 2018, in Advances in Neural Information Processing Systems, 31, 6114
  21. 2021, MichiGAN: sampling from disentangled representations of single-cell data using generative adversarial networks, Genome Biol., 22, 158, 10.1186/s13059-021-02373-4 · DOI · Paper page
  22. Luecken, M. D. & Theis, F. J. Current best practices in single-cell RNA-seq analysis: a tutorial. Mol. Syst. Biol. 15, (06/2019). · DOI · Paper page
  23. 2018, False signals induced by single-cell imputation, F1000Res., 7, 1740, 10.12688/f1000research.16613.1 · DOI · Paper page
  24. 2021, Biologically informed deep neural network for prostate cancer discovery, Nature, 598, 348, 10.1038/s41586-021-03922-4 · DOI · Paper page
  25. 10.1093/bioinformatics/btaa800 · DOI · Paper page
  26. Sohn, K. , Lee, H. & Yan, X. Learning Structured Output Representation using Deep Conditional Generative Models. in Advances in Neural Information Processing Systems (eds. Cortes, C. Lawrence, N. Lee, D. Sugiyama, M. & Garnett, R. ) vol. 28 (Curran Associates, Inc. 2015).
  27. Vaswani, A. et al. Attention Is All You Need. arXiv [cs.CL] (2017).
  28. 2019, Feature selection and dimension reduction for single-cell RNA-Seq based on a multinomial model, Genome Biol, 20, 295, 10.1186/s13059-019-1861-6 · DOI · Paper page
  29. 2020, Benchmarking principal component analysis for large-scale single-cell RNA-sequencing, Genome Biol., 21, 9, 10.1186/s13059-019-1900-3 · DOI
  30. 2018, Integrative analysis of single-cell genomics data by coupled nonnegative matrix factorizations, 115, 7723
  31. 10.1186/s12918-019-0699-6 · DOI
  32. 12/2018, Deep generative modeling for single-cell transcriptomics, Nat. Methods, 15, 1053
  33. Xu, C. et al. Probabilistic Harmonization and Annotation of Single-cell Transcriptomics Data with Deep Generative Models. https://biorxiv.org/lookup/doi/10.1101/532895 (2019). · DOI
  34. 2017, Reactome pathway analysis: a high-performance in-memory approach, BMC Bioinformatics, 18, 142, 10.1186/s12859-017-1559-2 · DOI
  35. Franzén, O. , Gan, L.-M. & Björkegren, J. L. M. PanglaoDB: a web server for exploration of mouse and human single-cell RNA sequencing data. Database 2019, (2019). · DOI · Paper page
  36. 2019, BioReader: a text mining tool for performing classification of biomedical literature, BMC Bioinformatics, 19, 57, 10.1186/s12859-019-2607-x · DOI · Paper page
  37. Burgess, C. P. et al. Understanding disentangling in β-VAE. arXiv [stat.ML] (2018).
  38. 2007, A kernel statistical test of independence, Nips, 20, 585
  39. Oetjen, K. A. et al. Human bone marrow assessment by single-cell RNA sequencing, mass cytometry, and flow cytometry. JCI Insight 3, (2018). · DOI · Paper page
  40. 2018, Comparison of clustering tools in R for medium-sized 10x Genomics single-cell RNA-sequencing data, F1000Research, 7, 1297, 10.12688/f1000research.15809.1 · DOI · Paper page
  41. 2019, A Bayesian mixture model for clustering droplet-based single-cell transcriptomic data from population studies, Nat. Commun., 10, 1649, 10.1038/s41467-019-09639-3 · DOI
  42. 10.1146/annurev.biochem.67.1.227 · DOI · Paper page
  43. 10.1016/j.cell.2015.12.032 · DOI · Paper page
  44. 10.3389/fimmu.2018.00053 · DOI
  45. 2017, Role of cellular metabolism in regulating type I interferon responses: Implications for tumour immunology and treatment, Cancer Lett., 409, 20, 10.1016/j.canlet.2017.08.037 · DOI
  46. 10.3389/fimmu.2016.00630 · DOI
  47. 10.1038/s41592-021-01336-8 · DOI · Paper page
  48. Cremer, C. , Li, X. & Duvenaud, D. Inference Suboptimality in Variational Autoencoders. arXiv [cs.LG] (2018).
  49. 2021, Sfaira accelerates data and model reuse in single cell genomics, Genome Biol., 22, 248, 10.1186/s13059-021-02452-6 · DOI · Paper page
  50. 2020, Interpretable factor models of single-cell RNA-seq via variational autoencoders, Bioinformatics, 36, 3418, 10.1093/bioinformatics/btaa169 · DOI
  51. 10.1186/s13059-017-1382-0 · DOI · Paper page
  52. 10.1038/s41467-020-17834-w · DOI
  53. 10.1016/j.immuni.2011.05.003 · DOI · Paper page
  54. 2020, The Complement C5a-C5aR1 GPCR Axis in COVID-19 Therapeutics, Trends Immunol., 41, 965, 10.1016/j.it.2020.09.008 · DOI
  55. 2021, RIG-I triggers a signaling-abortive anti-SARS-CoV-2 defense in human lung cells, Nat. Immunol., 22, 820, 10.1038/s41590-021-00942-0 · DOI
  56. 2021, Inference and analysis of cell-cell communication using CellChat, Nat. Commun, 12, 1088, 10.1038/s41467-021-21246-9 · DOI · Paper page
  57. 2016, Annexins family: insights into their functions and potential role in pathogenesis of sarcoidosis, J. Transl. Med, 14, 89, 10.1186/s12967-016-0843-7 · DOI
  58. 10.1183/13993003.00918-2021 · DOI
  59. 2021, Annexin A1 as a potential prognostic biomarker for COVID-19 disease: Case-control study, Int. J. Clin. Pract., 75, e14606
  60. 10.1038/s12276-020-00518-2 · DOI
  61. 2020, Cytokine Storm in COVID-19: The Current Evidence and Treatment Strategies, Front. Immunol, 11, 1708, 10.3389/fimmu.2020.01708 · DOI · Paper page
  62. 2021, CD81 marks immature and dedifferentiated pancreatic β-cells, Mol Metab, 49, 101188, 10.1016/j.molmet.2021.101188 · DOI
  63. 2020, Beta Cell Dedifferentiation Induced by IRE1α Deletion Prevents Type 1 Diabetes, Cell Metab., 31, 822, 10.1016/j.cmet.2020.03.002 · DOI · Paper page
  64. 2020, Single-cell RNA-seq with spike-in cells enables accurate quantification of cell-specific drug effects in pancreatic islets, Genome Biol., 21, 106, 10.1186/s13059-020-02006-2 · DOI
  65. 10.1038/s42255-020-0171-3 · DOI · Paper page
  66. 2021, Tutorial: guidelines for annotating single-cell transcriptomic maps using automated and manual methods, Nat. Protoc., 16, 2749, 10.1038/s41596-021-00534-0 · DOI
  67. 10.1016/j.csbj.2021.01.015 · DOI · Paper page
  68. 10.1093/nar/gkz1031 · DOI
  69. 10.1007/s11892-019-1194-6 · DOI
  70. 2017, The dynamic plasticity of insulin production in β-cells, Molecular Metabolism, 6, 958, 10.1016/j.molmet.2017.04.010 · DOI · Paper page
  71. 2018, Delta-like Ligand-4-Notch Signaling Inhibition Regulates Pancreatic Islet Function and Insulin Secretion, Cell Rep., 22, 895, 10.1016/j.celrep.2017.12.076 · DOI
  72. 10.1371/journal.pone.0027445 · DOI · Paper page
  73. 2019, Islet inflammation in type 2 diabetes. Semin. Immunopathol, 41, 501
  74. 10.3389/fgene.2017.00062 · DOI · Paper page
  75. 10.1186/s12859-019-2599-6 · DOI · Paper page
  76. 10.2337/db15-1633 · DOI
  77. Mustapha, S. et al. Current Status of Endoplasmic Reticulum Stress in Type II Diabetes. Molecules 26, (2021). · DOI
  78. 10.1038/s41581-019-0129-4 · DOI · Paper page
  79. 10.1002/1873-3468.13495 · DOI
  80. 2018, XBP1s activation can globally remodel N-glycan structure distribution patterns, 115, E10089
  81. 2020, Type 1 diabetes mellitus as a disease of the β-cell (do not blame the immune system?), Nat. Rev. Endocrinol., 17, 150
  82. 10.1101/2020.12.02.401182 · DOI
  83. 10.1186/s13059-017-1334-8 · DOI
  84. 2021, VEGA is an interpretable generative model for inferring biological network activity in single-cell transcriptomics, Nat. Commun, 12, 5684, 10.1038/s41467-021-26017-0 · DOI · Paper page
  85. 2021, Learning interpretable cellular and gene signature embeddings from single-cell transcriptomic data, Nat. Commun., 12, 5261, 10.1038/s41467-021-25534-2 · DOI
  86. 10.1101/2021.01.28.428664 · DOI · Paper page
  87. 10.1038/nmeth.4077 · DOI
  88. 2021, Causal integration of multi-omics data with prior knowledge to generate mechanistic hypotheses, Mol. Syst. Biol., 17, e9730, 10.15252/msb.20209730 · DOI · Paper page
  89. 2020, Knowledge-primed neural networks enable biologically interpretable deep learning on single-cell sequencing data, Genome Biol., 21, 190, 10.1186/s13059-020-02100-5 · DOI · Paper page
  90. 2018, Using deep learning to model the hierarchical structure and function of a cell, Nat. Methods, 15, 290, 10.1038/nmeth.4627 · DOI · Paper page
  91. Lotfollahi, M. , Litinetskaya, A. & Theis, F. pMultigrate: single-cell multi-omic data integration. ICML 2021 Workshop on Computational Biology (WCB) Proceedings Paper (2021). · DOI
  92. 2021, Joint probabilistic modeling of single-cell multi-omic data with totalVI, Nat. Methods, 18, 272, 10.1038/s41592-020-01050-x · DOI · Paper page
  93. 10.1101/2021.06.24.449733 · DOI
  94. 10.1038/s41586-021-03634-9 · DOI · Paper page
  95. 2020, CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes, Nat. Protoc., 15, 1484, 10.1038/s41596-020-0292-x · DOI · Paper page
  96. Chen, Y. , Lun, A. T. , & Smyth, G. K. (2016). From reads to genes to pathways: differential expression analysis of RNA-Seq experiments using Rsubread and the edgeR quasi-likelihood pipeline. F1000Research, 5. · DOI
  97. 10.1093/nar/gkv007 · DOI · Paper page
  98. Kingma, D. P. & Welling, M . Auto-encoding variational bayes. arXiv:1312.6114 (2013). · arXiv
  99. (2015), Learning structured output representation using deep conditional generative models, Advances in neural information processing systems, 28, 3483
  100. Ainsworth, S. K. , Foti, N. J. , Lee, A. K. C. & Fox, E. B . oi-VAE: Out-put interpretable VAEs for nonlinear group factor analysis. In Dy, J. & Krause, A . (eds.) Proceedings of the 35th International Conference on Machine Learning, vol. 80 of Proceedings of Machine Learning Research, 119–128 (PMLR, 2018).
  101. Lopez, R. , Regier, J. , Yosef, N. & Jordan, M. I . Information constraints on auto-encoding variational bayes. In NeurIPS (2018).
  102. Beck, A . First-Order Methods in Optimization (SIAM, Philadelphia, 2017). · DOI
  103. Yu, Y . On decomposing the proximal map. In Proceedings of the 26th International Conference on Neural Information Processing Systems - Volume 1, NIPS’13, 91–99 (Curran Associates Inc., Red Hook, NY, USA, 2013).
  104. Lotfollahi, M. et al. Mapping single-cell data to reference atlases by transfer learning. Nature Biotechnology 1–10 (2021). · DOI · Paper page
  105. (2018), Deep generative modeling for single-cell transcriptomics, Nature Methods, 15, 1053, 10.1038/s41592-018-0229-2 · DOI · Paper page
  106. Lücken, M. et al. Benchmarking atlas-level data integration in single-cell genomics. Nature Methods 19 (2022). · DOI · Paper page
  107. 10.1186/1465-6906-11-S1-I1 · DOI
  108. 10.1093/bioinformatics/btp616 · DOI · Paper page
  109. Chen, Y. , Lun, A. T. & Smyth, G. K . From reads to genes to pathways: differential expression analysis of rna-seq experiments using rsubread and the edger quasi-likelihood pipeline. F1000Research 5 (2016). · DOI
  110. 10.1093/nar/gkv007 · DOI · Paper page
  111. 10.1186/gb-2014-15-1-r1 · DOI
  112. 10.1093/nar/gkz1031 · DOI
  113. 10.1073/pnas.0506580102 · DOI · Paper page
  114. 10.1093/bioinformatics/btr260 · DOI · Paper page
  115. Oetjen, K. A. et al. Human bone marrow assessment by single-cell rna sequencing, mass cytometry, and flow cytometry. JCI insight 3 (2018). · DOI · Paper page
  116. Freytag, S. , Tian, L. , Lönnstedt, I. , Ng, M. & Bahlo, M . Comparison of clustering tools in r for medium-sized 10x genomics single-cell rna-sequencing data. F1000Research 7 (2018). · DOI · Paper page
  117. (2019), A bayesian mixture model for clustering droplet-based single-cell transcriptomic data from population studies, Nature Communications, 10, 1
  118. Kang, H. M. et al. Multiplexed droplet single-cell rna-sequencing using natural genetic variation. In Nature Biotechnology (2017). · DOI · Paper page
  119. Guo, C. et al. Single-cell analysis of two severe covid-19 patients reveals a monocyte-associated and tocilizumab-responding cytokine storm. Nature Communications 11 (2020). · DOI
  120. Jin, S. et al. Inference and analysis of cell-cell communication using cellchat. Nature Communications 12 (2021). · DOI · Paper page
  121. 10.1186/s13059-017-1382-0 · DOI · Paper page
  122. Virshup, I. , Rybakov, S. , Theis, F. , Angerer, P. & Wolf, F . anndata: Annotated data (2021). · DOI
  123. (2019), Panglaodb: a web server for exploration of mouse and human single-cell rna sequencing data, Database The Journal of Biological Databases and Curation, 2019, 46
  124. Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184 (2021). · DOI · Paper page
  125. Kang, J. et al. Efficient and precise single-cell reference atlas mapping with symphony. Nature Communications 12 (2021). · DOI
  126. Litviňuková, M. et al. Cells of the adult human heart. Nature 588 (2020). · DOI · Paper page
  127. (2020), A molecular cell atlas of the human lung from single-cell rna sequencing, Nature, 587, 1
  128. 10.1016/j.cell.2019.06.029 · DOI
  129. 10.1038/s41586-019-1684-3 · DOI
  130. Fischer, D. et al. Sfaira accelerates data and model reuse in single cell genomics. Genome Biology 22 (2021). · DOI · Paper page

Metadata

Authors
Lotfollahi, M., Rybakov, S., Hrovatin, K., Hediyeh-zadeh, S., Talavera-López, C., Misharin, A. V., & Theis, F. J.
Affiliations
  • Wellcome Sanger Institute (GB)
  • Helmholtz Munich (DE)
  • Walter and Eliza Hall Institute of Medical Research (AU)
  • Northwestern University (US)
Corresponding author
Theis, Fabian J.
ORCID
Venue
Nature Cell Biology
Publisher
openRxiv
Published
2022-02-07
Volume
25
Issue
2
Pages
337-350
ISSN
1465-7392 (ISSN-L)
DOI
10.1101/2022.02.05.479217
PMC
PMC9928587
OpenAlex
W4318981406
Cited by
100
References
123
Citations by year
2026: 11 · 2025: 32 · 2024: 35 · 2023: 20 · 2022: 2
FWCI
8.73
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Nature Cell Biology (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PuSH - Publication Server of Helmholtz Zentrum München (submittedVersion, cc-by)
Funders
  • Joachim Herz Stiftung
  • European Commission (874656, HORIZON2020)
  • Bundesministerium für Bildung und Forschung (01IS18036B, 01IS18036A)
  • Helmholtz Artificial Intelligence Cooperation Unit (ZT-I-PF-5-01)
  • National Institute on Aging (P01AG049665)
  • National Heart, Lung, and Blood Institute (R01HL153312)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene Regulatory Network Analysis
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
Keywords
gene programs, single-cell atlas, interpretability, deep learning, cell-type resolution, cross-species comparison
MeSH terms
Deep Learning, COVID-19, Humans, 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.

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