The Importance of Computational Modeling in Stem Cell Research
2020; Elsevier BV; Volume: 39; Issue: 2 Linguagem: Inglês
10.1016/j.tibtech.2020.07.006
ISSN0167-9430
Autores Tópico(s)Cancer Genomics and Diagnostics
ResumoIn the era of single-cell big data, computational modeling is a powerful tool to describe biological systems and generate predictions across different spatial and temporal scales. The steadily increasing amount of data allows the development of models that link different levels of biological organization, such as intracellular interactions, cellular behavior, and the behavior of cell populations.The development of single-cell–based mechanistic models is necessary to better characterize biological processes and generate more accurate predictions of cellular conversion factors, cell identity transcription factors, and cell–cell interactions relevant for tissue regeneration and homeostasis.We expect these models to accelerate the development of novel regenerative medicine strategies by guiding experimentalists in the design of stem cell transplantation and gene correction therapies. The generation of large amounts of omics data is increasingly enabling not only the processing and analysis of large data sets but also the development of computational models in the field of stem cell research. Although computational models have been proposed in recent decades, we believe that the stem cell community is not fully aware of the potentiality of computational modeling in guiding their experimental research. In this regard, we discuss how single-cell technologies provide the right framework for computational modeling at different scales of biological organization in order to address challenges in the stem cell field and to guide experimentalists in the design of new strategies for stem cell therapies and treatment of congenital disorders. The generation of large amounts of omics data is increasingly enabling not only the processing and analysis of large data sets but also the development of computational models in the field of stem cell research. Although computational models have been proposed in recent decades, we believe that the stem cell community is not fully aware of the potentiality of computational modeling in guiding their experimental research. In this regard, we discuss how single-cell technologies provide the right framework for computational modeling at different scales of biological organization in order to address challenges in the stem cell field and to guide experimentalists in the design of new strategies for stem cell therapies and treatment of congenital disorders. Stem cell research has witnessed a revolution in the last two decades after the discovery of induced pluripotent stem cells (iPSCs). This has opened new opportunities for studying human diseases; designing strategies for tissue regeneration, including cell transplantation (see Glossary) and developmental research. Rapid advances in single-cell approaches allow a detailed characterization of cellular phenotypes across tissues in different conditions, the discovery of new cellular subpopulations, and the reconstruction of single-cell trajectories in development and reprogramming. In particular, the increasing resolution of single-cell RNA sequencing (scRNA-seq) and the emergence of new technologies that generate other types of single-cell phenotypic omics data, such as epigenomes, proteomes, and spatial information, allow the systematic integration and analysis of these data, leading to a more comprehensive characterization of cell type classification, function, and interactions. Despite technical limitations, such as gene dropouts and low capture rates, the analysis of single-cell data attains high statistical power by considering a large number of individual samples and allows the identification of cellular subpopulations at an unprecedented resolution. Massive generation of these multiomics single-cell data enables the development of high-resolution computational models that are able to capture the collective behavior of genes at the molecular level or cells at the tissue level, thus providing an ideal framework to address key questions in the stem cell field. Indeed, computational models can generate novel predictions and provide new insights into biological mechanisms, guiding experimental research. In particular, systems biology models at different levels of complexity, including cellular, tissue, and even organ levels, can be developed to address relevant questions in stem cell research. For example, on the one hand, models at the cellular level, such as gene regulatory network (GRN)–based models, can improve the understanding of cellular differentiation and cellular conversion and can help to predict key transcription factors (TFs) and signaling molecules controlling such processes. On the other hand, models at the tissue level, including those based on cell–cell interaction networks, can be useful for elucidating general principles of tissue homeostasis and regeneration and for generating predictions of relevant cell–cell interaction events supporting the tissue regeneration capacity. In summary, computational models complement statistical data analysis by providing mechanistic insights into biological processes and by generating novel predictions that can guide experimental research. In particular, we believe that a number of biological questions and challenges in the field of stem cell research and regenerative medicine can be addressed with the help of computational models. These models have been shown to be useful in guiding experimental research, particularly in the design of cellular conversion strategies. Nevertheless, the time is ripe for the development of more sophisticated models, considering the continuous improvement and appearance of new single-cell technologies that are currently able to generate large amounts of different types of data. In this regard, we are now able to develop models at different scales of biological complexity, which are more representative of biological processes. For example, a model that considers stem cell–niche interactions is more appropriate for the in vivo prediction of cell fate determinants. Therefore, these models can be particularly useful in advancing regenerative medicine strategies, such as in vivo reprogramming, and stem cell transplantation and rejuvenation by overcoming limitations of current experimental protocols. Stem cell research comprises performing experiments and developing hypotheses that link different scales of biological organization, including intracellular interactions, cellular behavior, and the behavior of cell populations. The aim of multiscale computational modeling is to describe biological systems and generate predictions across these different spatial and temporal scales. The level at which the model should be constructed depends on the scientific question being addressed and the available input data. For example, computational methods that rely on the reconstruction and analysis of intracellular GRNs have been shown to be useful in modeling cellular conversion, enabling the identification of optimal sets of conversion factors (Box 1). Importantly, inference of cell type–specific GRNs is an essential step for these network-based methods; however, it is not always possible to accurately infer these GRNs, especially for newly characterized cell subtypes. In this regard, scRNA-seq is the ideal technology for capturing real interactions between genes in individual cells. Indeed, scRNA-seq captures the gene expression of thousands of individual cells in one experiment, which provides a large number of independent measurements that allow the extraction of information about gene expression heterogeneity across cells and gene–gene coexpression in individual cells. Hence, scRNA-seq allows the inference of cell type– or cell subtype–specific GRNs [1.Chan T.E. et al.Gene regulatory network inference from single-cell data using multivariate information measures.Cell Syst. 2017; 5: 251-267.e3Abstract Full Text Full Text PDF PubMed Scopus (164) Google Scholar, 2.Aibar S. et al.SCENIC: single-cell regulatory network inference and clustering.Nat. Methods. 2017; 14: 1083-1086Crossref PubMed Scopus (790) Google Scholar, 3.Papili Gao N. et al.SINCERITIES: inferring gene regulatory networks from time-stamped single cell transcriptional expression profiles.Bioinformatics. 2018; 34: 258-266Crossref PubMed Scopus (45) Google Scholar, 4.Stumpf P.S. MacArthur B.D. Machine learning of stem cell identities from single-cell expression data via regulatory network archetypes.Front. Genet. 2019; 10: 2Crossref PubMed Scopus (7) Google Scholar, 5.Matsumoto H. et al.SCODE: an efficient regulatory network inference algorithm from single-cell RNA-seq during differentiation.Bioinformatics. 2017; 33: 2314-2321Crossref PubMed Scopus (114) Google Scholar, 6.Woodhouse S. et al.SCNS: a graphical tool for reconstructing executable regulatory networks from single-cell genomic data.BMC Syst. Biol. 2018; 12: 59Crossref PubMed Scopus (25) Google Scholar, 7.Sanchez-Castillo M. et al.A Bayesian framework for the inference of gene regulatory networks from time and pseudo-time series data.Bioinformatics. 2018; 34: 964-970Crossref PubMed Scopus (44) Google Scholar, 8.Aubin-Frankowski P.-C. Vert J.-P. Gene regulation inference from single-cell RNA-seq data with linear differential equations and velocity inference.Bioinformatics. 2020; (Published online June 17, 2020. https://doi.org/10.1093/bioinformatics/btaa576)Crossref PubMed Scopus (12) Google Scholar, 9.Moerman T. et al.GRNBoost2 and Arboreto: efficient and scalable inference of gene regulatory networks.Bioinformatics. 2019; 35: 2159-2161Crossref PubMed Scopus (68) Google Scholar, 10.Deshpande A. et al.Network inference with Granger causality ensembles on single-cell transcriptomic data.bioRxiv. 2019; (Published online January 30, 2019. https://doi.org/10.1101/534834)Google Scholar, 11.Qiu X. et al.Inferring causal gene regulatory networks from coupled single-cell expression dynamics using Scribe.Cell Syst. 2020; 10: 265-274.e11Abstract Full Text Full Text PDF PubMed Scopus (28) Google Scholar], which constitutes an important step in the implementation of network-based methods for cellular conversion [12.Kamimoto K. et al.CellOracle: Dissecting cell identity via network inference and in silico gene perturbation.bioRxiv. 2020; (Published online April 21, 2020. https://doi.org/10.1101/2020.02.17.947416)Google Scholar,13.de Soysa T.Y. et al.Single-cell analysis of cardiogenesis reveals basis for organ-level developmental defects.Nature. 2019; 572: 120-124Crossref PubMed Scopus (71) Google Scholar] (Table 1). In particular, using a machine learning approach, it has been possible to identify different GRN configurations corresponding to different states of pluripotency, including naive and primed states [4.Stumpf P.S. MacArthur B.D. Machine learning of stem cell identities from single-cell expression data via regulatory network archetypes.Front. Genet. 2019; 10: 2Crossref PubMed Scopus (7) Google Scholar]. This knowledge is important not only for characterizing the regulatory program of different stem cell phenotypic states but also for devising new GRN-based strategies to direct stem cell conversion. Indeed, the availability of single-cell–based GRN inference methods has led to the optimization of various cell conversion protocols in the context of differentiation, transdifferentiation, and reprogramming [14.Liu X. et al.Single-cell RNA-seq of the developing cardiac outflow tract reveals convergent development of the vascular smooth muscle cells.Cell Rep. 2019; 28: 1346-1361.e4Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar, 15.Finnegan A. et al.Single-cell transcriptomics reveals spatial and temporal turnover of keratinocyte differentiation regulators.Front. Genet. 2019; 10: 775Crossref PubMed Scopus (20) Google Scholar, 16.Tran K.A. et al.Defining reprogramming checkpoints from single-cell analyses of induced pluripotency.Cell Rep. 2019; 27: 1726-1741.e5Abstract Full Text Full Text PDF PubMed Scopus (16) Google Scholar, 17.Ruan H. et al.Single-cell reconstruction of differentiation trajectory reveals a critical role of ETS1 in human cardiac lineage commitment.BMC Biol. 2019; 17: 89Crossref PubMed Scopus (16) Google Scholar]. For instance, a combination of small molecules was identified to increase the efficiency of embryonic fibroblast to pluripotent stem cell conversion by inferring and analyzing the GRN governing the corresponding reprogramming trajectory [16.Tran K.A. et al.Defining reprogramming checkpoints from single-cell analyses of induced pluripotency.Cell Rep. 2019; 27: 1726-1741.e5Abstract Full Text Full Text PDF PubMed Scopus (16) Google Scholar]. Similarly, single-cell–based network inference in the context of pre–B cell to macrophage transdifferentiation revealed the TF MYC as a crucial determinant of reprogramming efficiency [15.Finnegan A. et al.Single-cell transcriptomics reveals spatial and temporal turnover of keratinocyte differentiation regulators.Front. Genet. 2019; 10: 775Crossref PubMed Scopus (20) Google Scholar]. Moreover, a recent study demonstrated a critical role for ETS1 in cardiomyocyte differentiation based on the reconstruction of GRNs at different developmental time points [17.Ruan H. et al.Single-cell reconstruction of differentiation trajectory reveals a critical role of ETS1 in human cardiac lineage commitment.BMC Biol. 2019; 17: 89Crossref PubMed Scopus (16) Google Scholar]. Despite the recent progress in single-cell–based GRN inference, integration of scRNA-seq with other types of data, such as single-cell assay for transposase-accessible chromatin using sequencing (scATAC-seq) and DNA methylation profiling, would allow more accurate inference of cell type– or cell subtype–specific GRNs and therefore would advance conversion factor predictions (Figure 1).Box 1Applications of Different Modeling Types to Stem Cell ResearchThe development of computational models can aid in addressing key questions in stem cell research. However, the choice of the modeling framework is highly dependent on the specific research question (see Table 1 in main text).Boolean network models of GRNs work well for the identification of cell conversion factors. Although these models do not require the inference of kinetic parameters, they usually include regulatory interactions between large numbers of TFs. Moreover, this modeling framework allows the inference of cooperativity among TFs in regulation, enabling the identification of optimal combinations of TFs controlling cellular conversion.Continuous models of GRNs based on ordinary differential equations are suited for predicting the dynamic behavior of gene expression during biological processes such as cellular differentiation. Indeed, using linear ordinary differential equations is possible to infer the regulatory relationship among several tens of genes from time-series data corresponding, for example, to cellular differentiation. Further, these equations can be employed to predict the continuous dynamic behavior of gene expression in a quantitative manner following perturbation of specific TFs or signaling pathways.Probabilistic models, such as dynamic Bayesian networks or probabilistic Boolean networks, are adequate to simulate stochasticity in gene expression profiles and regulatory interactions in cellular systems, which is one of the determinants of cellular reprogramming efficiency. Hence, these models can be used to prioritize optimal combinations of TFs whose perturbations could induce cellular reprogramming with high efficiency.Directed graphs representing cell–cell communication networks can be inferred from ligand–receptor pair expression without the need to estimate receptor–ligand binding affinity. This type of descriptive model is appropriate to identify relevant features of the structure of cell–cell communication networks to describe how cells interact to guarantee tissue function in homeostasis. In addition, following this approach, it is possible to detect relevant cell–cell interactions that can impair tissue regeneration in disease or aging.Table 1scRNA-Seq-Based Mechanistic ModelsModeling frameworkModel typeApplicationBiological significanceDescriptive modelUndirected graphInference of gene coexpression networks from time-series data using information theoryPrioritization of functional modulesDirected graphInference of cell–cell communication networks from ligand–receptor pair expressionPrediction of tissue cell–cell interactionsLogical modelBoolean networkInference of Boolean model of GRNs based on coexpression modules and TF binding site predictionsPrediction of cell conversion factorsInference of Boolean model of GRNs based on Granger causality analysis of time-series dataPrediction of regulatory interactions governing cellular trajectoriesInference of Boolean logic rules of GRNsIdentification of cooperative TF regulationContinuous modelRegression-basedInference of quantitative model of GRNs based on ridge regression and partial correlationPrediction of gene expression dynamics during differentiationLinear ordinary differential equationsInference of quantitative model of GRNs based on differentiation trajectoriesInference of quantitative model of GRNs based on variation in cell cycle and cell stateProbabilistic modelDynamic Bayesian networkInference of probabilistic GRN model based on prior knowledgePrediction of reprogramming efficiency Open table in a new tab The development of computational models can aid in addressing key questions in stem cell research. However, the choice of the modeling framework is highly dependent on the specific research question (see Table 1 in main text). Boolean network models of GRNs work well for the identification of cell conversion factors. Although these models do not require the inference of kinetic parameters, they usually include regulatory interactions between large numbers of TFs. Moreover, this modeling framework allows the inference of cooperativity among TFs in regulation, enabling the identification of optimal combinations of TFs controlling cellular conversion. Continuous models of GRNs based on ordinary differential equations are suited for predicting the dynamic behavior of gene expression during biological processes such as cellular differentiation. Indeed, using linear ordinary differential equations is possible to infer the regulatory relationship among several tens of genes from time-series data corresponding, for example, to cellular differentiation. Further, these equations can be employed to predict the continuous dynamic behavior of gene expression in a quantitative manner following perturbation of specific TFs or signaling pathways. Probabilistic models, such as dynamic Bayesian networks or probabilistic Boolean networks, are adequate to simulate stochasticity in gene expression profiles and regulatory interactions in cellular systems, which is one of the determinants of cellular reprogramming efficiency. Hence, these models can be used to prioritize optimal combinations of TFs whose perturbations could induce cellular reprogramming with high efficiency. Directed graphs representing cell–cell communication networks can be inferred from ligand–receptor pair expression without the need to estimate receptor–ligand binding affinity. This type of descriptive model is appropriate to identify relevant features of the structure of cell–cell communication networks to describe how cells interact to guarantee tissue function in homeostasis. In addition, following this approach, it is possible to detect relevant cell–cell interactions that can impair tissue regeneration in disease or aging. Complementary computational approaches overcome the intrinsic complexity of GRN inference by solely extracting relevant gene expression patterns that identify cell identity TFs [18.D'Alessio A.C. et al.A systematic approach to identify candidate transcription factors that control cell identity.Stem Cell Rep. 2015; 5: 763-775Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar,19.Okawa S. et al.Transcriptional synergy as an emergent property defining cell subpopulation identity enables population shift.Nat. Commun. 2018; 9: 2595Crossref PubMed Scopus (8) Google Scholar]. Upregulation of these target cell identity TFs in the initial cell type can be used as a strategy for cellular conversion [19.Okawa S. et al.Transcriptional synergy as an emergent property defining cell subpopulation identity enables population shift.Nat. Commun. 2018; 9: 2595Crossref PubMed Scopus (8) Google Scholar]. Nevertheless, these methods do not provide details about gene regulation. Combining the predictions of these computational methods with GRN inference would allow the identification of optimal sets of conversion factors. Modeling stem cell–niche interactions would allow researchers to determine niche signals that maintain aberrant stem cell phenotypes in disease or aging and therefore to design strategies for counteracting the niche effect in these cells. To date, a few computational approaches have been developed that explicitly consider niche effects [20.Saçma M. et al.Haematopoietic stem cells in perisinusoidal niches are protected from ageing.Nat. Cell Biol. 2019; 21: 1309-1320Crossref PubMed Scopus (41) Google Scholar, 21.Yachie-Kinoshita A. et al.Modeling signaling-dependent pluripotency with Boolean logic to predict cell fate transitions.Mol. Syst. Biol. 2018; 14e7952Crossref PubMed Scopus (24) Google Scholar, 22.Mahadik B. et al.A computational model of feedback-mediated hematopoietic stem cell differentiation in vitro.PLoS One. 2019; 14e0212502Crossref PubMed Scopus (7) Google Scholar, 23.Thalheim T. et al.Linking stem cell function and growth pattern of intestinal organoids.Dev. Biol. 2018; 433: 254-261Crossref PubMed Scopus (18) Google Scholar]. However, most of these models have been assembled manually from experimentally validated interactions and are customized to stem cells in a particular niche. By contrast, recent computational approaches that use scRNA-seq to integrate signaling and GRNs provide a general framework of stem cell–niche interactions and have been able to infer the effect of niche signals on target gene expression [24.Browaeys R. et al.NicheNet: modeling intercellular communication by linking ligands to target genes.Nat. Methods. 2020; 17: 159-162Crossref PubMed Scopus (165) Google Scholar,25.Ravichandran S. et al.SigHotSpotter: scRNA-seq-based computational tool to control cell subpopulation phenotypes for cellular rejuvenation strategies.Bioinformatics. 2020; 36: 1963-1965Google Scholar]. In particular, predictions of key signaling molecules that mediate niche signals to maintain cellular phenotypes have been used for cellular rejuvenation by counteracting the effect of the aging niche [26.Kalamakis G. et al.Quiescence modulates stem cell maintenance and regenerative capacity in the aging brain.Cell. 2019; 176: 1407-1419.e14Abstract Full Text Full Text PDF PubMed Scopus (109) Google Scholar]. However, because these computational methods rely solely on scRNA-seq data for the integration of signaling and transcriptional regulatory networks, the identification of signaling molecules that mediate niche cues in stem cells to maintain their phenotypes remains a challenge. In this regard, combining scRNA-seq with bulk phosphoproteomics data from purified cell populations would facilitate the integration of signaling and transcriptional regulatory networks. Furthermore, advances in single-cell molecular profiling technologies, such as phosphoproteomics [27.Qin X. et al.Cell-type-specific signaling networks in heterocellular organoids.Nat. Methods. 2020; 17: 335-342Crossref PubMed Scopus (32) Google Scholar,28.Krishnaswamy S. et al.Systems biology: conditional density-based analysis of T cell signaling in single-cell data.Science. 2014; 346: 1250689Crossref PubMed Scopus (64) Google Scholar] and perturbation studies [29.Dixit A. et al.Perturb-seq: dissecting molecular circuits with scalable single-cell RNA profiling of pooled genetic screens.Cell. 2016; 167: 1853-1866.e17Abstract Full Text Full Text PDF PubMed Scopus (518) Google Scholar], would allow investigators to capture the heterogeneity in niche-induced signaling pathways, enabling the development of higher-resolution integrative network models. Modeling cell–cell communication mediated by receptor–ligand interactions using scRNA-seq has also been proposed to study crosstalk of different cell types in the context of development, differentiation, and disease [30.Vento-Tormo R. et al.Single-cell reconstruction of the early maternal-fetal interface in humans.Nature. 2018; 563: 347-353Crossref PubMed Scopus (624) Google Scholar, 31.Efremova M. et al.CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes.Nat. Protoc. 2020; 15: 1484-1506Crossref PubMed Scopus (364) Google Scholar, 32.Camp J.G. et al.Multilineage communication regulates human liver bud development from pluripotency.Nature. 2017; 546: 533-538Crossref PubMed Scopus (238) Google Scholar, 33.Skelly D.A. et al.Single-cell transcriptional profiling reveals cellular diversity and intercommunication in the mouse heart.Cell Rep. 2018; 22: 600-610Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar] (Table 1 and Box 1). These models combine prior knowledge of ligand–receptor complexes with a statistical framework to predict tissue-specific cell–cell communication networks via these molecular interactions. In particular, they can generate predictions of key cell–cell interactions and motifs responsible for maintaining tissue homeostasis and supporting tissue regeneration. In this regard, the single-cell based inference of cell–cell communication networks enabled the prediction of cell–cell interactions involved in maintaining tissue homeostasis [32.Camp J.G. et al.Multilineage communication regulates human liver bud development from pluripotency.Nature. 2017; 546: 533-538Crossref PubMed Scopus (238) Google Scholar, 33.Skelly D.A. et al.Single-cell transcriptional profiling reveals cellular diversity and intercommunication in the mouse heart.Cell Rep. 2018; 22: 600-610Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar, 34.Raredon M.S.B. et al.Single-cell connectomic analysis of adult mammalian lungs.Sci. Adv. 2019; 5eaaw3851Crossref PubMed Scopus (53) Google Scholar]. We expect that the comparison of these reference networks with cell–cell interactomes of pathological or injured tissues will allow the identification of dysregulated interactions and can guide the development of novel intervention strategies for restoring homeostasis and supporting tissue regeneration. In addition, computational models of cell–cell communication have provided insights into general principles underlying tissue homeostasis. For example, the analysis of cell–cell communication networks indicated the necessity of endocytosis for maintaining cell type proportions [35.Zhou X. et al.Circuit design features of a stable two-cell system.Cell. 2018; 172: 744-757.e17Abstract Full Text Full Text PDF PubMed Scopus (115) Google Scholar,36.Adler M. et al.Endocytosis as a stabilizing mechanism for tissue homeostasis.Proc. Natl. Acad. Sci. U. S. A. 2018; 115: E1926-E1935Crossref PubMed Scopus (20) Google Scholar]. Although current methodologies are able to infer cell–cell communication events between different cell populations, they cannot delineate their functional differences resulting from the spatial position of cells in the tissue. Therefore, these computational approaches can be combined with imaging-based technologies for spatial transcriptome reconstruction [37.Karaiskos N. et al.The Drosophila embryo at single-cell transcriptome resolution.Science. 2017; 358: 194-199Crossref PubMed Scopus (177) Google Scholar, 38.Halpern K.B. et al.Single-cell spatial reconstruction reveals global division of labour in the mammalian liver.Nature. 2017; 542: 352-356Crossref PubMed Scopus (390) Google Scholar, 39.Rodriques S.G. et al.Slide-seq: a scalable technology for measuring genome-wide expression at high spatial resolution.Science. 2019; 363: 1463-1467Crossref PubMed Scopus (488) Google Scholar] to enable the characterization of the complete interactome in a spatially resolved manner. For instance, single-molecule fluorescence in situ hybridization (smFISH) can be employed to detect genes characteristic of different tissue locations. On the basis of expression of these genes, the spatial heterogeneity within scRNA-seq data can be resolved. Models of tissue self-organization resulting from cell–cell interactions are valuable in the study of tissue homeostasis and regeneration. In particular, the combination of computational modeling, machine learning, and mathematical optimization has been employed to predict experimentally testable perturbations that generate desired multicellular spatial patterns in human iPSC colonies [40.Libby A.R.G. et al.Automated design of pluripotent stem cell self-organization.Cell Syst. 2019; 9: 483-495.e10Abstract Full Text Full Text PDF PubMed Scopus (13) Google Scholar]. Hence, this data-driven approach enables the prediction and control of spatial self-organization of heterogeneous populations of stem cells. An extension of this method to model systems composed of different cell types can potentially be used to study and control processes such as tissue homeostasis and regeneration. Additionally, they can characterize fundamental self-organized patterns for tissue regeneration, such as the existence of distinct stem cell GRN configurations governing different aspects of the cell's response to environmental cues [4.Stumpf P.S. MacArthur B.D. Machine learning of stem cell identities from single-cell expression data via regulatory network archetypes.Front. Genet. 2019; 10: 2
Referência(s)