Modularity map of the network of human cell differentiation

Abstract

Cell differentiation in multicellular organisms is a complex process whose mechanism can be understood by a reductionist approach, in which the individual processes that control the generation of different cell types are identified. Alternatively, a large-scale approach in search of different organizational features of the growth stages promises to reveal its modular global structure with the goal of discovering previously unknown relations between cell types. Here, we sort and analyze a large set of scattered data to construct the network of human cell differentiation (NHCD) based on cell types (nodes) and differentiation steps (links) from the fertilized egg to a developed human. We discover a dynamical law of critical branching that reveals a self-similar regularity in the modular organization of the network, and allows us to observe the network at different scales. The emerging picture clearly identifies clusters of cell types following a hierarchical organization, ranging from sub-modules to super-modules of specialized tissues and organs on varying scales. This discovery will allow one to treat the development of a particular cell function in the context of the complex network of human development as a whole. Our results point to an integrated large-scale view of the network of cell types systematically revealing ties between previously unrelated domains in organ functions.complex network | modular organization | self-similarity | stem cells T he cell differentiation process plays a crucial role in the prenatal development of multicellular organisms. Recent advances in the research on stem cell properties and embryonic development have uncovered several steps in the differentiation process (1-7). Single and multiple sequences of cell differentiation have been identified through in vivo observations of a particular embryo during early stages of development and through pathology studies of miscarriages during late stages of the process. While the identification of each cell differentiation step has been the subject of intense research, an integrated view of this complex process is still missing. Such a global view promises to reveal features associated with the large-scale modular organization of the cell types (5-12) with the purpose of discovering functional modules between cell types by using theoretical network analysis for community detection (9-11). In this letter, we take advantage of the current knowledge on the sequence of cell differentiation processes that is spread over a vast specialized literature (1-6, 13-27) (SI Appendix), to reveal and characterize the topological and dynamical features associated with the network of human cell differentiation (NHCD). I. ResultsWe construct the NHCD by systematically gathering the scattered information on the evolution of each cell type present in the embryo and fetus from a predecessor with a higher degree of differentiation potential into a more specialized type. The process of cell differentiation is then mapped onto a complex network that consists of 873...

Progress toward Quantitative Design Principles of Multicellular Systems

“…Multicellular systems can range in many sizes. Perhaps one of the most interesting aspects about multicellular systems, one that fascinates researchers and suggests a strategy for deconstructing multicellular systems into simpler elements, is that a multicellular system can be composed of several other sub-systems, each of which can also be multicellular [18,19]. For example, consider the human brain ( Figure 1A).…”

Architecture of the Florida power grid as a complex network

“…What we have measured and presented here is inherently geometric, revealing the spatial extension of the network in the (continuous) two-dimensional space. The other metric is topological [18,19], which is reminiscent of the Real Space Renormalization Group (RSRG) approach in statistical physics, viz. grouping vertices in the network according to their graphical distance (e.g., nearest-neighbors, second nearest-neighbors, etc.).…”

Complex networks modeled on the Sierpinski gasket

“…Using their techniques of fractal analysis on networks, they study the biological networks [29][30][31], such as function brain networks.…”