A groundbreaking study, recently published in the 'National Science Review' on June 17, 2022, has unveiled a new perspective on cell differentiation by incorporating the dynamics of cell birth and death rates into energy landscape theory. This research represents a significant stride in bioinformatics and cellular biology, providing a more nuanced understanding of the cellular differentiation process.
The study was conducted by a distinguished team of researchers: Shi Jifan, Young Associate Researcher at Fudan University's Laboratory of Basic Theory and Key Technologies of Intelligent Complex Systems; Professor Kazuyuki Kaihara from the University of Tokyo; Professor Li Tiejun from Peking University's School of Mathematics; and Professor Chen Luonan from the Shanghai Academy of Biological Sciences, Chinese Academy of Sciences.
The team's work is based on the analogy proposed by CH Waddington in the 1960s, which likens cell differentiation to small balls rolling down a hillside into various states(Figure 1). This analogy laid the foundation for using energy landscape theory to represent the dynamic properties of Gene Regulatory Networks (GRNs) in a more intuitive manner.
Fig 1: Waddington’s epigenetic landscape of cell differentiation. The process of cellular differentiation from stem cells into various somatic cell types can be likened to the trajectory of a ball descending from the peak of a mountain to a distinct state at its base. (figure from Waddintgon et al. Principles of development and differentiation, Macmillan 1966)
In their publication, the researchers proposed a novel decomposition theory for understanding the potential energy landscape during cell differentiation. This theory incorporates terms related to cell birth and death, revealing that the complex dynamics governed by GRNs can be broken down into two potential energy components, along with one non-gradient term. The potential energy landscape, U(x), represents different cell types with varying metastatic properties. In contrast, the multipotent potential energy landscape, V(x), reflects cellular plasticity and the overall directionality towards negative gradient descent during cell differentiation (Figure 2).
Fig 2: The potential energy landscapes U(x) and V(x) of cell types were constructed based on gene regulatory networks (GRNs) and the rates of cell birth and death R.
The study also addresses the distinct characteristics of low-dimensional and high-dimensional models by proposing a numerical method for decomposing potential energy landscapes. This method has been successfully applied in various systems, including convective diffusion processes, two-gene regulatory networks, and T-cell differentiation processes (FIG. 3).
Fig 3: Energy landscape decomposition during T cell differentiation. (A) Gene regulatory networks during T cell differentiation. (B) Potential energy landscape U of cell types in dimension reduction to dimension 2. The four potential Wells from top to bottom are ETP/DN1, DN2a, DN2b, and DN3 cell states in turn. (C) Multi-potential energy landscape V in dimension reduction to dimension 2, the value represents cell stemness in different states, and also reveals the differentiation direction of cells from top to bottom.
This theory not only advances the development of potential energy landscape theory but also provides an effective mathematical tool for modeling cell differentiation processes. It holds great potential for expanding our understanding of similar dynamic systems in biology and bioinformatics.
For further reading, the full paper can be accessed here: Energy landscape decomposition for cell differentiation with proliferation effect | National Science Review | Oxford Academic.
