A network-based approach to modelling cell-niche interactions and its application to studying salamander limb regeneration

细胞命运的决定是一个复杂的过程，涉及多个层次的调控。不仅细胞的内在遗传程序，而且外部影响，如由其生态位（微环境）施加的影响，可以诱导从一种细胞状态到另一种状态的转变。生态位效应是通过细胞内信号转导途径来解释的，这反过来又引起了细胞基因调控网络（GRN）的变化。关键的生态位成分包括与邻近细胞的直接相互作用、其他细胞分泌的因子、免疫控制和环境条件（如缺氧）。生态位相互作用的影响包括各种干细胞的分化倾向、上皮向间充质的转化、细胞迁移和再生。蝾螈是唯一一种四足动物，成年后的肢体可以再生所有的组织。了解蝾螈肢体再生的机制与包括人类在内的哺乳动物有直接关系，因为已经表明小鼠组织的再生利用了与蝾螈肢体再生中特征相似的细胞外基质（ECM）环境。因此，研究诱导蝾螈肢体再生的分子因素，从细胞生态位相互作用和解决哺乳动物再生两个方面都是至关重要的。虽然存在几种可以推断活性/非活性信号通路的工具，但它们不将信号通路与GRN整合，因此无法预测负责细胞状态转换的关键信号通路。在这里，我们提出了一种新的计算方法，集成信号通路和GRNs，并设计了一个系统的计算策略，预测关键信号通路和它们的靶基因，通过差分网络分析。我们将应用这种方法的蝾螈肢体再生系统和地址的信号通路的组合是负责再生。该项目的完成将丰富社区知识，并使再生在临床上更加有用。

Cell fate decision is a complex process involving several layers of regulation. Not only the intrinsic genetic programme of the cell, but also external influences such as those exerted by its niche (microenvironment), can induce a transition from one cellular state to another. The niche effect is interpreted through intracellular signal transduction pathways, which in turn brings about changes in the gene regulatory network (GRN) of the cell. Key niche components include direct interactions with neighbouring cells, secreted factors by other cells, immunological control, and environmental conditions such as hypoxia. The effects of niche interactions include differentiation tendency of various stem cells, epithelial-to-mesenchymal transitions, cell migration and regeneration. Salamander is the only tetrapod, where the adult limb functionally regenerates all constituent tissues. Understanding the mechanism of salamander limb regeneration has direct relevance to mammals including humans, as it has been shown that regeneration of mouse tissue utilized the extracellular matrix (ECM) environments that were similar to those characterized in salamander limb regeneration. Thus, investigating the molecular factors that induce salamander limb regeneration is crucial from both aspects of cell niche interactions and addressing regeneration in mammals. Although there exist several tools that can infer active/inactive signalling pathways, they do not integrate signalling pathways with a GRN, and are therefore unable to predict key signalling pathways responsible for cellular state transitions. Here we propose a novel computational method that integrates signalling pathways and GRNs, and devise a systematic computational strategy for predicting key signalling pathways and their target genes by means of differential network analysis. We will apply this method to the salamander limb regeneration system and address which combinations of signalling pathways are responsible for regeneration. The completion of this project will enrich community knowledge and enable regeneration to become clinically more useful.