From intra to intercellular regulatory networks that define cell type identity

项目摘要 细胞命运工程化，例如多能干细胞的定向分化或 体细胞类型之间的直接转换， 建模，药物筛选，并导致再生医学疗法。但我们的 不完全的基因突变阻碍了精确设计细胞命运的能力， 了解支配分化的细胞间和细胞内网络， 缺乏足够的计算工具来从数据中提取准确和可检验的假设， 来自单细胞组学技术的大量数据。初期的 在MIRA的资助下，我们开始通过开发新的 从单细胞RNA中定义细胞类型同一性的理论和计算方法- Seq（scRNA-Seq）数据，重点是出现的发育细胞类型 在中胚层发育和随后向滑膜细胞谱系的分化过程中， 关节作为这项工作的一部分，我们生成了发育中滑膜的scRNA-Seq数据， 联合，我们采用了多能干细胞向软骨细胞分化的方案，我们 发明了一种用于在单个细胞处评估细胞类型身份的通用平台 分辨率的水平。我们还开发了一种计算方法来推断动态 调节网络的准确性，并将它们与信号通路整合。现在我们 提议解决以下未回答的问题和未应对的挑战。第一、 我们将大大改进和扩展我们的计算方法，评估 细胞命运工程的结果，通过将它们扩展到更多的数据类型， 增加其结果的全面性，并通过预测不仅细胞身份， 而是功能。第二，我们会大幅改善和扩展监管网络 工具，以便他们的预测是统计校准，以便他们可以应用 染色质可及性和表达数据， 发现孤儿转录因子的结合位点基序。第三，我们将设计和 实验测试计算方法以产生可靠的细胞命运工程 不仅解释了转录网络， 这些通路为它们提供信息，并解释了时间动态。最后，我们将跟随 通过将我们的动态网络工具应用于体外原肠胚形成的观察， 表明一些信号通路通过重新连接 而不是直接影响效应靶基因的表达。 共同实现这些目标将有助于使细胞命运工程更加可靠 而且是可控的，它将揭示信号通路和细胞内 调控网络在发育过程中相互作用。

Project Abstract Cell fate engineering, for example the directed differentiation of pluripotent stem cells or the direct conversion among somatic cell types, holds great promise to improve disease modeling, drug screening, and to lead to regenerative medicine therapies. However, our ability to engineering cell fate with fidelity has been impeded by an incomplete understanding of inter- and intracellular networks that govern differentiation, and by the lack of adequate computational tools to distill accurate and testable hypothesis from the mountains of data coming from single cell omics technologies. In the initial period of funding under the MIRA, we started to address these challenges by developing novel theoretical and computational methods to define cell type identity from single cell RNA- Seq (scRNA-Seq) data with an emphasis on developmental cell types that emerge during mesoderm development and subsequent commitment to lineages of the synovial joint. As part of this work, we generated scRNA-Seq data of the developing synovial joint, we adopted a pluripotent stem cell-to-chondrocyte differentiation protocol, and we invented a generally applicable platform for assessing cell type identity at the single cell level of resolution. We also developed a computational method to infer dynamic regulatory networks accurately and to integrate them with signaling pathways. Now, we propose to address the following unanswered questions and unmet challenges. First, we will substantially improve and extend our computational methods that assess the outcomes of cell fate engineering by extending them to more data types, and thus increasing the comprehensiveness of its results, and by predicting not only cell identity but function. Second, we will substantially improve and extend our regulatory network tools so that their predictions are statistically calibrated and so that they can be applied to chromatin accessibility and expression data simultaneously with the intention of discovering binding site motifs of orphan transcription factors. Third, we will devise and experimentally test computational methods to generate reliable cell fate engineering recipes that account for not only transcriptional networks but also how signaling pathways inform them, and that account for temporal dynamics. Finally, we will follow up on observations from applying our dynamic network tool to in vitro gastrulation that indicates that some signaling pathways influence differentiation more by re-wiring network topology than by directly impacting expression of effector target genes. Collectively meeting these goals will help to make cell fate engineering more reliable and controllable, and it will shed light on how signaling pathways and intracellular regulatory networks interact during development.