Systematic Assessment of Combinatorial Transcription Factor Activity

项目概要 尽管转录组目录庞大，但我们对转录状态的理解仍然主要是描述性的。 在本提案中，我们寻求通过解决以下问题来推进该领域对细胞状态的预测性理解： 知识上的根本差距：驱动细胞转录的基因和调控网络是什么？ 状态？这些知识对于理解细胞状态的分子基础和操纵细胞状态至关重要 用于生物医学应用。转录因子（TF）协同驱动基因调控网络（GRN） 建立转录状态。值得注意的是，强制诱导 TF 可以通过支持来重新编程基因表达状态。 种植现有的GRN。因此，TF 和 GRN 是预测性理解细胞的构建模块。 转录状态。一个关键的挑战是，一般来说，TF 和 GRN 之间的关系是未知的 并且很难准确预测。这一挑战源于当前的几个问题：缺乏基本事实 GRN 源自实验 TF 扰动，依赖静态转录组数据库来推断 GRN TF 鸡尾酒预测，以及预测非线性 TF 行为的难度。直到我们能够理解如何 TF 协同影响 GRN，我们预测细胞状态 TF 驱动因素的能力仍然有限。我们的长期 术语目标是了解转录状态的分子基础，以便在细胞工程中应用。 为了实现这一目标，本提案的目标是生成一种独特的资源来直接测量 GRN TF 的简单组合，并使用此功能知识库来预测和基准复杂的 TF 用于转录重编程的鸡尾酒。我们假设实验衍生的 GRN 将改善 预测的 TF 鸡尾酒在转录重编程中的性能。我们的理由是这些研究将 1）为 TF 功能活动和实验衍生的 GRN 提供新颖且急需的资源 社区，2) 提供对转录状态分子驱动因素的见解，以及 3) 识别转录因子 为生物医学研究设计基因表达状态的潜力。我们提出以下具体目标： （目标1）测量转录因子的组合活性； （目标 2）改进计算框架 预测转录重编程的 TF 鸡尾酒； （目标 3）在初始单元中泛化 TF 驱动的 GRN 上下文和基准预测。这个提案是创新的，因为它将使用我们的单细胞平台Re- program-Seq 2.0，用于高通量转录重编程，生成独特的功能资源 TF 和 GRN 的活动。我们期望这一资源能够推动新的研究视野。这个提议意义重大 不能，因为它将通过量化组合 TF 活性来扩展我们对基因组功能的理解，例如 实验性地衍生 GRN，并提供设计基因表达状态的新工具。

PROJECT SUMMARY Despite vast catalogs of transcriptomes, our understanding of transcriptional state remains mostly descriptive. In this proposal, we seek to advance the field towards a predictive understanding of cell state by addressing the fundamental gap in knowledge: what are the genes and regulatory networks that drive a cell’s transcriptional state? This knowledge will be critical to understand the molecular basis of cell state and to manipulate cell state for biomedical applications. Transcription factors (TFs) cooperatively drive gene regulatory networks (GRNs) to establish transcriptional states. Notably, forced induction of TFs can reprogram gene expression states by sup- planting existing GRNs. Thus, TFs and GRNs are the building blocks to a predictive understanding of a cell’s transcriptional state. One key challenge is that, in general, the relationship between TFs and GRNs is not known and is difficult to accurately predict. This challenge arises from several current problems: a lack of ground truth GRNs derived from experimental TF perturbation, a reliance on static transcriptomic databases to infer GRNs for TF cocktail prediction, and the difficulty of predicting non-linear TF behaviors. Until we can understand how TFs cooperatively influence GRNs, our ability to predict the TF drivers of cell state will remain limited. Our long- term goal is to understand the molecular basis of transcriptional state for applications in cellular engineering. Towards this goal, the objective of this proposal is to generate a unique resource to directly measure GRNs for simple combinations of TFs, and to use this functional knowledgebase to predict and benchmark complex TF cocktails for transcriptional reprogramming. We hypothesize that experimentally-derived GRNs will improve the performance of predicted TF cocktails for transcriptional reprogramming. Our rationale is that these studies will 1) provide a novel and urgently needed resource of TF functional activity and experimentally-derived GRNs for the community, 2) provide insights into the molecular drivers of transcriptional state, and 3) identify TFs with potential to engineer gene expression states for biomedical research. We propose the following specific aims: (Aim 1) Measure the combinatorial activities of transcription factors; (Aim 2) Improve computational frameworks to predict TF cocktails for transcriptional reprogramming; (Aim 3) Generalize TF-driven GRNs across initial cell contexts and benchmark predictions. This proposal is innovative because it will use our single-cell platform Re- program-Seq 2.0 for high-throughput transcriptional reprogramming to generate a unique resource of functional activities for TFs and GRNs. We expect this resource to propel new research horizons. This proposal is signifi- cant because it will expand our understanding of genome function by quantifying combinatorial TF activity, ex- perimentally deriving GRNs, and providing new tools to engineer gene expression states.