A mathematical framework for interpreting the molecular code of the tumor suppressor p53

蛋白质的翻译后修饰(PTM)由于其快速的动力学在细胞信号和调控中起着重要的作用，因此对于动态的细胞信息处理是必不可少的。一个蛋白质分子可以有多个修饰位点，因此可以表现出不同模式的共生PTM(以下称为“修饰形式”)。这些模式的PTM充当蛋白质上的“分子密码”，提供有关细胞状况的信息，可被下游过程读取。肿瘤抑制基因P53是一种关键的细胞调节因子，它含有100多个修饰位点，使其能够整合不同的应激信号，并对数百个基因进行激活或抑制，从而决定不同的细胞命运。这个项目的总体目标是利用转录因子和肿瘤抑制因子p53作为细胞信号和调控的重要生物学例子来理解PTM编码。了解PTM编码和了解蛋白质的实际PTM状态对于理解细胞信息处理是必不可少的，然而，考虑到细胞中不仅存在一个而是一个群体的分子导致不同的调制形式的分布，这个问题变得更加复杂。目前用于估计模型分布的质谱学方法不能提供足够的数据/信息来重建高度修饰的蛋白质如P53的模型分布。例如，假设P53有大约100个修饰位点，并且都是二元的(例如，氨基酸残基要么被磷酸化，要么不被磷酸化)，那么蛋白质表现出2^100≈10^30种可能的修饰形式。这意味着重构模形分布的解存在于2^100维的空间中。我的目标是开发一种新的方法，绕过高维解空间来估计模形分布必须位于的区域。理想情况下，探索modform区域将有助于检测哪些站点最有可能被修改。这些预测将指导未来的实验，目标是确定决定细胞命运结果的普遍修饰位点。了解P53复杂的PTM编码将对细胞命运的决策以及对癌症发展和治疗的理解具有重要意义。

Post-translational modifications (PTMs) of proteins play a fundamental role in cellular signaling and regulation due to their fast dynamics and therefore are essential for dynamic cellular information processing. A single protein molecule can have multiple modification sites and thus can exhibit different patterns of co-occurring PTMs (hereafter called "modforms"). These patterns of PTMs act as a "molecular code" on that protein and confers information about the cellular condition that can be read by downstream processes. The tumor suppressor p53 is a key cellular regulator that harbors over 100 modification sites enabling it to integrate diverse stress signals and in response to activate or suppress hundreds of genes leading to different cell fate decisions. The overall aim of this project is to understand PTM encoding using the transcription factor and tumor suppressor p53 as an important biological example in cellular signaling and regulation.Elucidating PTM encoding and knowing the actual PTM state of a protein is essential to understand cellular information processing, however, the problem becomes more complicated considering that there is not only one but a population of molecules present in a cell leading to a distribution of different modforms. Current mass spectrometry methods that are used to estimate the modform distribution do not provide sufficient data/information to reconstruct the modform distribution of heavily modified proteins such as p53. Assuming, for instance, that p53 has approximately 100 modification sites and that all are binary (such as phosphorylation where an amino acid residue is either phosphorylated or not) then the protein exhibits 2^100 ≈ 10^30 possible modforms. This means that the solution of reconstructing the modform distribution exists in a space of dimension 2^100. I aim to develop a novel method that circumvents the high dimensional solution space to estimate a region in which the modform distribution must lie. Ideally, exploring the modform region will then help to detect which sites are most likely modified. These predictions shall guide future experiments with the goal to identify prevalent modified sites that determine cell fate outcome. Understanding the complicated PTM coding of p53 will have important implications for cell fate decision making and the understanding of cancer development and therapies.