Bioinformatics resources for genome-wide assessment of protein function at the proteoform level

我研究的长期目标是提高我们在高通量组学数据中检测翻译后蛋白质修饰的能力，并利用生物信息学和蛋白质组学方法在蛋白质形式水平上改善我们对蛋白质的功能理解。 现在人们普遍认识到，是蛋白质的变异，而不是基因的变异，构成了生物机械的多样性和复杂性。选择性剪接和翻译、RNA编辑和翻译后修饰(PTM)导致了被称为蛋白质形式的不同蛋白质分子。由同一基因编码的蛋白质形式可以在结构、功能和定位上有所不同。蛋白质形式的多样性并没有被经典的“一基因一蛋白质一功能”概念所捕捉到。目前的基因或典型的以蛋白质为中心的数据库将分子功能赋予基因/蛋白质，并且不区分蛋白质形式。这对于基因组和蛋白质组范围的高通量实验和从数据库中自动检索信息来说变得特别成问题。 显然需要能够从组学数据中更全面地识别蛋白质形式及其潜在的翻译后修饰的算法，以及反映蛋白质形式水平上的复杂性和功能关系的新的数据模型和数据库。在过去的几年里，我们通过开发数据模型、公共数据库以及网络推理算法和数据分析工具来解决这个问题，重点是一种特定的PTM蛋白酶介导的蛋白质切割加工成稳定的切割蛋白形式。 拟议的研究计划扩展了我们之前成功的工作，现在涵盖了所有的蛋白质形式。我们将使用将生物信息学与实验验证相结合的跨学科方法来解决三个具体目标： (1)获取蛋白质形式功能复杂性的数据库和工具 (2)从MS数据中识别蛋白质形式的统计和机器学习方法 (3)Proteoform函数去卷积和聚类 该计划为具有计算机和生物学背景的本科生和研究生提供出色的跨学科培训。这项工作有可能改变我们日常审问生物系统的方式。它将使高通量数据的功能分析的细节、精度和粒度达到新的水平，并支持生物技术、卫生和农业行业开发更具体的标记和化合物。

The long-term objective of my research is to increase our ability to detect post-translational protein modification in high-throughput omics data and improve our functional understanding of proteins at the proteoform level using bioinformatics and proteomics approaches. It is now widely recognized that variation in proteins, rather than of genes, underlies the diversity and complexity of biological machinery. Alternative splicing and translation, RNA editing and post-translational modification (PTM) result in distinct protein molecules called proteoforms'. Proteoforms encoded by the same gene can differ in structure, function and localization. The diversity of proteoforms is not captured by the classic "one geneone proteinone function" concept. Current gene or canonical' protein-centric databases assign a molecular function to a gene / protein and do not distinguish among proteoforms. This becomes particularly problematic for genome- and proteome-wide high-throughput experiments and automated information retrieval from databases. There is a clear need for algorithms enabling more comprehensive identification of proteoforms and their underlying post-translational modifications from omics data and for new data models and databases that reflect the complexity and functional relationships at the proteoform level. Over the past years, we have addressed this by developing data models, a public database as well as network inference algorithms and data analysis tools with a focus on one particular PTM protease mediated proteolytic processing of proteins into stable cleaved proteoforms. The proposed research program expands on our successful previous work to now encompass all proteoforms. We will use an interdisciplinary approach integrating bioinformatics with experimental validation to address three specific aims: (1) A Database and tools to capture the functional complexity of proteoforms (2) Statistical and machine learning approaches to identify proteoforms from MS data (3) Proteoform function de-convolution and clustering The program provides excellent interdisciplinary training for undergraduate and graduate students from computing and biology backgrounds. The work has the potential to transform how we routinely interrogate biological systems. It will enable a new level of detail, precision and granularity in the functional analysis of high-throughput data and support development of more specific markers and compounds in the biotechnology, health and agriculture industries.