A data analytics framework for mining the dark kinome

项目摘要 这项提议的总体目标是生成和实验测试未被研究的“暗”激酶模型。 进化和功能，并开发一个数据分析框架，用于假设的生成和测试 深色的近亲。我们的工作假设是对可用序列、结构和功能的综合挖掘 来自不同生物体(包括经过充分研究的和黑暗的激酶)的整个亲属组的数据将提供重要的 用于定义与暗蛋白激酶功能相关的序列和结构特征的上下文。作为初步准备 检验我们的假设，我们已经整合和概念化了与蛋白激酶相关的各种形式的数据 蛋白质激酶本体(ProKino)的结构、功能和进化，并成功地 演示了本体论框架在确定人类关键知识差距方面的应用 在发现与蛋白激酶调节相关的关键残基/基序方面。我们计划建造一座 对这些成功的研究实现了以下两个目标。Aim1将开发一种新颖的可比较 运动组学框架，在该框架中，暗蛋白激酶中的自然和疾病变体将被关联和可视化 PTMS和蛋白质-蛋白质相互作用的背景研究连接序列的关系， 结构、功能和调节。功能专业化的模型将在选定的 使用生化和基于细胞的分析的黑暗和假性激酶，并在人类和 机器可读格式，遵循可查找、可访问、可互操作和可重复使用(公平)的数据规则。 AIM2将构建一个独特的框架，用于对语义链接的蛋白激酶数据进行复杂的聚合查询 使用图形化、易于使用的界面从不同的来源和格式中获取数据。研究人员将与 根据熟悉和直观的数据视图构建和制定查询。基于知识地图的知识地图 将开发交互式查询界面，研究人员可以使用该界面与ProKinO进行交互 他们使用和理解的语义。ProKinO将正式与药物目标本体论Pharos联系起来 (DTO)和蛋白质本体论(PRO)，以扩大社区外联和用户基础。 预计拟议的研究将为知识发现提供统一的数据分析框架 暗染色体上假说的产生和提高照亮可药物基因组的能力 (IDG)联盟对暗蛋白激酶的生理作用做出准确的预测。建议数 ProKinO与DTO和PRO的集成将增强这些本体在药物发现中的应用 并提供用于为其他IDG目标构建数据实例化本体的开源软件，如ion- 频道和GPCR。这些结果，反过来，预计将加速对 可用药的“暗”蛋白质组，并解决了IDG将基因组数据转化为知识的倡议 药物发现。

Project Summary The overall goal of this proposal is to generate and experimentally test models of understudied “dark” kinase evolution and function, and to develop a data-analytics framework for hypothesis generation and testing on the dark kinome. Our working hypothesis is that integrative mining of available sequence, structure, and functional data on the entire kinome from diverse organisms (both well-studied and dark kinases) will provide important context for defining sequence and structural features associated with dark kinase functions. As a preliminary test of our hypothesis, we have integrated and conceptualized diverse forms of data related to protein kinase structure, function, and evolution in the form of the Protein Kinase Ontology (ProKinO), and successfully demonstrated the application of an ontological framework in identifying key knowledge gaps in the human kinome and in discovering key residues/motifs associated with protein kinase regulation. We propose to build on these successful studies to accomplish the following two aims. Aim1 will develop a novel comparative kinomics framework in which natural and disease variants in dark kinases will be correlated and visualized in the context of PTMs and protein-protein interactions to investigate the relationships connecting sequence, structure, function and regulation. Models of functional specialization will be experimentally tested in selected dark and pseudokinases using biochemical and cell-based assays and made publically available in human and machine-readable format, adhering to Findable, Accessible, Interoperable and Reusable (FAIR) data rules. Aim2 will build a unique framework for complex aggregate queries on semantically linked protein kinase data from disparate sources and formats using graphical, easy to use interfaces. Researchers will interact with the framework and formulate queries based on a familiar and intuitive view of the data. A knowledge map-based interactive query interface will be developed using which researchers can interact with ProKinO using semantics they use and understand. ProKinO will be formally linked with Pharos, the Drug Target Ontology (DTO) and the Protein Ontology (PRO) to expand community outreach and user base. The proposed studies are expected to provide a unified data analytics framework for knowledge discovery and hypothesis generation on the dark kinome and enhance the ability of the Illuminating the Druggable Genome (IDG) consortium to make accurate predictions about the physiological roles of dark kinases. The proposed integration of ProKinO with DTO and PRO will enhance the application of these ontologies in drug discovery and provide open source software for building data instantiated ontologies for other IDG targets such as ion- channels and GPCRs. These outcomes, in turn, are expected to accelerate the functional characterization of the druggable “dark” proteome and address the IDG initiative of translating genomic data into knowledge for drug discovery.