Semantic LAMHDI: Linking diseases to model organism resources

描述（由申请人提供）：我们对人类疾病的理解是通过使用模型系统来调查疾病机制和治疗的重要信息。由于确定疾病模型是如此重要，因此已经作出了重大努力，通过LAMHDI（将动物模型与人类疾病联系起来的门户网站）中的基因同源学和途径链接，汇集有关疾病和模式生物的信息，从而将研究人员与潜在的疾病模型联系起来（www.lamhdi.org）。然而，LAMHDI目前是不完整的，因为尚未包括体外生物学模型，并且无法基于与人类疾病的表型相似性来识别模型系统。此外，没有机制可以优先考虑疾病及其模型之间的无数关系，这使得确定候选模型和发现新的关系都很困难。这项工作的目标是促进疾病研究模型的识别，更好地利用现有的模式生物和有关它们的体外资源和数据，并提供发现疾病，表型和基因之间新关系的能力，这将进一步加深我们对疾病的理解。为此，我们建议：1)使用来自人类和模式生物的导入和对齐的表型数据，基于表型的语义相似性计算候选疾病模型。我们将包括表达数据，以改进基于特定解剖位置和/或基因型中存在的表达的表型搜索。2)扩大LAMHDI内疾病和体外模型系统之间的语义联系，包括eagle-i项目和外部来源的生物标本和细胞系等资源。这将允许研究人员根据表型或遗传基础确定候选的体外模型系统。3)创建一个发现工具，以改进搜索并发现疾病、模式生物和利用遗传、途径和表型关系的体外资源之间的新关系。为了汇集所需的不同数据，我们利用语义网技术和复杂的信息建模与计算算法相结合。可用性研究将为新的知识引导发现LAMDHI界面的迭代开发提供信息。

DESCRIPTION (provided by applicant): Our understanding of human disease is vitally informed by the use of model systems to investigate disease mechanisms and therapies. Because identifying disease models is so critical, a significant effort has been made to connect researchers to potential models of disease by bringing together information about diseases and model organisms via gene orthology and pathway links in LAMHDI, the initiative to Link Animal Models to Human Disease portal (www.lamhdi.org). However, LAMHDI is currently incomplete in that in vitro biological models are not yet included and one cannot identify model systems based on phenotype similarity to a human disease. Furthermore, there is no mechanism by which to prioritize the myriad of relationships between disease and their models, making both identification of candidate models and discovery of new relationships difficult. The goal of this work is to facilitate the identification of models for disease research, make better use of existin model organisms and in vitro resources and data about them, and provide the ability to uncover new relationships between disease, phenotypes and genes that will further our understanding of disease. To this end we propose to: 1) Enable computation of candidate disease models based on semantic similarity of phenotypes using imported and aligned phenotype data from humans and model organisms. We will include expression data to refine search of phenotypes based on presence of expression within a particular anatomical location and/or genotype. 2) Expand semantic linkage between diseases and in vitro model systems within LAMHDI, including resources such as biospecimens and cell lines from the eagle-i project and external sources. This will permit investigators to identify candidate in vitro model systems based on phenotype or genetic basis. 3) Create a discovery tool to refine searches and to uncover novel relationships between diseases, model organisms, and in vitro resources using genetic, pathway, and phenotype relationships. To bring together the disparate data required, we leverage semantic web technologies and sophisticated information modeling combined with computational algorithms. Usability studies will inform the iterative development of the new knowledge-guided discovery LAMDHI interface.