Integrating multiple types of genomics data to identify meaningful associations

复杂的人类特征和表型产生于跨越多个细胞“分辨率”的级联相互作用，包括基因组、调节组、蛋白质组以及基因调控和细胞网络。高通量技术的最新进展使我们能够以任何这些相互关联的分辨率收集基因组数据。这为从统计上整合这些数据提供了一个独特的机会，以便推导出将基因组序列与基因表达、基因及其产物之间的相互作用网络以及最终与复杂特征联系起来的级联模型。构建这些级联模型对于识别复杂的人类特征和表型背后的因果机制至关重要。*我们的计算生物学研究计划专注于开发统计和计算数据集成方法，以组合不同类型的生物数据。通过数据整合，我们得到了细胞的级联模型，推断了遗传变异导致复杂表型的详细生物学机制。特别是，拟议的研究重点是开发基于图形的数据集成方法，用于集成各种细胞分辨率的数据(而不会将数据折叠为“基因级别”的共同表示)。我们开发的算法和软件被生物学家用来推断不同类型的生物实体(如SNP、基因、非编码RNA)与其感兴趣的复杂特征之间的联系机制。**

Complex human traits and phenotypes arise from cascading interactions that span multiple cellular “resolutions” including the genome, the regulome, the proteome, and gene regulatory and cellular networks. Recent advances in high-throughput technologies allow us to collect genomic data at any of these interrelated resolutions. This offers a unique opportunity to statistically integrate these data in order to derive cascading models that link the genomic sequence to gene expression, to networks of interactions between genes and their products, and eventually to complex traits. Building these cascading models is essential for identifying causal mechanisms that underlie complex human traits and phenotypes. *** *** Our computational biology research program focuses on developing statistical and computational data integration methods for combining different types of biological data. Through data integration, we derive cascading models of the cell, inferring the detailed biological mechanisms by which genetic variation results in complex phenotypes. In particular, the proposed research focuses on developing graph-based data integration methods for integrating data at variety of cellular resolutions (without collapsing the data to a common representation at the “gene level”). The algorithms and software that we develop are used by biologists to infer the mechanisms connecting different types of biological entities (e.g., SNPs, genes, non-coding RNAs) to their complex trait of interest. **