There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | Momiao Xiong |
| Institution: | University of Texas School of Public Health |
| Department: | Biostatistics |
| Country: | |
| Proposed Analysis: | The next generation of genomic, sensing and image technologies will produce a deluge of DNA sequencing, image, behavioral, and clinical multiple phenotypes data with millions of features. Analysis of increasingly larger, more complex and more diverse genomic, molecular, space-temporal physiological and anatomical imaging data provides invaluable information for the holistic discovery of the genetic structure of disease and has the potential to be translated into a better understanding of the basic biomedical mechanisms, but also poses great methodological and computational challenges. The traditional paradigm for analyzing big and highly heterogeneous biological data is association or correlation analysis which is expected to be built among certain elements, such as genetic variation, genes and pathways. However, systematic genomic, physiological and imaging studies always need to uncover causal relationship among various molecular and morphological components, which form complex biological systems. It is a key to the success of big biological data analysis how we can identify causal relationship among genomic and morphological variation in big genomic, imaging and phenotype data. Goals of this application are to develop unified frameworks for systematic casual analysis of growing large, complex and diverse genetic and image data and novel statistical methods and computational algorithms for deciphering causal networks of genotype-phenotype, genotype-image, with next-generation sequencing data. As a consequence, the aims of this research are the following. Aim 1: We will develop novel statistical methods for genetic-image (including MRI, DTI and fMRI) association and interaction analysis with next-generation sequencing data and compare the performance of the developed novel methods with existing methods. Aim 2: We will develop a new framework and novel statistical tools for construction of statistic and dynamic causal genotype-image-phenotype networks with next-generation sequencing data to shift the paradigm of genetic studies of multiple phenotypes from association analysis to causal inference. Specifically, we will develop sparse functional SEMs with functional exogenous variables for multivariate trait association analysis and calculation of score function. We will formulate learning the optimal order of the nodes in the network as a mixed integer programming problem and develop efficiently greedy algorithms for solving integer program problem to search the structure of the network with the highest score function. We will apply the developed novel methods for causal inference to the image (MRI, DTI) - genomic analysis to infer causal genetics, image and Alzheimer’s disease relations. Aim 3: We use directed acyclic graphics as a tool to develop high-dimensional learning of causal networks for causal connectivity analysis for fMRI data alone, both fMRI and genotype data. Specifically, we develop sparse SEMs for vector time series, sparse SEMs with functional exogenous variables, functional endogenous variables and both functional endogenous and exogenous variables to calculate score functions for the connectivity analysis of fMRI, genotype and fMRI. The newly developed computational software and analytical tools will be implemented in cluster and cloud platforms, and will be publicly distributed to the scientific community during the course of this proposal. The ADNI dataset will be used as the discovery and method development dataset. |
| Additional Investigators |
