Data Structures, Algorithms and Tools for Ontological Discovery

Co-Principal Investigator:
Michael A. Langston, Department of Electrical Engineering and Computer Science, University of Tennessee
Abstract:
The main goal of this project is to develop and deliver a software system to classify behavioral traits and the substance abuse phenomena to which they are related through the connection of these phenotypes to shared biomolecular substrates, and to validate these classifications through tests of pleiotropic effects of specific gene perturbations. The classification problem has been historically troublesome for behavioral traits. Neurobiological and behavioral disorders each encompass a diverse array of underlying psychological and physiological conditions. Further, there is an incredible degree of interpretive subtlety to behavioral research methods, with many assays of the same construct each subject to different experimental biases. Rather than classify behaviors by external manifestations, we seek to classify them based on shared biomolecular substrates. To accomplish our objectives on a large scale, we aggregate, integrate and decompose data from diverse genomic studies of drugs and behavior. Convergence of evidence enables discovery of common and unique biological substrates of diverse behavioral disorders and thereby establishes relationships among behavioral characteristics and substance abuse related phenotypes. Our team applies developments in computer science, systems biology, bioinformatics and statistical genetics to address the data integration challenge posed in cross-species, multi-dimensional behavioral neuroscience for the Integrative Neuroscience Initiative on Alcoholism (INIA) Consortium. Our work reveals that to understand the relations among large sets of traits, there is a need for novel, advanced methods of data representation, professional data curation, innovative combinatorial algorithms, interfaces for processing the tremendous amount of data in a single biological domain. There is also a need for biological validation of the relations among biological attributes that are defined using our approach. Thus the primary focus of this project is to develop and validate an integrative method for large-scale analysis of relation-ships among behaviors and addiction related phenotypes.
Research Partners:
The PI for this project is Elissa J. Chesler at The Jackson Laboratory.
The other Co-PI is Erich J. Baker at Baylor University.
Relevant Site of Interest:
GeneWeaver