|
Abstract:
Technological advances in the high-throughput (HTP) generation of genomic and proteomic biological data have outpaced the development of computational tools to effectively analyze, visualize, and interpret this information.
Biological knowledge is characterized by complex representations, high dimensionality, imprecise or inconsistent theories, and incomplete information. As a result, HTP data analysis requires iterative and interactive approaches, and insight usually evolves with increased understanding of the domain. The increasing size and complexity of HTP datasets challenges the ability of software developers to create tools for discovering relevant biological relationships. Current techniques for integrating different types of HTP data do not allow a full appreciation of the information contained therein.
Through systematic knowledge management, analysis and intelligent use of information, we can significantly increase our understanding of biology in general and health and disease processes in specific. This requires new generation of computational tools.
The focus of this talk is to overview specific challenges, introduce machine-learning solutions, and present some evidence of their utility in HTP biological domains.