Integrated Kernel Learning for Genomic Data Mining and Prediction

• Activity Number: 510
• Type: Invited
• Date/Time: Wednesday, August 3, 2016 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Xuefeng Wang* and Zhenyu Zhang and Minqin Chen
• Companies: SUNY Stony Brook and SUNY Stony Brook and SUNY Stony Brook
• Keywords: Kernel learning ; Multi-platform ; prediction ; Cancer ; Genomics ; network

Abstract

Data mining of molecular data has important implications for clinical outcome prediction in cancer and other diseases. Linking high-throughput genomic profiling generated from multiple sources in building predictive models remains a challenge both conceptually and computationally. Multiple kernel learning has emerged as a promising machine learning technology to address this problem. This technology is particularly appealing for its generality on incorporating heterogeneous data. However, the efficiency and performance of variant algorithms when applied to genomic data remains to be investigated. In this talk, we discuss some important theoretical and practical aspects of this framework and speculate on how to build efficient supervised learning models based on publicly available cancer datasets. We demonstrate and benchmark the implementation of the proposed framework using data from The Cancer Genome Atlas (TCGA).