A GWAS study considering multiple genic/invironmental factors

2016 Fiscal Year Final Research Report

• Project/Area Number: 25700004
• Research Category: Grant-in-Aid for Young Scientists (A)
• Allocation Type: Partial Multi-year Fund
• Research Field: Statistical science
• Research Institution: Kyushu University (2016); Kyushu Institute of Technology (2013-2015)
• Principal Investigator: Saigo Hiroto 九州大学, システム情報科学研究院, 准教授 (90586124)
• Project Period (FY): 2013-04-01 – 2017-03-31
• Keywords: GWAS / SNP / variable selection / kernel methods / hypothesis testing / model selection

Outline of Final Research Achievements

(1)Traversing combinatorial space with brand-and-bound search and rigorous multiple testing correction. A combinatorial space spanned by combination of genes contains a large number of false positives. We devised to remove them by employing Tarone’s correction, which resulted in a faster search with higher statistical power.
(2)Estimating the size of combinatorial space by kernel methods: In our computational experiments equipped with polynomial kernels and kernel ridge regression, we have shown that detection of up-to-six degrees interaction was possible. By applying the same method to mouse genotype/phenotype data, we have successfully detected the region in a chromosome that harbors causal genes.

Free Research Field

Machine Learning, Bioinformatics