Presenter Profile
Shun Kodate
Assistant Professor
Tohoku University, Center for Data-driven Science and Artificial Intelligence
Shun Kodate is an Assistant Professor at Tohoku University's Unprecedented-scale Data Analytics Center in Japan, affiliated with the Yamada lab. His research centers on geographically weighted regression and large-scale data analysis, particularly animal gene coexpression databases. Kodate has contributed to notable publications, including COXPRESdb v8 in Nucleic Acids Research. With expertise in network science, he presented on detecting problematic transactions in consumer-to-consumer e-commerce networks at NetSci 2019. Kodate's work on analyzing Sleeping Beauty papers in life sciences earned him the 若手奨励賞 (Young Researcher Encouragement Award). Passionate about science communication, he explores creative ways like "Bioinfo Comics" to popularize NGS-related knowledge.
TALK TITLE
Inspection of metabolome-GWAS with pathway simulations
KEYWORDS
Metabolome, Metabolic pathway, Metabolome-GWAS, Pathway simulation
ABSTRACT
Recent technologies combining metabolome and genome-wide association studies (metabolome-GWAS or MGWAS) have made it possible to detect variants that significantly affect metabolite levels. However, MGWAS has problems such as unclear causal direction, unclear replicability and generalizability, and difficulty in achieving statistical significance. Therefore, we attempted to overcome these problems by pathway simulation. Specifically, we modified the enzyme activities corresponding to the genetic variants in the model and compared the metabolite level fluctuations with those in MGWAS. As a result, for the variants with small p-values, the positive and negative values of the fluctuations were almost consistent between the simulation and MGWAS. Some of the insignificant variants in MGWAS also showed larger fluctuations in the simulation. These results mean that the simulation supports the MGWAS results and provides guides to further research.