Mediation modeling and analysis forhigh-throughput omics data
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
There is a strong need for powerful unified statistical methods for discovering underlying genetic architecture of complex traits with the assistance of omics information. In this paper, two methods aiming to detect novel association between the human genome and complex traits using intermediate omics data are developed based on statistical mediation modeling. We demonstrate theoretically that given proper mediators, the proposed statistical mediation models have better power than genome-wide association studies (GWAS) to detect associations missed in standard GWAS that ignore the mediators. For each ofthe modeling methods in this paper, an empirical example is given, where the association between a SNP and BMI missed by standard GWAS can be discovered by mediation analysis.
Place, publisher, year, edition, pages
2015. , 17 p.
Mediation model, metabolite, SNP, BMI
IdentifiersURN: urn:nbn:se:uu:diva-256318OAI: oai:DiVA.org:uu-256318DiVA: diva2:825012
Master Programme in Statistics
2015-06-04, F332, Ekonomikum (plan 3), Kyrkogårdsg 10, Uppsala, 22:35 (English)
Xia, Shen, Dr.Yudi, Pawitan, Prof.Fan, Wallentin, Prof.