Gene-environment (GxE) interactions play a role in the development of many complex human diseases, including chronic inflammatory bowel diseases (IBD). Typically, the disease status (affected vs. non-affected) represents the primary phenotype of interest. In addition, a limited number of quantitative phenotypes such as clinical scores or laboratory values highlighting disease severity are also available. These data are complemented by high-dimensional molecular data as, for example, the abundance of microbial species in the gut (microbiome) or the concentration of nutritional metabolites in the blood (metabolome). The statistical methods required to properly analyze the interplay between genetics and environmental factors such as smoking on these entities have however not been developed sufficiently yet.
The aims of this project are to
* adapt existing statistical methods of GxE analysis to high-dimensional molecular phenotype data,
* evaluate the utility of these methods,
* apply the methods to real data on gene-smoking and gene-nutrition interaction, particularly to the microbiomes of healthy individuals and IBD patients
In a first step, existing methods for the GxE analysis of quantitative phenotypes will be adapted to, and evaluated for, different study designs such as case-only or case-control. In particular, generalized linear models are used. Dimensionality reduction methods, such as principal component analysis or machine learning, will be applied to both the genotype and phenotype data.
The PI’s institution hosts a PhD student from the second RTG generation, post-docs from the sysINFLAME and DFG FOR 2107 projects and an e:MED junior research group on systems medicine, all of whom work on the integration of high-dimensional molecular data, albeit following different statistical and machine learning approaches. The successful applicant will cooperate intensively with them on their own project.
The analysis of real data on GxE interactions will be performed in cooperation with other projects of the RTG.
Requirements for the position:
The successful candidate will have a master degree in statistics, mathematics, bioinformatics or a related field. Training in biology and computer science would be beneficial, but is not mandatory.