052_ Comorbidity analysis of schizophrenia patients using transcriptomic comorbid signatures
Research Question and Aims
In this study, we are investigating molecular mechanisms of comorbidity between schizophrenia and somatic diseases, such as type 2 diabetes and cardiovascular diseases. Using a large collection of publicly available transcriptomic datasets for these three disease classes, we have extracted transcriptomic signatures using non-negative matrix factorization (NMF). Using a reciprocal best-hit approach, we have reconstructed networks of disease signatures, identifying clusters in which several diseases are represented. We want to validate these comorbid signatures using the transcriptomic PsyCourse data combined to the relevant clinical variables associated to somatic risk factors, and validate that the comorbid signatures are associated with a higher risk for these diseases in the PsyCourse cohort.
1. Using the transcriptomic data from PsyCourse, we will apply NNLS to verify exposure of the patients to the comorbid signatures extracted from our training dataset
2. Association of the exposure to the comorbid signatures with relevant clinical variables which are predictive of disease risk for T2D and CVD will be verified. This will be performed using a multiple regression including covariates.
3. We will perform negative controls to verify that non-comorbid signatures are not/less associated to somatic risk factors.
OMICS dataset: RNA-seq transcriptomic on case samples (Raw dataset already available through Riya and preprocessing pipeline will be from ZI/Heidelberg)