kujala abstract

Abstract: The importance of lipids for cell function and health has been recognized, e.g., a disorder in the lipid spectrum of cells has been related to atherosclerosis caused cardiovascular disease. Differential network analysis provides a formal statistical methodology to examine differences in network structures under two biological conditions, and responds to the need for efficient and interpretable analysis tools in the fields of lipidomics. We provide a recipe to conduct permutation based statistical test on association scores resulted from partial least square regression on multiple imputed lipidomic data from the LUdwigshafen RIsk and Cardiovascular Health study, particularly paying attention to the left-censored missing values typical to most life science data sets resulting from mass spectrometry platform. With accordingly customized network analysis, we take full advantage of the data achieving useful information about the underlying biological process, find lipids that interact with each other, and recognize the most important differentially expressed lipids between two subgroups of CVD patients.