Kujala M, Nevalainen J, März W, Laaksonen R, Datta S
Differential network analysis with multiple imputation was applied to lipidomic data from the LURIC study, comparing CAD patients who experienced a fatal cardiovascular event with those who remained stable over a two-year follow-up. Left-censored missing values were addressed via multiple imputation combined with permutation testing on partial least squares regression scores. The approach identified lipids and lipid classes interacting differentially between the two subgroups, providing a methodological framework for incomplete high-dimensional lipidomic datasets.
PloS One, 30. März 2015 | PMID 25822937