Figure 2.

Different use cases for LipidSpace. (A) Confirmation of discoveries in published studies: (upper) LSM 18:1;O2 performs well to separate all measured samples based on their condition³² with an accuracy of 100%, (lower) main differences in lipid concentration of male and female (either taking contraceptives (CC) or not) plasma lipidomes²⁸ caused by PC, PE, and SM lipids. 23 lipids can achieve separation of data sets with an accuracy of 93%; (B) extended analyses on published studies: (left) feature analysis on Smpd1 knockout³² data reveal a correlation between the increase of PG and decrease of CL and a systematic shift of double bonds for Cer 18:x;O2 subclass, (right) six ethnicity-dependent studies show that SM had the most lipid species involved in the gender separation models; (C) QC of seven different human plasma experiments: (left) global structural space of all 1037 samples, (top right) hierarchical clustering of all samples shows similarities across four studies (experiment 1–4 from study,³⁴ ex. 5 from ref (27) 6 from ref (28), and 7 from ref (33)), (bottom right) study aggregated structural spaces reveal differences of lipid species compositions (at most on sn-position level) and concentrations among studies.