Skip to main content
. 2021 Oct 13;11:20322. doi: 10.1038/s41598-021-99586-1

Figure 6.

Figure 6

Influence of decreased number of lipid species used for MDA using UHPSFC/MS data in the discovery phase. (138 lipids: no exclusion, 91 lipids: common lipids from the discovery and qualification phase using UHPSFC/MS and shotgun MS, 29: only lipid species included, which are significant according to the Bonferroni correction for at least 10 from 15 models. SM 38:1 and SM 42:1 were excluded, because these variables were only significant for UHPSFC/MS data sets (10/10) and therefore a method bias cannot be excluded. 7: only lipid species included, which are significant according to the Bonferroni correction for at least 14 from 15 models. The average of the specificity (blue), sensitivity (red), and accuracy (green) for validation and training set as well as for both genders were calculated. (A) kidney cancer, (B) breast cancer, and (C) prostate cancer.