Skip to main content
. 2019 Apr 18;14(4):e0214121. doi: 10.1371/journal.pone.0214121

Table 1. Selected metabolites as identified using CFS, PLS-DA, COR-LVQ and their common compounds.

Feature selection schemes
Metabolite Panel CFS PLS COR-LVQ OL
Creatinine lysoPC.a.C16:1 Creatinine Creatinine
C14 C2 lysoPC.a.C16:1 C2
C2 Creatinine lysoPC.a.C20:3 C4
C4 lysoPC.a.C18:2 C2 lysoPC.a.C16:1
lysoPC.a.C16:1 lysoPC.a.C18:1 lysoPC.a.C18:2 lysoPC.a.C2:.3
lysoPC.a.C18:1 lysoPC.a.C20:3 C4 lysoPC.a.C28:1
lysoPC.a.C20:3 PC.aa.C24:0 C12:1 PC.aa.C24:0
lysoPC.a.C20:4 C6.C4:1.DC. EDTAca_N
lysoPC.a.C28:1 C4 C6.C4:1.DC.
PC.aa.C24:0 C10:1 Taurine
PC.aa.C36:4 C16:1 C16:2
PC.aa.C38:4 C12:1 C0
PC.aa.C42:4 C12 Putrescine
EDTAca_N C0 PC.aa.C24:0
Creatine lysoPC.a.C28:1 lysoPC.a.C28:1

CFS: correlation-based feature selection, PLS: Partial least squares regression. COR-LVQ: Correlation based Learning Vector Quantization earning. O: Overlapped panel