Skip to main content
Fig. 2 | Cancer & Metabolism

Fig. 2

From: Metabolic profiling reveals metabolic features of consolidation therapy in pediatric acute lymphoblastic leukemia

Fig. 2

Untargeted metabolomic cluster analysis identified significant plasma metabolite responses to high-dose methotrexate (HDMTX) therapy in pediatric acute lymphoblastic leukemia (ALL) patients. A Score plot showing sample values without outliers. After being fitted to an OPLS-DA model, features in the raw dataset are collapsed into 1 predictive component and i ≥ 1 orthogonal components. Observations (samples) are represented by scores, which are linear combinations of the original variables with weights defined by the loadings. For each 2 components (generally predictive (p1) and orthogonal (o1)), a score plot can be generated by putting them in X and Y axes. B Permutation test results, revealing greater prediction performance. In the OPLS-DA model, the higher the Q2Y, the better the prediction performance. The P value is equal to the proportion of Q2Yperm above Q2Y. C S-plot of the predictive (p1) and orthogonal (o1) components, with significant feature points highlighted by OPLS-DA (MV) and SAM (UV). D Heat map of average metabolite signal intensity for the differential metabolites selected by SAM analysis

Back to article page