Figure 4
MALDI-TOF-MS analysis for the GBM model. Left panel: comparison between (A) U251/ U251-V/U251-R, (B) LN229/ LN229-V/LN229-R, and (C) U251-R/LN229-R for the selection of discriminant peaks by the BDA algorithm. The algorithm outputs the t.score (group means vs. pooled mean), which is graphed in the x-axes of both graphs, and the peaks (m/z) are plotted on the y-axis. The sign of the t.score indicates the presence (positive) or the absence (negative) of that peak in that group. The groups are indicated with different colors. The differences have a 95% significance when the value of the t.score is equal to or greater than 2.5 (increased) and equal to or less than − 2.5 (diminished). Right panel: Spectra pattern comparison between (A) U251/ U251-V/U251-R, (B) LN229/ LN229-V/LN229-R, and (C) U251-R/LN229-R, was performed by the Hierarchical k-means clustering with a Principal Component Analysis visualization (HKmC-PCA). Each point on the graph corresponds to each experimental replicate. The label indicates an ID number of the spectrum and the cell line name. The spectra patterns were grouped in 3 clusters represented by different geometric symbols and colors. Lastly, the ellipses of each cluster represent 95% of confidence from the centroid, considering a normal distribution multivariate.
