Table 1.
Clinicopathological characteristics of the tumor patients.
Table 2.
The numbers of remaining features and their percentages after each data filter.
Figure 1.
PCA score plots with different normalisation methods.
Cancer subjects are labelled in red, controls in blue and QCs in green. The vector from diamond to 5-point star is labelled in black, to 4-point star in green and to inverted triangle in purple. (A–C) Normalisation to creatinine, MSTUS and osmolality respectively. (D) raw data without normalisation.
Figure 2.
OPLS-DA score plot with different normalisation methods.
Cancer subjects are shaped as squares and controls as circles. The training set is labelled in filled red and the test set in hollowed blue. The four selected subjects and the vectors are same shaped and coloured as Figure 1.
Figure 3.
LC-HRMS/MS results of isomers of C5H10N2O3.
Extracted ion chromatograms under 4 different LC-HRMS conditions (A) and the interpretation of their MS2 fragmentations (B).
Table 3.
The statistical results for biomarkers surviving testing against a new cohort of patients.
Figure 4.
ROC testing results of validated biomarker (UIBA) and the combined biomarker.