The spectra were normalized to the total ion current intensity in the m/z range over 2000–50,000 to modulate peak dimension. The peaks ranging between m/z 0 and m/z 2000 were eliminated

from analysis to avoid the interference of adducts, artefacts of the energy-absorbing molecules and other possible chemical contaminants. Biomarker Wizard Version 3.1 (BMW; Ciphergen) was used to identify corresponding peaks in each spectrum (peak clusters). The settings for autodetect peaks to cluster were as follows: signal-to-noise ratio was 5 and minimum peak threshold was 0% for the first pass; for cluster completion, cluster mass window was 0.3%, and signal-to-noise ratio for the second pass was 2. Also, BMW helps pick out differently expressed Panobinostat molecular weight peaks by evaluating the differences of peak intensities between groups by non-parametric Kruskal–Wallis test and Mann–Whitney test [23, 24]. Peak intensities were considered statistically significantly different at P-values below 0.05. Construction of classification tree model.  Construction of the Daporinad price classification tree model was based on a platform of bioinformatic, Biomarker Patterns Software Version 5.0 (BPS; Ciphergen), which was developed basing on the Classification

and Regression Trees decision tree system [25]. The BPS helps build a binary decision tree algorithm with the peak information of the training set, and that algorithm assigns each sample into one of the two nodes according to some rules established by the intensity of certain peaks [16–19]. Staurosporine order The data of differently expressed peaks generated by BMW between active TB and non-TB group were used in this proceeding. The BPS generated some classification tree models and evaluated the

error cost (represented as ‘relative cost’ in BPS) for each one. Of those models, the one with the lowest error cost was the best, and then this resulting model was applied to the data of the test set for evaluating the efficiency of classification. The spectra of 178 serum samples were detected by MALDI-TOF MS combined with WCX magnetic beads. This combination was particularly effective in resolving low molecular weight proteins and peptides, as shown in Fig. 1. Peak of m/z 48 was found differently expressed between active TB group and non-TB group (Table 2). Reproducibility was evaluated by performing an 8-spot assay (intra-assay), a 20-chip assay (interassay) and a 20-day assay with a mixed serum sample (age and sex matched, two patients with TB and two health volunteers), and the coefficient of variations for peak intensity of spectra were 1.7%, 1.9% and 13.3%, respectively. These data derived from averaging values for nine of the highest amplitude peaks as follows: m/z 4309, 4963, 5344, 7772, 7846, 8058, 8608, 9413 and 16105.