Fig. 3

Evaluation of the capacity of the proposed pipeline to identify potential circulating biomarkers. a Workflow used to identify potential oxidative stress biomarkers in plasma samples from Parkinson’s Disease (PD) and controls (Ctrl). The workflow used is an adaptation of the pipeline applied to the secretome samples (Fig. 2a) with the addition of a “targeted” data extraction using the spectral library obtained in the secretome analysis to complement the results from the sample-specific library. The selection of candidates was performed by combining statistical analysis and an informed selection of the mitochondrial-related proteins. b Venn diagrams comparing the libraries used in the extraction process considering only the 98 altered proteins. The mitochondrial-related proteins found changed are highlighted in blue. c PCA analysis using Clusterin and VPS35, two mitochondrial-related proteins involved in apoptotic mechanisms. The contribution of each principal component explaining the total variance is indicated on top of the graphic, indicating some degree of separation of the two groups of samples. d Linear discriminant analysis (LDA) using Clusterin and VPS35. The generated LDA discriminant function is indicated on top, with the statistical confidence. The graphic represents the distribution of the samples considering their discriminant score. In the y-axis, it is represented the method decision with the indication of the cutting point (dotted line) and the centroid values (mean values). The error bars indicate the mean value ± SD of 28 and 31 samples from Ctrl and PD groups, respectively. The number of well-classified and misclassified samples are shown on the right side of the plots, as well as the calculated percentages of specificity (spec) and sensitivity (sens). The insert plot illustrates the Receiver operating characteristic (ROC) curve using the discriminant function generated with the two mitochondrial-related proteins. The area under the curve (AUC) value is indicated on the bottom of the image, with the respective statistical confidence