Fig. 1
Schematic representation of the experimental workflow. Sixteen plasma samples of patients with new-onset T1D (i.e., between 5–21 days postdiagnosis) were depleted using ProteoMiner™ kit prior to an overnight digestion by trypsin. Resulting peptides were labelled using isobaric TMTpro 16plex™ before being pooled into a single sample and further separeted into 51 fractions using HILIC. Most orthogonal fractions of labelled-peptides were pooled together (i.e., 1 + 26, 2 + 27,[…], 25 + 51) resulting in a total of 25 DDA LC–MS/MS runs that were subsequently analyzed and quantified using Proteome Discoverer Software. Correlations between protein relative abundance and IDAA1C score was performed using linear mixed models. Various qualitative and statistical approaches were used to filter best candidates within significant proteins resulting in a total of 26 candidates. FDR false discovery rate, HILIC hydrophilic interaction liquid chromatography, IDAA1C insulin dose-adjusted A1C, ORA overrepresentation analysis, RF random forest, SVM support vector machine, T1D type 1 diabetes.
