Figure 4.

Validation phase data analysis

(A) Biomarker candidates were selected first based on statistical test with p value correction. Additional candidates were selected for validation based on the p value, number of samples in which the peptide was detected, spectral count, machine learning, and previous reports in the literature.

(B) Up to 5 peptides for each candidate protein were selected based on their physicochemical properties and probability ranking for a likely successful measurement by targeted proteomics.

(C) Cross-validated proteins across discovery (D) and validation (V) phases. Only significant, validated proteins are represented in the heatmap and are colored based on their regulation. Time points are represented by months prior (−) or post (+) seroconversion. Comparison I1: time point(s) before seroconversion of the group that remained in autoimmunity by the age of 6 years (IA group) paired against matched controls (n = 401). Comparison T1: time point(s) before seroconversion of the group that developed T1D by the age of 6 years (T1D group) paired against matched controls (n = 94). Comparisons I2 and T2 have the same group of individuals as I1 and T1, respectively, but after seroconversion. Comparisons I3 and T3 compare IA and T1D groups before vs. after seroconversion, respectively.