Figure 4

Multivariate correlation analyses according to Jo1⁺ patient selection including the proteomics data analysis of the samples. (A) PCA scores plot. The Jo1⁺ patients and/or patients with ILD diagnosis cluster together along component 2 (x-axis) and component 7 (y-axis). This indicates that there are factors that are characteristic for this patient group within the model. (B) OPLS-DA dot plot of the cross validated (tCV) scores (circles) and of predicted scores (diamonds) obtained from the model set to identify Jo1⁺ correlating features in the data set (all factors are listed in Supplementary Table 6). Both the Jo1⁺ and Jo1⁻ patient profiles of time point 1 that were treated as known (i.e. from which the model was based of) and unknown (the validation cohort of patients profiles analysed at other time points), are significantly different. (C) The factors that correlated positively or negatively with Jo1⁺ patients with 95% confidence in the OPLS-DA model.