Fig. 5

Models’ performance in the testing dataset. a PCA using the 37 selected transcripts in this independent dataset also demonstrated strong separation between samples from patients with IA and from controls. b Assessment of true model performance showed that models performed with an accuracy of 0.83–0.90. In this dataset all models had a sensitivity of 1. At 5% IA prevalence, the PPV ranged from 0.15 to 0.24 and NPV was 1 for all models. c ROC analysis showed that all models again had AUCs ≥ 0.95. d PCA using the 26 previously-identified transcripts demonstrated mediocre separation between IA and control cases. e Estimation of model performance in the testing cohort demonstrated that models performed with an accuracy of 0.83–0.93. Considering a 5% prevalence, PPV and NPV ranged from 0.15–0.52 and 0.99–1, respectively. f ROC analysis also showed inferior performance compared to newly identified transcripts (AUC range 0.84–0.97). (AUC = area under the ROC curve, C-V = cross validation, cSVM = cubic support vector machines, gSVM = Gaussian support vector machines, KNN = k-nearest neighbors, LOO = leave-one-out, NPV = negative predictive value, PCA = principal component analysis, PPV = positive predictive value, RF = random forests, ROC = receiver operator characteristic)