Machine learning–based analysis of differences in protein expression between VTE phenotypes. This figure displays the number of proteins selected by LASSO-regularized logistic regression in each model. The colors of lines indicate overlaps between models. All regression models were adjusted for age, sex, BMI, center, and medication intake. BMI, body mass index; CV-AUC, cross-validated AUC; λ ratio (LR) (unit on y-axis): λ ratio; N, number of selected proteins.