Figure 5. Protein abundance carries more information for the prediction of sampling probability than gene expression level.

Boxplots of the Spearman correlation coefficients resulting from one hundred 5× cross-validation runs for regression models that either include gene expression data or protein abundance data.