Relative K_D values are predicted using a convolutional neural network (CNN) that predicts binding affinity between any miRNA and any 12-nt sequence (McGeary, Lin et al., 2019). This CNN was trained on more than a million KD values measured for six AGO2-miRNA complexes, as well as on mRNA fold-change values observed after transfecting miRNAs into HeLa cells (McGeary, Lin et al., 2019).