Predicted occupancy of the miRNA on an mRNA target is calculated using the biochemical+ model, which uses predicted relative KD values for 12-nt sequences within the mRNA, further modified by other features of the sites within these sequences, including predicted structural accessibility, evolutionary conservation (PCT score), and 3'-pairing score (McGeary, Lin et al., 2019). TargetScan considers predicted relative KD values for 12-nt mRNA sequences encompassing one of the six canonical types to the miRNA (8mer, 7mer-m8, 7mer-A1, 6mer, 6mer-m8, 6mer-A1); values for 12-nt sequences with only noncanonical sites are not considered. The biochemical+ model penalizes 12-nt sequences mapping to ORFs and does not consider those mapping to 5' UTRs (McGeary, Lin et al., 2019).

TargetScan 8 predicts occupancy using three different estimates of the concentration of free AG0-miRNA. The highest concentration (4.1 x 10^-3 μM) estimates the concentration of free AG0-miRNA after efficient transfect of the miRNA into cultured cells (transfected miRNA); the middle concentration (5.5 x 10^-4 μM) estimates the concentration of free AG0-miRNA for a miRNA with very highly endogenous expression (high miRNA), and the lowest consentration (7.5 x 10^-5 μM) estimates the concentration of free AGO-miRNA for a miRNA express at a level that is more moderate yet sufficient to mediate repression (moderate miRNA).