In miRNA target prediction, typically two levels of information need to be modeled: the number of potential miRNA binding sites present in a target mRNA and the genomic context of each individual site. Single model structures insufficiently cope with this complex training data structure, consisting of feature vectors of unequal length as a consequence of the varying number of miRNA binding sites in different mRNAs. To circumvent this problem, we developed a two-layered, stacked model, named miSTAR (miRNA stacked model target prediction), in which the influence of binding site context is separately modeled. Using logistic regression and random forests, we trained a stacked model on a unique dataset of 7990 probed miRNA-mRNA interactions. The miSTAR model is described in Van Peer & De Paepe et al. (2016), Nucleic Acids Research.

The miSTAR webtool catalogues predictions for all human miRNAs in release 21 of miRBase and all human protein-coding transcripts in version 75 of Ensembl.