A new regression method for predicting likelihood of target mRNA down-regulation from sequence and structure features in microRNA/mRNA predicted target sites.
MirSVR is a new machine learning method for ranking microRNA target sites by a down-regulation score. The algorithm trains a regression model on sequence and contextual features extracted from miRanda-predicted target sites. In a large-scale evaluation, miRanda-mirSVR is competitive with other target prediction methods in identifying target genes and predicting the extent of their downregulation at the mRNA or protein levels. Importantly, the method identifies a significant number of experimentally determined non-canonical and non-conserved sites.
mirSVR scores are available at www.microRNA.org.
Betel D, Koppal A, Agius P, Sander C, Leslie C. (2010) Comprehensive modeling of microRNA targets predicts functional non-conserved and non-canonical sites. Genome Biology 11, R90. [article]
Subscribe to the miRNA blog
Thank you for subscribing.
Something went wrong.
I want to know how many mirSVR scores is good for miRNA target prediction, just like mirSVR scores is -1.1078 is good? thanks.
Hi Binghai,
The suggested mirSVR score cutoff is -0.1 or lower. This is based on the empirical distribution of the extent of target downregulation (measured as log-fold chage) that is expected given a mirSVR score. For scores closer to zero the probability of meaningful downregulation drops while the number of predictions rises sharply. So a target site with mirSVR score of -1.1 is expected to be a functional site.
Doron
I noticed that the miR-SVR scores are different if you look at predicted targets for the miRNA vs predicted miRNAs that target a specific protein. For example, The miR-SVR score for Ctnnb1 is -1.4 when I find it as a predicted target of miR-690 (mouse), but when I use the target mRNA tab (typing in Ctnnb1), the score comes out as -0.5365. Why is that? How can I tell if I have a good score when I am using the target miRNA tab?