A miRNA target prioritization method to rank the predicted target lists from commonly used target prediction methods. Leave-one-out cross validation has proved to be successful in identifying known targets, achieving an AUC score up to 0. 84. Validation in high-throughput data proved that mirTarPri was an unbiased method. Applying mirTarPri to prioritize results of six commonly used target prediction methods allowed us to find more positive targets at the top of the prioritized candidate list. In comparison with other methods, mirTarPri had an outstanding performance in gold standard and CLIP data. mirTarPri was a valuable method to improve the efficacy of current miRNA target prediction methods.
College of Bioinformatics Science and Technology, Harbin Medical University, Harbin, China
mirTarPri funding source(s)
This work was supported by the National Natural Science Foundation of China (Grant Nos. 31100948, 61073136, 91129710 and 61170154), National Science Foundation of Heilongjiang Province (Grant Nos. D201114, QC2009C23).