Correlation of DIANA-microT 3.0 prediction measures to protein repression. Fold changes are calculated for approximately 5,000 proteins after overexpression of a miRNA. The results for five miRNAs, as provided by Selbach et al., are used. The fold change and the miTG score is averaged in groups of 150 proteins sorted by fold change. (a) The correlation of several miRNA target prediction measures with protein production fold change induced by the same miRNAs. It may be observed that there is a trend for values of all the measures to increase as the level of downregulation increases. (b) The red line indicates the correlation between the anti-seed 2 occurrences on the 3'UTRs of downregulated genes with the protein production fold change of the corresponding genes using a linear regression. The blue line shows the corresponding correlation for a linear regressor based on a combination of the miTG score, the precision, the SNR and the anti-seed 2 frequency. The combined linear regressor correlates better with the protein production fold change than the regressor based solely on the anti-seed 2 frequency.
Maragkakis et al. BMC Bioinformatics 2009 10:295 doi:10.1186/1471-2105-10-295