About mLoc-mRNAL

Localization of mRNAs has several advantageous over targeting protein localizations, and thus often utilized for targeting location of proteins and their functions. Keeping in mind the importance of localization of mRNAs, this tool has been developed for predicting multiple sub-cellular localization of mRNAs. The mLoc-mRNA is capable of predicting the probabilities for each mRNA sequence to be predicted in nine different localizations that are cytoplasm, cytosol, endoplasmic reticulum, exosome, mitochondrion, nucleus, pseudopodium, posterior and ribosome. The developed model achieved area under ROC curves (auROC) of 78.13, 75.63, 75.54, 76.47, 98.98, 80.28, 76.73, 98.90 and 78.40% for the respective localization, while accuracies are measured following 5-fold cross validation. Further, auROC of 74.72, 76.30, 73.57, 76.49, 95.81, 78.75, 69.56, 98.45 and 77.91% are obtained while the model is evaluated using an independent test dataset. This tool will certainly supplement the future endeavor in the direction of mRNA localization study.

(Image source: Wang and Li 2012)

Please Cite:

Meher, P. K., Rai, A. and Rao, A. R. (2021). mLoc-mRNA: predicting multiple sub-cellular localization of mRNAs using random forest algorithm coupled with feature selection via elastic net. BMC Bioinformatics22(1), 1-2