Gene regulation is of utmost importance to cell homeostasis; thus, any dysregulation in it often leads to disease. MicroRNAs (miRNAs) are involved in posttranscriptional gene regulation and consequently, their dysregulation has been associated with many diseases.MiRBase version 21 contains microRNAs from about 200 species organized into about 70 clades. It has been shown that not all miRNAs collected in the database are likely to be real and, therefore, novel routes to delineate between correct and false miRNAs should be explored. We introduce a novel approach based on k-mer frequencies and machine learning that assigns an unknown/unlabeled miRNA to its most likely clade/species of origin. A simple way to filter new data would be to ensure that the novel miRNA categorizes closely to the species it is said to originate from. For that, an ensemble classifier of multiple two-class random forest classifiers was designed, where each random forest was trained on one species-clade pair. The approach was tested with different sampling methods on a dataset that was taken from miRBase version 21 and it was evaluated using a hierarchical F-measure. The approach predicted 81% to 94% of the test data correctly, depending on the sampling method. This is the first classifier that can classify miRNAs to their species of origin. This method will aid in the evaluation of miRNA database integrity and analysis of noisy miRNA samples.
Keywords: Categorization; Machine learning; miRNA.
© 2022. Springer Science+Business Media, LLC, part of Springer Nature.