Tailored machine learning models for functional RNA detection in genome-wide screens

The prediction of noncoding RNA and protein coding genetic loci has received<br> considerable attention in comparative genomics aiming in particular<br> at the identification of properties of nucleotide sequences that are<br> informative of their biological role in the cell. We present here a<br> software framework for the alignment-based training, evaluation and<br> application of machine learning models with user-defined<br> parameters. Instead of focusing on the one-size-fits-all approach of<br> pervasive \is annotation pipelines, we offer a framework for the<br> structured generation and evaluation of models based on arbitrary<br> features and input data, focusing on stable and explainable results.<br> Furthermore, we showcase the usage of our software package in a<br> full-genome screen of Drosophila melanogaster and evaluate<br> our results against the well-known but much less flexible program<br> RNAz.