Identification of Functional RNA Genes Using Evolved Neural Networks

Functional RNAs (fRNAs) play a key role in gene regulation, at both the transcriptional and translational levels. Identification of fRNA genes can be difficult, given that some classes of fRNAs (especially microRNAs) have short coding regions and do not use classical signals common to protein coding genes. This paper presents an approach to identify fRNA genes using evolved neural networks to discriminate between noncoding regions of genomes and regions that are likely to be fRNA coding. The results indicate that for human and C. elegans this approach can be used with considerable success.