Computational Identification and Functional Predictions of Long Noncoding RNA in Zea mays

Liang-Hu Qu,Susan Boerner,Karen M. McGinnis

doi:10.1371/journal.pone.0043047

Abstract

BackgroundComputational analysis of cDNA sequences from multiple organisms suggests that a large portion of transcribed DNA does not code for a functional protein. In mammals, noncoding transcription is abundant, and often results in functional RNA molecules that do not appear to encode proteins. Many long noncoding RNAs (lncRNAs) appear to have epigenetic regulatory function in humans, including HOTAIR and XIST. While epigenetic gene regulation is clearly an essential mechanism in plants, relatively little is known about the presence or function of lncRNAs in plants.Methodology/Principal FindingsTo explore the connection between lncRNA and epigenetic regulation of gene expression in plants, a computational pipeline using the programming language Python has been developed and applied to maize full length cDNA sequences to identify, classify, and localize potential lncRNAs. The pipeline was used in parallel with an SVM tool for identifying ncRNAs to identify the maximal number of ncRNAs in the dataset. Although the available library of sequences was small and potentially biased toward protein coding transcripts, 15% of the sequences were predicted to be noncoding. Approximately 60% of these sequences appear to act as precursors for small RNA molecules and may function to regulate gene expression via a small RNA dependent mechanism. ncRNAs were predicted to originate from both genic and intergenic loci. Of the lncRNAs that originated from genic loci, ∼20% were antisense to the host gene loci.Conclusions/SignificanceConsistent with similar studies in other organisms, noncoding transcription appears to be widespread in the maize genome. Computational predictions indicate that maize lncRNAs may function to regulate expression of other genes through multiple RNA mediated mechanisms.

Highlights

Expanding Roles for RNA in Eukaryotic Genomes A narrow interpretation of the central dogma would dictate that the ultimate contribution of RNA molecules to cellular function and phenotype is dependent upon their translation into proteins
Some long noncoding RNAs (lncRNAs) are likely precursor molecules that are processed into small RNAs, while others function as intact, long molecules that participate in a range of regulatory roles
The comparison of 3,122 mouse long noncoding RNAs with orthologous sequences in human and rat revealed selective forces acting on this set of noncoding transcripts similar to selective forces acting on coding transcripts

Summary

Introduction

Expanding Roles for RNA in Eukaryotic Genomes A narrow interpretation of the central dogma would dictate that the ultimate contribution of RNA molecules to cellular function and phenotype is dependent upon their translation into proteins. The shRNA category were classified based on the potential of their predicted precursors to form hairpin structures, and speculated to include currently unidentified miRNAs.The siRNAs would be predicted to generated by RNA-dependent RNA polymerase requiring pathways, including the RdDM pathway. These results suggest that noncoding RNAs may be an essential element of epigenetic gene regulation in plants, but relatively little is known about the abundance and activity of this type of molecule in plants. The use of machine learning algorithms, such as SVMs, has increased the accuracy of identification of protein coding potential to over 90 percent

A Computational Approach to Identifying Noncoding RNA in Plants

Results and Discussion

Materials and Methods