Computational model for ncRNA research
Xing Chen, Huang Li
Abstract
The explosion of research on non-coding RNAs (ncRNAs) in the past few decades has transformed the original notion of regarding such RNAs as ‘transcriptional noise’ [1, 2] before 1980s to gene expression regulators at transcriptional, RNA processing and translational levels [3–6]. It is evident from PubMed that each week new studies are published to reveal altered ncRNA expressions in diseases or discover novel non-coding transcripts [7]. Abundant in eukaryotes ranging from Homo sapiens to Caenorhabditis elegans [8, 9] and persistent in prokaryotes such as bacteria [10, 11], ncRNAs fall into two main classes according to transcript length [12, 13]: long ncRNAs (lncRNAs, with length > 200 bp) and small ncRNAs (sncRNAs, with length < 200 bp). Based on conformation and cellular function [14–17], the former can be further divided into linear RNAs and circular RNAs (circRNAs), whereas the latter can be grouped into more than a dozen subclasses, among which microRNAs (miRNAs) and piwi-interacting RNAs (piRNAs) have recently attracted considerable research attention.