A Review on Recent Computational Methods for Predicting Noncoding RNAs

Figure 1

Four popular categories of computational methods in predicating ncRNAs. (a) Homology-based methods, which compare a query RNA with known ncRNAs deposited in databases based on sequence or structure alignment; (b) de novo methods, which predict ncRNA from primary sequences or structure based on general principles that govern ncRNA folding or statistical tendencies of -mer features; (c) transcriptional sequencing and assembling based methods, which utilize next-generation sequencing and transcriptome data; and (d) RNA family specific methods, which predict specific ncRNA classes.

(a)

(b)

(c)

(d)