With the rapid growth of RNA sequences generated in the postgenomic age, it is highly desired to develop a flexible method that can generate various kinds of vectors to represent these sequences by focusing on their different features. This is because nearly all the existing machine-learning methods, such as SVM (support vector machine) and KNN (k-nearest neighbor), can only handle vectors but not sequences. To meet the increasing demands and speed up the genome analyses, we have developed a new web server, called "representations of RNA sequences" (repRNA).
Compared with the existing methods, repRNA is much more comprehensive, flexible and powerful, as reflected by the following facts:
(1) it can generate 11 different modes of feature vectors for users to choose according to their investigation purposes;
(2) it allows users to select the features from 22 built-in physicochemical properties and even those defined by users' own;
(3) the resultant feature vectors and the secondary structures of the corresponding RNA sequences can be visualized.