ARTEM: a method for RNA and DNA tertiary motif identification with backbone permutations

Abstract

Non-coding RNA functions are largely defined by their 3D structures, which consist of recurrent building blocks, tertiary motifs. The computational motif search problem remains largely unsolved, as standard approaches are restrained by sequence, interactions, or backbone topology. We present ARTEM 2.0, which enables automated, unrestrained searches of RNA and DNA structure databases to identify 3D motifs. We apply ARTEM for searching kink-turns, G-quadruplexes, GNRA tetraloops, and i-motifs. ARTEM outperforms existing methods and enables the discovery of novel motif variants. ARTEM opens a fundamentally new way of studying nucleic acid 3D folds and motifs and analyzing their correlations and variations.