Triplex Domain Finder: Detection of Triple Helix Binding Domains in Long Non-Coding RNAs

Abstract Long (>200vbps) non-coding RNAs (lncRNA) can act as a scaffold promoting the interaction of several proteins, RNA and DNA. Some lncRNAs interact with the DNA via a triple helix formation. Triple helices are formed by a single stranded RNA/DNA molecule, which binds to the major groove of a double helix following a canonical code. Recently, sequence analysis methods have been proposed to detect triple helices for a given RNA and DNA sequences. We propose the Triplex Domain Finder (TDF) to detect DNA binding domains in RNA molecules. For a candidate lncRNA and potential target DNA regions, i.e. promoter of genes differentially regulated after the knockdown of the lncRNA, TDF evaluates whether particular RNA regions are likely to form DNA binding domains (DBD). Moreover, the DNA binding sites from the predicted DBDs are used to indicate potential target DNA regions, i.e. genes with high binding site coverage in their promoter. The command line tool provides results on a user friendly and graphical html interface. A case study on FENDRR, an lncRNA known to form triple helices, demonstrates that TDF is able to recover both previously discovered DBDs and DNA binding sites. Source code, tutorial and case studies are available at www.regulatory-genomics.org/tdf .