Fig. 4

In silico modeling of Snhg15 binding to Ncam1 and Adamts14 promoter regions. a Snapshot of Snhg15 triple helix formation with the extended Ncam1 sequence (Ncam1-ext) during MD simulations. b Total interaction energy (LJ + CB) between DNA and RNA with phosphate and sugar backbone included as a function of time. When comparing Ncam1 and Ncam1-ext, we only considered the sequence without the extension (clipped Ncam1-ext sequence, see Additional file 3: Table S2) to maintain comparable energies for both models. c Number of hydrogen bonds between RNA and DNA. Please note that the hydrogen bonds counts have been smoothed using a running average with a window size of 5, as the curves would otherwise overlap too much. Again, we only considered the sequence without the extension (clipped Ncam1-ext sequence) to compare Ncam1 and Ncam1-ext. d Occurrence of in-register hydrogen bonded pairs between the RNA and the DNA–purine strand in Adamts14-1 along the simulation trajectory. Labels starting with a “D” indicate the DNA residue of the pair. e Occurrence of in-register hydrogen bonded pairs between the RNA and the DNA–purine strand in Ncam1-ext (clipped) along the simulation trajectory. Note that predicted interactions get more stable over time as indicated by less noise in the lower lanes. Labels starting with a “D” indicate the DNA residue of the pair. LJ: Lennard–Jones. CB: Coulomb