Thermodynamic prediction of RNA-DNA duplex-forming regions in the human genome

RNA-DNA hybrids can form in a physiological context, especially as a consequence of transcription, accompanied by the separation of the second strand of DNA. These structures seem to be important in regulating some aspects of transcription, and also form in the immunoglobulin switch domains. In some cases they are sufficiently large and stable to be directly visualised. I present a thermodynamic analysis of their formation, based on known experimental data. I then use this analysis to predict 28 700 regions in the genome likely to form RNA-DNA hybrids when the RNA strand is present, producing a list of regions for experimental analysis, as well as rationalizing the formation of RNA-DNA hybrids in previously known regions.