Only 2% of human DNA involves genes that code for proteins, the machinery of our cells. The DNA with genes is copied into mRNA read by ribosomes that synthesize then the respective proteins. However, 80% of the remainder DNA is also copied into RNA molecules that assume then manifold functions, one being that the RNA molecules bind to mRNA and silence thereby the respective genes. Frequently, gene-silencing RNA is activated with the help of proteins called double stranded RNA binding domains (dsRBDs). These domains bind to parts of gene-silencing RNA that happens to form double strands, similar but not identical to the double strands formed by DNA and discovered long ago by Watson and Crick. In fact, there exist small but characteristic differences between double stranded RNA, double stranded DNA and hybrid double strands made of RNA and DNA. In a recent study, computational biologists performed simulations using the molecular dynamics program NAMD to determine how dsRBDs recognize silencing RNA and discriminate between double stranded RNA, DNA, and hybrid double strands. The simulations revealed that dsRBDs and double stranded RNA fit together ideally like matching pieces of a puzzle, with mutually compatible shapes and electrostatic patterns. On the other hand, dsRBDs and double stranded DNA or hybrid double strands have poor fits due to changed and insufficiently flexible double strand forms. More here.