On the logic of DNA bilinearity