In experimenting with trying to create an inverse cantor dust model of RNA, which produces a string-heavy RNA as opposed to a loop-heavy RNA, I’ve found through turning up the gain on the model that the more strings you place close together, the more unstable the RNA becomes. However, adding distance only significantly contributes to stability up to a point. As such, this strategy likely may not apply to simpler RNA models. In this case, a string shall refer to any set of stacks, which may or may not contain loops, originating from a main loop of unpaired nucleotides.
My strategy is to consider a distance of 3 nucleotides between each string to be neutral, and bear a score of zero. The RNA will be penalized a point for each nucleotide less than 3 between each string, down to and including zero. The RNA will be granted a point for each nucleotide more than 3 between each string, up to and including 6.
For example, let ‘.’ represent an unpaired nucleotide, and let ‘()’ represent a string of stacks.
If .()(). Then the score is -3 or 3 penalizations.
If .()…(). Then the score is 0
If .()…(). Then the score is +3
If ()…() Then the score remains +3
If .()…().()…(). Then the score is -4, or 4 penalizations.