I do believe you have the values reversed for the FMN mimic scoring. It should be-
- A base is considered paired if SHAPE reactivity is less than 0.5 and unpaired if greater than 0.25
The problem I have with that definition is that if I manage to create a sequence that gives a consistent 0.4 SHAPE signal throughout all bases, then I win no matter the target structure! :P
Maybe for the MaxReward the paired base must be less than 0.5, the unpaired must be greater than 0.25, and the paired must be at least 0.25 less than the unpaired?
Well, I suppose a person could get get really lucky with a reactivity signal of .4 for every scored nucleotide! Although, with the current scoring scheme, there is currently an overlap where all bases between .25 and .5 are not scored :D
@ meechl - maybe. If there are any modifications to the scoring protocol, feel free to propose and implement them! I am sure we will change the way switches are scored several times, before we reach the end of this project.