Different level hardness of labs, takes different strategies.
I think we need a strategy for those labs whose structure forces them to need to go full moving. They can’t be solved with what works for a easy switch. Similar easy labs are very hard to solve going full moving strategy.
So EternaBot needs to be able to recognize when a switch lab puzzle is hard to solve. Hard labs happen when the switching parts starts to get too far apart. For Exclusion labs, things start to get harder as soon as the MS2 starts to get distanced from the one aptamer sequence. Potentially with just a 4 base distance as even more stressful than a longer distance between the switching elements.
-
Penalize multiloops - in any state
-
Penalize 0 GU’s and allow for double amount of GU’s to usual.
-
Penalize static parts, the more the harder penalty. Only allow for a single static hairpin somewhere in the middle of the sequence in the not hardest pressured labs.
-
Target both FMN’s at once, for turnoff (not with same turnoff sequence)
-
All stems switching - or leave a small hairpin static somewhere in the middle of the sequence, if the lab isn’t the hardest of hardest.
-
Extra turnoff sequences as switch elements are too far apart for them to get shared. Have 2-3 turnoff sequences, so the FMN bits and the MS2 don’t necessarily have to share.
Pressured designs, unconnected designs and normal easy designs