Intro
There seemingly is a difference in amount of repeats, between Turn on labs (same state) versus Turnoff labs (exclusion). So I have made strategies that aims to reward designs with a good percentage of base repeats.
There seems to be a relation between too high or too low amount of repeat bases and few clusters. Not sure what that is about, since we now got A ratio in our designs in a fairly good rate. So I don’t think the repeat A’s are to blame this time. 
The turn on labs seems to need more base repeats than turnoff labs. Those labs have open tail regions which may explain a part of it.
Strategy
In Turn on labs
- Reward designs that have a 40 to 50% base repeat.
- Decrease reward for 35 to 49 and 51 to 55%
- Penalize for higher or lower repeat ratio.
In Turnoff labs
- Reward designs that have a 30 to 40 % repeat bases.
- Decrease reward for 25-29% and 40 to 45%
- Penalize for higher and lower repeat ratio.
Strategy Background
The natural riboswitches showed a high amount of repeat bases. This is not fully the case for ours yet. Part of it may be due to limits of amount of repeat C’s and G’s, there are longer stretches of these in the natural riboswitches.
However while I expect the repeat bases to grow in prominence and importance, I have made a strategy for now.
What I basically wish it to judge is what amount of single bases there are compared to repeats of the same color. Like what is normal amount of single A’s to A repeat bases in top scorers.
I see turnoff labs behave differently from turn on labs so I have made separate strategies. Similar I see microRNA labs behave different from the other MS2 labs, so I have separated those too.
The amount of repeat in the microRNA designs are going to be dependent too of how much repeat there are in the microRNA sequence itself. But even there I think there will be a somehow good ratio.
Background articles
Different types of switches