I guess I need to make this a forum post and it might be a strategy market and has a feature request too
Strategy for labs where you have a series or more than one series of locked bases, and so few free choices.
Might work in other labs too but I’m guessing the variety of possiblities there is too huge.
step 1)
try anything you like (or ideally every combination) in the non locked bases
if you can find a natural shape that isn’t a line or the target shape then store those ‘misform’ shape notation (s) - this might be time consuming for players
[feature request] is to let a player ‘harvest the shape notation from the natural form’ good for puzzle building too.
Step 2)
go back to your best solutions
but take those base sequences
and paint them into puzzles with your ‘misform’ shape notations
Those sequences that look good in the lab already and do worst (most red where the arm’s shape notation does not match) in all the misformed shapes should be stronger candidates.
Even if the dotplot and meltplot and any bots don’t think that’s true - or maybe that’s what provides those values ? Hmmm maybe this might be a way of looking at what the dotplot is ? I’ll see if I can figure that out with one puzzle and lots of screenshots - if that is what the dotplot is I’ll make it into a guide to teh dotplot.
Step 3)
and maybe - just maybe 
You can add a blocking base (without changing a base from the lab locked set) to make more red in the misformed areas of the shape - and if that new sequence of bases still has the natural shape that matches the target shape for the lab - then you could look at using that in the lab rather than your previous ‘best’ solution?