I have found that C tends to be most prevalent at a specific spot in the neck area in the lab data. Like the marked nucleotide in the picture.
As Mat added when I discussed my forum post with him in chat: Eli - I do like my C there on most designs.
I won’t severly penalize the GC-pairs that turn opposite the most prevalent pattern, as they occour in good designs in 20-25% of the time. Mat added: Eli - yep - I have two 97% with “GC-pairs that turn opposite the most prevalent pattern”
So I will not rule them totally out as it would be a shame, just push the algorithm in the way that has highest success rate.
For theory behind see this post. Energy in neck aligned with energy spots in the design
I would like a strategy that says:
Give + 2 if marked base is a C (it’s basepair partner should be G)
Give +1 if marked base is a G (it’s basepair partner should be C)
Give 0 if end basepair is GU and AU basepairs. I know that a few time GU and AU- closing pairs will fly, but so far they are very rare.