Notes from Church lab folks (DBG, Gleb, Marc Lajoie)
- Start with wild type sequence and mutate N-terminal codons to remove secondary structure.
- Gleb has Python modules that overfit on their Recoli-57 project
- https://github.com/churchlab/getk/blob/master/src/getk/codon_usage_memex.py
- Factory caller at the bottom
get_ecoli_codon_usage_memex()will give you back a usage memex object.
- Factory caller at the bottom
- A messier version but with more codon usage calc
- Since you’re probably aiming for high expression, you probably just want to write a function to predict min free energy of your 5-prime mRNA
- You can do so using https://github.com/churchlab/getk/blob/master/src/getk/hybrid_ss_min_wrapper.py, which requires installing Unafold and specifically the hybrid_ss_min tool
- We typically pass (-30, +100) relative to start codon
- So generate some sequences and pick one for which predicted MFE is closer to 0 (less negative) as possible
- NOTE the https://github.com/churchlab/getk repo is still private and we have a license/patent situation going on so let me know if you pull out any parts into a separate open source library.
- DBG and Gleb still think it would be beneficial to get a real nucleotide sequence and tweak that rather than generate a sequence from scratch. Even if you have to tweak it a lot to fix GC content / codon usage.
- DBG: If you are moving from low GC to high GC, then wholesale codon opt makes more sense, but otherwise, I would just keep it WT and optimize the n-term folding.
- Marc: as a more general answer for arbitrary genes, you can also use dnaworks (I think that's open source and works on command line).