Features

1. Reverse translates input amino acid sequence to DNA.

2. Calculates the host’s per-AA codon usage profile – codons used less than a specified threshold (defaults to 10%) are dropped.

3. Compares the reverse-translated DNA sequence to the host profile, determines which codons are overused/underused.

4. Stochastically mutates codons according to host profile.

5. Ranks sequences by codon adaptation index relative to host

6. Processes DNA to remove unwanted features:

   • high GC content within a sliding window and across the entire sequence

   • unwanted restriction sites

   • alternate start positions (GA-rich regions 18 bp upstream of ATG/GTG/TTG)

   • 3-consecutive identical codons and 9-mer repeat chunks

   • areas with more than 4 (variable) consecutive identical bps (“local homopolymers”)

   • RNA hairpins, detected by looking for 10-mers with reverse complements (including wobble bases) in the sequence

   • RNA splice sites, detected by similarity to consensus donor and acceptor site sequences

The process is repeated from step 3 for a specified number of cycles (defaults to 1000) OR until the per-AA codon profile of current DNA and host profile matches (within tolerance).

Future work

• More advanced RNA-structure removal

  • CONTRAfold – overkill for now

  • nupack – overkill for now

0.9.2 (2019-02-06)

• First release on PyPI.

0.9.4 (2019-02-20)

• Full suite of tests added, bugs uncovered and fixed

• Adjustments to the packaging setup – actaully installable now

0.9.5 (2019-02-25)

• Adding support for RNA splice site detection and removal

0.9.6 (2019-02-28)

• Updating the way optimization failures are reported and displayed

• Parallelizing via a process pool

1.0.0 (2019-03-06)

• Added ability to use offline tables in addition to fetching from the internet

• Full suite of tests and documentation

• Tested on real-world sequences to