Scripts breakdown

---

• hg_intronGrabbing.py Grabs introns from the UCSC Genome Browser via API calls

• TF_assess_RE-removal.py Intelligently mutates out RE site(s) while maintaining transcription factor binding profile strength among categorized PWMs in the HOCOMOCO database, scoring with SARUS.jar

• codon_optimization.py Fancy and greedy codon optimizer. Matches fractions of codons per amino acid to a webscraped host genome for an input amino acid sequence, while scoring and sorting by:

  • Renyi entropy (measure of evenness of ATCG distribution)
  • CpG/ApU scores (calculated from count and desity)
  • chosen list of enzymes which do/don't cut in specified region(s)

• BioNeutral_RandDNAseq.py Functionially similar structure to the codon optimizer, but instead creates random sequence of input length, matching base-ratio of a webscraped host genome, and sorting against known structures and sites

• SpliceAl-multi-seq_inContext.py Takes the many-sequence output of the codon optimizer, and scores the top x sequences for forward (and optionially reverse orientation) splice sites within an input sequence context. Averages results from 5 AI models trained on splicing in the human genome, courtesy of Illumina.

• sites_grapher.py Interactively graphs accuracy of splice prediction against experimentally-derrived sites. Uses merged_sites.csv