Make tokenizer for gene sequences

[Lev1ty]

Create a tokenizer for base pairs of gene sequences.

• Encode base pairs
• Encode position of base pairs
  • With respect to sequence
  • With respect to transcription start site

[Lev1ty]

Gather wet lab expertise on position embeddings

[Lev1ty]

Deliverables

Encode base pairs

• [ ] Method to map a sequence of nucleotides to a sequence of ids
• [ ] Method to map one nucleotide to one id, i.e. the mapping A,C,G,T -> 0,1,2,3
• [ ] Method to collate offset sequences to for k-mer embeddings, e.g. ACGGGTCA -> {{0,1,2,2,2,3,1,0},{1,2,2,2,3,1,0},{2,2,2,3,1,0}} -> {{0,1,2},{1,2,2},{2,2,2},{2,2,3},{2,3,1},{3,1,0}} for k = 3

Encode position of base pairs

• Waiting on wet lab