Create a seed model

[stephaniesamm]

Create a seed model using local alignment of miRNA seed sequence to the target sequence.

Create a function for each of:

1. extract miRNA seed; 1-based positions 2-7 of miRNA (we call this a seed6mer)

INPUT: miRNA seq
OUTPUT: miRNA seed

2. reverse complement (rv) the extracted seed sequence

INPUT: miRNA seed
OUTPUT: rv of miRNA seed

3. align the rv sequence to the target sequence (you can use the Biopython package for a local alignment using the Smith-Waterman algorithm)

INPUT: rv of miRNA seed, target seq
OUTPUT: alignment score

4. apply functions above to each of the examples in the Hejret test set

INPUT: hejret test set (miRNA, target, label cols are enough)
OUTPUT: hejret test set with added column called seed6mer_alignment containing the alignment score for each row / pair of miRNA-target.

[stephaniesamm]

@VictoriaDeguara I thought about the alignment a bit more and what you need to do is called semi-global alignment not local aligment; this will align one sequence completely (the miRNA seed), while allowing the other sequence (the target) to have unaligned ends without penalty. In local alignment (Smith–Waterman), the algo finds the best matching subsequences anywhere within the two sequences. In global alignment (Needleman–Wunsch) the algo aligns both sequences from end to end.

[stephaniesamm]

@VictoriaDeguara You do not need to create a .gitignore file for every directory. Use the original one in the repo root, and keep adding what to ignore there. For example if you want to ignore all .venv/ dirs, I think **/.venv/ should work.