Reproducing germline assignment from biophi paper #25
Description
Hi David, thanks for the package!
I am exploring idea from BioPhi paper with germline content as a scoring function, however scores that I get with original function:
def get_seq_germline_content(seq, scheme='imgt', per_position=False):
chain = Chain(seq, scheme=scheme)
germline_chain = chain.find_merged_human_germline().renumber(scheme)
aligned = chain.align(germline_chain)
if per_position:
return [(pos, aa == bb) for pos, (aa, bb) in aligned]
num_identical = len(aligned) - aligned.num_mutations()
return num_identical / len(chain)
are different in a number of cases from the original results:
not that some human Abs have low score.
My guess is that some germlines are missing now (I use abnumber==0.4.3). I've tried uncommenting everything in germlines.py, but see very little difference. LMK if you have ideas/suggestions.