pyrepseq
pyrepseq copied to clipboard
andim
Python library for immune repertoire analysis
The current implementation uses group_weights**2 / sum(group_weights**2) for averaging over different conditioning groups. Instead we should used the same implementation as the probability of coincidence.
1. Construct a pc cross function for doing general across group coincidences and work out how to average this properly 2. Integrate pc joint and pc conditional to work with...
Currently the TCRdist calculation on e.g. CDR3A will fail when no TRAV gene is in the dataframe. It would be preferable if each function worked with only its required input.
It would be nice to better link custom pyrepseq functionality to the publications where they were first described. I can think of at least four papers - Yuta's Tidytcells paper...
While we shouldn't overtest, there are a few bits of code that might need more testing. To help decide if and where we want more testing, here are some suggestions:...
If strings at index `x` and `y` are sufficiently similar to one another at distance `d`, the current symdel implementation returns both `(x, y, d)` and `(y, x, d)`. Unless...