TODO:

[zktuong]

Pre-processing
- [ ] Do we run reassign_genotype for IGL/IGK and/or the J genes? No.

• Add in a step to rename the locus for multiple IGHV/J chain calls as ‘Multi’-whatever. Maybe after reassign_alleles
• shift create_germlines to preprocessing?
• retrieve identity score and length of constant region after assign_isotype
• • also return the constant region sequence call to allow for manual checks and adjustments?
• create a new column where constant genes are collapsed to isotypes

Visualization and analysis

• Tabulate V/J gene usage – should be fairly straight forward
• • Add an option to not sort the genes.
• • Add a plotting function to plot this as a barplot
• • • and heatmap
• Calculate mutation load
• • Add a wrapper to shazam's basic mutational load R function
    - - [ ] Add a plotting function to add this as a barplot, heatmap and/or violin plot
    This can be achieve using scanpy's plotting modules
• • add in a step to 'fix' the V/J gene usage for clones. Need to figure out how to run this without breaking the function.
• • Enable option to return the results of light chain and heavy chain mutations separately.
• • Add a wrapper to call IGoR?
• • or are there other methods?
• Add a spectratype plots to show distribution of CDR3 junction lengths

Finding clones

• Add in a wrapper to call Changeo’s DefineClones.py as an alternative way to call clones.
• • Not sure if its working yet?
• • Also port in SCOPer?
• Add a function using graph-based community detection to define sub-clones base on the eventual network
• • figure out a different way to calculate the BCR network/embedding using distance matrices from heavy, light and potential transcriptomics data

Network stats

• measures to describe the overall clonal structure in a single sample, or per clone.
• • Calculate gini indices (e.g. size of clones versus number of clones)
• • node connectivity (network stability)
• • • closeness-centrality and betweeness-centrality (highlight clonally expanded nodes!)
• • edge-betweeness (mutation rate/relationship)
• Add a random sampling function to calculate the above if needed to control for sample size.

Miscellaneous

• Add parsing function to convert the processed BCR changeo/airr tsv format back to a 10x-style format because there’s some tools like immunarch that can parse 10x format for their analyses and some of the tools included look quite cool.
• Add parsing function to convert processed TCR data changeo/airr tsv format back to a 10x-style format to integrate with scirpy
• Bulk integration

Documentation

• Update the documentation for the functions to ensure that they are descriptive enough for general use.
• Add expected type of parameter into functions so it's a bit clearer for some of the more complicated ones. - added 5334512
• Clean up some spaghetti code in _tools.py and _plotting.py
• update notebook 1 with reccomendations to stick to airr format