In ir.pl.repertoire_overlap() function the fraction param is not working

[llumdi]

When using the ir.pl.repertoire_overlap( ) function I get the same result with or without fraction=True.

The output is the clonotype size (nr of cells) in each sample, whereas what I would like to plot is the relative abundance of the clone in the sample. I would also like to plot the log10 scale (so that smaller frequencies are more visible). How can I do that?

An example of the current plot I am getting:

Thanks a lot for your support.

[grst]

Hi @llumdi,

thanks for raising this issue! The repertoire overlap plotting function is a bit of a mess at the moment and I need to rewrite it at some point (see #156). I'm afraid neither normalization nor log scaling currently works out of the box, but you can easily retrieve the individual clonotype sizes for each of your sample using the corresponding tl function:

df, _, _ = ir.tl.repertoire_overlap(adata, "sample", inplace=False)

df will the be a sample x clonotype data frame that you can use for making a plot yourself.

[llumdi]

Thanks a lot for the workaround. Indeed, retrieving the underlaying data already helps.
Another issue I faced is that I was not able to reproduce the heatmap grid coordinates as in the tutorial. To me, yticklabels=True, xticklabels=True don't do anything. How can I add them?
See:

[grst]

The grid in the tutorial is more a side-effect of calling sc.set_figure_params() at the beginning of the tutorial rather than there on purpose.

You can enable it by activating the scanpy default settings using

import scanpy as sc
sc.set_figure_params()

or by explicitly setting the grid option in matplotlib

import matplotlib.pyplot as plt
plt.rcParams.update({"axes.grid" : True})

[llumdi]

thanks for the help @grst. Strange that none of the two worked for me.
For the scatterplots, I report that I could generate the normalized version, thanks.

I would also like to report that in ir.tl.repertoire_overlap() function, the fraction=True param is also not working. I could only retrieve the nr of cells per clone. But would be very useful to be able to rertrieve the table with percentages directly.