Faster create_pseudobulks function.

[ghuls]

Faster create_pseudobulks function, by assuming that a lot of consecutive values will have the same value. Also streches of zeros are not stored at all.

Based on:
https://github.com/aertslab/single_cell_toolkit/blob/c15038ddf1322fd4957396c16bb7782ad2e6629e/fragments_to_bw.py#L263C1-L372C82

[jmschrei]

Thanks. I'll take a look soon. How much faster is it? Can you add in the documentation a reference to the code in single_cell_toolkit?

[ghuls]

It depends on how many ranges of the same value you have in a row.
For fragments_to_bw it was at least 10 times faster, if I remember correctly.

Do you have a pseudobulk bw file I can test with (as your distribution of values might be different than what I tested with in fragments_to_bw?

[jmschrei]

Thanks for looking into it. Any of the fragments files under scATAC_clusters.zip should be good for testing: https://zenodo.org/records/8313962