Reference for pearson residual normalization

[eroell]

Question Hey,

In 7.3 Analytic Pearson Residuals, Germain et al. is cited as a the recommender. While this benchmark is very comprehensive, here rather sctransform is recommended "We further evaluated normalization methods by investigating their impact on the separability of the subpopulations. [...] In general, sctransform systematically outperformed other methods."

Lause et al., on the other hand, recommend the approximate analytic pearson residuals, also over sctransform. The sc.pp.experimental.pearson_residuals function used in 7.3 Analytic Pearson Residuals also references Lause et al. in its documentation.

So wouldn't it be more Lause et al. that should be cited here?

I'd suggest to add a sentence about Germain et al.'s work where suitable and keep this reference.

Interested in what you think - I might be missing something here.

Could do a PR with this additional citation if you agree

[AnnaChristina]

I generally agree but we aimed to not base our recommendation based on claims from method authors, but only independent benchmarks.

Feel free to add one additional sentence, but potentially keep both citations.

[eroell]

Just a sidenote, did not forget, since not high priority will take a glance once I got some bandwith for the topic to make this right