Danila

Hey @yulijia, I think [you can still set environmental variables in R](https://stat.ethz.ch/R-manual/R-devel/library/base/html/Sys.setenv.html). Not sure if that is going to work but we'll make it even easier with the upcoming change...

Hey @MonicaSteffi, It can be just the case that the total number of features is the strongest signal after the appropriate normalisation applied. In practice, if e.g. principal components plots...

@ivirshup I don't think so, unless there's work towards https://github.com/scverse/anndata/issues/244. To follow the ideas in https://github.com/scverse/anndata/issues/706, seems like the steps would be: - [ ] add an attribute `._X_layer` to...

@ivirshup, my current thinking is to try to stay away from `.loc` and explore other alternatives. While `.loc` could help to make it more _pandas_-like, it doesn't seem to me...

@carmensandoval, there's no implicit copying so one should make a copy explicitly. Please see the following code for more details: ```py import numpy as np from anndata import AnnData from...

@cdpolt, is there are specific change ("new behavior") you're referring to? > Storing things in layers sequentially, I just end up with a bunch of layers that all are identically...

@yulijia Glad it got resolved!

Hey @jolvhull, If I understand the context correctly, the line `mdata[mdata.obs.sample == x]` returns a view into the part of the MuData object. You can read [more about views vs...

@ivirshup > I think this is one of the strongest cases for allowing dynamic naming of dimensions is for multimodal data I think I am not convinced about dynamic dim...

> axis: Literal[0, 1, 'obs', 'var'] Makes sense, here's one to track for MuData: https://github.com/scverse/mudata/issues/64. For `axis=-1`, it's mainly relevant only when instantiating a MuData object, and I would just...