Candace Liu

> @cliu72 agree with defaulting to 1024x1024 crops, though I'm curious to know the dimensions of the images that are causing these issues. > > For de-cropping, I prefer RxCx...

Hi @c0urtneyt! That function is here: https://github.com/angelolab/ark-analysis/blob/main/src/ark/phenotyping/post_cluster_utils.py#L174-L208 and was added in this pull request: https://github.com/angelolab/ark-analysis/pull/1095. A new version of ark was recently released, v0.7.1 (https://github.com/angelolab/ark-analysis/releases) - try installing this version...

Hi @janinemelsen! Thanks for bringing this to our attention. It seems like the default behavior of seaborn is to only show a subset of labels when it thinks there are...

@alex-l-kong can we add `yticklabels=True` as a permanent fix to `sns.clustermap` in `vizualize.draw_heatmap`? I think it makes sense for that to be the default behavior.

Hi @Nicholas-Bartelo! Sorry about this issue, like you suspect, it's probably a Windows line-ending error. In the lab, we have actually moved away from using docker because it's been causing...

@alex-l-kong I thought this was already addressed in this PR: https://github.com/angelolab/ark-analysis/pull/1114 (See point that says "Adds natsort calls throughout the Pixie pipeline to avoid issues with channel ordering." under "Pixel...

Hi @rahilgholami, Thanks for the question. In general, we don't recommend using Pixie for functional marker characterization, the reason being the combinatorial possibilities are massive. In your myeloid example, you...

Hmm so theoretically, you can have as many pixel metaclusters as you want (the default is 100 SOM clusters and 20 automatic metaclusters, but you can change the parameters to...