Clustering by reduction & gene instead of reduction & dimension

[rytpark]

Hi All,

In the pipeline FindNeighbors()%>%FindClusters, there is the option to specify features to cluster by instead of by dimension by setting features = "BCL", and dims = NULL, for instance. I'm currently thinking about clustering on the Harmony reduction, which is good for dataset integration. If I were to cluster on the harmony reduction while setting a specific set of features, what exactly would be happening, and is this a valid method? I'm an undergraduate student with limited experience, so I'm unfortunately having some difficulty interpreting the code for these functions.

Thanks!