Hi Lai Ping,
Thank you for reaching out!
pg.scatter(data, attrs=['Channel'], restrictions="Channel:s1,s2,s4,s6", basis='umap')
You can decide whether to show all the rest points or not by setting "show_background" parameter to True or False. If True, all the rest points will be shown in a grey faded color as background; if False, they won't appear in the resulting plot.
Moreover, you can add more restrictions. E.g.
pg.scatter(data, attrs=['Channel'], restrictions=["Channel:s1,s2,s4,s6", "louvain_labels:1,2"])
will only show points from s1, s2, s4, s6 samples AND belonging to Louvain cluster 1 and 2.
pg.scatter(data, attrs=['louvain_labels', 'Channel'], restrictions=["Channel:s1,s2,s4,s6", "louvain_labels:~1"])
will show points from the 4 samples and not belonging to Louvain cluster 1, and give you two UMAP plots colored by louvain_labels and Channel attributes, respectively.
Hope it helps!
Sincerely,
Yiming