Subsetting topic_model_de_analysis class for specific gene sets?

[BioParker]

Hi, apologies if this is a very basic issue and I've missed something, but I can't currently find a way to subset an object of class topic_model_de_analysis for a specific set of genes.

The basic set up is I want to use the run_homer() function to do motif finding on a set of genes that are common to a group of topics across different cell types. I know that the run_home() function itself allows you to subset based on a number of parameters, but not on a gene list, and I don't know how to subset a the topic_model_de_analysis object.

Any help is much appreciated

[pcarbo]

@BioParker Oops, I see that I forgot to add a description of the "subset" argument in run_homer().
The basic idea is to perform an enrichment analysis of motifs in which the "positive set" of genomic positions/features is determined by the "subset" argument; by default, it chooses genes (or genomic features) that show "significant" differences from the selected topic (k) to the other topics (where here the significance is based on the local false sign rate, or "lfsr").
One could imagine different ways to choose this subset; e.g., based on posterior estimates of the LFCs provided by "postmean".
I'm not sure if this directly answers your question, but may get you closer to an answer?