This is a dumb question but I just wanted to double check, this is independent of reference genome (mm9, mm10), correct?

This is something that you could accomplish fairly easily in any programming language e.g. R or Python. However the specific implementation would be heavily dependent on the structure of your data

Strategy for R:

Assuming you have a count table with genes on rows and samples on columns you can easily read this table in as a data.frame counts. Analogous you can read in the list of genes you want to keep (or remove, doesn't matter) into a vector keepgenes.

Then you can easily slice the counts data.frame by something like newdata <- counts[counts$gene %in% keepgenes, ]

If this strategy/pseudocode is unclear for you I can expand on this, but this is really basic R and you'll make your RNA-seq analysis far less painful if you dive in a R tutorial and learn yourself how to perform these tasks. Knowledge of at least one programming language is an enormous advantage.