Hello,
You can run ssGSEA using RNA-Seq data. This would generally work if your samples raw counts are not too different from one another. If counts vary too much, you may consider normalizing your data (e.g., so that they are between 0 and 1).
Perhaps you can run three different jobs trying these approaches looking for results that you can confirm are biologically significant.