Hi Matthias,
Based on the casing convention (all uppercase) I would suspect that these 25 genes are in the Human Gene Symbol namespace. Therefore to compute the enrichment of this set in your data would would need to either map your entire dataset into the human space using one of our chip files, or you would need to identify the mouse orthologs of the genes in this set and construct the corresponding mouse gene set. I would recommend doing the manual conversion into the mouse gene symbol space for this set as it is rather small.
We offer a human to mouse gene symbol conversion chip in our downloads repository. You would need to find the genes manually but it would ensure that your result symbols would match MSigDBs. The alliance of genome resources also provides a great service for this, which we use internally in construction of our files, and they have a web tool for lookups, which night help as well.
Do be aware though that if your expression dataset gene universe is missing more than 10 of the genes this would cause the set to fall below the minimum size threshold which would cause the set to be rejected by the software (unless you changed the threshold which we don't really recommend).
-Anthony
Anthony S. Castanza, PhD
Curator, Molecular Signatures Database
Mesirov Lab, Department of Medicine
University of California, San Diego