The tool is comparing the term in your file to the reference genome. If you are working at UseGalaxy.org, then mm39 would be a custom reference genome. Check the format of that fasta, and the custom “database” you created for it. Is the key unique, and it is assigned to your input data?
I can try the solution you mentioned, but I’m wondering if there is known problems about missing data in the mm39 genome in usegalaxy.eu? Would you happen to know how I can access the mm39 genome in usegalaxy.eu to dig into its annotation/sequence to find if those values are missing?