The short answer is no. However, there is a simple work-around and there is a good reason why we can't do it.

Workaround:
You can rename your CDS features in one file to CDS_Method1 or CDS_old and have the other file just have the regular CDS, then they will appear in separate tracks.

It was a deliberate decision to distinguish annotation items by type rather then by source.

Keeping track of where an individual annotation item comes from quickly becomes a tangled mess. Especially, when you allow users to create new features (came out of nowhere?) and allow them to merge them (source is now two files?). How do you store this information? Store two exons of a merged gene in one file and the last three exons in another file? This would also require some provision to transfer from one source to another and would expose a lot of under-the-hood mechanics to users.

If you need items in different tracks they need to be given different types, everything of the same types goes in the same track.

For data that you cannot edit, we can easily keep track of where it came from, so for those data types we do split out by source.