The error message seems to be more generic.
As suggested, collect backend objects (DSO changes) in one trnsport request and new query/ query chnages in seperate transport request and move it in sequence.
This makes easy to find out erroenous objects that causing transport failures.