Dear Scott,
sorry for the long delay to follow-up on your question. I wanted to notify you asap that you have probably run in a difficult problem, but wasn't completely sure, and wanted to check myself... So here we go:
Please correct me here if I got it wrong.
Assuming it is correct, the problem is that eQuilibrator does not know what to do with C00965, because it is not chemically defined -- it doesn't know how many units of glucose it should consider.
Your assumption might be that, starting from a specific chain length, it does not matter much whether n=20 or n=21. Right?
What you can do (respectively, I could do for you), is to explicitly provide a chemically defined reaction; say "UDP-glucose + 1,3-Glucose_(20) = UDP + 1,3-Glucose_(21)". Maybe it's better to take n=2->3 and n=3->4, but you know what I mean?
Do you know of any experimental measurements for that reaction? This might be very interesting! In the data used to train eQuilibrator, I did not find any (one-step) polymerisation reactions like this.
Hope this helps a bit -- happy to discuss further! Best,
Robert