We've processed several hundred sequences through our local I-TASSER installation, but we have a substantial number of them for which all or almost all of the EC number C-scores is 0.0600. All the active site predictions I've seen are NA, too, but that's expected to be common, I suppose.
Do you have any hints as to what could be causing this--is this something you have encountered before? Let me know what additional information I should provide.