On autoantibodies, the potential set we could have is probably millions. Science mostly has the ones associated with specific disease. Looking for those in ME, as others have suggested, is problematic at best.
The potential targets also number in the millions, probably, when you into account epitopes and protein conformation.
Its like looking for a needle in a haystack while wearing thick leather gloves and a blindfold.
The Maes experiment a few years back went the other way. He subjected blood from ME patients to various specific probable targets (six?), and found way too many antibodies binding to them. This did not identify the antibodies, but if replicated and not fundamentally flawed it shows we probably have a nonspecific autoantibody problem with a very broad target range. This is fundamentally different from other autoimmune diseases we understand.
There is some question about what that means though. I think Maes is of the view this is because too many of our proteins are damaged by oxidative and nitrosative stress, and hence our immune system is free to attack them ... but these proteins are attached to our own cells.