I made a lipid array on nitrocellulose membrane and am using chemiluminescence to detect whether or not my protein binds to the lipids. It works quite well. Except for my last protein: this protein produces a much higher background and all of the lipid spots are white against that background.

Has anyone ever observed white spots and know what its cause is?

Do they look like ghost-band i.e; hollow at centre.
and if so then clearly your protein bound is higher then usually you have been doing earlier or you should try increasing the primary/secondary antibody dilutions(as you have mentioned your background was high).
I hope this works for you, ask anything if not clear.
Wishes

No, they are not hollow in the middle. I've seen those before: they look like black donuts with a white center.

In this case, the protein does not appear to have bound to the lipid spot at all (which makes sense, because I don't expect it to bind in the first place). But instead of getting a pale (white background) with no gray/black spots to indicate protein-bound lipid, I have a dark gray background with prominent white lipid spots.
And ALL of the lipid spots are white (I forgot to mention that this is a lipid array whereby each lipid spot has an incrementally increased amount of lipid). Anyway, so it's not as if only the highest concentration lipid spots are white (meaning it's a reaction of the protein binding to the lipid), but something that happened to the membrane itself.

I'm wondering if protein degration might be causing the dark background (i.e. I'm seeing more non-specific background binding from the degradation products). But I have no idea why the spots would be white.

Oh, and I am already using a 5000 dilution for the antibodies and haven't had a problem with the other 6 domains I tried (I'm studying each of the 7 domains of a single protein as individually cloned GST recombinant proteins). Only this domain produced the white-spot reaction.