The material pelleting cannot be fat, since fat is lighter than aqueous solutions and will float on top; it is probably casein, the major milk protein.Like EDTA, citrate is a chelator that will remove free calcium from solutions. Since calcium is necessary for the formation of casein aggregates, this is probaly why your protocol includes EDTA to solubilize casein. However, milk contains about 25 mM Ca2+, thus, you will need at least that much EDTA or citrate to chelate Ca2+ completely. Changing chelators will probably not affect the result provided that you add enough of them.

Pierre Béguin - In the article, they mentioned removing the hardened fat layer after centrifugation, so I assumed that's what I was seeing in the pellet. Thank you for the clarification. As for the amount of EDTA, you are saying that I will need at least 25 mM EDTA to chelate the Ca2+? I'm currently using 0.5 M EDTA, which did not work. Or did I misunderstand? Thanks!

Do you mean that the final EDTA concentration in your yogurt mixture was 0.5 M ? This would mean 146 g/L free acid (which doesn't dissolve unless neutralized) or 168 g/L disodium salt. Or was it the concentration of the EDTA stock solution that you added in order to solubilize the casein in the yogurt ? I meant 25 mM final in the yogurt, regardless of the concentration of the stock solution.

Pierre Béguin - Yes, I meant the stock solution of EDTA is 0.5 M. 18 ml of the 0.5 M EDTA and 12 ml of TE buffer are added to 10 g of yogurt, so the final concentration of EDTA is still well over 25 mM. Perhaps I will try adding EDTA again to the pellet and see if it gets smaller.