Would also get weird calcium gradients in the cheese. Buttermilk is acidified to the isoelectric point of milk, to 4.6, or a little below. At that point, you have almost complete k-casein degradation and the bonds within micelles break up. If you add that to milk and set with rennet, you will get this weird blend of both fully lactic set and fully rennet set. Your gel will have all sorts of gradients in it, and during syneresis, drainage and calcium balance will be all over the place, like Francois said.
Curd technologies work for a reason - the science works. You can blend finished curds of similar moisture levels (eg. full lactic fully drained, full rennet fully drained), but not odd amalgams like 25% starters in the milk, unless you want to precipitate the proteins quickly with heat, like you do in tvorog variants, and let it find equilibrium over time as it acidifies in a lactic gel. The reason heat would work with a high starter rate is because b-lactoglobulin has a higher isoelectric point, at 5.2, and with heat, you will introduce enough destabilization to induce protein precipitation (let's remember that casein adsorbs lactoglobulin).
Basically, stick to what works and makes sense from a theoretical perspective. Sometimes, experimentation results in bits of brilliance, accidental or not. And other times, it doesn't. The truth is that for all the wide variety of cheeses out there, there are really three ways to make all of them: full rennet, semi-rennet, and full lactic. Everything fits within that continuum, and may use about the same starter amount to achieve each cheese.
