I found a paper suggesting Tween-20, but transformation efficiently is lowered by 30%... Thank you!

Similar questions and discussions