Dear Yolanda, Here is the basic idea of prime-and-switch phasing:
We calculate phases from a model. These phases are mostly correct, but may have incorrect peaks at the coordinates of some misplaced atoms in the model. Next we are going to go through our reflections one by one. We calculate a map without including one reflection. Then we look in the solvent region and ask "what phase of this reflection would make the solvent most flat?". That gives our new "map-based" phase for this reflection. (This is a slight simplification but it is very close to the real thing.) This process is iterated several times. Finally we calculate a map with our measured amplitudes and all these new phases. This is our prime-and-switch map. The central principle of prime-and-switch phasing is that we estimate phases using some feature of the electron density map (flatness of the solvent) that is essentially unrelated to the model. The reason it can work is that the model gets the phases pretty close to the right answer to start out with. Prime-and-switch phasing is identical to resolve statistical density modification, except that in regular density modification the starting phases are recombined with the phases from the map. You can tell if it is working by noticing if most of the density is remaining fairly strong for your model, but for certain parts the density is going away. A key point about prime-and-switch phasing is that in many cases, the density for ALL atoms decreases, but the density for incorrect atoms decreases much more quickly than that for correct atoms. I hope that helps! All the best, Tom T > Hi Thomas£¬Hi all, > I have used the amazing software "prime-and-switch" before,it is > quite helpful. > I have read through your article "Using prime-and-switch phasing to > reduce model bias in molecular replacement" , but it is hard for me to > understand. So what I am concerned is the basic idea of it. Could you > explain the key point with simple words? > I found that "prime-and-switch" also uses solvent flattening at the > beginning.Is it similar to ordinary solvent flattening? > By the way, how to know whether it is effective or not? Which > parameter should I concentrate after resolve running, except the electron > density? And how to evaulate those parameters? > > -- > Yolanda >
