Re: [modeller_usage] PDF score question

On 3/2/12 9:59 AM, albert wrote: > As far as we know the lower DOPE score, the better model. I am wondering > how about the PDF score? also the lower score, the better model? eg: > model A PDF=200 is better than model B PDF score=300.

The molecular PDF is simply the sum of restraint violations. So the lower the PDF, the less the restraints are violated. This typically means a good model, assuming that the restraints are an accurate description of reality (assessment scores such as DOPE are typically more reliable).

Ben Webb, Modeller Caretaker