On 1/27/10 8:06 AM, vidhya vasanth wrote: > In modeling, the best model can be judged by either low dope score or > low molpdf. But, what is a dope score? Why the best model is selected > based on low dope score??

See http://salilab.org/modeller/9v7/manual/node242.html and the Shen and Sali paper referenced there for information on the DOPE score. It is a pairwise atomistic statistical potential. In our benchmarks we demonstrated that it was far better at distinguishing "good" from "bad" models than the molpdf.

Ben Webb, Modeller Caretaker