That's interesting. I was not aware that procheck paid any attention
to anisotropic ADPs.
Try feeding your PDB file to the Parvati web server.
It will look for problems with your anisotropic model.
http://skuld.bmsc.washington.edu/parvati
On Friday 13 July 2007 01:15, Jiamu Du wrote:
> Dear All:
> I am refining a structure with TLS parameters in REFMAC5.
> When I run the Procheck, some warning appear like below:
>
> --- generate density ---
> WARNING: Determinant(U) =< 0 : -0.000475003297
> for atom: C , res: 18, CHAIN: 1
> U: 0.08410 0.31220 0.17130 0.02360 0.04920 -0.20800
> atom will not be used
> WARNING: Determinant(U) =< 0 : -0.000683392049
> for atom: O , res: 18, CHAIN: 1
> U: 0.08280 0.32110 0.15860 0.01930 0.05420 -0.20470
> atom will not be used
> WARNING: Determinant(U) =< 0 : -0.000691463996
> for atom: N , res: 19, CHAIN: 1
> U: 0.07950 0.30660 0.17030 0.01890 0.04760 -0.21570
> atom will not be used
>
> What does this warning mean? Does this mean something about the model
> quality?
> But I have found nothing unusual in this position, when I look at this part
> of my strucutre model with its well defined density.
>
> Is there anyone who have met the same warning? How to deal with it?
>
> Thanks
>
--
Ethan A Merritt Courier Deliveries: 1959 NE Pacific
Dept of Biochemistry
Health Sciences Building
University of Washington - Seattle WA 98195-7742
