Re: [phenixbb] TLS makes my r-factors shoot through the roof

Hi Morten, sorry for the trouble once again. Hm.. this seems like a case where I need the data to be able to reproduce the problem myself and fix it. Could you please send me all inputs so I can do it? I'll need data and model, and tell what exactly did you do, or just .eff file. Please send files off-list: directly to my email. I'm traveling right now so I might not be able to look right away, though I will do my best to address this problem asap (once I get the files). Thanks, Pavel On 5/29/12 7:39 AM, Morten Groftehauge wrote:

I like the new version but it does not seem to have helped much. Before it shot up by 3%, this time by 2.5%.

On 29 May 2012 12:23, Pavel Afonine mailto:[email protected]> wrote:

Hi Morten,

I'm sorry about the problem.

Are you using the latest Phenix? There was a bug in phenix.refine that I fixed a while ago, and the bug behavior was very similar to what I see happening in your case.

Could you please try the latest version and see if the problem goes away? Please let me know.

Pavel

On 5/29/12 3:56 AM, Morten Groftehauge wrote:

Hi Phenixbb,

3.2A data, 110 Wilson B, I41, twinned. Imperfect inverted repeat dsDNA with 4 protein monomers, NCS not used in refinement. If I don't include TLS I get a nice stable R-factor while RMS goes down. If I include TLS (whether as one group per chain or as chosen by Phenix) this happens:

Density disappears around some of the DNA bases as an obvious difference. If I run Refmac with TLS it just never finishes.

While I could just refine without TLS this does make me worried that there is something wrong with my model or my data (besides being terrible).

Cheers, Morten