Print

Print


Hi Rafael,

Phase combination is when you add the probability distributions from two different sources of phase information (in your case, your SAD phases and model phases; but this occurs in lots of other places as well...).  The idea is that combining two or more poor quality phase distributions will result in a map better than the inputs phases sets individually.

If I'm correctly understanding what you're trying to do, you can use the MLHL target in refmac (using your experimental HL coefficients and initial or current model as inputs).  I'd also recommend reading reference 3 from the refmac documentation.

It might also be worth trying to improve your experimental phases using the Se -> Su differences between your native and derivative datasets (this signal might be too weak to be useful, so it may not be worth trying too hard) in addition to the f'' from Se.

Regarding your refinement statistics...On the off chance that no one else mentions is, Rfree isn't supposed to be one of the things you try to optimize (and the Rwork/Rfree gap appears to be getting a bit large).  

Good luck,

Pete

-----Original Message-----
From: CCP4 bulletin board on behalf of aka akaka
Sent: Wed 1/2/2008 5:41 PM
To: [log in to unmask]
Subject: [ccp4bb] phase combination
 
Dear CCP4s

 I am trying to solve a structure for which I have two datasets: One comes from  a selenomethionine derivatized crystal at 3 A resolution and the second comes from a native one that gave 2.6A. I have obtained workable experimental phases by SAD using SOLVE/RESOLVE. With this I have built a model and obtained a calculated map using REFMAC and the native dataset. I cannot push the Rfree further than 34.5 (Rw:28) and given the low quality of the original density I ended up having 6.5% outliers. I was told there is an option with which I could start from scratch calculating a new map using "phase combination"  between the derivatized dataset and the model.

Could anyobody clarify further what phase combination is and which are the programs for that purpose?

Regards
Rafael (thats my real name)

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today it's FREE!
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/