Hi, the current version of Fo-Fo map calculation in PHENIX (phenix.fobs_minus_fobs_map) requires the unit cell parameters to be nearly identical, otherwise it would not run. So if it runs (according to Intekhab) then we can safely assume that this is the case. It's a good idea to make sure to contour the map at correct levels (or make sure the order in Fo1-Fo2 is right) as Dale suggested. Otherwise it is strange that the Fo-Fo map doesn't show anything, while residual map does show the ligand... Anyway, if unbiased Fo-Fc map shows the ligand then I don't see why one would struggle with something else instead of just using this map to build the ligand. Pavel. On 2/9/11 8:04 PM, Thomas C. Terwilliger wrote:

Hi Intekhab,

I would first examine the cell of your native and ligand-bound structures. If these are very similar (much less than 1% different) then I would be surprised by the results, and would try to make sure that everything is working as expected.

If the cells are rather different (>2%) then the results are plausible (though I don't exactly understand the "circular chunk of density"). In this case, I would use the 2mFo-DFc map from your refined ligand-bound structure as your map. It would be good to refine without the ligand (or even better to not put the ligand in at all until you have good density for it.)

If the cells are only somewhat different (1%) then either of the two above possibilities might hold.

Normally an omit map doesn't help to show a ligand if you did not put the ligand there in the first place. Omit maps are good for finding what is wrong with a model, not so good for finding things that are not in your model.

All the best, Tom T

...

...

Hi i am trying to calculate a difference map ( ligand-native ) using isomorphous difference map program in phenix. I used the reflection files of ligand and native and phase information of ligand derived data. But the difference map donot fits the ligand, and appears as a circular chunk of density at sigma level 5. I calculated an omit map that clearly showed the presence of my ligand at the specific position. Is there anything wrong in my calculation. What alternate ways are there to improve my difference map.

-- INTEKHAB ALAM LABORATORY OF STRUCTURAL BIOINFORMATICS KOREA UNIVERSITY, SEOUL _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

---

phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Hi Intekhab, I had a look at your data, thanks for providing it to Pavel. Here is what I see: 1. taking your starting model, removing the RIB ligand, and refining against the native.sca and ligand.sca data gives models that have good R/Rfree in each case (0.19/0.23 and 0.18/0.23), but differ by about 0.5A rms. The models have coordinated shifts relative to each other in which many atoms move together. 2. The largest positive density in your Fo-Fc map for the ligand.sca refinement is at the position of your ligand. The density is not great, and doesn't cover the entire ligand. 3. The Fo(ligand)-Fo(native) map phased with your starting model shows some density on parts of the ligand, and is considerably less clear than the Fo-Fc map above, as you pointed out. 4. The Fo-Fc map with native.sca shows a little density on part of the ligand. I would suggest that the Fo(ligand)-Fo(native) map is probably a relatively inaccurate picture of the ligand because it is a composite of all changes between native and ligand-bound structures. The Fo-Fc map based on ligand.sca refinement is probably your best picture of the ligand. The fact that this Fo-Fc map is not very clear despite the reasonable resolution (2.5 A) and good R/Rfree suggests that the ligand is not always in exactly the same orientation or location. All the best, Tom T

...

I tried both ways Fapo-F ligand as well as Fligand-Fapo. You are right the red density becomes negative when the F is reversed. The unit cell parameters are quite same.

For native: a=120.557 b=196.262 c=109.339 , alpha =90, beta=114.231 and gamma=90 ligand complexed data= a=119.952 b=196.084 c=109.206 , alpha =90, beta=114.116 and gamma=90

My 2Fo-Fc map as well as Fo-Fc is quite significant and i modeled the ligand uisng that.

Regards Intekhab Alam

On Thu, Feb 10, 2011 at 2:50 PM, wrote:

...

This is a long shot, but it's possible that you calculated a Fapo-Fligand map instead of the other way 'round. Normally you would have a positive peak surrounded by a negative ripple (due to series termination and other factors). If you get the F's reversed the negative ripple becomes positive and the ligand density becomes negative. What does you negative contour say?

Dale Tronrud

On 2/9/2011 7:33 PM, intekhab alam wrote:

...

Hi i am trying to calculate a difference map ( ligand-native ) using isomorphous difference map program in phenix. I used the reflection files of ligand and native and phase information of ligand derived data. But the difference map donot fits the ligand, and appears as a circular chunk of density at sigma level 5. I calculated an omit map that clearly showed the presence of my ligand at the specific position. Is there anything wrong in my calculation. What alternate ways are there to improve my difference map. -- INTEKHAB ALAM LABORATORY OF STRUCTURAL BIOINFORMATICS KOREA UNIVERSITY, SEOUL

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- INTEKHAB ALAM LABORATORY OF STRUCTURAL BIOINFORMATICS KOREA UNIVERSITY, SEOUL _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb