Search found 101 matches

by Hayds
2010.12.20 22:31
Forum: Primary Data Processing
Topic: autoGNOM
Replies: 3
Views: 2506

Hi RBelen,

does it work if you run it from the command line?

Haydyn
by Hayds
2010.12.17 15:48
Forum: Rigid Body Modelling
Topic: Bunch q range limit
Replies: 3
Views: 3302

Hi LC, yes that is a very good point. If you are confident of the rigid body fitting the SAXS data (eg. if you have data for this construct in the absence of linkers and it fits quite well) and would like to "ab initio" model the missing portion then you are absolutely correct: use the entire data r...
by Hayds
2010.12.17 09:35
Forum: Rigid Body Modelling
Topic: Bunch q range limit
Replies: 3
Views: 3302

Hi LC, With rigid body modeling we are primarily concerned with the relative positions of sub-units. The data range that helps drive this is from ~ 0 nm-1 to 2.5 - 3.0 nm-1 (basically distances between scattering centers such as: intra-subunit dimensions and inter-subunit). Anything above this range...
by Hayds
2010.12.08 19:14
Forum: Ab Initio Shape Determination
Topic: Volumetric difference Maps.
Replies: 3
Views: 2995

Hi David, if you have a curve for protein + ligand, and a separate curve with the protein only you can try MONSA? This will try a multiple phase ab initio bead model using both SAXS data sets. Of course you assume that at the level of resolution we work at the protein alone does not undergo a radica...
by Hayds
2010.12.07 10:24
Forum: Primary Data Processing
Topic: Buffer Subtraction
Replies: 9
Views: 7133

Hi Ian, I talked to Petr again and he gave me more info on what we do here at X33: for saxs detector we have also finer grid (0.003 vs 0.01) compared to waxs. The errors for radially averaged curves are estimated from Poisson statistics (as square roots from the detector counts) both for saxs and wa...
by Hayds
2010.12.06 23:03
Forum: Ab Initio Shape Determination
Topic: help with Gasbor
Replies: 10
Views: 6301

Hi RBelen, maybe try and shorten the range for scaling until both curves seem similar (keep in mind, however, scaling over a tiny number of points, say < 100 is not a good idea as this can be meaningless). You can use the 3 mg/ml data for modeling as long as the invariant parameters make sense. Chec...
by Hayds
2010.12.06 12:06
Forum: Ab Initio Shape Determination
Topic: help with Gasbor
Replies: 10
Views: 6301

Hi RBelen, The lowest conc. data looks quite noisy, maybe any estimates made from this are a bit dodgy. Do the curves overlay? Is any scaling needed to make them overlay well (if necessary, scale them in a range from ~ 0.08 to 0.16 ang-1, just to make sure you are not including effects from structur...
by Hayds
2010.12.04 21:02
Forum: Ab Initio Shape Determination
Topic: help with Gasbor
Replies: 10
Views: 6301

Correction, sorry:

at low angles the curve calculated from the model has higher intensity than the data, thus the model is too big/extended to describe the data.

Haydyn
by Hayds
2010.12.04 20:59
Forum: Ab Initio Shape Determination
Topic: help with Gasbor
Replies: 10
Views: 6301

Hi RBelen, The p(r) function looks a little strange, but the fit of the regularised curve to the data looks fine. Perhaps the optimal solution has not been obtained, As for the fit of the model generated by GASBOR to the SAXS data. The fit is poor. At low angles the data curve has higher intensity t...
by Hayds
2010.12.04 12:45
Forum: Rigid Body Modelling
Topic: pre_bunch not working, sequence and pdb file is fine
Replies: 24
Views: 12269

Hi again, just a thought. I think you can use CORAL (new version of BUNCH and SASREF developed by Maxim) in ATSAS 2.4 for the following case: eg. for PDB file with missing region: segment1-(missing linker)-segment2 define the segments and declare that segment1 and segement2 move in concert, allowing...
by Hayds
2010.12.04 12:39
Forum: Rigid Body Modelling
Topic: pre_bunch not working, sequence and pdb file is fine
Replies: 24
Views: 12269

Hi Jiao, if you have regions of a xtal structure that are missing you can break up the PDB file into segments, such that pre-bunch/bunch will fill in the gap with dummy residues. You can fix the segments if you require that they retain the same position and orientation relative to each other. Of cou...
by Hayds
2010.12.02 22:25
Forum: Rigid Body Modelling
Topic: pre_bunch not working, sequence and pdb file is fine
Replies: 24
Views: 12269

Hi again Abhay!

I found the problem. Your PDB file contains the following:

ENDMDL

Then another set of coordinates are listed. Delete these and you have the correct input (delete from line 2177 I think).

Best regards,

Haydyn
by Hayds
2010.12.02 22:15
Forum: Rigid Body Modelling
Topic: pre_bunch not working, sequence and pdb file is fine
Replies: 24
Views: 12269

Hi Abhay, the pdb file and the sequence file are exactly the same. This means that there is nothing for pre_bunch to do as it has no dummy residues to build. As a result the program exits (we probably need to code a better error message for this, sorry). If the actual sequence you are using has 100 ...
by Hayds
2010.12.02 22:04
Forum: Primary Data Processing
Topic: Buffer Subtraction
Replies: 9
Views: 7133

Hi Ian! Do people normally apply these corrections? The protein volume fraction correction actually works against the transmitted intensity correction in most of the cases I looked at since the Tsam/Tbuff is on average ~1.009 ( 9 datasets, stdev= 0.022). I talked to Petr Konarev here at X33 who is m...
by Hayds
2010.11.30 23:12
Forum: Primary Data Processing
Topic: problem about guinier plot
Replies: 2
Views: 3185

Hi Darrensu, Firstly, it looks like you are using the entire data range for the manual estimation of the Guinier region. This is incorrect. The linear dependence of ln(I) on s**2 is broken for globular (roughly spherical) particles at s*Rg ~ 1.3. AutoRg automates this procedure and tries to keep wit...