- Ifeffit - millenia.cars.aps.anl.gov

Atoms restraints
by Massimiliano Stola 15 Jul '04

15 Jul '04

Thank you for all the explanations about constraints for the histidine ring within Artemis. Well, I think that it will be easier, but boring, building several reasonable structures and test them one by one... I think I'll finally follow this way. I'll be interested as well in the possibility to simulate exafs data for several structures at once, like Stefano asked. It would be a nice feature for the next releases of the program! Regards Max

1 0

Chi^2 and R-factor reporting requirements
by Stanislav Stoupin 12 Jul '04

12 Jul '04

Hello all, I would like to hear your opinion about reporting Chi^2 and R-factor values in EXAFS studies. I found that Chi^2 and R-factor values are not provided by many EXAFS papers. However, according to standards of The International XAFS Society (http://ixs.iit.edu/) both these values should be reported. My understanding is that the reduced Chi^2 should be close to 1 in ideal case. In practice however, the value is always greater than 1 because of systematic errors. For example, my reduced Chi^2 values are around 300. However, the R-factor values are about 3% which doesn't look as "ugly" as Chi^2 reduced. On the other hand, ifeffit overestimates uncertainties using reduced Chi^2 value. Thus, all errors are taken into account. The final question: Is it appropriate to report just R-factor calculated by ifeffit and the obtained uncertainties (which look reasonable) without reporting the "ugly" Chi^2/reduced Chi^2 values? Stanislav Stanislav Stoupin ----------------------------------- Research Assistant/PhD student BCPS department Illinois Institute of Technology email: stousta(a)iit.edu voice: 312-567-7983

4 3

RE: [Ifeffit] multiple fits
by Kelly, Shelly D. 12 Jul '04

12 Jul '04

Hi Stefano, > > I have the following question: let's say I have a data set for which > I would like to try a long series of possible models and monitor the > result of the fits. The models are geometric variations of a > starting structure.I have all related feff.inp files for each model > ready. Is there a way to process the data so that I can get out the > related fits? I do not think I can do this using artemis, but maybe > there is a set of commands to be given from the command line? > > Thanks > Stefano The short answer is no. The long answer is sure. But the script that you need to write is long and complicated. You have to teach the computer how to model the structure for all the variations so that it will make the appropriate variables to determine in the fit, and also how to handle the output. These tools are not incorporated into ifeffit. I am responding to your request, because I did this as a graduate student. Back in the old days, I was using atoms, feff, and feffit. The program that I wrote moves the atoms in a unit cell, writes out a atoms.inp file, runs atoms, runs feff, makes a feffit.inp file and runs a fit, and then makes a spread sheet to compare the answers. If you are interested, I wrote is a small EXAFS paper about this script, and would be happy to share. S.D. Kelly, R. Ingalls, and E. A. Stern, "Determining crystalline atomic positions using XAFS, a new addition to the UWXAFS analysis package," Journal of Synchrotron Radiation. 8 311-313, 2001. In the end, I think that you will agree that you have a lot of work ahead of you. HTH Shelly

2 1

multiple fits
by Stefano Ciurli 12 Jul '04

12 Jul '04

Hello, I have the following question: let's say I have a data set for which I would like to try a long series of possible models and monitor the result of the fits. The models are geometric variations of a starting structure.I have all related feff.inp files for each model ready. Is there a way to process the data so that I can get out the related fits? I do not think I can do this using artemis, but maybe there is a set of commands to be given from the command line? Thanks Stefano -- ____________________________________________ Stefano Ciurli Professor of Chemistry Department of Agro-Environmental Science and Technology University of Bologna Viale Giuseppe Fanin, 40 I-40127 Bologna Italy Phone: +39-051-209-6204 Fax: +39-051-209-6203 "Fatti non foste a viver come bruti, ma per seguir virtute e canoscenza" Dante Alighieri - Inferno - Canto XXVI "Ihr seid bestimmt, nicht Tieren gleich zu leben, Nein, Tugend zu erringen und Erkenntnis" "Ye were not form'd to live the life of brutes, But virtue to pursue and knowledge high"

1 0

Ifeffit 1.2.6
by Matt Newville 09 Jul '04

09 Jul '04

Hi Everyone, I've posted Ifeffit 1.2.6 source code and Mac disk-image-installer today. This fixes two important bugs: 1. path lists in feffit() and ff2chi() could get polluted so that doing a second feffit() or ff2chi() using _fewer_ paths would some paths left over from the previous sum of paths. The path list is now initialized correctly. 2. reading of feff.inp with feff6 no longer fails because of line endings that don't match the standard for that platform. In addition, plot markers and arrows now support color, plot_arrow(...,color=red) plot_marker(...,color=blue) and the location of the legend in the plot window is now more customizable using either plot(...., key='data', key_x=0.9, key_y0=0.9,key_dy=0.1) or setting the variables &plot_key_x = 0.9 &plot_key_y0 = 0.9 &plot_key_dy = 0.1 where 'key_x' is the x-location of the legend, 'key_y0' is the y-location of the first key, and 'key_dy' is the vertical offset between keys. The Mac OS installer includes a pre-release of horae-037. I believe Bruce will be releasing new versions of his codes soon. When he does, we'll make a new installer for Windows as well as the usual automated update -- probably next week. Please let me know if there are any troubles with these versions, --Matt

4 5

RE: [Ifeffit] Atoms restraints
by Jensen, Mark P. 06 Jul '04

06 Jul '04

Here are my 2 cents worth on how to constraint the histidine ring within Artemis. Don't lose heart, Max. Both Scott and Bruce are right. There is no simple command to hold two scattering atoms in a particular relationship in Ifeffit, but the problem is easily solvable for this system partly because multiple scattering paths won't be too important and partly because the ring is quite rigid. (Actaully based on my experience with other organic ligands, the non-coordinated atoms in the histidine ring is probably not too important for EXAFS, but if you want to use the whole ring, this is how I would approach the problem using information external to the EXAFS data.) The histidine ring is a 5-membered (planar) aromatic ring, with 2 nitrogens and 3 carbons. That makes visualizing things a lot easier. Off-hand, I don't know what the bond lengths and bond angles within the ring are, but you could find that information from x-ray crystal structures of histidine complexes and it would be contained in the feff.inp file (which you could make using the procedure Shelly Kelly distributed to the list morning). But, somehow you have to build a feff.inp file and do the feff calculations so you have some paths to fit. Assuming that the absorbing atom , "M", is at coordinates (0,0,0) makes things simpler. Fortunatley that should be the case for a feff.inp file. You will have to define at least 17 new "set" variables in Artemis, one x, y, and z coordinate for each of the 5 atoms in the ring, as well as the angles theta and phi. The x,y,z coorinates have to be manually input from the feff.inp file. You will also need some other variables to hold intermediate coordinates, for example del_x, del_y, and del_z. As Scott pointed out, EXAFS is mostly sensative to the radial distance, and our case that's exactly what Artemist/Ifeffit is varying by adding del_r to the scattering path length. What you want to do is move the other 4 atoms of the histidine ring in conjunction with the del_r changes in the M-N bond length. First, convert the x,y,z coordinates of the coordinated nitrogen atom (I'll call these (xN, yN, zN) ) to spherical coordinates (r_eff, theta, phi). You need to put the values of theta and phi into the Guess,Set,Def parameters as set parameters because you will need them later to convert from spherical coordinates back to regular 3D-rectangular (x,y,z) coordinates. In the fitting, the changes in the M-N single scattering path length (i.e., the M-N bond distance) will not effect the two angles theta and phi for the coordinated N. However, changing the M-N distance will change theta and phi (relative to the absorbing atom) for the other 4, so there is some algebra to do. Now this is where the whole approach might fail. I don't know if it is possible to to incorporate the delr or the delr+reff value from a specific feff path into an algebraic expressions in the Guess,Set,Def panel. Bruce or Matt or someone else will have to comment on this and explain how to do it. However, assuming that there is a straigforward way to do it, then... for each iteration of the fit in Ifeffit/Artemis there is a particular M-N distance for the scattering path of the coordinated histidine nitrogen, Artemis calls this "r" (for each path) and it is equal to reff+delr for each path (also using the Artemis nomenclature). Using theta and phi (which haven't changed for the coordinated N) and "r", a new set of coordinates in regular space (x_new,y_new,z_new) for the coordinated nitrogen can be calculated within Artemis using the standard equations for spherical coordinates. (Like I said, if we can't access "r" from the alegbraic expressions in Guess,Set,Def this won't work, though). Then, taking the difference of the current position and the original position of the coordinated nitrogen in x,y,z coordinates, we get (del_x,del_y,del_z). In other words del_x = (x_new - xN) can be defined as a DEF parameter in the Guess,Set,Def panel. This single set (del_x,del_y,del_z) vector is going to be added to the (x,y,z) coordinates of each of the 4 other atoms in the histidine ring to get their current coordinates in the next and final step. Finally, to get the change in the radial distance of each of the 4 other ring atoms, put the expression [(x_orig+del_x)^2+(y_orig+del_y)^2+(z_orig+del_z)^2]^0.5 - [x_orig^2+y_orig^2+z_orig^2]^0.5 in the "delr" blank (where [x_orig^2+y_orig^2+zorig^2]^0.5 is actually reff for the path) for each of the 4 feff paths of the 4 non-coordinated ring atoms. In this equation, "x_orig" is the original x-coordinate of the non-coordinated ring atoms and "del_x" was calculated from the new position of the coordinated nitrogen atom above. Things can get trickier if you want to incorporate multiple scattering paths, but I think that woiuld probably not be a fruitful line of inquiry anyway. If you stick to singlescattering paths, that's all there is to it! ;) Good Luck, Mark Jensen ------------------------------ Mark P. Jensen Chemistry Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439-4831 USA 630-252-3670 630-252-7501 FAX mjensen(a)anl.gov > -----Original Message----- > From: scalvin(a)slc.edu [mailto:scalvin@slc.edu] > Sent: Friday, July 02, 2004 9:19 AM > To: XAFS Analysis using Ifeffit > Subject: Re: [Ifeffit] Atoms restraints > > If I understand your question correctly, the answer is no. > You cannot tell Ifeffit or Artemis to constrain the distance > between two atoms with a simple command. The reason is > simple--EXAFS doesn't directly contain that information; it > only contains it implicitly (maybe). For direct-scattering > paths, EXAFS only provides radial information. > Multiple-scattering does provide some information along other > axes, but this causes a greater effect in some cases than others. > > What Ifeffit and Artemis do allow is for you to constrain the > parameters related to individual scattering paths in pretty > much any way you desire. > So, for example, you can constrain the distance to the direct > scattering path to atom A, the direct scattering path to atom > B, and the multiple-scattering path that hits A and B in such > a way that the distance between A and B remains constant. But > this takes careful geometrical thinking on your part. > > Incidentally, restraints are different from constraints. > "Constraints" are forces and cannot vary. "Restraints" > penalize the fit if it drifts too far, but allows some variation. > > Hope that helps. > > --Scott Calvin > Sarah Lawrence College > > > So I understand that I can just tell Ifeffit/Artemis: > distance between > > Atom#1 > > and Atom#2 is equal to "x" Angstroms, and that will be > forced for all > > the next calculations. Did I understand well?? > > If this is true I believe that this way is easier than > build a lot of > > structures to test (If you have more than one histidine the > number of > > structures increases...). > > > > Anyway, I had a look at many .inp the examples but I did not > > understand well how to give this command to the program. > Can you write > > me an example? > > It would be much easier for me to start from that example > and add many > > atom-atom distance restraints to keep a rigid structure. > > > > > _______________________________________________ > Ifeffit mailing list > Ifeffit(a)millenia.cars.aps.anl.gov > http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit >

3 2

athena: larger data visualisation window size
by biquard xavier 06 Jul '04

06 Jul '04

Hi ! I have been using athena under LinuX with great satisfaction for a long time now. Recently, I changed my old portable computer for a brand new one: my Xwindow screen size went from 1024x768 to 1920x1250, but all windows opened by athena for data visualisation have the same old size of around 870x670: I cannot take advantage of my new screen size! As I compiled athena from scratch (horae-021), I wonder if I should not modify athena files set_properties.pl and draw_properties.pl that seems to me to be defining window size. If anyone knows for sure how to do it, please let me know! cheers, Xavier -- ------------------------------------------------------------------- Biquard Xavier Tél CEA : 0438785413 CEA-Grenoble Fax : 0438785138 DSM/DRFMC/SP2M/NRS mailto:xbiquard@cea.fr 17, av des Martyrs 38054 Grenoble Cedex 9 ------------------------------------------------------------------- 2nd chance : Tél ESRF1 : 0476882546 Tél ESRF2 : 0476882598 -------------------------------------------------------------------

2 1

updating the faq with OSX-specific information
by Bruce Ravel 05 Jul '04

05 Jul '04

Hi folks, The last email got me thinking about the documentation for the various installers and updaters and such. I would like to ask one of the OSX users on the list to create a new FAQ entry in section 2 explaining how to update Athena and Artemis from source code once Ifeffit has been installed from the OSX disk image. (That is, I'd like an expert OSX user to walk the novice OSX user through the process of building from source and using sudo to install the newly built code.) Similarly, section 2 needs an entry explaining how the unix updater works. I plan to write that in the coming days. A section 2 entry from Carlo or one of the other debian users would probably be a good idea as well. Thanks, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu Code 6134, Building 3, Room 405 Naval Research Laboratory phone: (1) 202 767 2268 Washington DC 20375, USA fax: (1) 202 767 4642 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

6 9

ifeffit and Mac OSX
by Matt Newville 03 Jul '04

03 Jul '04

Hi Raja, On Mon, 7 Oct 2002, rajasekaran swaminathan wrote: > Hi Matt, > > I tried again over the past few days to install the binary > installer on my mac (OS X.2). I couldn't find an attached aquaterm > / pgplot demo. The Aquaterm program opens, but I cannot see any > windows come out from that application. I mean, it is just like a > blank application. When I type ifeffit on Terminal app, it doesnt > recognise the presence of ifeffit in the ifeffitosx directory. i.e. > it says that the command is not found. The executable works > perfectly well on a Windows machine, and I was wondering whether > there are any such equivalents for a Macintosh machine. I also dont > think that the installer comes with athena/artemis/tkatoms etc.. Do > you have any other recommendations to resolve this issue? Also,if > it is possible, can I get a CD of these programs for a Macintosh > equivalent to an executable version for the Windows? > > I appreciate your help in this regard. I have used the binary installer once on a OS 10.1 system. It worked for me (after following the simple included instructions, of course) and plotting with Aquaterm did work. I haven't tried 10.2. I know the Mac binary installer does not include athena, artemis, or tkatoms. I believe it only provides the command-line ifeffit. I'm not sure I can answer many other questions about it. I don't know enough about it or 10.2 v. 10.1 issues to know how well it would work with 10.2. If the aquaterm/pgplot demos don't work, I'd guess it's a more serious problem and that the binary installer may not work. Is the Fink / X-Windows / PGPLOT installation procedure too big of a hassle? I know that can work, and that athena will work too. I also believe that the instructions are simpler for 10.2 than 10.1 (especially with respect to getting perl/Tk to work), but haven't tried this myself. I'm copying this message to the ifeffit mailing list, hoping that someone else may have better answers to these questions. How necessary is a complete binary installer for Mac OSX? Thanks, --Matt

4 4

protein data base to feff.inp file
by Kelly, Shelly D. 02 Jul '04

02 Jul '04

Hello All, I just updated my little web page to share the protein data base file converter that I wrote. You can find the perl program and some instructions at: http://www.mesg.anl.gov/exafs.html If you use the program, let me know how it goes. Shelly > -----Original Message----- > From: Kelly, Shelly D. > Sent: Friday, July 02, 2004 9:55 AM > To: XAFS Analysis using Ifeffit > Subject: RE: [Ifeffit] Atoms restraints > > > Hi Max, > > > > > Anyway, I had a look at many .inp the examples but I did not > > understand well > > how to give this command to the program. Can you write me an > > example? It would be much easier for me to start from that > > example and add many > > atom-atom distance restraints to keep a rigid structure. > > I do not understand this question. What kind of .inp file are you > looking at? Is this the input to feff? The feff.inp file needs to > have a list of atoms with their corrdinates in x, y, and z. > For a hystatine you will want to make a feff.inp file from > the crystal structure. From this list feff makes a theory > that has the contriution to the EXAFS data for each path of > the photo-electron in a different feffxxxx.dat file. Then > Ifeffit/Athena can be used to fit these paths to the data. > During the fitting process you can change the distance > between the atoms by a little bit (<0.5 Angstroms). To > change the distance during the fitting process you need to > define delr for each path included in the fit. > > I am attaching a catalase protien feff.inp file that I made > from a 7CAT.pdb file. > Hum...I see that I wrote a perl program to help do this. I > must appologize that I forgot that I had this program. > Others have asked for it in the past. > > The proceedure that I used to create this file goes like this. > 1) Get your hands on a crystal structure with the same core > hole atom. (i.e. Fe-oxide for my case) > 2) Run atoms inside of Artemis to generate a feff.inp file. > 2b) save the feff.inp file to disk. > 2c) open feff.inp file in a text editor and remove the ipot > and x,y,z coordinates. > 3) Get the pdb file for the protein, or use mine for an > example 7CAT.pdb. > 4) Save the pdf2fi.pl file to disk and the 7CAT.pdb file to disk. > 4b) make the pdf2fi.pl executable by typing "chmod 777 pdf2fi.pl" > 4c) type "pdf2fi.pl 7CAT.pdb" This will generate a > 7CAT.pdb-feff.inp file. This file has the ipot and the x,y,z > coordinates for the atoms in the pdb file. > 4d) To use the program on your own system, open pdf2fi.pl and > edit a few of the lines at the top. Don't worry it is easy > to do and you will not mess it up. You have to tell the > program what atoms are in the pdb file and which one is the core. > 5) paste the file from 4c into the feff.inp file from 2. > 6) Save the feff.inp file to disk and open it in Artemis and run feff. > > > HTH > Shelly >

1 0