From owner-structural-nmr@net.bio.net Sun May 01 23:00:00 1994 Path: biosci!biosci!not-for-mail From: kristoff@net.bio.net (David Kristofferson) Newsgroups: bionet.structural-nmr Subject: Re: Hello? Date: 2 May 1994 15:13:07 -0700 Organization: BIOSCI International Newsgroups for Biology Lines: 23 Distribution: bionet Message-ID: <2q3ttj$otb@net.bio.net> References: <2pioib$dat@news.uni-c.dk> <2piugg$lff@lyra.csx.cam.ac.uk> Reply-To: biosci-help@net.bio.net NNTP-Posting-Host: net.bio.net I have received a couple of messages to our help address (biosci-help@net.bio.net) indicating that a couple of sites do not seem to be getting messages in their newsreaders. I have run tests at the sites accessible to me and can confirm that the news and mail distributions are working correctly. These sites are here in California, Bethesda, Maryland (NCBI), and Daresbury in the U.K. In addition I see traffic on a daily basis in the newsreaders both here and at NCBI (I don't have access to news at Daresbury). Given the limited number of complaints to date, the most likely reason might be a problem with the news distribution to the couple of locales that have complained. However, if this problem is more widespread than we currently think, please let us know by contacting biosci-help@net.bio.net (assuming you somehow see this message 8-). Sincerely, Dave Kristofferson BIOSCI/bionet Manager biosci-help@net.bio.net From owner-structural-nmr@net.bio.net Sun May 01 23:00:00 1994 Newsgroups: bionet.structural-nmr Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!sun4nl!news.nic.surfnet.nl!ruu.nl!RUUCI5!rasmus From: rasmus@RUUCI5.NoSubdomain.NoDomain (Rasmus Fogh) Subject: Peak integration and message style Message-ID: <1994May2.092838.11977@cc.ruu.nl> Sender: rasmus@RUUCI5 (Rasmus Fogh) Organization: ACCU Date: Mon, 2 May 1994 09:28:38 GMT Lines: 52 -- wchutt@alex.monsanto.com (Bill C Hutton) Wrote : > -- In about six months a means will exist to dramatically improve the volume : > integrals - or more properly the amplitude estimates - of NOESY cross-peaks. : > In addition, each cross-peak amplitude will have a statistically rigorous : > uncertainty assigned to it as well... and by the way, overlap (unless the peak : > seperation is on the order of the linewidth or less) will no longer be a : > problem. : > How can this be you ask... see J. Magn. Reson. 98, 483-500 (1992) and Magn. : > Reson. Med. 9, 282-287 (1989) for some clues. : > You heard it here first. Copy of flaming reply deleted >What sort of information would you like? >The software does exist... it just happens to be in an alpha form and is >constantly being improved. The approach does not rely on the descreet FT and is >much too complicated to describe via a post. The literature does exist for a >reason. Since you are curious (but perhaps a long way from a library) I will >give you another clue - J. Biomol. NMR, 3, 515-533 (1993). >The problem of cross-peak amplitude estimates has been solved. Remainder deleted Flame wars are a waste of time (and the polite tone of Bill Huttons reply is appreciated). Mea culpa. But because this is a young newsgroup, I think that some discussion of message style is still relevant; after all we could be setting the tone for the future. IMHO posts should give enough information to enable you to evaluate immediately what is going on and whether it is of interest to you - also if you do not know all about it already. I find this sadly lacking in the Bill Hutton posts. So, what sort of information would I like ? For a start, at least some idea of what approach is being used. Even a one-line description would help, 'Time-domain fitting using Bayesian probability theory', for instance. Then some idea of what it can do - now, not in six months. Was it tested, on what, how large systems can it handle, what is the error rate? A representative example could be given in a few lines. Without this kind of information claims that you will soon have solved this or that problem are impossible to evaluate and not very interesting (except possibly for posters named R. R. Ernst). After all, didn't Pons and Fleischmann claim to have solved the world's energy problem? My library is down the corridor, but I object to people being coy and giving 'clues' instead of just telling me what they have to say. I doubt that I am the only one. Rasmus Fogh (Protein Structural) NMR Group Utrecht University, Utrecht, Holland From owner-structural-nmr@net.bio.net Sun May 01 23:00:00 1994 Path: biosci!LAPLACE.CSB.YALE.EDU!abonvin From: abonvin@LAPLACE.CSB.YALE.EDU (Alexandre Bonvin) Newsgroups: bionet.structural-nmr Subject: NMR data treatment: Rsym Date: 2 May 1994 11:08:25 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 92 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <9405021804.AA10141@laplace.csb.yale.edu> NNTP-Posting-Host: net.bio.net In a previous posting I wrote in answer to Per Kraulis: > I like the idea of using an Rsym to define the quality of the spectra. > Is the following definition correct? > > Rsym = Sum | Aij - Aji | / Sum [0.5 * | Aij + Aji |] Per Kraulis answered with: ... > The crystallographers define their Rsym slightly differently from the > above, namely: > > Rsym = Sum | I - | / Sum > > where I is the reflection intensity, and is the average for the > respective symmetry-related intensities. I would suggest that we use > this measure, just to avoid making the term "Rsym" ambiguous. Note that there is only a factor two between the two definitions since only pairs of symmetry related peaks are expected in 2D NOE spectra. > One problem with the Rsym as defined above is that it will be dominated by > strong cross peaks. Maybe one should consider some other measure that > deals with this, or maybe give Rsym as a function of average > intensity. Also, one should give the number of NOE cross peaks > contributing to the Rsym, and the number of NOE cross peaks for which > the symmtry-related peaks could not be measured, for whatever reason. > All this needs testing, and I hope someone out there feels up to it... To deal with the problem of strong peaks one could define a sixth root residual, R6sym, similar to the definiton used to judge the quality of the fit between computed and experimental NOE data ("distance like" measure of the quality). It should have the advantage to give some more weight to the weak NOEs. R6sym = Sum | I**1/6 - **1/6 | / Sum **1/6 Depending on the Rfactor definition used to assess the quality of the NMR structures one of these two definitions could be used. Here are some numbers for a protein I worked on in Prof. Kaptein's group in Utrecht: Arc repressor (dimer, 2x53 aa): =============================== Exp. data for mixing time # 1 (100 ms) Total number of peaks = 1951 Number of symmetrical peaks for Rsym = 1784 Av. rel. err. between both side of diagonal = 0.367 Rsym = 0.130 R6sym = 0.031 Exp. data for mixing time # 2 (150 ms) Total number of peaks = 1971 Number of symmetrical peaks for Rsym = 1804 Av. rel. err. between both side of diagonal = 0.354 Rsym = 0.132 R6sym = 0.030 Exp. data for mixing time # 3 (200 ms) Total number of peaks = 1975 Number of symmetrical peaks for Rsym = 1808 Av. rel. err. between both side of diagonal = 0.338 Rsym = 0.127 R6sym = 0.028 These values should thus give lower limits to the achievable Rfactors for this particular structure (the final Rfactors for Arc after relaxation matrix refinement with IRMA and DINOSAUR (see J. Mol. Biol., 1994, vol 236, pp 328-341) were 0.35 (Xray definition) and 0.085 (sixth root definition)). It would be interesting to compare results for proteins of different sizes. There are also quite a number of DNA structures refined with relaxation matrix calculations to very low Rfactors what will be suited for such an analysis... anyone interested ? Alexandre =============================================================================== | Alexandre Bonvin PhD | Phone: (203) 432-5066 | | Dept. Mol. Biophys. & Biochemistry | Fax: (203) 432-6946 | | Yale University, 266 Whitney Avenue | Email: abonvin@laplace.csb.yale.edu | | New Haven CT 06511, USA | | =============================================================================== From owner-structural-nmr@net.bio.net Sun May 01 23:00:00 1994 Newsgroups: bionet.structural-nmr Path: biosci!daresbury!trane.uninett.no!sunic!pipex!howland.reston.ans.net!EU.net!Germany.EU.net!news.dfn.de!news.dfn.de!scsing.switch.ch!rzusuntk.unizh.ch!cweber From: cweber@oci.unizh.ch (Christoph Weber) Subject: Re: non contacts Message-ID: <1994May2.121414.18207@rzu-news.unizh.ch> Sender: newsadm@rzu-news.unizh.ch (CNEWS ADMINISTRATION) Organization: University of Zurich, Switzerland X-Newsreader: TIN [version 1.2 PL2] References: <199404272139.OAA11115@net.bio.net> <1994Apr28.080508.8789@cc.ruu.nl> Distribution: bionet Date: Mon, 2 May 1994 12:14:14 GMT Lines: 21 I've always been under the impression that backcalculation and refinement against the original NOESY data amounts to implicit inclusion of lower bounds. If that's correct (experts out there, please let us know), then this is probably the most justified way of including lower limits. Alternatively, if combined use of NOESY and ROESY data clearly indicate spin diffusion as the major pathway of magnetization transfer, I believe it's pretty safe to estimate lower bounds of the distance involved. I'm not sure how much lower bounds help a protein structure. Does anyone have experiences to share? Just my two cents. Cheers, Christoph Christoph Weber email: cweber@oci.unizh.ch OCI Uni Zurich phone: +41 1 257 4925 Winterthurerstr. 190 FAX: +41 1 361 9895 CH-8057 Zurich, Switzerland From owner-structural-nmr@net.bio.net Sun May 01 23:00:00 1994 Path: biosci!daresbury!not-for-mail From: pjk@ciclid.csb.ki.se (Per Kraulis) Newsgroups: bionet.structural-nmr Subject: Re: NMR data treatment: wrong problem? Date: 2 May 1994 13:25:56 +0100 Lines: 119 Sender: daemon@mserv1.dl.ac.uk Distribution: bionet Message-ID: <2q2rgk$1fo@mserv1.dl.ac.uk> Original-To: str-nmr@dl.ac.uk Posted-Date: Mon, 2 May 94 14:24:03 CET My original posting about NMR data treatment (April 29) has generated some discussion. Excellent! My comments so far: Bryan Finn wondered whether the various automatic assignment procedures being discussed would allow a more objective measure of assignment correctness. I think this might be correct. But as far as I know, there have not been any serious attempts in this direction yet. I have some vague ideas of my own how to check assignments objectively. For instance, the methods for calculating chemical shifts from structure could be used for this. Rather than using chemical shifts calculations for structure refinement (with its problems), one could use it as a way of highlighting unlikely chemical shifts for specific spins, thus potentially spotting problems. I haven't tested this, so I have no idea whether it would be useful. This area is entirely open, and I hope someone out there has been inspired by this discussion... Another problem in this connection: we all assume that erroneously assigned NOEs will show up as consistently violated distance constraints in 3D structure calculations. Has anyone really tested whether this is true? Or could the structure be "plastic" enough in some cases to distribute the error among many distance restraints, thus hiding the error? Is the free R-factor analysis (Brunger et al., Science, 1993, vol 261, pp 328-331) maybe the answer to this? Alexandre Bonvin wrote: > Are the differences between both side of the diagonal only due to > integration errors? In reality the transfer pathways are different. > So there may be more in there than only integration errors. Taking > the average (if both peaks are well resolved) seems to me the safest way. Bill C Hutton wrote: > The differences are real and they are artifacts of how the data was > collected. Let me clear up one misunderstanding. The fact that 2D spectra cannot be perfectly symmetrical (due to different relaxation pathways, or whatever) is not a valid criticism against an Rsym. I have received the same reaction when talking to other NMR people, namely that 2D spectra are not strictly symmetrical, the implication being that Rsym is not a very good idea. But this is wrong. Consider the situation: we have two measurements of the same NOE interaction. Which one do we use in our calculation? There is no a priori reason to use one rather than the other. We have (not yet) any practical, numerical grip on the effects underlying asymmetry. So it seems obvious to me that one should use the average value. Alexandre Bonvin is exactly right when he says that it is the safest way. And the Rsym then becomes a natural way of estimating the error. The fact that the differences in the NOE from either side of the diagonal may, in part, be due to true physical differences, rather than to mere computational integration error, is simply another reason for trying to remove such effects. We are interested in the NOE intensity, and we want to remove the effects of different relaxation pathways, or asymmetrical data collection. Thus, let's take the average. This problem becomes even more important when we are dealing with 15N- and 13C-separated NOE data from 3D/4D spectra. There, yet another modulation is expected for symmetry-related cross preaks, namely the 1J-coupling transfer. If such data are to be used for relaxation matrix refinement, then this must be dealt with somehow. As an aside, Bill C Hutton's final comment has left me mystified: > Use the data from the side with the highest resolution. Punt the others. I do not understand this. We have one cross peak at (F1=a, F2=b), another at (F1=b, F2=a). Which one is from the side with the highest resolution? Both have one chemical shift from either of the high- and low-resolution dimension. So we can't choose cross peak on that basis. Or have I misunderstood something? Bill C Hutton promises that the integration problem will be solved in 6 months using some kind of Bayesian analysis of NMR signals. Fine. Time will tell. Along this line, I may add that Maximum Entropy algorithms are said to have been developed to do basically the same thing, namely to estimate the errors in position and amplitude of signals (NMR as well as other). I know of no details. And similarly, we haven't seen the real thing yet... Alexandre Bonvin wrote: > I like the idea of using an Rsym to define the quality of the spectra. > Is the following definition correct? > > Rsym = Sum | Aij - Aji | / Sum [0.5 * | Aij + Aji |] The crystallographers define their Rsym slightly differently from the above, namely: Rsym = Sum | I - | / Sum where I is the reflection intensity, and is the average for the respective symmetry-related intensities. I would suggest that we use this measure, just to avoid making the term "Rsym" ambiguous. Alexandre Bonvin rightly points out that a difference of typically 20 % between symmetrical cross peaks does not necessarily give an Rsym of 20 %. I was wrong in suggesting that it did. Sorry about that. One problem with the Rsym as defined above is that it will be dominated by strong cross peaks. Maybe one should consider some other measure that deals with this, or maybe give Rsym as a function of average intensity. Also, one should give the number of NOE cross peaks contributing to the Rsym, and the number of NOE cross peaks for which the symmtry-related peaks could not be measured, for whatever reason. All this needs testing, and I hope someone out there feels up to it... -- Dr. Per Kraulis Center for Structural Biochemistry phone: +46 8 608 9266 Karolinska Institute fax: +46 8 608 9290 NOVUM email: pjk@ciclid.csb.ki.se 141 57 Huddinge SWEDEN ----------------------------------------------------------- "Nothing is more difficult than simplicity." Unattributed. ----------------------------------------------------------- From owner-structural-nmr@net.bio.net Mon May 02 23:00:00 1994 Newsgroups: bionet.structural-nmr Path: biosci!daresbury!trane.uninett.no!sunic!news.lth.se!news.lu.se!heimdall!bryan From: bryan@fkem2.lth.se (Bryan Finn) Subject: Re: NMR data treatment: wrong problem? Message-ID: <1994May3.165207.22070@nomina.lu.se> Sender: news@nomina.lu.se (USENET News System) Nntp-Posting-Host: heimdall.fkem2.lth.se Reply-To: bryan@fkem2.lth.se Organization: Physical Chemistry 2, Lund University, Sweden References: <2q5t93$fgo@columba.udac.uu.se> Date: Tue, 3 May 1994 16:52:07 GMT Lines: 36 In article fgo@columba.udac.uu.se, gerard@rigel.bmc.uu.se (Gerard Kleijwegt) writes: > ...Another option is to collect duplicate > datasets ("multiple xtals", although you would probably > have to use the same sample ?) and to check how well > the two sets of NOE volumes "merge". This might give > a much more realistic estimate of the errors. I agree that multiple datasets would be one way to check the reproducibility of one's data, but not just on the same sample. If the data is truly valid, it should be possible to reproduce one week or one year later. Several papers on relaxation have used multiple spectra at a single time point to estimate the errors (see e.g. Stone et al. Biochemistry (1992) 31,4394). NMR spectrometer time is not unlimited but a double spectrum at a single mixing time might be worth checking. Bryan --- ___________________________________________________________________ | | | Bryan Finn | | Department of Physical Chemistry 2 Tel: +46-46-108254 | | Chemical Center Fax: +46-46-104543 | | University of Lund | | POB 124 | | S-221 00 Lund Sweden e-mail: bryan@freja.fkem2.lth.se | |_________________________________________________________________| | | | "It is unworthy of excellent men to lose hours like slaves in | | the labor of calculation." -- Leibniz | |_________________________________________________________________| From owner-structural-nmr@net.bio.net Mon May 02 23:00:00 1994 Path: biosci!agate!overload.lbl.gov!dog.ee.lbl.gov!ihnp4.ucsd.edu!swrinde!cs.utexas.edu!howland.reston.ans.net!pipex!sunic!columba.udac.uu.se!rigel.bmc.uu.se!gerard From: gerard@rigel.bmc.uu.se (Gerard Kleijwegt) Newsgroups: bionet.structural-nmr,bionet.xtallography Subject: Re: NMR data treatment: wrong problem? Date: 3 May 1994 16:14:27 GMT Organization: Uppsala University Lines: 92 Distribution: bionet Message-ID: <2q5t93$fgo@columba.udac.uu.se> References: <2q2rgk$1fo@mserv1.dl.ac.uk> NNTP-Posting-Host: rigel.bmc.uu.se Xref: biosci bionet.structural-nmr:87 bionet.xtallography:898 In article <2q2rgk$1fo@mserv1.dl.ac.uk>, pjk@ciclid.csb.ki.se (Per Kraulis) writes: |> |> One problem with the Rsym as defined above is that it will be dominated by |> strong cross peaks. Maybe one should consider some other measure that |> deals with this, or maybe give Rsym as a function of average |> intensity. Also, one should give the number of NOE cross peaks |> contributing to the Rsym, and the number of NOE cross peaks for which |> the symmtry-related peaks could not be measured, for whatever reason. |> All this needs testing, and I hope someone out there feels up to it... In protein xtallography, careful papers include a table of the following items as a function of resolution (could be as a function of NOE intensity for you NMR guys): - Rsym (a.k.a., internal Rmerge) - completeness (in our case, which percentage of the possible reflections were observed; in your case, which percentage of the NOEs, expected on the basis of the refined structure, was actually observed ? of course, the latter criterion would mix model and data) - multiplicity (how often was a reflection measured) - percentage of the data with F > 2 (or 3) sigma(F); alternatively, or / (<..>=average) Someone else mentioned Axel Brunger's free Rfactor; this statistic is rapidly becoming the major tool for judging whether or not a structure was overfitted. Another is simply the data-to-parameter ratio (although I can't imagine that this will be particularly popular in the NMR community ... ;-); it turns out that many low-resolution X-ray protein structures have been refined with more parameters than observations ... (GJ Kleywegt & TA Jones, to be published - sorry, Rasmus). If anyone thinks of a procedure to estimate sigmas of NOE volumes, remember that you should also test if the obtained sigmas are a good estimate of the true sigmas (i.e., the distribution of the number of NOEs with (int(j)-meanint)/sigma(j)= -5,-4,...,+5 should be normal with average zero and standard deviation 1). As for Alexandre's R6, is there any statistical justification for using this (e.g., are you using an objective function with sixth powers of I in your refinement ?), or is it just cosmetic (i.e., to get lower values) ? It might be worthwhile to use a correlation coefficient instead of Rfactors (contrary to R, a CC is independent of the averages and scales of the two arrays you're comparing). We use this in xtallography as well (Axel Brunger's PC-refinement, for instance, and in so-called real-space electron-density averaging procedures). Also, since the "NMR Rsym" will probably mostly involve a maximum of 2 observations, you might want to consider to calculate a "CCsym", i.e. the correlation coefficient between the volumes 'below' and 'above' the diagonal. Another option is to collect duplicate datasets ("multiple xtals", although you would probably have to use the same sample ?) and to check how well the two sets of NOE volumes "merge". This might give a much more realistic estimate of the errors. A question (just out of curiosity): why do so few NMR-protein structure papers contain a Ramachandran plot ? Note that I'm x-posting this to bionet.xtallography (it's time for some x-pollination). --Gerard ****************************************************************** Gerard J. Kleywegt ___ Department of Molecular Biology | | /\ Biomedical Centre /\ -- || Uppsala University || || || Box 590, S-751 24 || || || Uppsala, SWEDEN || \/ -- -- |__| E-mail: gerard@xray.bmc.uu.se ****************************************************************** "He's probably pining for the fiords ..." ****************************************************************** "Visit famous Uppsala, home of the ... er ... people who live in Uppsala !" ****************************************************************** The opinions in this mail/post are fictional. Any similarity to actual opinions, living or dead, is purely coincidental. ****************************************************************** From owner-structural-nmr@net.bio.net Mon May 02 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!nac.no!nntp-oslo.uninett.no!usenet From: ibfls@nlh10.nlh.no Newsgroups: bionet.structural-nmr Subject: Macintosh? Date: 3 May 1994 10:37:18 GMT Organization: agricultural university of norway Lines: 5 Distribution: world Message-ID: <2q59gu$ii3@ratatosk.uninett.no> Reply-To: Lars Skjeldal NNTP-Posting-Host: ibf-ls.nlh.no Does anyone have any experience with any Macintosh software within protein NMR? Lars Skjeldal Agricultural University of Norway e-mail: ibfls@nlh10.nlh.no From owner-structural-nmr@net.bio.net Mon May 02 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!howland.reston.ans.net!pipex!warwick!lyra.csx.cam.ac.uk!bjd12 From: bjd12@cus.cam.ac.uk (Ben Davis) Newsgroups: bionet.structural-nmr Subject: Re: Welcome II Date: 3 May 1994 08:59:35 GMT Organization: U of Cambridge, England Lines: 30 Distribution: bionet Message-ID: <2q53pn$1d2@lyra.csx.cam.ac.uk> References: <2po2u8$sf8@mserv1.dl.ac.uk> NNTP-Posting-Host: apus.cus.cam.ac.uk X-Newsreader: TIN [version 1.2 PL2] Paul Driscoll (driscoll@bioch.ox.ac.uk) wrote: : P.s. Is there anyone out ther at all getting the newsgroup via USENET in the : U.K. 'cos we shurely are not? Yup - we're getting it here - and have been since it started up. : Paul C. Driscoll : Royal Society University Research Fellow : Department of Biochemistry, : University of Oxford, : South Parks Road, : Oxford OX1 3QU. UK. : tel. (44) 865 275799 : fax. (44) 865 275259 : E-mail: driscoll@bioch.ox.ac.uk -- ______________________________________________________________________________ Ben Davis, MRC Protein Function and Design, Cambridge, UK ______________________________________________________________________________ "They can make me do it, but they can't make me do it with dignity." From owner-structural-nmr@net.bio.net Tue May 03 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!columba.udac.uu.se!rigel.bmc.uu.se!gerard From: gerard@rigel.bmc.uu.se (Gerard Kleijwegt) Newsgroups: bionet.structural-nmr Subject: Re: NMR data treatment: wrong problem? Date: 4 May 1994 18:46:44 GMT Organization: The tall, hunky Dutchmen society Lines: 124 Distribution: bionet Message-ID: <2q8qik$u3g@columba.udac.uu.se> References: <9405041520.AA13671@laplace.csb.yale.edu> NNTP-Posting-Host: rigel.bmc.uu.se Keywords: data quality, model quality, fit of model and data, proteins, xray In article <9405041520.AA13671@laplace.csb.yale.edu>, abonvin@LAPLACE.CSB.YALE.EDU (Alexandre Bonvin) writes: |> In bionet.structural-nmr Gerard Kleijwegt writes: |> > - completeness (in our case, which percentage of |> > the possible reflections were observed; in your |> > case, which percentage of the NOEs, expected |> > on the basis of the refined structure, was |> > actually observed ? of course, the latter |> > criterion would mix model and data) |> |> And that's a major problem at the time being! The measure of |> completeness will be very dependent of the model used for the |> calculations. We might not know exactly what's the internal |> dynamics of the molecule. Each case will be different and it |> will be very difficult to compare results from various proteins. Nevertheless, knowing: - how many of the expected NOEs based on the model were actually observed/not observed - how many of the expected "non-NOEs" based on the model were nevertheless observed (i.e., nr of violations) is interesting in itself, since it tells something about the degree to which your model(s) explains your data (note that the R-factor is necessarily limited to observations included in the refinement). |> > As for Alexandre's R6, is there any statistical justification |> > for using this (e.g., are you using an objective function |> > with sixth powers of I in your refinement ?), or is |> > it just cosmetic (i.e., to get lower values) ? |> |> Some groups do use a sixth root function in their refinement, some |> don't. The choice of the Rsym definition should depend on what |> function was used for refinement (or just give both to make everyone |> happy). This is not true, since Rsym says something about the quality (internal consistency, at least) of your DATA. This is completely independent of the model and the refinement. I still think that using the sixth power is artificial in this respect, although there may be a case for it in the case of R(Iobs,Icalc). In XRAY, we have the same situation: low-resolution Is tend to be much higher than high-resolution Is. |> function should give more weight to weak NOEs, corresponding to |> long distances, that are expected to be more important for |> determining the 3D structure. Are you saying that: long-range NOE = long-distance NOE ? Why ? |> > A question (just out of curiosity): why do so few |> > NMR-protein structure papers contain a Ramachandran |> > plot ? |> |> May-be because they quite often look much more awfull for NMR structures |> than for Xray ones, especially for undetermined floppy ends and loops... |> But again, we are measuring in solution where conformational averaging can |> take place and much more mobility is expected than in crystals: so, should |> almost perfect Ramachandran plots for NMR structures be expected? That's as maybe, but I think that the individual models should still obey sensible protein stereochemistry. I.e., if you make a Ramachandran plot for each of the models separately, each plot should look normal. Only the Ramachandran plot of an averaged structure would be expected to look peculiar, perhaps. In any event, outliers should be inspected closely (and mentioned in papers). In our experience, outliers tend to be interesting for one of two reasons: - errors in the model - an interesting place in the structure In the case of NMR, "underdetermination" due to lack of NOEs might be third option. |> They are nice tools to assess the stereochemical quality of |> Xray protein structures like PROCHECK for example (Morris et al., PROTEINS: |> Structure Function & Genetics 12 , 345-364) that could be used more often |> for NMR structures too. All protein structures submitted to PDB |> are now first checked with PROCHECK (at least if I am correct). PROCHECK as is can be used for individual NMR models just as well as for XRAY structures. For your information, work is in progress to create a tailor-made version of PROCHECK for NMR structure ensembles (by Roman Laskowski in Janet Thornton's group, with input from the Utrecht group). By the way, note that by now we are discussing three fundamentally different types of 'quality check' (Per's initial mail was mainly about the first type): (1) quality of the data (Rsym, multiplicity, assignment verification etc.) (2) quality of the model (PROCHECK, Ramachandran plots etc.) (3) extent to which the model explains the data (R, free R, "completeness" in the NMR sense etc.) |> Alexandre |> |> =============================================================================== |> | Alexandre Bonvin PhD | Phone: (203) 432-5066 | |> | Dept. Mol. Biophys. & Biochemistry | Fax: (203) 432-6946 | |> | Yale University, 266 Whitney Avenue | Email: abonvin@laplace.csb.yale.edu | |> | New Haven CT 06511, USA | | |> =============================================================================== --Gerard ****************************************************************** Gerard J. Kleywegt ___ Department of Molecular Biology | | /\ Biomedical Centre /\ -- || Uppsala University || || || Box 590, S-751 24 || || || Uppsala, SWEDEN || \/ -- -- |__| E-mail: gerard@xray.bmc.uu.se ****************************************************************** "He's probably pining for the fiords ..." ****************************************************************** "Visit famous Uppsala, home of the ... er ... people who live in Uppsala !" ****************************************************************** The opinions in this mail/post are fictional. Any similarity to actual opinions, living or dead, is purely coincidental. ****************************************************************** From owner-structural-nmr@net.bio.net Tue May 03 23:00:00 1994 Path: biosci!LAPLACE.CSB.YALE.EDU!abonvin From: abonvin@LAPLACE.CSB.YALE.EDU (Alexandre Bonvin) Newsgroups: bionet.structural-nmr Subject: Re: NMR data treatment: wrong problem? Date: 4 May 1994 08:19:39 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 90 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <9405041520.AA13671@laplace.csb.yale.edu> NNTP-Posting-Host: net.bio.net In bionet.structural-nmr Gerard Kleijwegt writes: > > In protein xtallography, careful papers include a table > of the following items as a function of resolution > (could be as a function of NOE intensity for you NMR > guys): > > - Rsym (a.k.a., internal Rmerge) > - completeness (in our case, which percentage of > the possible reflections were observed; in your > case, which percentage of the NOEs, expected > on the basis of the refined structure, was > actually observed ? of course, the latter > criterion would mix model and data) And that's a major problem at the time being! The measure of completeness will be very dependent of the model used for the calculations. We might not know exactly what's the internal dynamics of the molecule. Each case will be different and it will be very difficult to compare results from various proteins. > > As for Alexandre's R6, is there any statistical justification > for using this (e.g., are you using an objective function > with sixth powers of I in your refinement ?), or is > it just cosmetic (i.e., to get lower values) ? Some groups do use a sixth root function in their refinement, some don't. The choice of the Rsym definition should depend on what function was used for refinement (or just give both to make everyone happy). There is no statistical justification for using a sixth root function, but it's not just cosmetic. It comes from the dependence of the NOEs on distances (I ~ 1/r**6). The sixth root function should give more weight to weak NOEs, corresponding to long distances, that are expected to be more important for determining the 3D structure. > It might be worthwhile to use a correlation coefficient > instead of Rfactors (contrary to R, a CC is independent > of the averages and scales of the two arrays you're > comparing). We use this in xtallography as well (Axel > Brunger's PC-refinement, for instance, and in so-called > real-space electron-density averaging procedures). > Also, since the "NMR Rsym" will probably mostly > involve a maximum of 2 observations, you might want to > consider to calculate a "CCsym", i.e. the correlation > coefficient between the volumes 'below' and 'above' > the diagonal. Another option is to collect duplicate > datasets ("multiple xtals", although you would probably > have to use the same sample ?) and to check how well > the two sets of NOE volumes "merge". This might give > a much more realistic estimate of the errors. > > A question (just out of curiosity): why do so few > NMR-protein structure papers contain a Ramachandran > plot ? May-be because they quite often look much more awfull for NMR structures than for Xray ones, especially for undetermined floppy ends and loops... But again, we are measuring in solution where conformational averaging can take place and much more mobility is expected than in crystals: so, should almost perfect Ramachandran plots for NMR structures be expected? They are nice tools to assess the stereochemical quality of Xray protein structures like PROCHECK for example (Morris et al., PROTEINS: Structure Function & Genetics 12 , 345-364) that could be used more often for NMR structures too. All protein structures submitted to PDB are now first checked with PROCHECK (at least if I am correct). Alexandre =============================================================================== | Alexandre Bonvin PhD | Phone: (203) 432-5066 | | Dept. Mol. Biophys. & Biochemistry | Fax: (203) 432-6946 | | Yale University, 266 Whitney Avenue | Email: abonvin@laplace.csb.yale.edu | | New Haven CT 06511, USA | | =============================================================================== From owner-structural-nmr@net.bio.net Tue May 03 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!umdac!Peter.Lundberg From: peterl@umdix.umdc.umu.se (Peter Lundberg) Newsgroups: bionet.structural-nmr Subject: Re: If bionet.structural-nmr is not found at your site Date: 4 May 1994 08:35:43 GMT Organization: University of Umea Lines: 12 Sender: -Not-Authenticated-[5799] Message-ID: <2q7mov$8og@news.umu.se> References: NNTP-Posting-Host: plgmac.chem.umu.se X-Posted-From: InterNews 1.0@news.umu.se. Xdisclaimer: No attempt was made to authenticate the sender's name. The problem at this site was that NEW news groups didn't show up in the listing. After contact with the (local) university system (network) manager this problem was rectified. I suppose it depends on how the local news server is set up whether you get access to new groups, or not. ======================================================= Peter Lundberg Ph. (+46-90) 16 59 73 Physical Chemistry Fax. (+46-90) 16 77 79 U of Umea peterl@umdix.umdc.umu.se S-901 87 Umea, Sweden ======================================================= From owner-structural-nmr@net.bio.net Tue May 03 23:00:00 1994 Path: biosci!NET.BIO.NET!biosci-help From: biosci-help@NET.BIO.NET (BIOSCI Administrator) Newsgroups: bionet.structural-nmr Subject: If bionet.structural-nmr is empty at your site Date: 3 May 1994 23:08:37 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 35 Sender: daemon@net.bio.net Distribution: bionet Message-ID: Reply-To: biosci-help@net.bio.net NNTP-Posting-Host: net.bio.net > > One question for you: can you see any reason a priori that > a site would have problems with bionet.structural-nmr while maintaining > proper distribution of the other bionet groups? > > Thanks, > Kevin > I have discussed this with the systems staff here and we think that the only way to resolve this issue is to have your news administrator check with the manager at the USENET news site immediately downstream from yours (they should know how to do this). Newsgroup creation messages go to a "control" newsgroup and at many sites these have to be acted on by a systems administrator manually. It is quite possible that you are not getting news to this group because, even though the group was set up at your site as a result of getting the newgroup message via "control", the site feeding you has not set it up properly even though they may have acted on other bionet USENET groups. The volume of newsgroup control messages is large and any time that a manual step is inserted, the probability of running in to a problem increases. Dave Mack in our systems group here assures me that all news sites that we feed directly are getting news from us, and I have run several tests personally of both the mail and news distributions. Unfortunately there is no way that we can diagnose this kind of remote new problem. You will have to ask your systems manager to get involved and check with your closest newsfeed site as mentioned above. Sincerely, Dave Kristofferson BIOSCI/bionet Manager biosci-help@net.bio.net From owner-structural-nmr@net.bio.net Tue May 03 23:00:00 1994 Path: biosci!biosci!not-for-mail From: kristoff@net.bio.net (David Kristofferson) Newsgroups: bionet.structural-nmr Subject: Re: If bionet.structural-nmr is not found at your site Date: 4 May 1994 16:11:00 -0700 Organization: BIOSCI International Newsgroups for Biology Lines: 46 Distribution: bionet Message-ID: <2q9a24$lin@net.bio.net> References: <2q7mov$8og@news.umu.se> NNTP-Posting-Host: net.bio.net In article <2q7mov$8og@news.umu.se>, Peter Lundberg wrote: >The problem at this site was that NEW news groups didn't show up in the >listing. After contact with the (local) university system (network) >manager this problem was rectified. I suppose it depends on how the >local news server is set up whether you get access to new groups, or >not. The jist of most of these problems seems to be that the news distribution can get disrupted either at one's site or at a site downstream when the news system in question is set up to require the manual intervention of the system news administrator to create new newsgroups. The volume of new USENET newsgroups is large enough that many sites are fearful to allow automatic newgroup creations. A possible solution to the problem is the following. Check with the person who administers USENET news at your site and see if they can allow automatic creation of groups in the bionet heirarchy (and any other hierarchies that your site would always find of interest). This can be done for some news programs such as INN but may not be possible at your site if you are using older news distribution software (note that I am talking here about the software that manages the news transport, not your front-end newsreading program.). While you are at it, you might also want to see what kind of message expiration policy they employ. At least one problem that was reported to me appeared to be caused by messages being deleted from the news system before the person got a chance to read them. News users should be aware that the expiration period can be set separately for each newsgroup. For example, because we maintain the BIOSCI/bionet archives here, we do not expire any of the bionet groups (except for bionet.molbio.genbank.updates). This is the extreme case, but users at your site should work with the news adminstrator to come up with a reasonable expiration policy dependent both upon the frequency with which you read news and the availability of disk space in your news directory. Sincerely, Dave Kristofferson BIOSCI/bionet Manager biosci-help@net.bio.net From owner-structural-nmr@net.bio.net Tue May 03 23:00:00 1994 Path: biosci!LAPLACE.CSB.YALE.EDU!abonvin From: abonvin@LAPLACE.CSB.YALE.EDU (Alexandre Bonvin) Newsgroups: bionet.structural-nmr Subject: Re: NMR data treatment: wrong problem? Date: 4 May 1994 13:08:48 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 53 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <9405042009.AA14320@laplace.csb.yale.edu> NNTP-Posting-Host: net.bio.net Gerard writes: .... > This is not true, since Rsym says something about the > quality (internal consistency, at least) of your DATA. > This is completely independent of the model and the > refinement. I still think that using the sixth power > is artificial in this respect, although there may be > a case for it in the case of R(Iobs,Icalc). In XRAY, > we have the same situation: low-resolution Is tend > to be much higher than high-resolution Is. > Ok, if you are just interested in the quality of the data then I completely agree with your remarks. When considering NMR refinement, however, the Rsym will give some lower limit to the achievable Rfactors and, in this respect, using a Rsym definition similar to the definition used to judge the quality of the model (R, R1/6 or whatever) seem logical to me. As you pointed out in your mail these are two different subjects. .... > > Are you saying that: long-range NOE = long-distance NOE ? > Why ? > What I mean is that weak NOEs = long distances (assuming uniform mobility for the entire structure). Long-range NOEs can still have very high intensities and correspond to short distances. In the later case of course these NOEs will be important for the 3D structure. What the 1/6 power is doing is somewhat "flattening" the NOE function (~ 1/r dependence instead of 1/r**6), still giving more weight to intense NOEs. Anyway, from my experience, both definitions (R and R1/6) give quite similar results in terms of distinguishing between bad and good structures (at least when you are reasonably close to the right answer). Alexandre =============================================================================== | Alexandre Bonvin PhD | Phone: (203) 432-5066 | | Dept. Mol. Biophys. & Biochemistry | Fax: (203) 432-6946 | | Yale University, 266 Whitney Avenue | Email: abonvin@laplace.csb.yale.edu | | New Haven CT 06511, USA | | =============================================================================== From owner-structural-nmr@net.bio.net Thu May 05 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!uknet!lyra.csx.cam.ac.uk!mole.bio.cam.ac.uk!smb18 From: smb18@mole.bio.cam.ac.uk (Simon Brocklehurst (Bioc)) Newsgroups: bionet.structural-nmr Subject: Re: NMR data treatment: wrong problem? Date: 6 May 1994 10:55:11 GMT Organization: University of Cambridge, England Lines: 38 Distribution: bionet Message-ID: <2qd7mf$q5@lyra.csx.cam.ac.uk> References: <9405041520.AA13671@laplace.csb.yale.edu> NNTP-Posting-Host: mole.bio.cam.ac.uk abonvin@LAPLACE.CSB.YALE.EDU (Alexandre Bonvin) writes: >> A question (just out of curiosity): why do so few >> NMR-protein structure papers contain a Ramachandran >> plot ? >May-be because they quite often look much more awfull for NMR structures >than for Xray ones, especially for undetermined floppy ends and loops... >But again, we are measuring in solution where conformational averaging can >take place and much more mobility is expected than in crystals: so, should >almost perfect Ramachandran plots for NMR structures be expected? There should be correlation between backbone r.m.s.d. and the "quality" of the Ramachandran plot, just as there is a correlation between X-ray structure resolution and plot "quality". High quality n.m.r. structures _do_ have "nice" Ramachandran plots for all members of the ensemble, and the _restrained_ minimized average structrure. In other words conformational averaging in solution does not imply high populations of states possessing poor stereochemistry. In n.m.r. structures, poor Ramachandran plots and low r.m.s.d.s almost certainly indicate that the model is a bit inaccurate (arising from errors in the restraints). This will not usually matter for rationalizing/planning protein engineering expriments etc. It will matter for computational analyses/structure prediction work. -- Simon !----------------------------------------------------------------------- Simon M. Brocklehurst Cambridge Centre for Molecular Recognition Department of Biochemistry University of Cambridge Cambridge UK E-mail: s.m.brocklehurst@bioc.cam.ac.uk From owner-structural-nmr@net.bio.net Thu May 05 23:00:00 1994 Newsgroups: bionet.structural-nmr Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!howland.reston.ans.net!cs.utexas.edu!convex!news.duke.edu!solaris.cc.vt.edu!spcuna!news.columbia.edu!psinntp!psinntp!psinntp!newsserver.pixel.kodak.com!kodak!r06a661 From: r06a661@bcc26.kodak.com (Martin J. Dellwo) Subject: Re: NMR data treatment: wrong problem? Message-ID: <1994May3.022203.8809@kodak.rdcs.kodak.com> Sender: news@kodak.rdcs.kodak.com (USENET NEWS) Reply-To: mdellwo@Kodak.COM Organization: Sterling Winthrop Pharmaceuticals Research Division References: <2pqse3$5tg@mserv1.dl.ac.uk> Distribution: bionet Date: Tue, 3 May 94 02:22:03 GMT Lines: 37 In article <2pqse3$5tg@mserv1.dl.ac.uk>, Per Kraulis discussed basic data handling. First, I would like to say the points raised were very good. In particular we have problems measuring good peak intensities, simply because of difficulty in getting sufficiently flat baselines in spite of the many data massaging tools at our disposal... I'd like to comment on the following point: >An example of what we can learn from the X-ray people: They calculate a >parameter R-sym to give an idea of how large the errors are in their >measured reflection intensities. R-sym is a measure of how different the >intensities of symmetry-related reflections are. These should ideally >be zero, but are actually around 5-10 % for typical data. This idea >can be used in NMR: compare the intensities of symmetry-related NOE >cross peaks, as a measure of how large the volume integration errors >are. In a recent paper (Widmer, Widmer & Braun, J. Biomol. NMR, 1993, >vol 3, pp 307-324) this was done, and errors of 20 % were found (30 % >for weak peaks). Interestingly, these authors chose to use the >*strongest* of the two symmetry-related NOE peaks in their refinement. >No justification was given. Unfortunately, you can't rely on NOE crosspeaks being symmetric because of incomplete relaxation between FID acquisitions. We've done simulations showing that incomplete relaxation in proteins can be quite severe, and subsequently leads to asymmetric intensities (see recent JMR). This can be accounted for in simulations or corrected for if you can measure the diagonal peaks in a 0 mixing time NOESY. Similar asymmetry can occur if solvent presaturation leads to saturation transfer to various spins. Guy Monteliano has shown experimentally that this can also be quite severe. -- Martin J. Dellwo (610) 983-7396 mdellwo@Kodak.COM Biophysics and Computational Chemistry Sterling Winthrop Pharmaceuticals Research Division From owner-structural-nmr@net.bio.net Thu May 05 23:00:00 1994 Path: biosci!AVERY.MED.VIRGINIA.EDU!rj8w From: rj8w@AVERY.MED.VIRGINIA.EDU (Roman Jerala) Newsgroups: bionet.structural-nmr Subject: Re: NMR data treatment Date: 6 May 1994 12:29:02 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 64 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <199405061929.AA24003@avery.med.Virginia.EDU> NNTP-Posting-Host: net.bio.net Per has thrown into the arena several problems of the NMR structure determination. I would like to address the problem of assignment which has not attracted the attention of others in the grou up to now. >We have no objective measures at all to assess the correctness of >assignments. It is *not* possible to say that this is not a problem. I .. >equilibria in the spectra of almost every protein he as worked with. I >can add that I have also seen examples of "too many cross peaks" in >many spectra that I have worked with. It is most likely a wide-spread >phenomenon. And yet we in the NMR community do not state this in our >papers, nor really do we know how to deal with it? I remember a talk >given by Gerhard Wagner about "ugly features in spectra", meaning >mostly these strange peaks. But that's all the public discussion I >have ever heard about this problem. I think that direct comparison with crystallographic R factor by including only the intensity of the NOE crosspeaks is inappropriate and insufficient. The basic difference between the X-ray and NMR data is that the position of Xray diffraction spots reflects only the parameters of the crystal cell and contains essentially no information about the structure of the molecule. Comparison of observed and calculated intensities is adequate measure of the quality of the model. On the other hand, the position of crosspeaks (xpk) in the NMR spectra contain the majority of the information about the structure. This is reflected also in the fact that sorting of NOE xpk intensity into only a few categories is sufficient to obtain good structures. Using only the intensity of assigned xpks to estimate the quality of the model does not reflect the information content of assignment of NMR spectra. The quality of the NMR structure depends in large extent on the completeness and accuracy of the assignment of xpks. The measure of the quality of the model should to my opinion include the comparison between the observed NOEs (% of the assigned xpks) and predicted from the ensamble of model structures. Roman P.S. Does anybody know what is the highest % of all observed NOE xpks that was assigned and used in determination of 3D structure of a protein ? -- Roman Jerala rj8w@virginia.edu Department of Pharmacology University of Virginia, Charlottesville From owner-structural-nmr@net.bio.net Thu May 05 23:00:00 1994 Newsgroups: bionet.structural-nmr Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!sun4nl!news.nic.surfnet.nl!ruu.nl!RUUCI10!rasmus From: rasmus@RUUCI10.NoSubdomain.NoDomain (Rasmus Fogh) Subject: Re: NMR data treatment: wrong problem? Message-ID: <1994May6.172312.26534@cc.ruu.nl> Sender: rasmus@RUUCI10 (Rasmus Fogh) Organization: ACCU References: <9405041520.AA13671@laplace.csb.yale.edu> <2qd7mf$q5@lyra.csx.cam.ac.uk> Distribution: bionet Date: Fri, 6 May 1994 17:23:12 GMT Lines: 50 smb18@mole.bio.cam.ac.uk (Simon Brocklehurst (Bioc)) writes : >abonvin@LAPLACE.CSB.YALE.EDU (Alexandre Bonvin) writes: >>> A question (just out of curiosity): why do so few >>> NMR-protein structure papers contain a Ramachandran >>> plot ? >>May-be because they quite often look much more awfull for NMR structures >>than for Xray ones, especially for undetermined floppy ends and loops... >>But again, we are measuring in solution where conformational averaging can >>take place and much more mobility is expected than in crystals: so, should >>almost perfect Ramachandran plots for NMR structures be expected? > There should be correlation between backbone r.m.s.d. and the >"quality" of the Ramachandran plot, just as there is a correlation >between X-ray structure resolution and plot "quality". >High quality n.m.r. structures _do_ have "nice" Ramachandran plots for >all members of the ensemble, and the _restrained_ minimized average >structrure. >In other words conformational averaging in solution does not imply high >populations of states possessing poor stereochemistry. In n.m.r. >structures, poor Ramachandran plots and low r.m.s.d.s almost certainly >indicate that the model is a bit inaccurate (arising from errors in the >restraints). This will not usually matter for rationalizing/planning >protein engineering expriments etc. It will matter for computational >analyses/structure prediction work. My two cents worth : Certainly high-quality NMR structures should have "nice" Ramachandran plots, but does a "bad" plot mean that the model is inaccurate, or that some backbone dihedrals are not very well defined? Only well-defined angles in funny parts of the Ramachandran plots should count towards that conclusion, methinks. As for the 'undetermined floppy ends and loops', would it not be better to conclude that this part of the structure is undefined, and then exclude it from consideration (and from the Ramachandran plot)? I have a couple of times come across parts of the a structure that a) were undefined, b) contained very few restrained c) contained very high (sic!) levels of restraint violations, d) contained high concentrations of "bad" Ramachandran plot residues. My best explanation for this strange phenomenon is that maybe that is what you get when you sample over a too large conformational space for essentially unconstrained residues using an incorrect (=vacuum) force field. Can anyone suggest what is really going on ? -- Rasmus Fogh (Protein Structural) NMR Group Utrecht University, Utrecht, Holland From owner-structural-nmr@net.bio.net Mon May 09 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!pipex!howland.reston.ans.net!wupost!news.miami.edu!usenet.ufl.edu!nervm.nerdc.ufl.edu!NEMNM From: NEMNM@nervm.nerdc.ufl.edu (Marcus N. Morgan) Newsgroups: bionet.structural-nmr Subject: this is a test, please ignore Date: Tue, 10 May 94 13:07:10 EDT Organization: University of Florida, NERDC Lines: 1 Message-ID: <16FB2B888S86.NEMNM@nervm.nerdc.ufl.edu> NNTP-Posting-Host: nervm.nerdc.ufl.edu this is a test. please ignore. thank you. From owner-structural-nmr@net.bio.net Tue May 10 23:00:00 1994 Path: biosci!daresbury!s-crim1!mbarg From: mbarg@s-crim1.dl.ac.uk (A.R. Gargaro) Newsgroups: bionet.structural-nmr Subject: xplor question Date: 11 May 1994 13:43:02 GMT Organization: SERC Daresbury Lab, Warrington, U.K. Lines: 15 Sender: mbarg@s-crim1 (A.R. Gargaro) Distribution: world Message-ID: <2qqnd6$ap6@mserv1.dl.ac.uk> NNTP-Posting-Host: s-crim1.dl.ac.uk Dear all, I notice that the xplor sa.inp and refine.inp protocols do not appear (from my reading of them anyway) to scale the noe force constant in any way,other than by increasing the asymptote before the dynamic stages of the calculation. My question is does this have any effect on the final structures that are calculated from these methods by comparison to those people who gradually increase the noe force constant slowly during the dyanimc stages of the calculation. Many thanks in advance Angelo Gargaro Department of Biochemistry University of Bristol Bristol UK From owner-structural-nmr@net.bio.net Wed May 11 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!howland.reston.ans.net!xlink.net!scsing.switch.ch!elna.ethz.ch!torda From: torda@igc.ethz.ch (Andrew Torda) Newsgroups: bionet.structural-nmr Subject: Re: Request for restraint lists Date: 12 May 1994 20:37:25 GMT Organization: Computational Chemistry, ETH, Zuerich Lines: 17 Message-ID: <2qu425$cl1@elna.ethz.ch> References: <9405121422.AA09350@sneezy.abbott.com> NNTP-Posting-Host: hermitage.ethz.ch rob@sneezy.abbott.com (Robert Powers Meadows) wrote [... looking for restraint data ..] >published the structures. I've looked in Brookhaven, no help there, and >who wants to pay $$$ for a couple of restraint lists from BioMagResBank? When you say "looked in Brookhaven, no help there" do you mean no restraints or none of the right type ? There is a bit more than 20 set there. Have a look at the following URL URL: gopher://pdb.pdb.bnl.gov:70/0/0/FTP/nmr_restraints/README or just ftp into the appropriate directory. -Andrew -- This posting is not merely my opinion. It is the official position of my employer, the Swiss Federal Institute of Technology. (Andrew Torda, Computational Chemistry, ETH, Zurich, torda@igc.chem.ethz.ch) From owner-structural-nmr@net.bio.net Wed May 11 23:00:00 1994 Path: biosci!sneezy.abbott.com!rob From: rob@sneezy.abbott.com (Robert Powers Meadows) Newsgroups: bionet.structural-nmr Subject: Request for restraint lists Date: 12 May 1994 07:22:43 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 28 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <9405121422.AA09350@sneezy.abbott.com> NNTP-Posting-Host: net.bio.net Dear Netters, I am performing some tests on a new (to the NMR community) method for calculating the 3-dimensional structures of small peptides and proteins from NMR data. It appears that the protocol will be limited, at least in my hands, to systems of +/- 35 residues and less. Although the protocol is rapid and samples lots of conformational space, it may be ultimately restricted in scope to these relatively small systems. Unfortunately, here at Fesik Labs, we don't do many smaller proteins. I'd like to apply this technique to a real (not simulated) system, especially one that may be sampling a number of different conformations, but we just don't have any of this size. Thus my question. Would it be possible to get from any of you a restraint list for your 10-30 residue peptides? I could complete the refinement in an evening and send you the results, if you wish. Of course, this would only be acceptable if you were near completion, or had completed and published the structures. I've looked in Brookhaven, no help there, and who wants to pay $$$ for a couple of restraint lists from BioMagResBank? Thanks for any help. Robert P. Meadows, Ph.D. rob@sneezy.abbott.com Computational Chemistry and Molecular Sciences Abbott Laboratories, Abbott Park, IL 60064 (708) 937-5317 From owner-structural-nmr@net.bio.net Thu May 12 23:00:00 1994 Path: biosci!DGGPI2.CHEM.PURDUE.EDU!bruce From: bruce@DGGPI2.CHEM.PURDUE.EDU (Bruce Luxon) Newsgroups: bionet.structural-nmr Subject: Morass Software Update Date: 13 May 1994 12:16:00 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 81 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <9405131916.AA10628@dggpi2.chem.purdue.edu> NNTP-Posting-Host: net.bio.net -=-=-=-=--=-=-=-=-=-=-=-=-= MORASS version 2.1 -=-=-=-=-=-=-=--=-=-=-=-=- ANNOUNCING NEW MORASS SOFTWARE NOW AVAILABLE ON INTERNET - May 13, 1994 o MORASS - Multiple Overhauser Relaxation AnalysiS and Simulation - uses a full hybrid matrix eigenvalue/eigenvector solution to the Bloch equations to derive cross-relaxation rates and interproton distances. o MORASS analyzes 2D NMR NOESY data from oligonucleotides and proteins to evaluate cross-relaxation rates from which interproton distances are obtained. These are output in a format suitable for use as distance constraints in molecular dynamics calculations. o MORASS 2.1 will now produce output suitable for display using the graphics package GRASP. With GRASP, NOESY constraint differences are displayed with a .pdb file as pair-wise interactions (constraint differences being the difference between the theoretical NOESY derived interproton distances and the distance calculated from the .pdb file). The constraint diffs are shown as color-coded cylinders between the two sites and the width of the cylinders are proportional to the magnitude of the difference. o GRASP modified to accept MORASS input is available from Dr. Anthony Nicholls at Columbia Universiy (nicholls@cuhhca.hhmi.columbia.edu). o MORASS 2.1 comes with new, complete and up to date documentation. o Older MORASS input files should run error-free on version 2.1 . o MORASS code compiles on SGI Irix 4.04 & 4.0.5H, IBM RS6000, DEC Ultrix & VAX. It may compile on others but I'm personally not familiar with them. MORASS IS AVAILABLE FROM - 1. FTP ANONYMOUS to dggpi2.chem.purdue.edu. Tar file is in /pub/morass. 2. GOPHER to infomeister.osc.edu (Ohio Supercomputer Center, Ohio State U.) -> Software / Morass 3. GOPHER to nmrfam.wisc.edu (Univ. of Wisconsin, Madison) -> Other NMR Gopher Servers and NMR Software Sites / Morass 4. GOPHER to micro.ifas.ufl.edu (University of Florida, Gainsville) -> NMR Software / Morass TO RUN MORASS YOU NEED TO GET - o morass21.tar.Z and YOU MAY ALSO WISH TO GET - o ../flatwell/flatwell.tar.Z - AMBER flatwell potential code o cshift/cshift.tar.Z - Chemical shift routine o ../util/util/tar.Z - AMBER and MORASS utilities o testout/testout.tar.Z - output files from the test input CONTENTS of .tar FILE INCLUDE - o Source code in Fortran o Makefiles for SGI, IBM RS6K and DEC Ultrix o Documentation in LaTeX format (also in /docs) o Sample input files o Sample output files o GRASP - MORASS .rgb sample files (also in /grasp) Note that the DOCS and GRASP files are available separately as noted but the .tar file contains EVERYTHING. BRUCE ____________________________________________________________________________ | | | |"You can't always get what you want, | Dr. Bruce A. Luxon | | But if you try sometimes, | Chemistry Department | | You just might find | Purdue University | | You get what you need ..." | W. Lafayette, IN 47907 | | | (317)494-5289; Fax (317)494-0239 | | Mick Jagger | bruce@dggpi2.chem.purdue.edu | |______________________________________|___________________________________| From owner-structural-nmr@net.bio.net Fri May 13 23:00:00 1994 Path: biosci!SCRIPPS.EDU!jyao From: jyao@SCRIPPS.EDU (Jian Yao) Newsgroups: bionet.structural-nmr Subject: (none) Date: 13 May 1994 19:52:17 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 1 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <9405140051.AA01502@wright.Scripps.EDU> NNTP-Posting-Host: net.bio.net subscribe str-nmr From owner-structural-nmr@net.bio.net Sun May 15 23:00:00 1994 Path: biosci!LILY.CHEM.WESLEYAN.EDU!phbolton From: phbolton@LILY.CHEM.WESLEYAN.EDU (Philip H. Bolton) Newsgroups: bionet.structural-nmr Subject: shigemi & water Date: 16 May 1994 11:04:39 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 18 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <9405161905.AA12818@lily.chem.wesleyan.edu> NNTP-Posting-Host: net.bio.net Has anyone tried out the shigemi "symmetrical nmr microtubes" and if so did you obtain significantly better water lineshape-water suppression? ---------------------------------------------------------------------------- --------------------- Philip H. Bolton Professor of Chemistry Chemistry Department Wesleyan University Middletown, CT 06459 phbolton@lily.chem.wesleyan.edu ---------------------------------------------------------------------------- --------------------- From owner-structural-nmr@net.bio.net Mon May 16 23:00:00 1994 Path: biosci!PUCC.PRINCETON.EDU!PMCDONNE%ALLOY.BITNET From: PMCDONNE%ALLOY.BITNET@PUCC.PRINCETON.EDU Newsgroups: bionet.structural-nmr Subject: C-F coupling Date: 17 May 1994 10:01:53 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 10 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <199405171701.KAA19340@net.bio.net> NNTP-Posting-Host: net.bio.net Has anyone observed a 6 bond C-F coupling? I have a 2-p-fluoro-ethene side chain and think I observe a coupling of 2.6 Hz between the first carbon and the fluorine. Patty McDonnell Pharmaceutical Research Institute Spring House, PA 19477 pmcdonne@alloy.bitnet From owner-structural-nmr@net.bio.net Mon May 16 23:00:00 1994 Path: biosci!PUCC.PRINCETON.EDU!PMCDONNE%ALLOY.BITNET From: PMCDONNE%ALLOY.BITNET@PUCC.PRINCETON.EDU Newsgroups: bionet.structural-nmr Subject: address for fts Date: 17 May 1994 06:31:44 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 11 Sender: daemon@net.bio.net Distribution: bionet Message-ID: <199405171331.GAA08257@net.bio.net> NNTP-Posting-Host: net.bio.net Does anyone know the address or phone number for FTS? I want to order a temperature control unit for my NMR. The phone number that I have for FTS is incorrect and I am having difficulties getting the correct phone number and address. Thanks. Patty McDonnell Pharmaceutical Research Institute Spring House, PA 19477 pmcdonne@alloy.bitnet From owner-structural-nmr@net.bio.net Mon May 16 23:00:00 1994 Path: biosci!parc!decwrl!envoy.wl.com!reeve.research.aa.wl.com!mcmag2.chemistry.aa.wl.com!user From: reily@aa.wl.com (Dr. Michael Reily) Newsgroups: bionet.structural-nmr Subject: Re: shigemi & water Followup-To: bionet.structural-nmr Date: 17 May 1994 13:06:35 GMT Organization: Parke-Davis Lines: 38 Distribution: bionet Message-ID: References: <9405161905.AA12818@lily.chem.wesleyan.edu> NNTP-Posting-Host: mcmag2.chemistry.aa.wl.com In article <9405161905.AA12818@lily.chem.wesleyan.edu>, phbolton@LILY.CHEM.WESLEYAN.EDU (Philip H. Bolton) wrote: > Has anyone tried out the shigemi "symmetrical nmr microtubes" and if so did > you obtain significantly better water lineshape-water suppression? > > ---------------------------------------------------------------------------- > --------------------- > Philip H. Bolton > Professor of Chemistry > > Chemistry Department > Wesleyan University > Middletown, CT 06459 > > phbolton@lily.chem.wesleyan.edu > > ---------------------------------------------------------------------------- > --------------------- My experience with the shigemi tubes has been good. I had a terrible problem with getting the bubbles completely out of the space between the plunger and the sample, largely due to the propensity of my sample to froth as I moved the plunger up and down. As a result, I wound up using the tubes without the plunger portion, allowing the use of about 150ul less solvent. If you are working with non-surfactive molecules, using the tubes as they were meant to be used should be no problem. Michael D. Reily Senior Research Associate Parke Davis Pharmaceutical Reseach Division Warner-Lambert Company 2800 Plymouth Road Ann Arbor, Michigan 48105 313-996-7128 phone 313-998-2716 fax reily@aa.wl.com From owner-structural-nmr@net.bio.net Mon May 16 23:00:00 1994 Path: biosci!SBMM1.UCSB.EDU!BLASKO From: BLASKO@SBMM1.UCSB.EDU (Andrei Blasko) Newsgroups: bionet.structural-nmr Subject: P31 Date: 17 May 1994 11:27:56 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 8 Sender: daemon@net.bio.net Distribution: world Message-ID: <940517112718.202015c3@SBMM1.UCSB.EDU> NNTP-Posting-Host: net.bio.net Has anyone observed a P31 signal of a cyclic acyl phosphate 6 membered ring at -11 ppm? Andrei Blasko University of California Santa Barbara,CA 93106 (805)893-2234 e-mail blasko@sbmm1.ucsb.edu From owner-structural-nmr@net.bio.net Tue May 17 23:00:00 1994 Newsgroups: bionet.structural-nmr Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!uknet!bcc.ac.uk!bsm.bioc.ucl.ac.uk!mcdonald From: mcdonald@bsm.bioc.ucl.ac.uk (Ian McDonald) Subject: Histidine Chemistry Message-ID: <1994May18.204338.26681@ucl.ac.uk> Date: Wed, 18 May 1994 20:43:38 GMT Organization: University College London Followup-To: Poster Lines: 15 I am trying to discover which proto-isomer of Histidine is more common. Ie amongst Histidine side-chains in basic form, which nitrogen is more likely to carry the hydrogen. I am more interested in chemical studies. I know this is off-topic, but I'm asking it because I thought you might be likely to know. Please email replies, I'll post a summary. Thanks Ian From owner-structural-nmr@net.bio.net Mon May 23 23:00:00 1994 Path: biosci!WPI.EDU!samster From: samster@WPI.EDU (Sam Han) Newsgroups: bionet.structural-nmr Subject: unsubscribe Date: 23 May 1994 18:13:26 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 3 Sender: daemon@net.bio.net Distribution: world Message-ID: Reply-To: Sam Han NNTP-Posting-Host: net.bio.net unsubscribe From owner-structural-nmr@net.bio.net Mon May 23 23:00:00 1994 Path: biosci!BCCHEM.BC.EDU!markman From: markman@BCCHEM.BC.EDU Newsgroups: bionet.structural-nmr Subject: does realy short range mean low info Date: 23 May 1994 17:00:08 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 53 Sender: daemon@net.bio.net Distribution: world Message-ID: <94052318541176@bcchem.bc.edu> NNTP-Posting-Host: net.bio.net I have read the content of all the months I have not been reading the news group. I have enjoyed the disscussion on quality of nmr structures and the new ideas about the way to calculate quality measurements I would want to note something that seems to be 'unproven-truth' of nmr: that a long range interaction is "more informative" than the "short range" ones. I would argue that as for X-ray, so does in nmr the "solution" or calculation of primery structure is sensitive to different things than refinement. While for the primary structure I would agree that the long range interaction are important, for refinement, the short range interaction would have more info on side chain conformation than the long range. More so, the information from short range interaction many time constraint one conformation of SC with less possible conformers, as it is defining conformation in a defined system, and with one possible orientation, while in long range interactions, the distance is between two independent systems each have its position unknown, I.e. more degrees of freedom. In that sense, and in that sense only, the long range interactions carry less information on the structure, because they have less known about them. For example interaction (NOE) of HD of lysine XX with MeD of Lue YY carry very little positional information, and in some way should be regarded as low resolution data, while the distance between HN and HD in Lys XX and the HN MeD in Lue YY each can define the conformation of the whole residue. In X-ray it is clear today that all data points are important, for either the short range info (6.0 ang and lower) for sec structure elements and for the shape of the molecule's envelope (6.0 ang and more) and one should consider putting them in for certain applications. I think there was a lot of jumping to conclusion by the descision to give weak noes more weights. More over, a general application of bias toward the weak data have less sense, both in term of the physical reality (the reliability of weak noes is less than of strong) and in terms of information contents, The info in the long range (noe) deppend on the info from short range and on the quality of assessing side chain conformations. Moreover, short distance - long range noe, is definitly more informative then long distance - long range noe. It is still remained to check this claim. Many time it seems like just a phylosophical problem, but it seem to be important when we come to assess a weighting scheme for any R factor. It maybe that in nmr we have to consider not only distance and intensity, but also context, i.e. what protons does the noes come from? After all while in X-ray each reflection have info on all the structure, nmr data is context dependent. It sure deppend on the assignment. Well... these are my thoughts on this issue If things go the way they are I am not getting to see the responses to this, unless sent also directly to me at MARKMANO@bcvms.bc.edu. or markman@bcchem.bc.edu Ofer Markman , Boston College From owner-structural-nmr@net.bio.net Mon May 23 23:00:00 1994 Path: biosci!TELEMANN.PSE.UMASS.EDU!charlie From: charlie@TELEMANN.PSE.UMASS.EDU (leonard charles dickinson) Newsgroups: bionet.structural-nmr Subject: (none) Date: 24 May 1994 05:19:59 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 20 Sender: daemon@net.bio.net Distribution: world Message-ID: <9405240819.AA21584@telemann.pse.umass.edu> NNTP-Posting-Host: net.bio.net May 24, 1994 My initial interest in this network was for applications to synthetic polymer structures. The general content of the network traffic though of some interest does not merit the time it takes to keep up with the brisk flow of information. Thus I request that my name be removed from the mail listing. Thanks and good luck for what appears to be a most useful tool for the protein/nucleic acid NMR structuralist. +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Dr. L.Charles Dickinson voice(413)545-3141 NMR Director fax -0082 Department of Polymer Science and Engineering University of Massachusetts Amherst MA 01003 USA charlie@telemann.pse.umass.edu +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ From owner-structural-nmr@net.bio.net Mon May 23 23:00:00 1994 Path: biosci!TELEMANN.PSE.UMASS.EDU!charlie From: charlie@TELEMANN.PSE.UMASS.EDU (leonard charles dickinson) Newsgroups: bionet.structural-nmr Subject: (none) Date: 24 May 1994 05:19:54 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 2 Sender: daemon@net.bio.net Distribution: world Message-ID: <9405240818.AA21582@telemann.pse.umass.edu> NNTP-Posting-Host: net.bio.net NNTP-Posting-Host: net.bio.net unsubscribe ******************************** Charles L. Borders, Jr. Department of Chemistry The College of Wooster Wooster, OH 44691 ******************************** From owner-structural-nmr@net.bio.net Tue May 24 23:00:00 1994 Path: biosci!agate!howland.reston.ans.net!pipex!uknet!EU.net!ub4b!idefix.CS.kuleuven.ac.be!rc1.vub.ac.be!is3e!philstas From: philstas@vub.ac.be (Stas Philippe) Newsgroups: bionet.general,bionet.software,bionet.structural-nmr,bionet.xtallography,bionet.molbio.proteins,sci.techniques.xtallography Subject: Re: CALL FOR VOTES: MOLECULAR-MODELLING/bionet.molec-model Date: 25 May 1994 09:15:34 GMT Organization: Brussels Free Universities (VUB/ULB), Belgium Lines: 70 Distribution: bionet Message-ID: <2rv4vm$jsi@rc1.vub.ac.be> References: <2rqscc$mo5@net.bio.net> NNTP-Posting-Host: is3e.vub.ac.be X-Newsreader: TIN [version 1.2 PL2] Xref: biosci bionet.general:9298 bionet.software:8295 bionet.structural-nmr:115 bionet.xtallography:939 bionet.molbio.proteins:1987 sci.techniques.xtallography:385 The newsgroup molecular-modelling will discuss all modelling techniques , excuding X-ray and Nmr techniques (see other newsgroups) Voting is now open on the MOLECULAR-MODELLING/bionet.molec-model newsgroup and will run through 24:00 hrs Pacific Time on 21 June 1994. Please send your vote to either of the following addresses: Address Location Network ------- -------- ------- biovote@daresbury.ac.uk U.K. JANET biovote@net.bio.net U.S.A. Internet/BITNET PLEASE BE SURE TO FOLLOW THE FORMAT BELOW - WE OFTEN RUN MORE THAN ONE VOTE AT A TIME SO A SIMPLE "YES" OR "NO" MESSAGE WITHOUT THE NEWSGROUP NAME MAY BE AMBIGUOUS. Your vote should contain a single line: YES on MOLMODEL if you favor creation of the newsgroup above or NO on MOLMODEL : if you think that the creation of this newsgroup will adversely affect : the BIOSCI system. If you are simply not interested in participating : in this newsgroup, please don't cast a NO vote, but instead just don't : vote at all. : The newsgroup proposal must receive at least 80 YES votes to pass and : the number of YES votes must be greater than the number of NO votes by : at least 40. Discussion of the newsgroup proposal is now closed and : we strongly discourage posting any messages in other forums about the : fact that a CALL FOR VOTES has been issued. ---------------------------------------------------------------------- Proposal to establish MOLECULAR-MODELLING/bionet.molec-model Proposed USENET name: bionet.molec-model Proposed Mailing list name: MOLECULAR-MODELLING Proposed e-mail addresses: molmodel@net.bio.net molmodel@daresbury.ac.uk Discussion Leader: Philippe Stas philstas@vub.ac.be Dept. of Cellular Immunology Free University of Brussels tel: 32-2-359.03.58 Paardenstraat 65 Fax: 32-2-359.03.59 B1640 St.Genesius Rode Charter: This will be a non-moderated newsgroup about the physical and chemical aspects of molecular modelling, including discussions about algorythms and the applications behind the software. The interest of the group will be more towards modelling outcomes, pitfalls, methodology and techniques, rather than towards discussions on bugs or system errors. Discussion about all kinds of molecular modelling techniques, such as homology modelling, molecular dynamics, diffuse dynamics, electrostatic calculations, surface analysis, will be included in the newsgroup, excluding discussions about X-tal or NMR techniques (see other groups). The newsgroup will not be restricted on the program or platform used. A FAQ will be compiled and posted monthly. ---------------------------------------------------------------------- : Sincerely, : Dave Kristofferson : BIOSCI/bionet Manager : biosci-help@net.bio.net From owner-structural-nmr@net.bio.net Tue May 24 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!howland.reston.ans.net!spool.mu.edu!uwm.edu!post.its.mcw.edu!news.doit.wisc.edu!f181-196.net.wisc.edu!user From: Kckim@students.wisc.edu (Konrad C. Kim) Newsgroups: bionet.structural-nmr Subject: Biotechnology job Followup-To: bionet.structural-nmr Date: 25 May 1994 21:56:45 GMT Organization: Univeristy of Wisconsin Lines: 52 Distribution: world Message-ID: NNTP-Posting-Host: f181-196.net.wisc.edu KONRAD C. KIM 16 White Deer Ct. Huntington, NY 11743 (516)549-5205 EDUCATION: University of Wisconsin-Madison: College of Letters and Sciences, B.S Zoology 1988-92Õ College of Agriculture and Life Sciences, B.S. Genetics 1993-94Õ Related Courses taken: Biochemistry, Cell Biology, Organic Chemistry, Endocrinology, Cell Physiology, Neurobiology, Microbiology, Soil Microbiology and Biochemistry, Human Genetics WORK EXPERIENCE: Junior Lab Technician, 6/92-6/93 University of Minnesota Hospital and Clinics Department of Therapeutic Radiology Radiation Oncology Radiation Biology Section Supervisor: Beth Auger. Assistant Professor Ph.D. Principal Investigator: Changwon Song. Professor Ph.D. 424 Harvard St. SE Minneapolis, MN 55455 612-626-6090 Biological Assistant, 1/92-6/92 Veterans Memorial Middelton Hospital Department of Neurology Supervisor: Jim Vann. Senior Lab Scientist and Manager 1134 Terrace Av. Madison WI 53715 SKILLS: Protein Isolation: SDS-PAGE, Affinity chromotography, Dialysis/Salting out, Agglutination Tissue Culture: Mark I Cesium Irradiation Chamber, Fsa2 and SCK Cell Types Genetics: Southern, Western, and Northern Blots, Gel Electrophoresis, PCR AWARDS AND HOBBIES: Eagle Scout University of Wisconsin All-String Orchestra 1988-92Õ References on request From owner-structural-nmr@net.bio.net Tue May 24 23:00:00 1994 Path: biosci!daresbury!not-for-mail From: gharden@phoenix.oulu.fi (Grahame Harden) Newsgroups: bionet.structural-nmr Subject: Listings Date: 25 May 1994 23:03:43 +0100 Lines: 15 Sender: daemon@mserv1.dl.ac.uk Distribution: bionet Message-ID: <2s0hvv$41v@mserv1.dl.ac.uk> Original-To: str-nmr@dl.ac.uk Georges asked, How do I get access to some of the past discussions on nmr on this list? Thanks Quite simply... point your GOPHER client to net.bio.net and look in the Structural NM sub-menu. Cheers Grahame gharden@phoenix.oulu.fi From owner-structural-nmr@net.bio.net Tue May 24 23:00:00 1994 Path: biosci!BIOC01.UTHSCSA.EDU!raman From: raman@BIOC01.UTHSCSA.EDU (C.S.RAMAN) Newsgroups: bionet.structural-nmr Subject: Re: past discussions.. Date: 25 May 1994 14:50:33 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 30 Sender: daemon@net.bio.net Distribution: world Message-ID: <9405252151.AA06343@bioc01.uthscsa.edu> References: <2s0fso$1rj@mserv1.dl.ac.uk> NNTP-Posting-Host: net.bio.net Georges: > > How do I get access to some of the past discussions on nmr on this list? Here is the URL for locating past discussion: Type=1 Name=STRUCTURAL-NMR Path=1/STRUCTURAL-NMR Host=gopher.bio.net Port=70 If you have further questions, feel free to contact me. Cheers -raman -- C.S.Raman UNIX Programming & Administration SPARC & SGI Systems raman@bioc01.uthscsa.edu - INTERNET Department of Biochemistry raman@mintaka.chpc.utexas.edu - CHPC UTHSCSA c.raman@launchpad.unc.edu 7703 Floyd Curl Dr. (210) 567-6623 [Tel] San Antonio, TX 78284-7760 (210) 567-6595 [Fax] ****************************************************************************** If a man's wit be wandering, let him study the Mathematics -Francis Bacon ****************************************************************************** From owner-structural-nmr@net.bio.net Tue May 24 23:00:00 1994 Path: biosci!daresbury!not-for-mail From: Mr Georges L Friedli Newsgroups: bionet.structural-nmr Subject: past discursions Date: 25 May 1994 22:27:52 +0100 Lines: 4 Sender: daemon@mserv1.dl.ac.uk Distribution: bionet Message-ID: <2s0fso$1rj@mserv1.dl.ac.uk> Original-To: str-nmr@dl.ac.uk How do I get access to some of the past discussions on nmr on this list? Thanks Georges From owner-structural-nmr@net.bio.net Wed May 25 23:00:00 1994 Path: biosci!WHITESWS.MC.DUKE.EDU!dave From: dave@WHITESWS.MC.DUKE.EDU (david w hoffman) Newsgroups: bionet.structural-nmr Subject: unsubscribe Date: 26 May 1994 09:08:28 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 2 Sender: daemon@net.bio.net Distribution: world Message-ID: <9405261610.AA04659@whitesws.mc.duke.edu> NNTP-Posting-Host: net.bio.net unsubscribe From owner-structural-nmr@net.bio.net Wed May 25 23:00:00 1994 Path: biosci!daresbury!not-for-mail From: " (Ton Rullmann)" Newsgroups: bionet.structural-nmr Subject: Re: References needed Date: 26 May 1994 15:21:42 +0100 Lines: 47 Sender: daemon@mserv1.dl.ac.uk Distribution: bionet Message-ID: <2s2b9m$7d8@mserv1.dl.ac.uk> Reply-To: rull@ruuci9.chem.ruu.nl Original-To: str-nmr@dl.ac.uk Therese Malliavin wrote: > I know that methods were proposed for nOe build-up curve > fitting by using polynomials. I need precise references about > these methods. I know that some ones were published by R. Boelens > and others in J.Mag.Res. > Could anyone help me? > Boelens et al., J.Mag.Res. 82, 290-308 (1989) describe initial rate analysis of the NOE buildup curve. (It is not polynomial fitting, so maybe you are thinking of something else, but this is what we sometimes use.) The buildup is fitted to : noe_at_time_t = final + [initial - final]*exp(-k*t), where 'initial' and 'final' are estimates for the initial and final values of the noe. These values can be set by hand (while viewing the curve on a graphics screen), or automatically (e.g. set initial = 0, and final = (1+delta)*noe_at_largest_mixing_time). The fitting is done by computing the rate constant k as: sum{ ln((noe - final)/(initial - final)) } -k = ___________________________________________ sum{ t } where the sums are taken over all mixing times t. If for a given mixing time noe < initial, then this mixing time is excluded from the summations. However, the fitting does not make much sense for buildups that are heavily influenced by spin diffusion. Some of these have an 's-shape'. Our automatic fitting procedures try to detect this (slope of first line segment smaller than of second segment). The slope of the first line segment of the buildup might then be used as (the lower limit to) the buildup rate. Of course, one should inspect all curves visually, and correct for spin diffusion by using relaxation matrix theory. I hope this is of some help. -- | Ton Rullmann NMR Spectroscopy | | Bijvoet Center for Biomolecular Research | Tel. : int+31.30.533641 | | Utrecht University, Padualaan 8, | Fax : int+31.30.537623 | | 3584 CH Utrecht, The Netherlands | Email: rull@nmr.chem.ruu.nl | From owner-structural-nmr@net.bio.net Wed May 25 23:00:00 1994 Path: biosci!internet!biosci!not-for-mail From: kristoff (David Kristofferson) Newsgroups: bionet.structural-nmr Subject: UNSUBSCRIBING, BIOSCI ARCHIVES, ADDRESS DATABASE & BIOSCI FAQ Date: 26 May 1994 02:00:29 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 289 Sender: daemon@net.bio.net Distribution: world Message-ID: <199405260900.CAA21933@net.bio.net> NNTP-Posting-Host: net.bio.net Four important items follow: How to cancel e-mail subscriptions to BIOSCI newsgroups, BIOSCI archive searching, the BIOSCI FAQ, and the BIOSCI User Address Directory form. If you have not yet listed yourself in our BIOSCI user directory, please take a few minutes to complete and return the form below. If your personal information has changed since you listed yourself, please send us an updated form. Sincerely, Dave Kristofferson BIOSCI/bionet Manager biosci-help@net.bio.net **** How to cancel a BIOSCI e-mail subscription **** If you want to cancel your e-mail subscription to this group, PLEASE DO NOT POST YOUR UNSUBSCRIBE REQUEST TO THE NEWSGROUP ADDRESS (NOR REPLY TO A MESSAGE POSTED TO THE NEWSGROUP)!!! This would send your request to all of the readers of the newsgroup, but it might still not be seen by the BIOSCI staff - thus you would annoy many people and possibly not accomplish your goal anyway. IF YOU ARE LOCATED IN THE AMERICAS OR PACIFIC RIM COUNTRIES, please send a message to biosci@net.bio.net Instructions will be returned automatically, so the contents of your message do not matter. IF YOU ARE LOCATED IN EUROPE, AFRICA OR CENTRAL ASIA, please send a message to MXT@dl.ac.uk containing the word help in the body of the message to retrieve e-mail server instructions. Any text placed on the Subject: line of your message will be ignored, so be sure to put the "help" command in the body of the message. If you need personal assistance, a BIOSCI staff member can be contacted at either of the following addresses. Please contact the address designated for your location. Support Address Location --------------- -------- biosci@daresbury.ac.uk Europe, Africa, and Central Asia biosci-help@net.bio.net Americas and the Pacific Rim **** SEARCHING BIOSCI ARCHIVES **** The easiest way to search the BIOSCI archives is to use gopher software and connect over the Internet to net.bio.net, the U.S. BIOSCI computer. We maintain three indexes which are searchable from the main gopher menu on net.bio.net: (1) an index of all BIOSCI postings; (2) an index of individual journal article references from the Table of Contents postings on the BIO-JOURNALS newsgroup; and (3) an index of BIOSCI users including regular mail and e-mail addresses, phone/FAX numbers, research interests, and newsgroup participation. E-mail users can search the BIOSCI archives by using our waismail e-mail server. For instructions send the message help to waismail@net.bio.net. Leave the Subject: line blank (anything entered on the Subject: line is ignored). WAIS software can also be used to search the archives as described in the BIOSCI FAQ (see below). **** BIOSCI FREQUENTLY ASKED QUESTIONS (FAQ) SHEET **** New users of BIOSCI/bionet may want to read the "Frequently Asked Questions" or "FAQ" sheet for BIOSCI. The FAQ provides details on how to participate in these forums and is available for anonymous FTP from net.bio.net [134.172.2.69] in pub/BIOSCI/doc/biosci.FAQ or for retrieval by gopher to net.bio.net, port 70. It may also be requested by sending e-mail to biosci-help@net.bio.net (use plain English for your request). The FAQ is also posted on the first of each month to the newsgroup BIONEWS/bionet.announce immediately following the posting of the BIOSCI information sheet. **** BIOSCI USER ADDRESS DIRECTORY **** Please take this opportunity to add your name and address information to the BIOSCI User Address Database if you have not already done so. Below is the address form that we would like each reader of the BIOSCI/bionet newsgroups to complete and return if you would like to be listed in our database. The database serves as a directory that enables biologists, who are currently using (or even just reading) the BIOSCI newsgroups, to look up e-mail addresses and other information about our users. The address database is reindexed nightly for gopher, WAIS, and waismail access. The raw unindexed data is available for FTP from net.bio.net and is atomized sufficiently to allow import into your local RDBMS should you so desire. Please carefully follow the instructions for completing the form below and return it to either of the following two addresses (whichever is more convenient for you). Thanks in advance for taking the time to complete and return the form. Addresses for returning forms Location Network ----------------------------- -------- ------- biovote@net.bio.net U.S.A. Internet/BITNET biovote@daresbury.ac.uk U.K. JANET MAKING SURE THAT YOUR INFORMATION IS CURRENT This notice will be mailed bimonthly to each newsgroup. You should check our gopher, WAIS source or waismail e-mail server from time-to-time to see if your address information is still up-to-date. Using Gopher to complete the form --------------------------------- If you don't want to use a text editor, you can also use Dan Jacobson's gopher site to fill out the address database form as follows. Otherwise skip this section on gopher and proceed to the instructions for filling out the form below. > To add yourself to the database just point your > gopher client at merlot.gdb.org and select the following: > > --> 15. Searching For Biologists/ > > --> 9. E-mail Addresses of Biosci-Bionet Users/ > > --> 1. Add (or Correct) Your Address to the BIOSCI User Address > Data.. > > > And fill out the form. or Rob Harper's gopher site in Europe as follows: > Europeans can point their gopher client at gopher.csc.fi and add their > information to the database. All entries will be mailed directly to > Dave for incorporation in a wais source. > > The path to the questionare is as follows. > > ---> 10. Finnish EMBnet BioBox/ > > ---> 8. FAQ Files/ > > FAQ Files > > 1. EMBnet: Information. > 2. EMBnet: Internet resources guide. > 3. A Biologist's Guide to Internet Resources/ > 4. All FAQs (Frequently Asked Questions) Searches and Archives/ > --->5. Bionauts Address Database (questionaire) IMPORTANT INSTRUCTIONS - PLEASE READ CAREFULLY Please enter all responses after the : on each line, leaving one (1) blank space after the : (i.e., before the start of your text). Please do NOT extend your responses past the end of each line (80 characters) or alter any of the field identifiers such as "first name: ". Several lines are provided at the end of the form for comments, but, please adhere to the line length restriction. On the date: line, please enter the date in the DD-MM-YY format, e.g., 05-05-93 for 5 May 1993. This line will tell others when the information was last updated. Please be sure to include the 0's for single digit days or months, e.g., 05-05-93, not 5-5-93. Note that the "e-mail network: " line below is for specifying, e.g., "Internet," "BITNET," "EARN," "JANET," or whatever other network that your computer may be on. If you are uncertain about any field, please feel free to leave it blank, but please DO NOT DELETE the field identifier from the form! In the first field below, "New information or Update ...", please enter "N" if this is the first time that you have registered in the directory or "U" if you are correcting a listing that you sent to us previously. The comment: lines may be used for anything that you like but PLEASE DO NOT DELETE THEM FROM THE FORM OR ALTER THEM. One suggested use is to list the names of the newsgroups in which you participate. Please use the MAILING LIST name (see below - the latest version of the list can be requested from biosci@net.bio.net) instead of the USENET name even if you don't participate by e-mail. WAIS might get confused by the periods in the USENET names. This allows one to retrieve via WAIS or waismail the list of participants in a particular group. For example: comment: ARABIDOPSIS PLANT-BIOLOGY BIONEWS On the comment: lines use these names below ---- NOT the USENET names below MAILING LIST NAME USENET Newsgroup Name ----------------- --------------------- ACEDB-SOFT bionet.software.acedb AGEING bionet.molbio.ageing AGROFORESTRY bionet.agroforestry ARABIDOPSIS bionet.genome.arabidopsis BIOFORUM bionet.general BIO-INFORMATION-THEORY bionet.info-theory BIONAUTS bionet.users.addresses BIONEWS bionet.announce BIO-JOURNALS bionet.journals.contents BIO-MATRIX bionet.molbio.bio-matrix BIO-SOFTWARE bionet.software CHROMOSOMES bionet.genome.chromosomes COMPUTATIONAL-BIOLOGY bionet.biology.computational DROSOPHILA bionet.drosophila EMBL-DATABANK bionet.molbio.embldatabank EMPLOYMENT bionet.jobs GDB bionet.molbio.gdb GENBANK-BB bionet.molbio.genbank GENETIC-LINKAGE bionet.molbio.gene-linkage HIV-MOLECULAR-BIOLOGY bionet.molbio.hiv HUMAN-GENOME-PROGRAM bionet.molbio.genome-program IMMUNOLOGY bionet.immunology INFO-GCG bionet.software.gcg JOURNAL-NOTES bionet.journals.note METHODS-AND-REAGENTS bionet.molbio.methds-reagnts MOLECULAR-EVOLUTION bionet.molbio.evolution NEUROSCIENCE bionet.neuroscience N2-FIXATION bionet.biology.n2-fixation PHOTOSYNTHESIS bionet.photosynthesis PLANT-BIOLOGY bionet.plants POPULATION-BIOLOGY bionet.population-bio PROTEIN-ANALYSIS bionet.molbio.proteins PROTEIN-CRYSTALLOGRAPHY bionet.xtallography RAPD bionet.molbio.rapd SCIENCE-RESOURCES bionet.sci-resources TROPICAL-BIOLOGY bionet.biology.tropical VIROLOGY bionet.virology WOMEN-IN-BIOLOGY bionet.women-in-bio YEAST bionet.molbio.yeast Listing newsgroups on the comment: line is optional, of course. Thanks again for your cooperation! --------------- please cut here and return portion below --------------- New information or Update to old record (enter N or U): date (DD-MM-YY): first name: middle initial: family name: job title: e-mail address: e-mail network: phone number: FAX number: institution: address1: address2: address3: city: state/province: country: postal code: research interest: research interest: comment: comment: comment: comment: comment: From owner-structural-nmr@net.bio.net Wed May 25 23:00:00 1994 Path: biosci!daresbury!not-for-mail From: terez@tome.cbs.univ-montp1.fr (Therese Malliavin) Newsgroups: bionet.structural-nmr Subject: References needed Date: 26 May 1994 08:18:47 +0100 Lines: 26 Sender: daemon@mserv1.dl.ac.uk Distribution: bionet Message-ID: <2s1ign$lpe@mserv1.dl.ac.uk> Original-To: str-nmr@dl.ac.uk Hello Netters, I am looking for bibliographic information. I know that methods were proposed for nOe build-up curve fitting by using polynomials. I need precise references about these methods. I know that some ones were published by R. Boelens and others in J.Mag.Res. Could anyone help me? Thank you in advance. ********************************************************************* _/_/_/ _/_/_/ _/_/_/ Therese E. Malliavin _/ _/ _/ _/ Centre de Biochimie Structurale _/ _/_/_/ _/_/_/ Faculte de Pharmacie _/ _/ _/ _/ 15, av. Ch. Flahault _/_/_/ _/_/_/ _/_/_/ F-34060 Montpellier terez@cbs.univ-montp1.fr Tel: (33) 67 04 38 42 Fax: (33) 67 52 96 23 ********************************************************************** From owner-structural-nmr@net.bio.net Thu May 26 23:00:00 1994 Path: biosci!SCRIPPS.EDU!case From: case@SCRIPPS.EDU (David Case) Newsgroups: bionet.structural-nmr Subject: Re: NMR R and R6th quality factors Date: 27 May 1994 07:34:37 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 40 Sender: daemon@net.bio.net Distribution: world Message-ID: <9405271434.AA19712@lanczos.Scripps.EDU> NNTP-Posting-Host: net.bio.net > laub@algol.rm.fccc.edu (Paul Laub) wrote: > > I am interested in assessing the quality of protein structures generated > using NOE-derived restraints by back calculation. I am calculating both the > R and R**(1/6) factor and am wondering whether anyone has an opinion as to > which is more representative of quality. I am aware of a paper from TL James' > lab suggesting that R**(1/6) is better (PD Thomas et al. PNAS 88: 1237 (1991)). > > In most cases, the ranking of quality factors from different > structures or from different stages in refinement are the same for each > quality factor. However I have found exceptions. > It would be interesting to analyze the exceptions. Gonzalez et al JMR 91:659(1991) also report that most indices refine in parallel, and that has been my experience, too. But it seems to me that none of the r-factors usually used are effective in preventing small observed NOE peaks from "disappearing" during refinement (i.e. by allowing distances corresponding to weak NOE cross-peaks to become quite large.) About the best I've been able to do is to include both (conservative) distance bounds (as in conventional refinements) and an R-factor-like quantity. This has the qualitative effect of adding lower bounds to many distances. Peter Wright and I discuss some of this (pretty sketchy) in our article in "NMR in Proteins", edited by Clore and Gronenborn (MacMillen, 1993). ...dave case ===================================================================== David A. Case | internet: case@scripps.edu Dept. of Molecular Biology, MB1 | bitnet: case%scripps.edu@sdsc The Scripps Research Institute | fax: 619-554-3789 10666 N. Torrey Pines Rd. | phone: 619-554-9768 La Jolla CA 92037 USA | ===================================================================== From owner-structural-nmr@net.bio.net Thu May 26 23:00:00 1994 Path: biosci!CHEM.PURDUE.EDU!David From: David@CHEM.PURDUE.EDU Newsgroups: bionet.structural-nmr Subject: Re: NMR R and R6th quality factors Date: 27 May 1994 14:41:02 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 64 Sender: daemon@net.bio.net Distribution: world Message-ID: <199405272140.OAA09449@net.bio.net> NNTP-Posting-Host: net.bio.net Commenting on: >Hello, > > I am interested in assessing the quality of protein structures generated >using NOE-derived restraints by back calculation. I am calculating both the >R and R**(1/6) factor and am wondering whether anyone has an opinion as to >which is more representative of quality. I am aware of a paper from TL James' >lab suggesting that R**(1/6) is better (PD Thomas et al. PNAS 88: 1237 (1991)). > > In most cases, the ranking of quality factors from different >structures or from different stages in refinement are the same for each >quality factor. However I have found exceptions. > > What are your opinions ? > >Sincerely, Paul Laub >-- >Paul Laub/Institute for Cancer Research/7701 Burholme Ave./Phila. PA 19111 USA >sisyphus@astatine.fold.fccc.edu (215)728-2840 and -2748 fax: 728-3574 >"He only earns his freedom and existence who daily conquers them anew." -Goethe Commenting on various R-factors. In our lab we also like to compare a percent RMS value - root mean square value of the percentage error between the calculated (v_calc) and the experimental NOE volumes (v_exp), v_calc - v_exp)/v_*, where v_* is either v_calc or v_exp. This way a 20% error in the volume of a small peak is equally weighted with a 20% error of a large peak.This gives a much better weighting of the importance to the weak, long distance peaks than is possible with a simple R factor which is overwhelmed by the geminal-type short distances. Another advantage is that it gives a better feel for the errors in the data. What does a 6% error in R**(1/6) mean? Also by propagation of errors a 24% error in our %RMS volume implies an ca. 24%/6 or 4% error in the distances. We use these %RMS errors to set the errors in the flatwell NOE distance potential for restrained MD during our iterative MORASS hybrid matrix refinement. As Dave Case points out, generally all measures, RMS volume error, R-factor and R**(1/6) give parallel numbers for quality of data if everything is in order. The RMS volume error is the most sensitive one however since it is pretty easy to get some 200% errors in very small peaks. At some point we should get together to decide what are the best measures of accuracy and precision in NMR structures. __________________________________________________________________ Dr. David Gorenstein Chemistry Department Purdue University W. Lafayette, IN 47907 USA Internet: David@chem.purdue.edu Office: (317) 494 7851 FAX: (317) 494 0239 From owner-structural-nmr@net.bio.net Thu May 26 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!EU.net!uunet!spool.mu.edu!uwm.edu!msuinfo!netnews.upenn.edu!taurus.fccc.edu!laub From: laub@algol.rm.fccc.edu (Paul Laub) Newsgroups: bionet.structural-nmr Subject: NMR R and R6th quality factors Date: 27 May 1994 12:19:30 GMT Organization: Fox Chase Cancer Center, Philadelphia, PA Lines: 19 Message-ID: <2s4ogi$h47@taurus.fccc.edu> NNTP-Posting-Host: algol.rm.fccc.edu Hello, I am interested in assessing the quality of protein structures generated using NOE-derived restraints by back calculation. I am calculating both the R and R**(1/6) factor and am wondering whether anyone has an opinion as to which is more representative of quality. I am aware of a paper from TL James' lab suggesting that R**(1/6) is better (PD Thomas et al. PNAS 88: 1237 (1991)). In most cases, the ranking of quality factors from different structures or from different stages in refinement are the same for each quality factor. However I have found exceptions. What are your opinions ? Sincerely, Paul Laub -- Paul Laub/Institute for Cancer Research/7701 Burholme Ave./Phila. PA 19111 USA sisyphus@astatine.fold.fccc.edu (215)728-2840 and -2748 fax: 728-3574 "He only earns his freedom and existence who daily conquers them anew." -Goethe From owner-structural-nmr@net.bio.net Sat May 28 23:00:00 1994 Path: biosci!bloom-beacon.mit.edu!spool.mu.edu!howland.reston.ans.net!europa.eng.gtefsd.com!newsxfer.itd.umich.edu!nntp.cs.ubc.ca!unixg.ubc.ca!bear!logan From: logan@bear.ubc.ca (Logan Donaldson) Newsgroups: bionet.structural-nmr Subject: xplor 3.1 -> IRIX 5.2 Date: 29 May 1994 04:59:53 GMT Organization: University of British Columbia, Vancouver, B.C., Canada Lines: 13 Sender: logan@bear (Logan Donaldson) Distribution: world Message-ID: <2s97g9$g6e@nntp.ucs.ubc.ca> NNTP-Posting-Host: bear.biochem.ubc.ca Has anyone in NMRland compiled xplor 3.1 on IRIX 5.2? We currently have all of the fortran libraries and development libraries installed with the exception of the Power Fortran release. When the installation script is run, things progress nicely until the libI77_mp library is called. Any suggestions??? Logan Donaldson Department of Biochemistry University of British Columbia logan@otter.biochem.ubc.ca Vancouver BC, V6T 1Z3 From owner-structural-nmr@net.bio.net Sat May 28 23:00:00 1994 Path: biosci!bloom-beacon.mit.edu!spool.mu.edu!howland.reston.ans.net!pipex!sunic!news.uni-c.dk!unidhp.uni-c.dk!carlmk From: carlmk@unidhp.uni-c.dk (Mogens Kjaer) Newsgroups: bionet.structural-nmr Subject: Re: xplor 3.1 -> IRIX 5.2 Date: 29 May 1994 12:25:32 GMT Organization: News Server at UNI-C, Danish Computing Centre for Research and Education. Lines: 34 Distribution: world Message-ID: <2sa1js$dor@news.uni-c.dk> References: <2s97g9$g6e@nntp.ucs.ubc.ca> NNTP-Posting-Host: unidhp.uni-c.dk X-Newsreader: TIN [version 1.2 PL1] Logan Donaldson (logan@bear.ubc.ca) wrote: : Has anyone in NMRland compiled xplor 3.1 on IRIX 5.2? : We currently have all of the fortran libraries and development libraries : installed with the exception of the Power Fortran release. When the installation : script is run, things progress nicely until the libI77_mp library is called. : Any suggestions??? : Logan Donaldson : Department of Biochemistry : University of British Columbia logan@otter.biochem.ubc.ca : Vancouver BC, V6T 1Z3 The solution is probably the same as for 4.0.5 running without the Power stuff. Remove everything that has to do with the power switches during the compilation and linking. Anyway, running X-PLOR parallelized is a very, very bad thing. If your running multiple MD simulations, e.g. to calculate a structure from a set of restraints, you are much better off by doing two calculations simultanously using different starting conditions than to run one job parallelized at a time. Mogens -- Mogens Kjaer Pronto Software - Development and Distribution Carlsberg Research Center Gamle Carlsberg Vej 10 DK-2500 Valby Denmark Phone: + 45 33 27 53 25 Fax: + 45 33 27 47 08 Email: carlmk@unidhp.uni-c.dk From owner-structural-nmr@net.bio.net Sat May 28 23:00:00 1994 Path: biosci!daresbury!not-for-mail From: Serge Iourine Newsgroups: bionet.structural-nmr Subject: NMR --> PostScript ? Date: 29 May 1994 08:14:13 +0100 Lines: 10 Sender: daemon@mserv1.dl.ac.uk Distribution: bionet Message-ID: <2s9fc5$7qs@mserv1.dl.ac.uk> Original-To: str-nmr@dl.ac.uk Dear netters, Does anybody know whether it's possible to convert output data from NMR spectrometer (we are using Varian 300MHz) into the PostScript format? Regards, S.Iourine Dept. of Chemistry U. of Natal, Durban From owner-structural-nmr@net.bio.net Sun May 29 23:00:00 1994 Path: biosci!daresbury!not-for-mail From: thep@risc1.lrm.fi.cnr.it (Pornthep Sompornpisut) Newsgroups: bionet.structural-nmr Subject: problem on the residue's name in the library Date: 30 May 1994 19:31:23 +0100 Lines: 33 Sender: daemon@mserv1.dl.ac.uk Distribution: bionet Message-ID: <2sdbdr$abd@mserv1.dl.ac.uk> Original-To: str-nmr@dl.ac.uk Dear bio-nmr -ers Please forgive me if this message interupts you. Since I have a doubt about some typical data sets from the DG (Distance Geometry) calculation. The program provides the library in which contains several data sets. I was looking at it and I found that some residues which are different in term of the charges, for example ARG/ARG+ or GLU/GLU- etc., are provided in order to choose as the kind of residue be considering. I afraid that if I define the wrong one in the input file (for instance, instead of ARG+ , I use ARG), it will be a big problem. And then I will get the different structure. Therefore, I have some question are; - Are there any serious problem if I define the wrong residue's name which is not corresponded to the real system? (As far as I looked the library, there are just a few differences in term of proton donor/acceptor.) - These problem can be corrected by the molecular dynamic calculation. Is that right? - Are there any pKa value of the amino acids which consist of labile proton? How can I get them? or, does anybody have this value in hand? Please inform me. I will be appreciate if I get any response. Thank you very much Pornthep From owner-structural-nmr@net.bio.net Mon May 30 23:00:00 1994 Newsgroups: bionet.structural-nmr Path: biosci!daresbury!trane.uninett.no!sunic!news.lth.se!news.lu.se!heimdall!bryan From: bryan@fkem2.lth.se (Bryan Finn) Subject: Re: xplor 3.1 -> IRIX 5.2 Message-ID: <1994May31.114015.3186@nomina.lu.se> Sender: news@nomina.lu.se (USENET News System) Nntp-Posting-Host: heimdall.fkem2.lth.se Reply-To: bryan@fkem2.lth.se Organization: Physical Chemistry 2, Lund University, Sweden References: <2sa1js$dor@news.uni-c.dk> Date: Tue, 31 May 1994 11:40:15 GMT Lines: 30 In article dor@news.uni-c.dk, carlmk@unidhp.uni-c.dk (Mogens Kjaer) writes: > ... Anyway, running X-PLOR parallelized is > a very, very bad thing. If your running multiple MD simulations, e.g. > to calculate a structure from a set of restraints, you are much better > off by doing two calculations simultanously using different starting > conditions than to run one job parallelized at a time. > Okay, I'll bite, why is running Xplor parallelized a very, very bad thing? Bryan --- ___________________________________________________________________ | | | Dr. Bryan Finn | | Department of Physical Chemistry 2 Tel: +46-46-108254 | | Chemical Center Fax: +46-46-104543 | | University of Lund | | POB 124 | | S-221 00 Lund Sweden e-mail: bryan@freja.fkem2.lth.se | |_________________________________________________________________| | | | "It is unworthy of excellent men to lose hours like slaves in | | the labor of calculation." -- Leibniz | |_________________________________________________________________| From owner-structural-nmr@net.bio.net Mon May 30 23:00:00 1994 Path: biosci!daresbury!trane.uninett.no!sunic!news.uni-c.dk!unidhp.uni-c.dk!carlmk From: carlmk@unidhp.uni-c.dk (Mogens Kjaer) Newsgroups: bionet.structural-nmr Subject: Re: xplor 3.1 -> IRIX 5.2 Date: 31 May 1994 12:51:29 GMT Organization: News Server at UNI-C, Danish Computing Centre for Research and Education. Lines: 65 Message-ID: <2sfbsh$kv8@news.uni-c.dk> References: <2sa1js$dor@news.uni-c.dk> <1994May31.114015.3186@nomina.lu.se> NNTP-Posting-Host: unidhp.uni-c.dk X-Newsreader: TIN [version 1.2 PL1] Bryan Finn (bryan@fkem2.lth.se) wrote: : In article dor@news.uni-c.dk, carlmk@unidhp.uni-c.dk (Mogens Kjaer) writes: : > ... Anyway, running X-PLOR parallelized is : > a very, very bad thing. If your running multiple MD simulations, e.g. : > to calculate a structure from a set of restraints, you are much better : > off by doing two calculations simultanously using different starting : > conditions than to run one job parallelized at a time. : > : : Okay, I'll bite, why is running Xplor parallelized a very, very bad thing? : Bryan : --- : ___________________________________________________________________ : | | : | Dr. Bryan Finn | : | Department of Physical Chemistry 2 Tel: +46-46-108254 | : | Chemical Center Fax: +46-46-104543 | : | University of Lund | : | POB 124 | : | S-221 00 Lund Sweden e-mail: bryan@freja.fkem2.lth.se | : |_________________________________________________________________| : | | : | "It is unworthy of excellent men to lose hours like slaves in | : | the labor of calculation." -- Leibniz | : |_________________________________________________________________| Hi Bryan! This is why running anything parallelized is BAD: When you run a program in parallel mode, not all of the code will run in parallel. Only some specific loops may be parallelized. The time spent switching between serial and parallel mode, including syncronization, etc., can sometimes be more than what you gain from running in parallel. On some machines, the CPU's not used when running in serial mode is not released for other jobs, but are actually running active loops. Also, when XPLOR is used for structure calculations, you normally make 20-40 runs with different initial value of the random number seed. You get much more out of a parallel machine by doing these jobs in parallel, 2 or 4 at a time, depending on how many processors you have, and what else is running on the machine. Consider using NQS to balance the load on the processors. I've heard of people having a FORTRAN program running for two days in serial mode; compiling it in parallel mode made it run for four days instead! That is just my 2 oerer... (US: two cents... :-) ) -- Mogens Kjaer Pronto Software - Development and Distribution Carlsberg Research Center Gamle Carlsberg Vej 10 DK-2500 Valby Denmark Phone: + 45 33 27 53 25 Fax: + 45 33 27 47 08 Email: carlmk@unidhp.uni-c.dk From owner-structural-nmr@net.bio.net Mon May 30 23:00:00 1994 Path: biosci!NODDY.CM.UTEXAS.EDU!dave From: dave@NODDY.CM.UTEXAS.EDU (david w. hoffman) Newsgroups: bionet.structural-nmr Subject: (none) Date: 31 May 1994 09:28:55 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 13 Sender: daemon@net.bio.net Distribution: world Message-ID: <9405311629.AA00361@noddy.cm.utexas.edu> NNTP-Posting-Host: net.bio.net Dear NMR-netters, I am looking for software that does linear prediction, for use in Fourier transforming under-digitized 3D data. I am hoping to find software of the non-commercial (free) variety. Thanks. Dave Hoffman Dept of Chemistry and Biochemistry Univ. of Texas at Austin dave@noddy.cm.utexas.edu From owner-structural-nmr@net.bio.net Mon May 30 23:00:00 1994 Path: biosci!PINES1.LJCRF.EDU!elya From: elya@PINES1.LJCRF.EDU (Elya S. Kurktchi) Newsgroups: bionet.structural-nmr Subject: What Macintosh to use? Date: 31 May 1994 10:56:31 -0700 Organization: BIOSCI International Newsgroups for Molecular Biology Lines: 22 Sender: daemon@net.bio.net Distribution: world Message-ID: <9405311750.AA22679@pines1.ljcrf.edu> NNTP-Posting-Host: net.bio.net Hi. What is a good, cheap Macintosh that I can buy that will allow me do relatively simple NMR calculations? This is going to be for my own personal use, but if it is going to cost $4K or more, I can get a sparc5 (which I am to do the calculations). Basically, I am looking for a MAC that can be upgraded to at least 64MB of RAM and is fast enough to run Word, FrameMaker, and some NMR apps. Thanks in advance. Elya. ---------------------------------------------------------------------------- | Elya S. Kurktchi Scientific Computing Center | | Network Systems Manager La Jolla Cancer Research Foundation | | Email: elya@ljcrf.edu 10901 North Torrey Pines Road | | Phone: (619) 455-6480 x631 La Jolla, CA 92037 | | Fax : (619) 450-3432 "X-Ray Protein Crystallography" | ---------------------------------------------------------------------------- From owner-structural-nmr@net.bio.net Tue May 31 23:00:00 1994 Path: biosci!agate!spool.mu.edu!howland.reston.ans.net!EU.net!sunic!news.funet.fi!ousrvr.oulu.fi!eha From: eha@phoenix.oulu.fi (Esa Haapaniemi) Newsgroups: bionet.structural-nmr Subject: Re: What Macintosh to use? Date: 1 Jun 1994 06:49:24 GMT Organization: University of Oulu, Finland Lines: 31 Distribution: world Message-ID: <2shb1k$ick@ousrvr.oulu.fi> References: <9405311750.AA22679@pines1.ljcrf.edu> NNTP-Posting-Host: zombie.oulu.fi Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Newsreader: TIN [version 1.2 PL1] : Hi. What is a good, cheap Macintosh that I can buy that will allow : me do relatively simple NMR calculations? This is going to be for : my own personal use, but if it is going to cost $4K or more, I can : get a sparc5 (which I am to do the calculations). : Basically, I am looking for a MAC that can be upgraded to at least : 64MB of RAM and is fast enough to run Word, FrameMaker, and some : NMR apps. I'm serious when I advice you to look at Amiga4000/40 w. Emplant. It is upgradeable to 256 M of memory and all usable on Mac side (Emplant emulates Quadra 900 in that configuration, except maths is some 113 % of Quadra). For faster machine you could upgrade the Amiga to 50 MHz 040, or hopefully some RISC, and then it could be some 2x faster than ANY Mac. Still for best NMR program on PC (personal computer, no SGI or alike included) is Perch. That is currently available only on IBM compatibles, but versions for Amiga and SGI are under work. And ofcource the manufacturer of the Emplant has promiced to have his 486 emulation (66 MHz 486 in A4000/40) to be ready during summertime. Hmm... ;) Esa Haapaniemi University of Oulu Department of Chemistry Finland P.S. I don't have other connection to Perch except we have been "betatesting" it for several years !-)