Hi Ian, The halting of relax and restarting using the scripting mode was really only backup system. This is not a clean way of running the model-free protocol as the data from the currently running model-free protocol iteration are lost. All data from previous iterations are stored in the saved 'results.bz2' files. When re-running the protocol, it will first check if any results files are present and if so, then it will attempt to load the results files from previous iterations and then restart the calculations for each diffusion tensor. This may result is slightly different results than if the calculation was not interrupted as converged diffusion models are reoptimised (though this should not change the results significantly).
Note that all of the user interfaces (prompt, script, GUI) use the same code for the analysis. Therefore this resumption behaviour you know from the scripting mode should be active in the GUI mode as well. I haven't tested this so I don't know how well this works. I would recommend saving your state prior to clicking on 'Execute relax', and then load this state when you would like to restart the calculation. Clicking on 'Execute relax' should then run the same analysis code, find the last results files, and continue from there. However I would recommend you use your Mac's "Safe sleep" mode if you would like to turn your Mac off. This will allow you to resume the relax once you turn your computer back on. Regards, Edward On 28 October 2012 13:34, Robertson, Ian <[email protected]> wrote: > Hi Edward, > > I have a question that's an extension of Ravi's. > > I'm using my laptop to run relax and I'd like to be able to halt the > calculation and then return to it later to avoid leaving my computer on the > entire time. If I remember correctly, it was possible for me to do this when > running relax from the command line. However, when I stop and start the > calculation in the GUI it always returns to the beginning of the > calculation. Furthermore, while the calculation is running, the "save" and > "save as" buttons are shaded out. I reviewed the mailing lists but couldn't > see this brought up anywhere. > > I'm running relax 2.1.2 on lion, although I've also noted this issue on > earlier versions of relax. > > Regards, > > Ian > > > On 2012-10-25, at 3:19 PM, Edward d'Auvergne wrote: > > Hi Ravi, > > I will answer your questions below: > > (1) I am running this program on macpro computer with i7 processor and 4 Gb > > run. When the program is running in GUI mode, it slows down the computer. is > > it normal or is there a way to make it faster without slowing down the > > computer? > > > A full model-free analysis - i.e. execution of the full protocol and > not simply running the basic model-free models - can require a long > time to calculate. This is explained in a few places in the relax > manual, for example: > > http://www.nmr-relax.com/manual/d_Auvergne_protocol_script_mode_execution.html > > and: > > http://www.nmr-relax.com/manual/new_protocol_in_GUI.html > > Essentially this is a long calculation which can take up to a couple > of weeks to complete, depending on the system, and you can use Gary > Thompson's multi-processor code to speed things up if you need > (http://www.nmr-relax.com/manual/multi_processor_framework.html). > > Depending on your operating system, you can use priorities to allow > your computer to run at a reasonable speed during an intensive > number-crunching calculation. You can renice processes using, for > example, 'top' in GNU/Linux and Mac OS X systems. In MS Windows you > can use the task manager to reset the priorities. > > > (2) My protein is 152 amino acid long. it seems the program is running for > > last few days and many more models to complete. Usually how long does it > > take to complete a full automated run? > > > As it says in the manual, this is not possible to predict but it can > take up to a couple of weeks. If you are lucky, then only a few days > are required. This really depends on what types of internal dynamics > you have as well as the quality of the R1, R2, and NOE relaxation > data. You can check the quality of the data using the Sebastien > Morin's consistency testing analysis in relax > (http://www.nmr-relax.com/manual/Consistency_testing.html). It might > be worth running the relaxation curve-fitting and NOE calculation > through relax as well to be sure of the error estimates: > > http://www.nmr-relax.com/manual/Relaxation_curve_fitting.html > http://www.nmr-relax.com/manual/Calculating_NOE.html > > These are simply links to the relevant chapters of the relax user > manual. What software did you use to calculate your R1, R2, and NOE > data? My paper: > > d'Auvergne, E. J. and Gooley, P. R. (2008). Optimisation of NMR > dynamic models II. A new methodology for the dual optimisation of the > model-free parameters and the Brownian rotational diffusion tensor. J. > Biomol. NMR, 40(2), 121-133. > (http://dx.doi.org/10.1007/s10858-007-9213-3). > > talks about this as well. See figure 2 for an idea of how long global > convergence of the combined model-free optimisation and model-free > model selection problem can take. > > > (3) what are differences between local_tm models and model-free models? What > > i could see is that local-tm models have tm with it whereas model-free > > models dont! > > > This is again described in the manual: > > http://www.nmr-relax.com/manual/model_free_models.html > > However a much better description is given in some of my published papers: > > d'Auvergne E. J., Gooley P. R. (2007). Set theory formulation of the > model-free problem and the diffusion seeded model-free paradigm. Mol. > Biosyst., 3(7), 483-494. (http://dx.doi.org/10.1039/b702202f). > > and: > > d'Auvergne, E. J. and Gooley, P. R. (2008). Optimisation of NMR > dynamic models II. A new methodology for the dual optimisation of the > model-free parameters and the Brownian rotational diffusion tensor. J. > Biomol. NMR, 40(2), 121-133. > (http://dx.doi.org/10.1007/s10858-007-9213-3). > > Note that these papers also describe the protocol that you are using. > The Lipari-Szabo model-free theory is only part of the problem. On > top of this is the bigger problem of the internal bond motions verses > external global tumbling. A number of protocols have been developed > over time to solve this problem (see the manual and the above papers > for a list of these protocols), and in relax via the GUI you are using > the new model-free protocol that I developed in the group of Prof. > Paul Gooley. You can click on the 'About' button in the GUI for more > information. > > > (4) I want to see the results of local_tm run which seems to be over with > > aic and tm0-tm9 folders created having results.bz2 in it? How to check the > > results, for example, i want to see the overall correlation time of my > > protein, how to see the results.bz2 file and which folder to check? How to > > look for other parameters? > > > You will have to wait until the analysis is finished. To understand > what is possible, you will need to minimally read the two references > given above. But to properly understand the complexity of the > analysis you are performing, you should really read all the relevant > references listed in the citations chapter of the user manual > (http://www.nmr-relax.com/manual/Citations.html, though the citations > here are not properly built so you should see the PDF version of the > manual http://download.gna.org/relax/manual/relax.pdf). > > After reading these, you will note that there is no overall > correlation time for the local tm models - and this is the power of > these models and the key to the start of the protocol I developed. > You will also see that you need to wait until the protocol has > completed to obtain the information that you are after. Until then > relax will not be responsive and you will not be able to do anything. > I would also recommend looking at Figure 1 of my 2008b paper. > > > Since, i am trying use relax for 1st time, these questions might be naive of > > its kind. Sorry for any inconvenience and looking forward to hear from your > > side. > > > This isn't a problem. If you have any other questions or need > pointers to the right documentation, user manual section or reference, > don't hesitate to ask. > > Regards, > > Edward > > _______________________________________________ > relax (http://www.nmr-relax.com) > > This is the relax-users mailing list > [email protected] > > To unsubscribe from this list, get a password > reminder, or change your subscription options, > visit the list information page at > https://mail.gna.org/listinfo/relax-users > > > Ian Robertson, Ph.D. > Professor Malcolm Irving's Lab > Randall Division of Cell & Molecular Biophysics > 3.26 New Hunt's House > Guy's Campus > King's College London > [email protected] > _______________________________________________ relax (http://www.nmr-relax.com) This is the relax-users mailing list [email protected] To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users
