Menu
▾
▴
scipion-users — Mailing list for Scipion users
You can subscribe to this list here.
| 2016 |
Jan
(2) |
Feb
(13) |
Mar
(9) |
Apr
(4) |
May
(5) |
Jun
(2) |
Jul
(8) |
Aug
(3) |
Sep
(25) |
Oct
(7) |
Nov
(49) |
Dec
(15) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2017 |
Jan
(24) |
Feb
(36) |
Mar
(53) |
Apr
(44) |
May
(37) |
Jun
(34) |
Jul
(12) |
Aug
(15) |
Sep
(14) |
Oct
(9) |
Nov
(9) |
Dec
(7) |
| 2018 |
Jan
(16) |
Feb
(9) |
Mar
(27) |
Apr
(39) |
May
(8) |
Jun
(24) |
Jul
(22) |
Aug
(11) |
Sep
(1) |
Oct
|
Nov
|
Dec
|
| 2019 |
Jan
(4) |
Feb
(5) |
Mar
|
Apr
(1) |
May
(21) |
Jun
(13) |
Jul
(31) |
Aug
(22) |
Sep
(9) |
Oct
(19) |
Nov
(24) |
Dec
(12) |
| 2020 |
Jan
(30) |
Feb
(12) |
Mar
(16) |
Apr
(4) |
May
(37) |
Jun
(17) |
Jul
(19) |
Aug
(15) |
Sep
(26) |
Oct
(84) |
Nov
(64) |
Dec
(55) |
| 2021 |
Jan
(18) |
Feb
(58) |
Mar
(26) |
Apr
(88) |
May
(51) |
Jun
(36) |
Jul
(31) |
Aug
(37) |
Sep
(79) |
Oct
(15) |
Nov
(29) |
Dec
(8) |
| 2022 |
Jan
(5) |
Feb
(8) |
Mar
(29) |
Apr
(21) |
May
(11) |
Jun
(11) |
Jul
(18) |
Aug
(16) |
Sep
(6) |
Oct
(10) |
Nov
(23) |
Dec
(1) |
| 2023 |
Jan
(18) |
Feb
|
Mar
(4) |
Apr
|
May
(3) |
Jun
(10) |
Jul
(1) |
Aug
|
Sep
|
Oct
(1) |
Nov
(3) |
Dec
(5) |
| 2024 |
Jan
(2) |
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2025 |
Jan
(1) |
Feb
|
Mar
|
Apr
(5) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Dennis T. <dt...@pr...> - 2017-03-07 17:33:01
|
Using xmipp extract particles to extract particles from MRC aligned frames, the coordinates are imported EMAN *.box files do I have to invert Y or X or both. It doesn't seem to extract real particles. Thanks, Dennis -- Dr. Dennis Thomas Cryo-EM Facility Manager Center for Integrative Proteomics Research Rutgers University Phone: 848-445-5302 |
|
From: Pablo C. <su...@bc...> - 2017-03-07 17:12:33
|
Hi Sergey, I can't see anyway you can get the. pos file. Unless you get into the details of the sql.
I'll check with the team if there is a workaround/hack to this and get back to you.
I'll also annotate this as something useful to implement.
All the best, Pablo.
Scipion team
Activo Mar, 7 Marzo at 1:21 PM
, Sergey Nazarov <ser...@un...> Escrito:
Hi,
I would like to ask if it is possible to save coordinate files in .pos or any other format after Scipion-extract coordinates? Now they saved only in .sql format, but I would like to save them as normal text files for further usage.
Best,
Sergey Nazarov
Biozentrum
|
|
From: Sergey N. <ser...@un...> - 2017-03-07 12:21:41
|
Hi, I would like to ask if it is possible to save coordinate files in .pos or any other format after Scipion-extract coordinates? Now they saved only in .sql format, but I would like to save them as normal text files for further usage. Best, Sergey Nazarov Biozentrum |
|
From: Pablo C. <su...@bc...> - 2017-03-07 09:43:30
|
Hi Dmitry, May I recomend you to switch to release-1.1 branch. Reasons are:
Now devel and release-1.1 are almost similar. Differences are that release has more activity and bugfixes are pushed first to release-1.1 and after a while they end up un devel. Like in this case for the picking bug that is not yet in devel but soon will be.
You will also help us to test the future release while using it.
To do so, follow Jose Miguel steps but switch to release-1.1:
git pull # and this should bring the changes...and release-1.1 branch if not yet local.
git checkout release-1.1 #This switches you to release-1.1 branch
git pull # Again, this might not be needed, but I use to do it to be sure I have the latest changes.
scipion install -j N # replace N by an integer, it means how many CPUs you want to use to compile
scipion install is optimized and will only compile what has changed and EM packages are not reinstall. It should not take long.
Let us know how it goes?
All the best, Pablo.
Activo Lun, 6 Marzo at 5:08 PM
, Dmitry Semchonok <sem...@gm...> Escrito:
Dear colleagues, If there an easy way to update the xmipp - particle picking protocol in Scipion_dev (or any other protocol inside Scipion dev that was bug fixed) without reinstalling everything? Sincerely, Dmitry ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________ scipion-users mailing list sci...@li... https://lists.sourceforge.net/lists/listinfo/scipion-users
|
|
From: Jose M. de la R. T. <del...@gm...> - 2017-03-06 22:52:19
|
Hi Dmitry, If you have installed Scipion using git, it is easy to update with latest changes. You only need to do: *git pull* and this should bring the changes...and then: *scipion install -j N* then, this command should not install everything from scratch, just recompile the changed code. Hope this helps, Jose Miguel. On Mon, Mar 6, 2017 at 5:08 PM, Dmitry A. Semchonok <sem...@gm...> wrote: > Dear colleagues, > > If there an easy way to update the xmipp - particle picking protocol in > Scipion_dev (or any other protocol inside Scipion dev that was bug fixed) > without reinstalling everything? > > Sincerely, > > Dmitry > > > > ------------------------------------------------------------ > ------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users > |
|
From: Dmitry A. S. <sem...@gm...> - 2017-03-06 16:08:47
|
Dear colleagues, If there an easy way to update the xmipp - particle picking protocol in Scipion_dev (or any other protocol inside Scipion dev that was bug fixed) without reinstalling everything? Sincerely, Dmitry |
|
From: Pablo C. <su...@bc...> - 2017-03-06 07:16:58
|
Hi Juha, I had these error before but was fixed (or we believe so). It look suspicious that (no branch) after Scipion. I guess you are using source code right?
Could you try to checkout "release-1.1" branch and install.
We have changed the installation "engine", and it might not be necessary, but there is a chance you might need to review the path for your packages. You can try "scipion config" to see if there are missing variables and run scipion config --update (to update your config with those present in the template but missing in your <HOME>/.config/scipion/scipion.conf
If checking out release-1.1 does not fix your problem....please send us your "project.sqlite" and will have a look to it.
Sorry for any inconvenience,
All the best, Pablo.
Activo Dom, 5 Marzo at 1:51 PM
, Juha Huiskonen <ju...@st...> Escrito:
Hi Jose Miguel,
Here's the complete error message: Scipion (no branch) (2017-02-06) c1fec5b >>>>> python /gpfs0/apps/well/scipion/20170209/scipion/pyworkflow/apps/pw_manager.py Library path does not exists: /usr/local/cuda-7.5/lib64 Library path does not exists: /usr/local/cuda-7.5/lib64 can't compare datetime.datetime to NoneType
Best wishes,
Juha
On Sat, Mar 4, 2017 at 10:09 AM, Jose Miguel de la Rosa Trevin <del...@gm...> wrote:
Hi Juha,
What is the Scipion version you are running? There is a more verbose error message for this issue? Maybe this could give us a better idea where the problem is coming from...and figure out a way to solve it.
Bests,
Jose Miguel
On Fri, Mar 3, 2017 at 8:56 PM, Juha Huiskonen <ju...@st...> wrote:
Hello developers,
I have a problem, I can't open my project. Instead an error is printed:
can't compare datetime.datetime to datetime.date
Is there any way to fix this? I tried import the project to create a copy of it, but I still get the same error when I try to open it.
Best wishes,
Juha
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________
scipion-users mailing list
sci...@li...
https://lists.sourceforge.net/lists/listinfo/scipion-users
|
|
From: Juha H. <ju...@st...> - 2017-03-05 12:51:14
|
Hi Jose Miguel, Here's the complete error message: Scipion (no branch) (2017-02-06) c1fec5b >>>>> python /gpfs0/apps/well/scipion/20170209/scipion/pyworkflow/apps/pw_manager.py Library path does not exists: /usr/local/cuda-7.5/lib64 Library path does not exists: /usr/local/cuda-7.5/lib64 can't compare datetime.datetime to NoneType Best wishes, Juha On Sat, Mar 4, 2017 at 10:09 AM, Jose Miguel de la Rosa Trevin < del...@gm...> wrote: > Hi Juha, > > What is the Scipion version you are running? There is a more verbose error > message for this issue? Maybe this could give us a better idea where the > problem is coming from...and figure out a way to solve it. > > Bests, > Jose Miguel > > > On Fri, Mar 3, 2017 at 8:56 PM, Juha Huiskonen <ju...@st...> > wrote: > >> Hello developers, >> >> I have a problem, I can't open my project. Instead an error is printed: >> can't compare datetime.datetime to datetime.date >> >> Is there any way to fix this? I tried import the project to create a copy >> of it, but I still get the same error when I try to open it. >> >> Best wishes, >> Juha >> >> >> ------------------------------------------------------------ >> ------------------ >> Check out the vibrant tech community on one of the world's most >> engaging tech sites, SlashDot.org! http://sdm.link/slashdot >> _______________________________________________ >> scipion-users mailing list >> sci...@li... >> https://lists.sourceforge.net/lists/listinfo/scipion-users >> >> > |
|
From: Jose M. de la R. T. <del...@gm...> - 2017-03-04 10:09:13
|
Hi Juha, What is the Scipion version you are running? There is a more verbose error message for this issue? Maybe this could give us a better idea where the problem is coming from...and figure out a way to solve it. Bests, Jose Miguel On Fri, Mar 3, 2017 at 8:56 PM, Juha Huiskonen <ju...@st...> wrote: > Hello developers, > > I have a problem, I can't open my project. Instead an error is printed: > can't compare datetime.datetime to datetime.date > > Is there any way to fix this? I tried import the project to create a copy > of it, but I still get the same error when I try to open it. > > Best wishes, > Juha > > > ------------------------------------------------------------ > ------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users > > |
|
From: Jose M. de la R. T. <del...@gm...> - 2017-03-04 10:03:27
|
Hi Juha, Thanks for your feedback. Could you do a small experiment to detect if the exported EMX file is correct? Could you import back the EMX file into Scipion and try to reconstruct only with Relion (there is a protocol for that) and check if it makes sense or not. Take into account that the map probably will not be exactly the same that originally came out from Relion, since it is filtered. Anyway we will also check about this issue. Bests, Jose Miguel On Fri, Mar 3, 2017 at 9:46 AM, Juha Huiskonen <ju...@st...> wrote: > Hello > > I was wondering if the EMX export should work and if others have managed > to use it with cryoSparc? > > I exported my dataset and imported EMX files to cryoSPARC after running > Relion 3D refinement through Scipion. Import to CryoSparc seemed to work > (it created a gallery of the images) but results from the first ab initio > run were nonsensical. > > Next I tested importing MRC files and a corresponding STAR file (manually > converting the same particle stacks to MRC and providing them together with > a corresponding Relion STAR file from the refinement run). After the same > run in cryoSparc, the results were as expected. > > Any ideas? Apologies if this is likely to be an issue on the cryoSPARC > side, in this case I would naturally contact those developers. > > Best wishes, > Juha > > > > ------------------------------------------------------------ > ------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users > > |
|
From: Juha H. <ju...@st...> - 2017-03-03 19:57:12
|
Hello developers, I have a problem, I can't open my project. Instead an error is printed: can't compare datetime.datetime to datetime.date Is there any way to fix this? I tried import the project to create a copy of it, but I still get the same error when I try to open it. Best wishes, Juha |
|
From: Juha H. <ju...@st...> - 2017-03-03 09:09:28
|
Hello I was wondering if the EMX export should work and if others have managed to use it with cryoSparc? I exported my dataset and imported EMX files to cryoSPARC after running Relion 3D refinement through Scipion. Import to CryoSparc seemed to work (it created a gallery of the images) but results from the first ab initio run were nonsensical. Next I tested importing MRC files and a corresponding STAR file (manually converting the same particle stacks to MRC and providing them together with a corresponding Relion STAR file from the refinement run). After the same run in cryoSparc, the results were as expected. Any ideas? Apologies if this is likely to be an issue on the cryoSPARC side, in this case I would naturally contact those developers. Best wishes, Juha |
|
From: Carlos O. S. <co...@cn...> - 2017-03-02 06:21:07
|
Dear Dennis, you may use the create subset tool (at the bottom of the left menu). The big set would be the original, full set of unbinned images, the small subset would be the ones coming from your 2D classification. This tool will create the corresponding subset of large images. Kind regards, Carlos Oscar On 02/03/2017 2:17, Dennis Thomas wrote: > Hi > > I use binned and filtered images for align cl2d and classify cl2d. > Then disable some classes and create a subset of "good" particles. How > do I apply this to the unbinned images? > > Dennis > > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > > > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users -- ------------------------------------------------------------------------ Carlos Oscar Sánchez Sorzano e-mail: co...@cn... Biocomputing unit http://biocomp.cnb.csic.es National Center of Biotechnology (CSIC) c/Darwin, 3 Campus Universidad Autónoma (Cantoblanco) Tlf: 34-91-585 4510 28049 MADRID (SPAIN) Fax: 34-91-585 4506 ------------------------------------------------------------------------ |
|
From: Pablo C. <su...@bc...> - 2017-03-02 03:24:43
|
Hi Dmitry,
This is already fixed in release-1.1 branch which will soon become the release. I'll push changes into dev today, so it will be fixed too.
Thanks for reporting.
Activo Mie, 1 Marzo at 7:15 PM
, Dmitry Semchonok <sem...@gm...> Escrito:
Dear colleagues,
I think I found 1 bug in the particle picking protocol in the Scipion dev
version - sometimes the shift+click as a delete key just does not work.
Sincerely,
Dmitry
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
scipion-users mailing list
sci...@li...
https://lists.sourceforge.net/lists/listinfo/scipion-users
|
|
From: Dennis T. <dt...@pr...> - 2017-03-02 01:17:51
|
Hi I use binned and filtered images for align cl2d and classify cl2d. Then disable some classes and create a subset of "good" particles. How do I apply this to the unbinned images? Dennis -- Dr. Dennis Thomas Cryo-EM Facility Manager Center for Integrative Proteomics Research Rutgers University Phone: 848-445-5302 |
|
From: Dmitry A. S. <sem...@gm...> - 2017-03-01 18:15:31
|
Dear colleagues, I think I found 1 bug in the particle picking protocol in the Scipion dev version - sometimes the shift+click as a delete key just does not work. Sincerely, Dmitry |
|
From: Carlos O. S. <co...@cn...> - 2017-03-01 16:36:28
|
Dear Dmitry, you may use 2D -> Align -> Align with CL2D Kind regards, Carlos Oscar On 03/01/17 11:17, Dmitry A. Semchonok wrote: > Dear colleagues, > > Is there already a protocol that can centered the picked particles? > > Sincerely, > Dmitry > > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users > -- ------------------------------------------------------------------------ Carlos Oscar Sánchez Sorzano e-mail: co...@cn... Biocomputing unit http://biocomp.cnb.csic.es National Center of Biotechnology (CSIC) c/Darwin, 3 Campus Universidad Autónoma (Cantoblanco) Tlf: 34-91-585 4510 28049 MADRID (SPAIN) Fax: 34-91-585 4506 ------------------------------------------------------------------------ |
|
From: Dennis T. <dt...@pr...> - 2017-03-01 16:20:07
|
Hi, When I reach the point of doing initial reconstructions after aligning with xmipp cl2d and classifiying with xmipp cl2d and select "good" particles. I get failures when running EMAN easy refine and Xmipp projection matching. Below are the error messages for each: using xmipp projection matching it fails with the following error : Traceback (most recent call last): 00467: File "/opt/scipion/pyworkflow/protocol/protocol.py", line 167, in run 00468: self._run() 00469: File "/opt/scipion/pyworkflow/protocol/protocol.py", line 211, in _run 00470: resultFiles = self._runFunc() 00471: File "/opt/scipion/pyworkflow/protocol/protocol.py", line 207, in _runFunc 00472: return self._func(*self._args) 00473: File "/opt/scipion/pyworkflow/em/packages/xmipp3/protocol_projmatch/protocol_projmatch.py", line 202, in filterVolumeStep 00474: runFilterVolumeStep(self, iterN, refN, constantToAddToFiltration, **kwargs) 00475: File "/opt/scipion/pyworkflow/em/packages/xmipp3/protocol_projmatch/projmatch_steps.py", line 665, in runFilterVolumeStep 00476: if (filterInPxAt > 0.5): 00477: UnboundLocalError: local variable 'filterInPxAt' referenced before assignment 00478: Protocol failed: local variable 'filterInPxAt' referenced before assignment 00479: FAILED: filterVolumeStep, step 18 00480: 2017-03-01 09:38:58.098115 00481: Cleaning temporarly files.... 00482: ------------------- PROTOCOL FAILED (DONE 18/36) Using EMAN2 easy process I get the following errors 01507: executing e2simmx.py refine_01/projections_02_even.hdf refine_01/projections_02_even.hdf refine_01/proj_simmx_02_even.hdf --shrink=2 --align=rotate_translate_flip --aligncmp=sqeuclidean:normto=1 --cmp=sqeuclidean --saveali --force --verbose=-1 --parallel=thread:30 01508: in check 01509: Matrix size 76 x 76 into 96 regions -> 11 x 8 blocks -> 88 subprocesses (6 x 9 each) 01510: 87/88 01511: 88 tasks submitted 01512: 01513: All simmx tasks complete 01514: Cleaning up thread server. Please wait. 01515: Finding 24 dissimilar classification centers 01516: Sort references 01517: Associating references with centers 01518: Averaging each center 01519: Traceback (most recent call last): 01520: File "/opt/scipion/software/em/eman2.12/bin/e2simmx2stage.py", line 235, in <module> 01521: main() 01522: File "/opt/scipion/software/em/eman2.12/bin/e2simmx2stage.py", line 147, in main 01523: avg=EMData(args[0],i[0]) 01524: IndexError: list index out of range 01525: Error running: e2simmx2stage.py refine_01/projections_02_even.hdf sets/inputSet_even.lst refine_01/simmx_02_even.hdf refine_01/proj_simmx_02_even.hdf refine_01/proj_stg1_02_even.hdf refine_01/simmx_stg1_02_even.hdf --saveali --cmp frc:snrweight=1:maxres=14.0 --align rotate_translate_flip --aligncmp ccc --ralign refine --raligncmp ccc --shrinks1 2 --mask refine_01/simmask.hdf --parallel thread:30 01526: Traceback (most recent call last): 01527: File "/opt/scipion/pyworkflow/protocol/protocol.py", line 167, in run 01528: self._run() 01529: File "/opt/scipion/pyworkflow/protocol/protocol.py", line 211, in _run 01530: resultFiles = self._runFunc() 01531: File "/opt/scipion/pyworkflow/protocol/protocol.py", line 207, in _runFunc 01532: return self._func(*self._args) 01533: File "/opt/scipion/pyworkflow/em/packages/eman2/protocol_refineasy.py", line 274, in refineStep 01534: self.runJob(program, args, cwd=self._getExtraPath()) 01535: File "/opt/scipion/pyworkflow/protocol/protocol.py", line 960, in runJob 01536: self._stepsExecutor.runJob(self._log, program, arguments, **kwargs) 01537: File "/opt/scipion/pyworkflow/protocol/executor.py", line 56, in runJob 01538: env=env, cwd=cwd) 01539: File "/opt/scipion/pyworkflow/utils/process.py", line 51, in runJob 01540: return runCommand(command, env, cwd) 01541: File "/opt/scipion/pyworkflow/utils/process.py", line 65, in runCommand 01542: check_call(command, shell=True, stdout=sys.stdout, stderr=sys.stderr, env=env, cwd=cwd) 01543: File "/opt/scipion/software/lib/python2.7/subprocess.py", line 540, in check_call 01544: raise CalledProcessError(retcode, cmd) 01545: CalledProcessError: Command '/opt/scipion/software/em/eman2.12/Python/bin/python /opt/scipion/software/em/eman2.12/bin/e2refine_easy.py --input=sets/inputSet.lst --model=../../000970_XmippProtFilterVolumes/extra/output_volume.vol --targetres=14.000000 --speed=5 --sym=d2 --iter=4 --mass=150.000000 --apix=2.100000 --classkeep=0.900000 --m3dkeep=1.000000 --parallel=thread:30 --threads=30 --classrefsf --m3dpostprocess=filter.lowpass.gauss' returned non-zero exit status 1 01546: Protocol failed: Command '/opt/scipion/software/em/eman2.12/Python/bin/python /opt/scipion/software/em/eman2.12/bin/e2refine_easy.py --input=sets/inputSet.lst --model=../../000970_XmippProtFilterVolumes/extra/output_volume.vol --targetres=14.000000 --speed=5 --sym=d2 --iter=4 --mass=150.000000 --apix=2.100000 --classkeep=0.900000 --m3dkeep=1.000000 --parallel=thread:30 --threads=30 --classrefsf --m3dpostprocess=filter.lowpass.gauss' returned non-zero exit status 1 01547: FAILED: refineStep, step 2 01548: 2017-02-28 18:49:02.599250 01549: Cleaning temporarly files.... 01550: ------------------- PROTOCOL FAILED (DONE 2/3) -- Dr. Dennis Thomas Cryo-EM Facility Manager Center for Integrative Proteomics Research Rutgers University Phone: 848-445-5302 |
|
From: Dmitry A. S. <sem...@gm...> - 2017-03-01 10:18:24
|
Dear colleagues, Is there already a protocol that can centered the picked particles? Sincerely, Dmitry |
|
From: ashutosh s. <ash...@gm...> - 2017-02-19 23:57:47
|
Dear Carlos Thank you for the explanation. Best Regards Ashutosh On Sun, Feb 19, 2017 at 3:20 PM, Carlos Oscar Sorzano <co...@cn...> wrote: > Dear Ashutosh, > > On 17/02/2017 2:53, ashutosh srivastava wrote: > > Hi Everyone > > Is it possible to increase the number of iterations in an existing CL2D > classification run if it does not converge? > > Note that CL2D is not an optimization algorithm, the algorithm tends to > maximize the similarity between the class representative and the class > images, but it may depart from a strictly monotonic optimization. > > I performed a CL2D classification for ~23000 particles for 50 iterations > and it took quite a lot of time even on a cluster. > > Normally, the way of using the algorithms is with multiple levels (with > different number of classes at each level, 4, 8, 16, 32, ...) and > relatively few iterations at each level (10 should suffice for practical > purposes, and even smaller numbers, like 5, usually work although they are > not recommended). > > Visualizing the convergence I realized that the classification does not > seem to be converged with CL2D similarity of 0.4622 and CL2D changes 8814 > for the 50th iteration. I believe I need to run the classification for > larger number of iterations. However, there does not seem to a way to > continue the same run for further 50 or 100 iterations. > > This behavior of 1/3 of the particles changing classes is normal in CL2D > when the number of classes is relatively high. It does not mean that > classes are bad or that you should remove classes, it means that each class > has another very close class in such a way that particles have a hard time > to decide with which one to go (remind that the algorithm is not aimed at > maximizing a functional, although in general, it tends to maximize the > similarity between the class representatives and images). > > At the level of protocol there is not an easy option to continue from some > iteration, when using the program there is. A trick could be to use the > "Additional parameters" field and provide the parameter --ref0 with the > path to the xmd file of the last iteration of the previous classification. > > Kind regards, Carlos Oscar > > I am using Scipion v1.0.1. > Thank you > > Best Regards > Ashutosh > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > > > > _______________________________________________ > scipion-users mailing lis...@li...https://lists.sourceforge.net/lists/listinfo/scipion-users > > > -- > ------------------------------------------------------------------------ > Carlos Oscar Sánchez Sorzano e-mail: co...@cn... > Biocomputing unit http://biocomp.cnb.csic.es > National Center of Biotechnology (CSIC) > c/Darwin, 3 > Campus Universidad Autónoma (Cantoblanco) Tlf: 34-91-585 4510 > 28049 MADRID (SPAIN) Fax: 34-91-585 4506 > ------------------------------------------------------------------------ > > |
|
From: Carlos O. S. <co...@cn...> - 2017-02-19 06:20:59
|
Dear Ashutosh, On 17/02/2017 2:53, ashutosh srivastava wrote: > Hi Everyone > > Is it possible to increase the number of iterations in an existing > CL2D classification run if it does not converge? Note that CL2D is not an optimization algorithm, the algorithm tends to maximize the similarity between the class representative and the class images, but it may depart from a strictly monotonic optimization. > I performed a CL2D classification for ~23000 particles for 50 > iterations and it took quite a lot of time even on a cluster. Normally, the way of using the algorithms is with multiple levels (with different number of classes at each level, 4, 8, 16, 32, ...) and relatively few iterations at each level (10 should suffice for practical purposes, and even smaller numbers, like 5, usually work although they are not recommended). > Visualizing the convergence I realized that the classification does > not seem to be converged with CL2D similarity of 0.4622 and CL2D > changes 8814 for the 50th iteration. I believe I need to run the > classification for larger number of iterations. However, there does > not seem to a way to continue the same run for further 50 or 100 > iterations. This behavior of 1/3 of the particles changing classes is normal in CL2D when the number of classes is relatively high. It does not mean that classes are bad or that you should remove classes, it means that each class has another very close class in such a way that particles have a hard time to decide with which one to go (remind that the algorithm is not aimed at maximizing a functional, although in general, it tends to maximize the similarity between the class representatives and images). At the level of protocol there is not an easy option to continue from some iteration, when using the program there is. A trick could be to use the "Additional parameters" field and provide the parameter --ref0 with the path to the xmd file of the last iteration of the previous classification. Kind regards, Carlos Oscar > I am using Scipion v1.0.1. > Thank you > > Best Regards > Ashutosh > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org!http://sdm.link/slashdot > > > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users -- ------------------------------------------------------------------------ Carlos Oscar Sánchez Sorzano e-mail:co...@cn... Biocomputing unithttp://biocomp.cnb.csic.es National Center of Biotechnology (CSIC) c/Darwin, 3 Campus Universidad Autónoma (Cantoblanco) Tlf: 34-91-585 4510 28049 MADRID (SPAIN) Fax: 34-91-585 4506 ------------------------------------------------------------------------ |
|
From: ashutosh s. <ash...@gm...> - 2017-02-17 01:53:49
|
Hi Everyone Is it possible to increase the number of iterations in an existing CL2D classification run if it does not converge? I performed a CL2D classification for ~23000 particles for 50 iterations and it took quite a lot of time even on a cluster. Visualizing the convergence I realized that the classification does not seem to be converged with CL2D similarity of 0.4622 and CL2D changes 8814 for the 50th iteration. I believe I need to run the classification for larger number of iterations. However, there does not seem to a way to continue the same run for further 50 or 100 iterations. I am using Scipion v1.0.1. Thank you Best Regards Ashutosh |
|
From: Carlos O. S. <co...@cn...> - 2017-02-16 21:56:21
|
Dear Dmitry, On 02/16/17 12:01, Dmitry Semchonok wrote: > Dear colleagues, > > I would like to ask about 2 movie alignments - correlation and optical. > > When we do the movie alignment using these 2 protocols isn't there a > danger that the signal becomes weaker due to several interpolations? The correlation protocol does not interpolate the images, it simply writes the shifts associated to each frame. Then, optical alignment compose these global shifts with the local shifts and perform a single interpolation at the very end, which is performed with cubic B-splines trying to achieve the highest possible quality. Cheers, Carlos Oscar > > > Sincerely, > Dmitry > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > > > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users -- ------------------------------------------------------------------------ Carlos Oscar Sánchez Sorzano e-mail: co...@cn... Biocomputing unit http://biocomp.cnb.csic.es National Center of Biotechnology (CSIC) c/Darwin, 3 Campus Universidad Autónoma (Cantoblanco) Tlf: 34-91-585 4510 28049 MADRID (SPAIN) Fax: 34-91-585 4506 ------------------------------------------------------------------------ |
|
From: Carlos O. S. <co...@cn...> - 2017-02-16 21:53:03
|
Dear Dmitry, On 02/16/17 11:34, Dmitry Semchonok wrote: > Dear colleagues, > > When you use *xmipp3 - optical alignment* what values you are using > for the GPU? In principle, there is no difference in the parameters when using GPU or CPU. > > Is there also an option to run that process using the Threads/MPI? We have not performed a scaling experiment, it would be interesting if you try and share the results. It would not matter if you use threads or Mpi, but use just one of them, like threads=2 and mpi=1 or threads=1 and mpi=2 (during development we always used threads=mpi=1). > > > And do you *enable* /Use previous movie alignment t/o SUM frames > option if you*run**protocol*//upon/xmipp3 - correlation alignment? > / /Yes. Kind regards, Carlos Oscar / > / > / > Sincerely, > Dmitry > > > > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > > > _______________________________________________ > scipion-users mailing list > sci...@li... > https://lists.sourceforge.net/lists/listinfo/scipion-users -- ------------------------------------------------------------------------ Carlos Oscar Sánchez Sorzano e-mail: co...@cn... Biocomputing unit http://biocomp.cnb.csic.es National Center of Biotechnology (CSIC) c/Darwin, 3 Campus Universidad Autónoma (Cantoblanco) Tlf: 34-91-585 4510 28049 MADRID (SPAIN) Fax: 34-91-585 4506 ------------------------------------------------------------------------ |
|
From: Dmitry S. <sem...@gm...> - 2017-02-16 11:01:16
|
Dear colleagues, I would like to ask about 2 movie alignments - correlation and optical. When we do the movie alignment using these 2 protocols isn't there a danger that the signal becomes weaker due to several interpolations? Sincerely, Dmitry |
1 message has been excluded from this view by a project administrator.
