aligning tilt series fails (Cannot find .xf or cannot find path)

423 views
Skip to first unread message

Jesse Hansen

unread,
Jul 24, 2024, 6:02:24 AM7/24/24
to Warp
Sorry for so many posts!  Assuming if there are issues with warptools you'd like to know about them. 

I ran through the tutorial data and it worked fine, so the installation is okay.  Now I'm trying with my data and running into new issues.  The frameseries steps worked okay, I could create tomostars, but now when I try alignment with etomo I get the output below:

2024-07-24 09:50:37.227 Tracking patches to make alignment model (running xcorr_pt.com)   [brt2]
2024-07-24 09:52:33.490 Successfully finished xcorr_pt.com   in 01:56.3   [brt3]
2024-07-24 09:52:33.491
2024-07-24 09:52:33.492 Reached step 5
2024-07-24 09:52:33.493
2024-07-24 09:52:33.493 Completed dataset TS_014_ts_021 through step 5  in 02:12.5   [brt4]
2024-07-24 09:52:33.494 Batchruntomo finished with data set, Wed Jul 24 11:52:33 2024  [brt12]
2024-07-24 09:52:33.494
2024-07-24 09:52:33.494 Batch run finished; no failures occurred
2024-07-24 09:52:33.515 Calculating projection parameters from patch tracking results in (path)/warp_tiltseries/tiltstack/TS_014_ts_021
2024-07-24 09:52:33.893 Python PID: 283498
2024-07-24 09:52:34.525 Running align.com ... Error executing align.com
2024-07-24 09:52:34.526 ERROR: tiltalign -StandardInput: exited with status 134
2024-07-24 09:52:34.535 Finished for (path)/warp_tiltseries/TS_014_ts_021.tomostar
2024-07-24 09:52:34.535 Execution took 133.835 seconds
2024-07-24 09:52:34.536 


then afterwards outputs: 
Failed to import alignments:
Could not find TS_014_ts_021.xf



So I tried also with AREtomo and get a different error below that it can't find the stack.  Clearly it has created the stack, and when I check that folder location I see the stack is there, so I'm guessing this is a bug?  

2024-07-24 09:58:28.846 Received "TomoStack", with 2 arguments, for GPU #0, 10759 MB free:
2024-07-24 09:58:39.780 Created tilt stack for (path)/warp_tiltseries/TS_014_ts_021.tomostar
2024-07-24 09:58:39.780 Execution took 10.924 seconds
2024-07-24 09:58:39.780
2024-07-24 09:58:39.817 Received "TomoAretomo", with 2 arguments, for GPU #0, 10663 MB free:
2024-07-24 09:58:39.821 Executing AreTomo2 in (path)/warp_tiltseries/tiltstack/TS_014_ts_021 with arguments: -InMrc TS_014_ts_021.st -AngFile TS_014_ts_021.rawtlt -VolZ 0 -OutBin 0 -TiltAxis 85.00001 -1 -AlignZ 124 -TiltCor 1 -OutImod 1 -DarkTol 0 -OutMrc TS_014_ts_021_aligned.mrc -Gpu 0 -Patch 5 4
2024-07-24 09:58:39.855 System.ComponentModel.Win32Exception (2): An error occurred trying to start process 'AreTomo2' with working directory '(path)/warp_tiltseries/tiltstack/TS_014_ts_021'. No such file or directory
2024-07-24 09:58:39.855    at System.Diagnostics.Process.ForkAndExecProcess(ProcessStartInfo startInfo, String resolvedFilename, String[] argv, String[] envp, String cwd, Boolean setCredentials, UInt32 userId, UInt32 groupId, UInt32[] groups, Int32& stdinFd, Int32& stdoutFd, Int32& stderrFd, Boolean usesTerminal, Boolean throwOnNoExec)
2024-07-24 09:58:39.855    at System.Diagnostics.Process.StartCore(ProcessStartInfo startInfo)
2024-07-24 09:58:39.855    at WarpWorker.WarpWorker.EvaluateCommand(NamedSerializableObject Command) in (path)/src/warp/WarpWorker/WarpWorker.cs:line 377


best

Jesse Hansen

Jesse Hansen

unread,
Jul 24, 2024, 9:38:13 AM7/24/24
to Warp
Well I'm talking to myself now, but I figured I'd post this in case anyone encounters in the future.   

I figured out the AREtomo issue.  I used the flag --exe and pointed to a downloaded version of AREtomo2.  I was also careful to load the right CUDA beforehand.   

I guess this suggests that the built-in AREtomo (or some CUDA ver) may be the issue?  

best,

Jesse

Pranav Shah

unread,
Jul 24, 2024, 10:12:58 AM7/24/24
to Jesse Hansen, Warp
A lot can be achieved by simply talking out loud ;)

WarpTools does not bring with it any AreTomo/Etomo binaries. These
need to be installed as external binaries and be available in the
$PATH environment variable in your session. It is not a bad idea for
AreTomo2 and WarpTools to be compiled with the same CUDA versions
as it helps to avoid dependency issues of the nature you are
experiencing. But as far as I understand, there is no absolute
requirement for the two binaries to share the same CUDA versions.

Best,
Pranav
--
Pranav Shah
Postdoctoral Research Fellow.

Division of Structural Biology,
Wellcome Trust Centre for Human Genetics,
University of Oxford,
Roosevelt Drive, Oxford OX3 7BN,
UK
> --
> You received this message because you are subscribed to the Google Groups "Warp" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to warp-em+u...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/warp-em/9c9ef7fb-cce2-4a9d-a7ec-b8165e505eaen%40googlegroups.com.

Alister Burt

unread,
Jul 25, 2024, 9:09:01 AM7/25/24
to Pranav Shah, Jesse Hansen, Warp
Hi Jesse,

As Pranav said, we don't distribute Etomo/AreTomo and users are expected to make sure the binaries are on the PATH and can run correctly.

If everything ran fine with the tutorial data using IMOD something in your data/parameters must be making it fail on your data - try opening the corresponding .edf files in eTomo and seeing if everything works as expected from there...

If there are issues that you think are bugs, please report them in the GitHub issues at https://github.com/warpem/warp/issues as it's easier to track, this discussion list should be kept for more general usage questions. I'm aware the line between these things isn't always clear when the question is being asked, don't worry too much about getting it right :-)

Cheers and good luck processing!

Alister

Reply all
Reply to author
Forward
0 new messages