how to import one tilt stack from SerialEM

577 views
Skip to first unread message

Juha Huiskonen

unread,
Nov 20, 2020, 5:09:18 PM11/20/20
to Warp
Hi, 

I have an unaligned stack (MRC) and an MDOC file from SerialEM. How can I import them to Warp 1.0.9 for processing.  

To clarify, I have one image per tilt and all of the images are in one MRC file. 

Best wishes
Juha

Vasilii Mikirtumov

unread,
Nov 20, 2020, 5:24:50 PM11/20/20
to Juha Huiskonen, Warp
Hi,

Change the input format to *.tomostar and press import tilt series from imod. 

V

--
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/1fb90696-6f0f-4c11-8d64-555d818e6241n%40googlegroups.com.

Juha Huiskonen

unread,
Nov 20, 2020, 5:50:39 PM11/20/20
to Warp
Hi, thanks for the quick reply.  I have done that.  I have tried to give the correct folder names on the next page that opens, but nothing appears in the list.

Vasilii Mikirtumov

unread,
Nov 20, 2020, 6:13:42 PM11/20/20
to Juha Huiskonen, Warp
Hi,

I think Warp will only see stacks in case you have created them in Warp beforehand. Normally, you start with raw frames, preproccess them in Warp, then create stacks (by changing to .tomostar and create stacks for imod), then align them outside and import back to Warp. If you don't have frames, just do the same with individual mrc images from your unaligned stack, as it's explained on quick start tilt series web page.

Hth,
V

Alister Burt

unread,
Nov 21, 2020, 4:10:31 AM11/21/20
to Vasilii Mikirtumov, Juha Huiskonen, Warp
Hi Juha/all,

Warp expects an mdoc file for the individual frames for subsequent stack creation rather than the stack itself, I've attached one so you have an example.

For data coming from Tomo5 as individual frames I wrote a script 'mdocspoofer' which can make these for you https://github.com/alisterburt/mdocspoofer

The frames directories from Tomo spit out filenames which look like:
basename_001[-0.01]_fractions.mrc
Where 001 is a set of three numbers defining the order of collection, the tilt angle is in square brackets directly after.

A solution which would get your data into the Warp/M pipeline is thus:
  1. split your stacks into individual mrc files matching the filenames above with a different basename per tilt series, correct order and tilt angle in the filename
  2. run mdocspoofer on the directory containing all of your individual tilt images (or modify/roll your own of course)
  3. process the data in Warp as 2D images first (seems redundant but Warp was designed with frames in mind)
  4. follow the user guide for tomography processing
Good luck!

Alister


ts01.mrc.mdoc

teg...@gmail.com

unread,
Nov 21, 2020, 7:22:07 AM11/21/20
to Warp
Hi Juha,

Vasilii and Alister are absolutely correct! You'll get better results for alignments in M and denoising in Warp/Noise2Map by starting with tilt movies, so that's what the pipeline expects. If you start with a stack of aligned averages, you'll have to unpack them, pretend they are movies, and have Warp produce averages (just copies of the original files) first. This is very inconvenient for old data that have been stored as stacks. I hope we will get used to keeping the original tilt movies in the future – as TIFFs, they are often smaller than the fp32 aligned averages.

Cheers,
Dimitry

Juha Huiskonen

unread,
Nov 21, 2020, 10:47:27 AM11/21/20
to Warp
Thanks Dimitry et al! Makes sense and thanks for the instructions 
Reply all
Reply to author
Forward
0 new messages