Creating Relion-5 compatible star files (particle and tomogram) from Warp to re-extract particles with Relion

[mpm896]

Hello,

I've been using Warp for tomogram processing and particle extraction and Relion-5 for subtomogram averaging. I want to re-extract particles with Relion (for use in other programs such as cryoDRGN-ET, which requires 2D particle stacks without CTF premultiplication), however there seems to be issues with how Warp constructs the tomograms.star file so the extracted particles are just black volumes. Has anyone developed a workaround for this? Thanks!

Matt

[sporebyte]

any progress on this @mpm896 ? I'm having the same issue

[delarosatrevin]

Hi Matt and Sonja, The main problem is not the tomograms.star, but the underlying geometry of the tilt-series alignment (either from the Aretomo2 wrapper or Etomo) and how Warp reconstructs the tomogram. If the same geometry is not properly converted to Relion, then the particle extraction will fail to locate the particles in the tilt-series. I'm working with Sjors and Bogdan to better integrate Warp and Relion, but we are still working on the conversion and waiting for some clarification from Dimitry.

…

On Tue, Jul 7, 2026 at 7:40 AM Sonja V ***@***.***> wrote: *sporebyte* left a comment (3dem/relion#1333) <#1333 (comment)> any progress on this @mpm896 <https://github.com/mpm896> ? I'm having the same issue — Reply to this email directly, view it on GitHub <#1333?email_source=notifications&email_token=ACAM6WHHJI3ZHHOGHLWVK6T5DTVVJA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIOJQGM4DQOBTGY4KM4TFMFZW63VKON2WE43DOJUWEZLEUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#issuecomment-4903888368>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACAM6WGDYBYZKPASZUVFRL35DTVVJAVCNFSNUABEKJSXA33TNF2G64TZHM4DAMJSGYYTMMR3JFZXG5LFHM2DCMBSHA2TKNBWGSQXMAQ> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/ACAM6WDY4RLEDLR5VUAYUOL5DTVVJA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIOJQGM4DQOBTGY4KM4TFMFZW63VKON2WE43DOJUWEZLEUVSXMZLOOSVGM33PORSXEX3JN5ZQ> and Android <https://github.com/notifications/mobile/android/ACAM6WEFZP774VKI5A6H4Z35DTVVJA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIOJQGM4DQOBTGY4KM4TFMFZW63VKON2WE43DOJUWEZLEUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ>. Download it today! You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[mpm896]

Great to know, I figured it had to do with geometries and differences in how Relion, Warp, Etomo handle and interpret them but I just don't know enough to deal with this. Thank you for working on it!

[mpm896]

@sporebyte , we ended up temporarily working around the issue by de-premultiplying (dividing??) the CTF from particle stacks so that they could be used in cryoDRGN-ET. I don't know exactly how this was implemented as a coworker helped out and used Claude to assist with some of the technicalities, but the results looked nice so it seemed to work well 😄