Relion5 Ctf estimation： BUG: MDin should contain a rlnMicrographPreExposure label

[kexinliu0001]

hi! I'm doing CtfFind using Relion5. I did Import and MotionCorr as the tutorial, but when I'm doing CTF estimation with input tilt series corrected_tilt_series.star, it gave the error information "BUG: MDin should contain a rlnMicrographPreExposure label":

I checked the star files, and the star file for each tomogram contains the rlnMicrographPreExposure

How can I solve it?

Thanks!!!

[2019C]

Hello, I am running into the same error message in RELION 5 tomo CTF estiomation job. Have you figured out the problem?

BUG: MDin should contain a rlnMicrographPreExposure label

Thanks, Chen

[2019C]

Hi,

My problem was fixed. It turned out that the CTFFIND4 couldn't run on the cluster. The solution is to add "export LC_ALL=C" to my ~/.bashrc.

Hope it helps. Chen

[Li19970731]

I have added "export LC_ALL=C" to my ~/.bashrc, but this bug still here.

[xingxingjiukunkun]

"I have the same issue, have you resolved it?

[AssmannG]

HI, I had the same issue and could also resolve it with exporting LC_ALL=C inside the relion SLURM submission templates ...seems as if the $LANG setting that was set on the cluster was not not the standard "en_US.UTF-8". Accoringly, CtfFind failed.... you can check that with $ locale from where you are running the jobs...

[HDuyvesteyn]

Hiya, I am also encountering the same issue, even after trying exporting LC_ALL=C in my slurm submission script with relion-5.0-beta-3 and ctffind-4.1.14.

I tried an earlier version of ctffind4 (I am afraid that I cannot find exactly which version at present, as was using someone else's installation) and this ran absolutely fine, so might be an issue with the 2019 version of ctffind4?

[MXXXZ]

Thanks @HDuyvesteyn , I changed the ctffind version from 4.1.14 to 4.1.5 and the bug disappeared, so it is indeed likely a bug in the new version.

[huwjenkins]

@HDuyvesteyn - did you try the version ctffind_4.1.14_centOS7.tar.gz from https://grigoriefflab.umassmed.edu? That works on our cluster but the other ctffind_4.1.14 binaries require a either a wrapper script that does:

#!/bin/bash
LC_ALL=C /path/to/bin/ctffind-4.1.14 $@

~~which I guess means~~ or exporting LC_ALL=C in the SLURM job submission script ~~also didn't work for me (sorry I don't remember)...~~

edited as I just tested and that does work.

[biochem-fan]

Repaired in #1212 by @huwjenkins.