Relion error: CTF::initialise: ERROR

777 views

Skip to first unread message

Cryo-EM in the cloud

unread,

Oct 1, 2015, 8:50:50 PM10/1/15

to Cryo-EM in the cloud

Hello,

I got the following error at the end of my standard error file (ending in .e?) and my job crashed:

CTF::initialise: ERROR: CTF initialises to all-zero values. Was a correct STAR file provided?
File: src/ctf.cpp line: 169

And, accompanying this error, I got this warning in the standard output file (ending in .o?):

 + Slave    90 runs on host            = node002
 + Slave    91 runs on host            = node002
 Estimating initial noise spectra
   0/   0 sec ............................................................~~(,_,">
WARNING: There are only 4 particles in group 87
WARNING: You may want to consider joining some micrographs into larger groups to obtain more robust noise estimates.
         You can do so by using the same rlnMicrographName for particles from multiple different micrographs in the input STAR file.
         It is then best to join micrographs with similar defocus values and similar apparent signal-to-noise ratios.
 Estimating accuracies in the orientational assignment ...
000/??? sec ~~(,_,">                                                          [oo]--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 1 in communicator MPI_COMM_WORLD
with errorcode 1.


NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------

Where does this error come from?

Thanks!

Michael Cianfrocco

unread,

Oct 1, 2015, 9:01:19 PM10/1/15

to Cryo-EM in the cloud

That error occurs when the groups of particles are too small. You can read more on the Relion wiki here:

http://www2.mrc-lmb.cam.ac.uk/relion/index.php/FAQs#How_can_I_group_micrographs_in_order_to_have_more_particles_per_group.3F

You can either use the Relion GUI to open a _model.star file to regroup the particles, OR, if you haven't run any 2D/3D classification yet, you can use their manual instructions: