 |
 |
Hi Dieter, First of all: it is now recommended to NOT use relion_preprocess from the command line. Just use the entire preprocessing pipeline from the GUI as explained in the tutorial: this will reduce errors. Probably some of the particles are somewhat strange: have you looked at them? If norm correction remains a problem after throwing away bad particles, you can always switch it off using --dont_check_norm (you can always see the options of a program by typing its name without options on the CL. The program for 2D/3D classification and 3D auto-refine is called relion_refine). HTH, S > Hi Sjors, > > I am running 2D-classifications on a negative-stain dataset of > mediocre quality and low resolution (some thousand particles). The first > run is fine and stops at the default of 25 iterations. However, when I > select the good particles and run a second time (with just some hundreds > of particles) relion always stops at different iterations (depending on > the chosen parameters) with messages like the one below followed by a > segmentation fault. > > It *does not help* to make *even just one single group*. Does it mean > that the data set is simply too bad or is there a way to get through? > Should I remove the particle(s) it complains about one after the other? > I also tried regularization parameters between 1.2 and 1.8 and various > values for the Limit Resolution (between 8 and 30 A). Or does it simply > need a minimum number of particles to do a meaningful norm correction? > > 0.17/1.65 min ......~~(,_,"> WARNING: norm_correction= 16.7549 for > particle 384 in group 1; Are your groups large enough? > mymodel.current_size= 30 mymodel.ori_size= 38 part_id= 384 > coarse_size= 30 > DIRECT_A2D_ELEM(exp_metadata, my_image_no, METADATA_NORM)= 16.7549 > mymodel.avg_norm_correction= 0.690068 > exp_wsum_norm_correction[ipart]= 1.87198 > old_norm_correction= 8.65915 > wsum_model.avg_norm_correction= 51.5414 > group_id= 0 mymodel.scale_correction[group_id]= 1 > mymodel.sigma2_noise[group_id]= > 0.11602 > 0.099363 > 0.023142 > ................. > -------------------------------------------------------------------------- > mpirun noticed that process rank 2 with PID 11287 on node blaaslinpc > exited on signal 11 (Segmentation fault). > -------------------------------------------------------------------------- > > BTW: I noticed that the output of relion_preprocess is always > fileX.*mrcs.mrcs* regardless of whether a file extension is given or not. > > Thanks, best, Dieter > > ------------------------------------------------------------------------ > Dieter Blaas, > Max F. Perutz Laboratories > Medical University of Vienna, > Inst. Med. Biochem., Vienna Biocenter (VBC), > Dr. Bohr Gasse 9/3, > A-1030 Vienna, Austria, > Tel: 0043 1 4277 61630, > Fax: 0043 1 4277 9616, > e-mail: [log in to unmask] > ------------------------------------------------------------------------ > > -- Sjors Scheres MRC Laboratory of Molecular Biology Francis Crick Avenue, Cambridge Biomedical Campus Cambridge CB2 0QH, U.K. tel: +44 (0)1223 267061 http://www2.mrc-lmb.cam.ac.uk/groups/scheres