[cryoelectron microscope] relion4.0 pipeline command summary (self use)

This article will be sorted out RELION4.0 Commands used in each step of the tutorial , Easy to view and use related API.
Suggested supporting experimental process RELION4.0 subtomo Tutorial parsing Look at .

1. Import tomograms

relion_tomo_import_tomograms 
--i input/tomograms_descr.star --o ImportTomo/job001/tomograms.star 
--angpix 1.35 --voltage 300 --Cs 2.7 --Q0 0.07 
--ol input/order_list.csv --flipZ --hand 1 
--pipeline_control ImportTomo/job001/

Corresponding source code ：import_tomograms.cpp

2. Import particles

relion_tomo_import_particles 
--i input/coords_5tomos.star --o ImportTomo/job002/ 
--t ImportTomo/job001/tomograms.satr 
--pipeline_control ImportTomo/job002/

Corresponding source code ：import_particles.cpp

3. establish pseudo subtomo

relion_tomo_subtomo_mpi 
--i ImportTomo/job002/optimisation_set.star --o PseudoSubtomo/job003/
--theme classic 
--b 192 --crop 96 --bin 4 --float16 --j 24
--pipeline_control PseudoSubtomo/job003/

Corresponding source code ：subtomo_mpi.cpp

4. establish initial model

relion_refine 
--i PseudoSubtomo/job003/particles.star 
--o InitialModel/job004/run
--iter 100 
--grad 
--denovo_3dref 
--ctf 
--K 1 
--sym C1 
--flatten_solvent 
--zero_mask 
--dont_combine_weights_via_disc 
--pool 16 --pad 1 
--skip_gridding 
--particle_diameter 230 
--oversampling 1 
--healpix_order 1 
--offset_range 6 --offset_step 2 
--auto_sampling 
--tau2_fudge 4 
--j 8 --gpu '0' 
--pipeline_control InitialModel/job004/

# dont_combine_weights_via_disc:  adopt MPI The network sends a large array of weighted sums , Instead of writing large files to disk .
# skip_gridding:  skip M Mesh subdivision in step 
# tau2_fudge:  Regularization parameters （ Greater than 1 The value of gives more weight to the data ）
# oversampling： Adaptive oversampling sequence to speed up computation （0： No oversampling ,1:2x,2:4x)
# healpix_order：3D Angle sampling on sphere （ Before oversampling ） Of Healpix The order .


relion_align_symmetry 
--i InitialModel/job004/run_it100_model.star --o InitialModel/job004/initial_model.mrc 
--sym C6 --apply_sym --select_largest_class 
--pipeline_control InitialModel/job004/
# apply_sym:  Apply symmetry to  map
# select_largest_class：  from  model.star  Select the largest class in , By default, the one with the best symmetry will be selected  map.

Corresponding source code ：ml_optimiser.cpp, align_symmetry.cpp

5. Refine 3D

relion_refine_mpi
--o Refine3D/job005/run 
--auto_refine 
--split_random_halves
--ios PseudoSubtomo/job003/optimisation_set.star
--ref InitialModel/job004/initial_model.mrc
--ini_hign 50 
--dont_combine_weights_via_disc 
--pool 16 --pad 2
--skip_gridding 
--auto_ignore_angles 
--auto_resol_angles 
--ctf 
--particle_diameter 230 
--flatten_solvant 
--zero_mask 
--oversampling 1 
--healpix_order 2 
--auto_local_healpix_order 4 
--offset_range 5 
--offset_step 2 
--sym C6 
--low_resol_join_halves 40 
--norm --scale 
--j 8 --gpu ""
--pipeline_control Refine3D/job005/

# auto_refine： In automatic thinning , Update the angle sampling based on the sampling required by the resolution , Make the convergence speed faster .

Corresponding source code ：src/apps/refine(_mpi).cpp, The main algorithm implementation is located in src/ml_optimiser(_mpi).cpp.

6. reconstruct particle

relion_tomo_reconstruct_particle_mpi
--i PseudoSubtomo/job003/optimisation_set.star
--p Refine3D/job005/run_data_z2.75/star
--theme classic
--o ReconstructParticleTomo/job006/
--b 256 --crop 128 --bin 2 
--j 24 --j_out 24 --j_in 1 --sym C6
--pipeline_control ReconstructParticleTomo/job006/

Corresponding source code ：reconstruct_particle_mpi.cpp

---

$$Author:Chier$$