Relion 3.1 3D refine error: one of your half sets has no segments.

[Jake4484]

I imported particles from a select 2D class in Cryosparc. I then ran multiple 2D/3D classifications with no with no issues. I get an error with 3D auto refine that states one of your half sets has no segments and asks if rlnRandomSubset is set to 1 or 2 in your particles Star file .The data is for a 100 angstrom diameter particle with a 200 box size. Here is part of the Star file being used for Refinement:

version 30001

data_optics

loop_
_rlnVoltage #1
_rlnSphericalAberration #2
_rlnAmplitudeContrast #3
_rlnOpticsGroup #4
_rlnImagePixelSize #5
_rlnImageDimensionality #6
_rlnImageSize #7
300.000000 2.700000 0.070000 1 1.060000 2 200

version 30001

data_particles

loop_
_rlnImageName #1
_rlnMicrographName #2
_rlnCoordinateX #3
_rlnCoordinateY #4
_rlnAnglePsi #5
_rlnOriginXAngst #6
_rlnOriginYAngst #7
_rlnDefocusU #8
_rlnDefocusV #9
_rlnDefocusAngle #10
_rlnPhaseShift #11
_rlnCtfBfactor #12
_rlnRandomSubset #13
_rlnClassNumber #14
_rlnOpticsGroup #15
_rlnGroupNumber #16
_rlnAngleRot #17
_rlnAngleTilt #18
_rlnNormCorrection #19
_rlnLogLikeliContribution #20
_rlnMaxValueProbDistribution #21
_rlnNrOfSignificantSamples #22
000116@J774/extract/FoilHole_1159715_Data_31204767_31204768_20190201_1934_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204767_31204768_20190201_1934_aligned_DW.mrc 2156.000000 3060.000000 60.615960 3.974408 0.794408 29223.787109 28601.591797 25.456038 0.000000 0.000000 1 1 1 1 28.244671 85.321598 0.671937 17981.327251 0.657452 16

Is there something I should change on this to get it work?

[biochem-fan]

Look at the 13th column (rlnRandomSubset). The values there should be either 1 or 2. If not, delete this column.

[Jake4484]

Isnt the 13th column a one for rlnRandomeSubset?

[biochem-fan]

rlnRandomSubset (without e).

[Jake4484]

Yeah, sorry the e was a typo on my part. I mean it looks like a one is in the correct spot though right?

[biochem-fan]

You showed me only one row, so I don't know if it really contains only 1 and 2. Did you check?

[Jake4484]

Yeah I checked and it looks fine from what I can tell. Here is more of it:
000027@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 1074.000000 1828.000000 15.645364 -1.32559 1.854407 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 -56.76078 81.666363 0.668370 17504.979429 0.086946 38
000161@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 1090.000000 2175.000000 22.746588 -0.26559 -3.44559 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 -50.10469 95.593959 0.685574 17726.083163 0.261871 34
000024@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 3200.000000 2399.000000 143.099817 -3.44559 3.974407 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 -56.03130 91.199049 0.721454 17825.602786 0.218381 113
000059@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 858.000000 1745.000000 -172.63626 -3.44559 2.914408 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 -45.44609 80.880869 0.660900 17698.170603 0.177964 17
000046@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 1414.000000 248.000000 46.288876 3.974408 9.274408 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 61.502677 175.094153 0.648444 17199.075695 0.511167 25
000033@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 1252.000000 844.000000 -23.57131 -0.26559 2.914408 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 -164.82812 53.422966 0.754781 17909.753263 0.297193 31
000264@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 1152.000000 1853.000000 91.524650 -3.44559 9.274408 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 151.396176 104.786385 0.660234 17485.838194 0.430846 132
000040@J774/extract/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW_particles.mrcs J764/imported/FoilHole_1159715_Data_31204773_31204774_20190201_1935_aligned_DW.mrc 3362.000000 2151.000000 151.669963 -1.32559 6.094408 29848.451172 29340.681641 14.153535 0.000000 0.000000 1 1 1 2 -79.28633 83.609763 0.668099 17754.115808 0.406712 22

[biochem-fan]

Did you check the number of 1s and 2s? Is it more-or-less balanced?

[Jake4484]

All of them are ones at the 13th column

[biochem-fan]

OK, that is the problem. Please report this bug to whoever wrote this CryoSPARC-to-RELION converter.

Meanwhile, you should remove the column to let RELION reassign half sets. In this case, you should low pass filter the initial model as much as possible and start refinement by global search to avoid overfitting.

[Jake4484]

Okay sounds good I will give that a try. I really appreciate the help with it. I'll let you know what I come up with

[Jake4484]

It seems to run to just fine by deleting the column. I will add this issue to pyem, so that it can get fixed. Is there a setting I should change to start refinement by global search at the start of the refinement?

[biochem-fan]

I will add this issue to pyem

Thanks!

Is there a setting I should change to start refinement by global search at the start of the refinement?

The default is global search.

[Jake4484]

Not a problem, thanks for your help! I really appreciate it!