Poor 2D Class average resolutions

Troubleshooting
1

Hi, I am trying to solve a heptameric oligomer of ~350 kDa; however, despite the CTF fit resolutions being well within 7 Å. The 2D class averages look very blurry (with fewer features visible) and range from 15 to 25 Å. I have a mixture of GroEL contaminations with my protein of interest; however, both proteins are stuck in that low-resolution range. Here I am attaching the reference micrograph and 2D class averages. pixel size is 1.17Å, 20 frames,320pixel box size.

J81_2d_classes_for_iteration_20
J81_2d_classes_for_iteration_201920×1537 481 KB

200kV, K2 camera without energy filter.

J102_2d_classes_for_iteration_20
J102_2d_classes_for_iteration_201920×963 307 KB

J116_non_doseweighted_aligned_micrograph_for_014809511092617908769_20260610_1559_a000_g000_h070_d001
J116_non_doseweighted_aligned_micrograph_for_014809511092617908769_20260610_1559_a000_g000_h070_d001479×463 131 KB

(gold grid 0.6)

**I have another protein sample data that I collected from the same setup. Unfortunately, that sample is also getting stuck at low resolution.

J48_2d_classes_for_iteration_20
J48_2d_classes_for_iteration_201920×1537 510 KB

Quantifoil grid, pixel 1.17Å, box size 320pixel

I did not see any burning in the images either. The defocus range is also well within the range of -1.25 to -2.75. I have tried 40-frame data collection, with gain preprocessed and unprocessed. The thon rings were excellent for both samples. Any suggestions appreciated!

2

Hi,

A few initial thoughts:

I’d be a little worried if your CTF fits are around 7 A – do you mean that they are all below 7A fit for the samples? Even on our Tundra, most CTF fits are around 4-6A, and folks for Krios data sets often only take data that were fit to below 4 or 5A.

Also, defocus sounds like it’s a little high. You lose higher-resolution information the higher your defocus is; we usually collect at -0.5 to -2 or so on our Tundra and smaller on the Krios. We only use around -3 for screening. However I wouldn’t think that this alone would cause you to have issues at the 2D stage.

What dose are you using to collect your data? Can you show some of the CTF fit graphs and the overview graphs of CTF scatterplot?

Are you downsampling your boxes at all? That can influence things as well.