Dear All,
To follow up on this topic. I also ran into the similar issue with one of sample with small size. Protein size 65A, pixel size 0.671, boxsize 384->192, defocus 0.4 to 2 um.
I have tried
Your first run with force max on is somewhat promising. Try to increase iteration to 40 and use a circular mask ~1.2x the diameter of your particle
Thanks for quick reply. I am still trying.
What confused me is that many mentioned to turn force max off, which significantly improve the quality of 2D. How come I got pure spheres, not sure where is wrong.
Best Regards,
Hi Jamon - if you turn force/max off, you will need a lot more iterations for it to converge. This picture is from iteration 8 - how did it look after 20 iterations? I would usually run for 40 iterations with force/max off.
Good morning
following on the thread here: I have a small complex (~65 kDa), 2D classes look good, and the ab initio produces a reasonable model. But I fail to go past 5-6 Ang during refinement, and the maps looks either featureless or (if I look at the sharpened map) a total mess (over-refinement is a word that comes to mind). I tried masking and local refinement and again the maps don’t make any sense. Any suggestion will be appreciated
Giovanna
Hi all,
I have same issue with 2D classification and Ab-initio to build model in my project as shown here.
My protein is not too small, around 200 kDa, the size should be around 100 A including the membrane micell. I tried to build 3D with Ab-initio but I could not get any reasonable secondary structures like helix or transmembrane domain. I includedd some 2D classification and select 2D here, could you please give me some advice to further select 2D and ab-initio?
Hi Nguyen,
This is too small and featureless to have 200kDa of ordered mass - these look like empty micelles or maybe a very small transmembrane protein. Do you expect a significant extramembrane domain?
Cheers
Oli
Hi,
Thank Oli for your comments!
The cicular mask diameter is set 150A. My expected particles are a protein complexes. Yes, the lower classes look much smaller than 100A. But, do you think the upper 2D classes might belong to transmembrane proteins or complex?
And I do not expect a significant extramembrane domain or they will be small.
I think the upper classes are unlikely to contain 200kDa of protein - they likely have a TM component of some sort, but not your complete complex IMO
Thanks a lot, Oli!
Actually, I tried to capture the early stages of the final complex. So, it might be reasonable if I see something smaller. I will try to figure out what is the upper classes, protein or only empty micell. ANW, I might have to more optimize my protocol to reduce the heterogeneity of my samples since I have a lot of noise particles as you see in the lower class.
Thanks
Hi Daniel,
I used 0.01% of LMNG, I did not concentrate the sample so the detergent concentration will not go higher.
I see. How many transmembrane passes are expected?
Maybe the protein-containing particle are on the carbon?
Hi Daniel,
Thanks and sorry for very late response! I expected a protein complex of 200 kDA, however, it looks like that I do not have it. It might be much smaller. I will look at the carbon coated grids to see whether my particles are there or not!
Thanks,
Hi Nguyen,
They do look like empty micelle. Try to keep LMNG detergent concentration below 0.01%.
Could you figure it out in the end? whether your 2D classes were empty micelles or they had your protein in it???
Hi Nguyen,
These are all empty micelles. I am facing the same problem. My protein is a 160 kDa ABC transporter. I could capture them once but subsequently, i got these micelles. You need to lower the detergent just the same as your CMC say 0.01% for DDM should be fine. Also, you could try nanodisc/amphiphol to get rid of these micelles. The worse part is these micelles elute with your protein and there are few tedious methods of getting rid of them. see this
Cryo-EM: spinning the micelles away - PMC (nih.gov)
best
Uddipan
Micelles usually concentrate along with the protein. The solution is to use ultra low CMC detergent mixtures. Many proteins remain soluble in LMNG even to infinite dilution (i.e. no detergent in SEC buffer), for others 0.0001% total detergent (where some of it is LMNG) is enough.
Does the 200 kDa mass only include protein? If no extramembrane domain, the micelles should be much larger. this is too small. Lan
