Fail to reconstruct reasonable volume for EMPIAR10028

ml-struct-bio / cryofire

Amortized inference for ab initio heterogeneous cryo-EM reconstruction

https://cryofire.cs.princeton.edu/

GNU General Public License v3.0

31 stars 5 forks source link

Fail to reconstruct reasonable volume for EMPIAR10028 #1

Open AlenCarral opened 12 months ago

AlenCarral commented 12 months ago

I strictly followed the steps provided in the README file. However, I encountered an issue where the reconstruction protein model appears to have a hollow structure, whether w/wo gt-pose: left: wo gt-pose right with gt-pose

ps: I have already validated the input data by running cryodrgn and cryodrgn2 using exact the same set of data, they produced reasonable structures:

Has anyone encountered a similar situation before?Or can someone provide some ideas or suggestions?

smallelephant9516 commented 11 months ago

I think it can be solved by multiplying a negative sign to the ctf function or the image in the fourier space.

1320harry commented 11 months ago

I also encountered the same situation. Did you resolve the problem?

LucunJi commented 1 month ago

Same situation, and after multiplying -1 to the reconstructed volume, I get a seemingly correct result.