Hello.
I have a complex problem, but to spare the details here is a summary of what we could use some help with.
Trying to work through a cryoEM dataset of a ~230 kDa multi-protein complex with an additional transient binding partner (~32 kDa protein). The core complex refines reasonably well, but there is extra density that consistently stays low resolution and poorly defined, which may be the 32 kDa transient binding protein.
What is interesting: 1.) the density does not appear in control samples without the extra added 32 kDa transient binding partner, 2.) 2D class averages show smeared/ghosted density that could be due to continuous flexibility, and 3.) 3D reconstruction can recover the density, but refinement methods are not improving the density map to a point where known models of the transient binding partner(~32 kDa) can be rigid body fit into the density map.
These are the cryoSPARC jobs that have been tried so far: 1.) Blob picking 2.) 2D classification, 3.) Ab initio, 4.)Heterogeneous refinement, 5.) NU refinement, 6.) Template picking, 6.) reiterate the previous steps up to NU-refinement, 7.) Local refinement, 8.) 3D classification, and 9.)3DFlex refine. I can share plots of completed jobs if requested.
We want to determine whether we’re dealing with: flexibility/transient binding, low occupancy, or simply a sample problem that processing alone won’t solve.
For anyone who has dealt with small flexible peripheral density before (aka small dynamic protein binding partner), do you have any advice? And how do you decide when it’s time to stop pushing processing and go back to biochemistry/sample prep?
Thank you.