Hi,
I don’t know if it’s important or just a display issue; in previous versions of cryosparc the map displays icons in the output of the job always showed detergent belt being on the top of the viewer.
In the newer version of cryosparc (now I’m using 4.4.1; I can’t trace it back to which version changed but it might be v4), in one conformation of the protein the detergent belt is up, and in the other conformation the belt is down.
So far it didn’t bother me as I know what I’m looking for, but I was wondering if it could have an impact on particle alignments? (My guess is not, but I’d like to be sure). More specifically, when I’m trying to align particles that are intermediates between these 2 extreme conformations.
Thank you
Vincent
Hi @vincent! Could you show us some example volumes which are in the “up” and “down” conformations?
Hi @rposert ,
appologies, I wrote too fast, I was running NU jobs at the same time so I confused the two, my question is a 3DFLEX question, which is what I was getting to.
I collected several turnover condition datasets, in which I can clearly make out discrete conformations and visualize very nice conformational changes around each conformation using 3DVA.
Now, I want to use 3DFLEX to connect these 2 discrete conformations, and I believe I have particles in my dataset that are intermediate states, not well resolved through NU refinement but features lead me to believe they can be the missing link.
Thus, I am testing to see if non-linear exploration of latent space can give me the Holy Grail of seeing the whole conformation change.
J201: I gather all the particles, and perform NU refinement against conformation A (so particles of A+B+“intermediate” refined against conformation A), then 3DFLEX train (I have optimized many of the parameters described in the video, mesh rigidity, train rigidity and latent strength). In this case, the detergent belt is “up”.
J219: same particles, but NU refinement performed against conformation B, then 3DFLEX train with more or less the same parameters as J201 but different mesh of course. In that case, the detergent belt is “down”.
It turns out that the “up” position corresponds to the place where the most variability is expected if NU refinement is performed (coincidence?). BUT, for the whole conformational change to occur correctly, the detergent belt will need to stay in the same place, and the protein would move around it sort of speak.
Thus far, the movements are rather limited and resemble a lot the ones that can be observed in 3DVA around each given conformation and don’t go further. So my procedure might not be the way to go. But I’m wondering if this “up” or “down” position would keep the flexibility in one orientation, which will not work in my case as the flexibility is expected to change as the conformation change, and the detergent belt is helpful in normalizing the protein position during this transition.
I should also mention that there is no compositional heterogeneity in the dataset. The whole protein is visible in each conformation.
Thanks for the help.
Vincent
When you performed the two NU refinements, did you leave Symmetry Alignment on? Sometimes that flips the volume along the Z-axis…
Yes it was on. I’ll give that a try, thanks.
