I have been struggling a lot with one of my datasets (~18k mics). I am working on a four-protein complex in which three proteins form a well-resolved core that is consistently visible in both 2D class averages and 3D reconstructions. However, the fourth component, a helicase, remains poorly resolved throughout processing.
I can occasionally see weak density that may correspond to the helicase in some 2D classes and during early 3D classification, but the feature becomes very weak, fragmented, or noisy after further refinement. In several reconstructions, the helicase density is only apparent at lower contour thresholds, while increasing the contour causes it to disappear. At the same time, the core complex remains well resolved.
I suspect there may be compositional heterogeneity, conformational flexibility, or partial occupancy of the helicase, but I am not sure how to distinguish or point out between these possibilities.
Has anyone encountered a similar situation? What would you recommend to make the reconstruction better?
This is a very common phenomenon to happen with larger protein complexes that exhibit flexibility. A recent preprint has a great processing workflow that I think shows a clear way to resolve this issue. In short, you will want to lowpass filter the map to 10-15 Å. This will help to define the region where your helicase is located to make mask creation easier. Once you do that, you can make a mask around the unresolved region and then perform a 3D-classification and then local refinement. You might find a couple conformations of the helicase this way. In the preprint below, they dont show a lowpass filter step, but it can be useful in this scenario.
Hi @kstachowski , thanks for the suggestion! I actually tried pretty much exactly the workflow you described. I low-pass filtered the map to around 10-15 Å to help define the helicase region, created a mask around the unresolved density, and then performed 3D classification followed by local refinement. Unfortunately, it still doesn’t seem to work, the helicase density remains poorly resolved, and I haven’t been able to separate out any distinct conformations. I’m wondering if there might be something else I’m missing or if the flexibility is simply too extensive. I’d appreciate any additional suggestions you might have.
