Run summovie on particle stacks

Submitted by tarek on Mon, 2016-03-07 08:56

Forums

Unblur & Summovie

Dear Unblur experts,

I want to give exposure filtering a try for a ribosomal test data set.
Since I have already pre-aligned particle stacks, I thought I can skip the alignment procedure within unblur and just run summovie with shifts 0,0.
According to your eLife 2015 paper, the filter function was developed for a VP6 sample and also applied to proteasomal particles, which both only consist of proteins.

First of all I'd like to know if you redetermine the filter function for different biological samples.
I can imagine that the nature of the sample has an influence on e.g. radiation sensitivity.

My second question is concerning the input for unblur/summovie. Do you have experience with filtering of particle stacks instead of whole image stacks?

Best,
tarek

timgrant Mon, 2016-03-07 09:56

Hi Tarek,

The filter function was estimated on the DLP VP6, and it is indeed possible that different proteins exhibit different radiation sensitivities, however from what we have seen so far it seems to be a fairly good approximation for the other proteins we have tried.

Ribosomes however, are a different story due to their high RNA content. In this case, the filter will likely weight well for the ribosomal proteins, however it will tend to over weight the RNA which seems to be less radiation sensitive. This means that you may end up with a situation where the resolution of the proteins is better with the filter, but the RNA is worse. The FSC appears to be mainly driven by the RNA so your FSC measured resolution may decrease with the exposure filter (depending on the total exposure amount).

In theory summovie won't differentiate between movies and particle stacks, it will just sum whatever you give it, and so you could use your prealigned particle stacks and 0.0 shifts as you describe. However, there would either need to be one stack for each particle, or one stack will all particles where the frames for each particle are consecutive in the stack, so you may need to rearrange things a little bit first.

Cheers,

Tim

tarek Mon, 2016-03-07 10:49

Hi Tim,

thanks a lot for the quick response.
Rearranging stacks should not be a big problem, however the different outcome for RNA and protein parts of the sample make things more complicated.
Would you expect the overall resolution to improve at all by exposure filtering or is the manual summation of frames the better approach in that case?

Best,
tarek

timgrant Tue, 2016-03-08 18:27

Hi Tarek,

The different outcome for the RNA and protein parts of the sample is not a feature of the exposure filter, it will be a feature of any kind of weighting / summing different numbers of frames. Essentially your sample is composed of two different materials with different radiation sensitivities, and so one cannot weight them both optimally with a single filter.

As the RNA density is quite dominant in ribosomes, if you primarily care about what gives you the best FSC then you should probably try and optimally filter the RNA, which will require higher exposure values than are accounted for by the exposure filter, however, in this case your protein density may suffer.

Perhaps you could make different reconstructions for the protein / RNA parts of the ribosome to aid in the interpretation of your structure?

Tim

tarek Thu, 2016-03-17 03:32

Hi Tim,
thank you for your ideas.
Since I am lacking experience with quantitative exposure effects, I will simply try what works best.
From your 2015 eLife paper the effect on proteins is quite convincing, however we need to find out how the RNA behaves.

Best,
tarek