Dear Qiu-Xing,
You can extract particles and downscale them at the same time in the
extraction tab. Just downscale to a smaller image size. It does not need
to be a factor of 2 or 4. You can use any (even) image size for the
downscaling. It is a cropping operation in the Fourier domain, so this
should not affect the lower frequencies at all.
HTH,
S
> Dear Sjors,
> For one of our projects that use a large box size, we wanted to run our
> jobs in bin 2 or bin 4 in the earlier runs and then use the full size in
> the last. Can we do that in the command line or need to do the binned
> version spearately?
> Thanks.
> Qiu-Xing
>
>
>
>
>
> On 11/10/14, 8:31 AM, "Sjors Scheres" <[log in to unmask]> wrote:
>
>>Hola Marta (CC ccp-em),
>>Good question. Before iteration 1 starts, relion will calculate:
>>1) the power spectra of all individual particles, and their average
>>2) the average image of all unaligned particles
>>3) the power spectrum of this average image
>>The initial tau estimate is taken from 3). The initial sigma is
>> calculated
>>from 1) minus 3). The reason for this subtraction is to remove some power
>>in the lowest frequencies of the individual particles' power spectra,
>>which is due to signal, not noise. These are crude approximations to
>> sigma
>>and tau, but that doesn't really matter as the iterative estimation
>> inside
>>the refinement will quickly converge to better values.
>>HTH,
>>S
>>
>>
>>
>>> Dear Sjors,
>>> I am a bit confused on how the power of the noise as well as the power
>>of the signal is calculated in relion. I can understand how it is learnt
>>from the data in the iterative process, but how is it given initially?
>>For example in the first initial 2D classification when no model is
>>given how is the power of the signal estimated for the first iteration?
>>Similarly, how is the noise calculated? Is it done by taking the gray
>>values outside the mask in the boxed particles and averaging together
>>those that belong to the same group?
>>> And the signal?
>>> Si crees que esto pueda ser de public interest, lo pregunto en la mail
>>list. Pero ahí todas las preguntas me parecen más bien prácticas.
>>> Gracias
>>> Marta
>>
>>
>>--
>>Sjors Scheres
>>MRC Laboratory of Molecular Biology
>>Francis Crick Avenue, Cambridge Biomedical Campus
>>Cambridge CB2 0QH, U.K.
>>tel: +44 (0)1223 267061
>>http://www2.mrc-lmb.cam.ac.uk/groups/scheres
>
>
> ________________________________
>
> UT Southwestern Medical Center
> The future of medicine, today.
>
>
--
Sjors Scheres
MRC Laboratory of Molecular Biology
Francis Crick Avenue, Cambridge Biomedical Campus
Cambridge CB2 0QH, U.K.
tel: +44 (0)1223 267061
http://www2.mrc-lmb.cam.ac.uk/groups/scheres
|