Hi Rubbiya,
I think we need a small clarification about your problem, since each
answer you got this far aims to solve different issues. Do you mean that
your particles have duplicated entries in the star file (exact same line
twice), or that the particles have been picked twice (slightly different
coordinates were the boxes overlap over the same particle)?
In the first case the answer proposed by Joshua is perfectly fine.
Your particle stacks should not have duplicates in this case, and even
if they did any particle without a line in the star would be ignored in
the processing.
In the second case I don't see any fix using relion, other that start
over from the picking step and use more stringent settings to avoid that
kind of overlapping, like Michael suggested. I doubt relion has any way
to tell too similar particles, since ideally in single particle EM you
should have several groups of identical particles with as high as
possible SNR. You probably would need a external software to find pixel
by pixel overlapping areas between different images, and that seems way
more problematic that picking again 56k particles, less without
duplicates.
Hope that helps.
Regards,
Jesus Gomez de Segura
On 2016-08-02 04:58, Rubbiya Ali wrote:
> Hi everyone,
>
> My dataset has around ~56000 particles (negatively stained). I can
> see a large number of particles have duplicates. I am just wondering
> is there any automatic way (in Relion) to delete duplicates? If not
> any preferable way to deal with it? I checked old posts to find a way
> to deal with bad particles/duplicates, but could not find any related
> thread.
>
> Regards,
>
> Rubbiya Ali
> PhD I Hankamer Group I
> Institute for Molecular Bioscience I Queensland Bioscience Precinct I
> The University of Queensland I Brisbane Queensland 4072 I Australia
> Tel: +61 7 334 62015
> Email: [log in to unmask]
> Skype: rubbiya.ali
|