Dear Jesus,

We have checked the volume which we used for subtraction to make sure that we have used the right volume. It also only happens with data on which during a 3D classification or 3D refinement a mask is applied. Otherwise the subtraction goes fine.

Best wishes,

Laura
________________________________________
Van: Jesus GOMEZ [[log in to unmask]]
Verzonden: woensdag 10 augustus 2016 12:05
Aan: Bezouwen, L.S. van (Laura)
CC: [log in to unmask]
Onderwerp: Re: [ccpem] signal subtracting

Dear Laura Deniz,

  You describe an image with "a mask around it". Does that mean that all
the noise around the particle is gone? In that case sounds like you
subtracted that noise and the first thing I would check is the volume
you used for subtraction, make sure it's the right one representing only
the density you want to subtract (with a Gaussian edge) and no noise or
regions you are interested in. Hope that helps, but if it doesn't maybe
posting an actual image of the wrong looking particles might help people
recognize what went wrong.

  Best regards,

  Jesus Gomez de Segura




On 2016-08-05 13:49, Bezouwen, L.S. van (Laura) wrote:
> Dear all,
>
>  We work on a dataset where we want to focuss on two different
> regions, individually. Therefore we decided to use the option signal
> subtracting (Bai et al. 2015 eLife). For the first region the signal
> subtraction worked fine. But for the second region (the rest, with
> the
> other masked map) we started to have problems.
>
>  We started with a dataset from refinement, and after substraction of
> the masked map, we realised that the orginal file was adjusted. The
> first image looks strange, it is filtered and there is put a mask
> around it. Then we took an other dataset and we retried it. But the
> same occured. We realized that during refinement a mask was used, the
> mask is covering the area that now needs to be subtracted. We wonder
> does a mask during refinement influence the data.star file, or the
> subtraction. If so, how can we subtract that area? Or what else could
> have happend.
>
>  Best wishes,
>
>  Laura and Deniz