[phenixbb] Local sharpening error

Ricardo Righetto ricardorighetto at gmail.com
Thu Jan 18 14:00:31 PST 2018


Hi Tom,

Interesting approach. I agree that providing the sequence gives a more
accurate volume estimation. May I suggest that alternatively the user could
provide the MW and from that the volume could be estimated? This is how
it's done in FREALIGN to calculate a volume-normalized FSC (Single-Particle
Wiener filter).

ps: I just ran the local sharpening with default parameters, providing the
actual sequence and number of ncs_copies for my protein. It didn't take
long but my volume is small (96x96x96), however it seems to be very
over-sharpened in this particular case. The global sharpening on the other
hand looks really nice!

Thank you very much for the help.

Best wishes,


--
Ricardo Diogo Righetto

2018-01-18 22:16 GMT+01:00 Tom Terwilliger <
tterwilliger at newmexicoconsortium.org>:

> Hi Ricardo,
>
> Yes, perfectly good question. The sequence is to guess the fraction of
> volume occupied by the molecule. That in turn is used to try and segment
> the map and figure out where the important parts of the map are located.
> Those are where local sharpening is done.
>
> I should note that in my experience local sharpening in auto_sharpen only
> occasionally helps beyond global sharpening and just takes a long time...
>
> All the best,
> Tom T
>
> On Thu, Jan 18, 2018 at 1:59 PM, Ricardo Righetto <
> ricardorighetto at gmail.com> wrote:
>
>> OK Tom, thanks a lot for the very quick reply! I'll try that.
>>
>> But now I'd like to understand - why is the sequence file necessary at
>> all? I thought the local sharpening approach just did the same as the
>> global sharpening, only sliding a small box throughout the volume. It seems
>> then that this is not the case?
>>
>> Best wishes,
>>
>>
>> --
>> Ricardo Diogo Righetto
>>
>> 2018-01-18 21:53 GMT+01:00 Tom Terwilliger <tterwilliger at newmexicoconsort
>> ium.org>:
>>
>>> Hi Ricardo,
>>>
>>> I'm sorry, yes you do have to supply a sequence file for local
>>> sharpening.  I will fix the documentation.  If you want you can just put in
>>> a dummy sequence and supply a value for "solvent_content=xxx" and that will
>>> accomplish the same thing.
>>>
>>> Let me know if that doesn't do it!
>>> all the best,
>>> Tom T
>>>
>>> On Thu, Jan 18, 2018 at 1:42 PM, Ricardo Righetto <
>>> ricardorighetto at gmail.com> wrote:
>>>
>>>> Hi,
>>>>
>>>> I am trying to do local sharpening with phenix.auto_sharpen, with the
>>>> following command:
>>>>
>>>> phenix.auto_sharpen map.mrc resolution=4.0 local_sharpening=True
>>>>
>>>> Then it runs the global sharpening first as expected, but when it gets
>>>> to the local sharpening it crashes with the following error:
>>>>
>>>> Sorry: Please specify a sequence file with seq_file=myseq.seq
>>>>
>>>> Nowhere it says that I would need a .seq file in order to do local
>>>> sharpening (global sharpening works fine). Any ideas?
>>>>
>>>> Thanks,
>>>>
>>>> --
>>>> Ricardo Diogo Righetto
>>>>
>>>> _______________________________________________
>>>> phenixbb mailing list
>>>> phenixbb at phenix-online.org
>>>> http://phenix-online.org/mailman/listinfo/phenixbb
>>>> Unsubscribe: phenixbb-leave at phenix-online.org
>>>>
>>>
>>>
>>>
>>> --
>>> Thomas C Terwilliger
>>> Laboratory Fellow, Los Alamos National Laboratory
>>> Senior Scientist, New Mexico Consortium
>>> 100 Entrada Dr, Los Alamos, NM 87544
>>> <https://maps.google.com/?q=100+Entrada+Dr,+Los+Alamos,+NM+87544&entry=gmail&source=g>
>>> Email: tterwilliger at newmexicoconsortium.org
>>> Tel: 505-431-0033 <(505)%20431-0033>
>>>
>>>
>>
>
>
> --
> Thomas C Terwilliger
> Laboratory Fellow, Los Alamos National Laboratory
> Senior Scientist, New Mexico Consortium
> 100 Entrada Dr, Los Alamos, NM 87544
> <https://maps.google.com/?q=100+Entrada+Dr,+Los+Alamos,+NM+87544&entry=gmail&source=g>
> Email: tterwilliger at newmexicoconsortium.org
> Tel: 505-431-0033
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20180118/ef06b5f6/attachment.htm>


More information about the phenixbb mailing list