Re: [ccp4bb] Ideas for improving refinement with anisotopic data.

2020-06-22 Thread vincent Chaptal 

Dear Matthew,

at your resolution, 1.7A, you are seeing a large difference in 
resolution limits. I'd be curious to know how much "anisotropy" you have 
in the single number anisoB. Could you share it?
As to building felxible loops or peptide, I suspect you anisoB to be 
mild (<40A2 for your resolution) and therefore the effect on map 
artefacts to also be mild. Including the high res data is definitely 
your best shot at getting better maps.


Good luck with your modeling.
Vincent

Le 21/06/2020 à 19:11, Gerard Bricogne a écrit :

Dear Matthew,

  You seem to have been unlucky with this data collection. STARANISO
would be prepared to see significant data extending to 2.0A along a*, 2.7A
along b*, and 1.7A along c*. The latter limit, however, involves some guess
work via the fitting by an ellipsoid of a cut-off surface that is very much
perturbed by the fact that you have a both a cusp and a truncation by the
detector edges in that direction. Without those, it looks quite probable
that you would have good solid data extending beyond 1.6A - if only you
could redo the experiment in such a way as to actually measure them ;-) .

  As Eleanor said, it is a bad idea to make decisions about data on the
basis of the "downstream" criteria you were mentioning. STARANISO is not
just a numerical program that estimates numbers and suggests some sensible
actions that it will carry out for you: it is also a 3D visualisation tool
that shows you much more about your data and its possible shortcomings than
any Table1-style numbers ever will. If there are indeed deficiencies in the
data, like here, these 3D pictures will often point out what shortcomings in
your experiment are responsible for them. In that case, a new experiment, if
feasible, is always better than any surgery on the deficient data.

  Please feel free to get in touch off-line if you would like more
details, such as patterns of outliers that we have observed.


  With best wishes,

  The STARANISO developers.

--
On Sun, Jun 21, 2020 at 05:33:53PM +0100, Eleanor Dodson wrote:

*Please* dont throw good 1.8A data for the sake of statistics!
You should see more detail along certain directions
You will publish your structure providing honest details of the anisotropy
(I hope..) but it is the map quality that matters ..
Eleanor

On Sun, 21 Jun 2020 at 15:16, Matthew Snee <
matthew.s...@postgrad.manchester.ac.uk> wrote:


Hi everyone.

I have an x ray structure that I am finishing up, and there are a few
ambiguous regions where the peptide is poorly resolved.

The data is highly anisotrophic, and requires truncation to around 2.4A to
achieve acceptable merging stats, although there is data in the "good"
direction going as high as 1.8-1.9A (determined by merging stats when using
elliptical cutoff).

I have tried feeding my integration outputs to STARANISO with both the
elliptical and spherical cutoffs, but neither produce better results than
Xia2 Dials, as both need to be truncated further than 2.3A before they give
the same refinement stats (i.e it's best to just let refmac/phenix.refine
deal with the anistropy).

As I understand it, anisotrophy can lead to loss of high resolution detail
because weak observations from the high res shells in the bad direction are
down-weighted, So any tricks to improve map quality (or conversely refining
data with poor completeness) would be appreciated.

I'm not holding out hope that I can deposit anything better than 2.3A, but
Improving the maps might really help me with some of these troublesome
loops :)

Cheers

Matthew.

Sent from Outlook Mobile 

--

To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/


--

Vincent Chaptal, PhD

Director of GdR APPICOM

Drug Resistance and Membrane Proteins Lab


MMSB -UMR5086

7 passage du Vercors

69007 LYON

FRANCE

+33 4 37 65 29 01

http://www.appicom.cnrs.fr

http://mmsb.cnrs.fr/en/





To unsubscribe from the CCP4BB list, click the following link:

Re: [ccp4bb] Ideas for improving refinement with anisotopic data.

2020-06-21 Thread Gerard Bricogne 
Dear Matthew,

 You seem to have been unlucky with this data collection. STARANISO
would be prepared to see significant data extending to 2.0A along a*, 2.7A
along b*, and 1.7A along c*. The latter limit, however, involves some guess
work via the fitting by an ellipsoid of a cut-off surface that is very much
perturbed by the fact that you have a both a cusp and a truncation by the
detector edges in that direction. Without those, it looks quite probable
that you would have good solid data extending beyond 1.6A - if only you
could redo the experiment in such a way as to actually measure them ;-) .

 As Eleanor said, it is a bad idea to make decisions about data on the
basis of the "downstream" criteria you were mentioning. STARANISO is not
just a numerical program that estimates numbers and suggests some sensible
actions that it will carry out for you: it is also a 3D visualisation tool
that shows you much more about your data and its possible shortcomings than
any Table1-style numbers ever will. If there are indeed deficiencies in the
data, like here, these 3D pictures will often point out what shortcomings in
your experiment are responsible for them. In that case, a new experiment, if
feasible, is always better than any surgery on the deficient data.

 Please feel free to get in touch off-line if you would like more
details, such as patterns of outliers that we have observed.


 With best wishes,

 The STARANISO developers.

--
On Sun, Jun 21, 2020 at 05:33:53PM +0100, Eleanor Dodson wrote:
> *Please* dont throw good 1.8A data for the sake of statistics!
> You should see more detail along certain directions
> You will publish your structure providing honest details of the anisotropy
> (I hope..) but it is the map quality that matters ..
> Eleanor
> 
> On Sun, 21 Jun 2020 at 15:16, Matthew Snee <
> matthew.s...@postgrad.manchester.ac.uk> wrote:
> 
> > Hi everyone.
> >
> > I have an x ray structure that I am finishing up, and there are a few
> > ambiguous regions where the peptide is poorly resolved.
> >
> > The data is highly anisotrophic, and requires truncation to around 2.4A to
> > achieve acceptable merging stats, although there is data in the "good"
> > direction going as high as 1.8-1.9A (determined by merging stats when using
> > elliptical cutoff).
> >
> > I have tried feeding my integration outputs to STARANISO with both the
> > elliptical and spherical cutoffs, but neither produce better results than
> > Xia2 Dials, as both need to be truncated further than 2.3A before they give
> > the same refinement stats (i.e it's best to just let refmac/phenix.refine
> > deal with the anistropy).
> >
> > As I understand it, anisotrophy can lead to loss of high resolution detail
> > because weak observations from the high res shells in the bad direction are
> > down-weighted, So any tricks to improve map quality (or conversely refining
> > data with poor completeness) would be appreciated.
> >
> > I'm not holding out hope that I can deposit anything better than 2.3A, but
> > Improving the maps might really help me with some of these troublesome
> > loops :)
> >
> > Cheers
> >
> > Matthew.
> >
> > Sent from Outlook Mobile 
> >
> > --
> >
> > To unsubscribe from the CCP4BB list, click the following link:
> > https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> >
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 
> This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing 
> list hosted by www.jiscmail.ac.uk, terms & conditions are available at 
> https://www.jiscmail.ac.uk/policyandsecurity/



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] Ideas for improving refinement with anisotopic data.

2020-06-21 Thread Eleanor Dodson 
*Please* dont throw good 1.8A data for the sake of statistics!
You should see more detail along certain directions
You will publish your structure providing honest details of the anisotropy
(I hope..) but it is the map quality that matters ..
Eleanor

On Sun, 21 Jun 2020 at 15:16, Matthew Snee <
matthew.s...@postgrad.manchester.ac.uk> wrote:

> Hi everyone.
>
> I have an x ray structure that I am finishing up, and there are a few
> ambiguous regions where the peptide is poorly resolved.
>
> The data is highly anisotrophic, and requires truncation to around 2.4A to
> achieve acceptable merging stats, although there is data in the "good"
> direction going as high as 1.8-1.9A (determined by merging stats when using
> elliptical cutoff).
>
> I have tried feeding my integration outputs to STARANISO with both the
> elliptical and spherical cutoffs, but neither produce better results than
> Xia2 Dials, as both need to be truncated further than 2.3A before they give
> the same refinement stats (i.e it's best to just let refmac/phenix.refine
> deal with the anistropy).
>
> As I understand it, anisotrophy can lead to loss of high resolution detail
> because weak observations from the high res shells in the bad direction are
> down-weighted, So any tricks to improve map quality (or conversely refining
> data with poor completeness) would be appreciated.
>
> I'm not holding out hope that I can deposit anything better than 2.3A, but
> Improving the maps might really help me with some of these troublesome
> loops :)
>
> Cheers
>
> Matthew.
>
> Sent from Outlook Mobile 
>
> --
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

[ccp4bb] Ideas for improving refinement with anisotopic data.

2020-06-21 Thread Matthew Snee 
Hi everyone.

I have an x ray structure that I am finishing up, and there are a few ambiguous 
regions where the peptide is poorly resolved.

The data is highly anisotrophic, and requires truncation to around 2.4A to 
achieve acceptable merging stats, although there is data in the "good" 
direction going as high as 1.8-1.9A (determined by merging stats when using 
elliptical cutoff).

I have tried feeding my integration outputs to STARANISO with both the 
elliptical and spherical cutoffs, but neither produce better results than Xia2 
Dials, as both need to be truncated further than 2.3A before they give the same 
refinement stats (i.e it's best to just let refmac/phenix.refine deal with the 
anistropy).

As I understand it, anisotrophy can lead to loss of high resolution detail 
because weak observations from the high res shells in the bad direction are 
down-weighted, So any tricks to improve map quality (or conversely refining 
data with poor completeness) would be appreciated.

I'm not holding out hope that I can deposit anything better than 2.3A, but 
Improving the maps might really help me with some of these troublesome loops :)

Cheers

Matthew.

Sent from Outlook Mobile



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/