Re: [ccp4bb] Multi-lattice crystal data processing strategy for structure solution

[Rezaul Karim]

Hi Devbrat,Have you tried indexing using XDS multi-lattice processing? It is a very useful indexing process for multi-lattice and/or 2nd lattice interference with the major one. Best,Rezaul Karim, Ph.D.Structural Biology ScientistNurix Therapeutics Inc.Sent from my iPhoneOn Nov 15, 2023, at 6:22 PM, Devbrat Kumar  wrote:Dear Prof JohnThank you for your valuable time and insight. I will do as you have suggested.Warm Regards-Devbrat Kumar On Wed, Nov 15, 2023 at 11:28 PM John Bacik  wrote:
Hi Devbrat, here are a couple of other things to try:- When screening crystals use rastering to find regions of the sample where multiple lattices may be less problematic. If multiple lattices are observed, often regions on the crystal(s) close to the edge will not be as affected by twinning/multiple lattices. Also try using fine slicing if you are not already.- Try using AlphaFold to generate a model for the MR template.All the best,John





On Wednesday, November 15, 2023 at 06:10:06 AM CST, Phil Evans  wrote:



Is the space group really P2? P21 is MUCH more commonPhil> On 14 Nov 2023, at 15:55, Devbrat Kumar  wrote:> > sed data were integrated with the data reduction tool AIMLESS in the CCP4i2 suite.To unsubscribe from the CCP4BB list, click the following link:https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1This 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



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

Re: [ccp4bb] Multi-lattice crystal data processing strategy for structure solution

[Devbrat Kumar]

Dear Prof John

Thank you for your valuable time and insight. I will do as you have
suggested.

*Warm Regards-*
*Devbrat Kumar*





On Wed, Nov 15, 2023 at 11:28 PM John Bacik <
b45abf420e1f-dmarc-requ...@jiscmail.ac.uk> wrote:

>
> Hi Devbrat, here are a couple of other things to try:
>
> - When screening crystals use rastering to find regions of the sample
> where multiple lattices may be less problematic. If multiple lattices are
> observed, often regions on the crystal(s) close to the edge will not be as
> affected by twinning/multiple lattices. Also try using fine slicing if you
> are not already.
>
> - Try using AlphaFold to generate a model for the MR template.
>
> All the best,
> John
>
> On Wednesday, November 15, 2023 at 06:10:06 AM CST, Phil Evans <
> p...@mrc-lmb.cam.ac.uk> wrote:
>
>
> Is the space group really P2? P21 is MUCH more common
> Phil
>
> > On 14 Nov 2023, at 15:55, Devbrat Kumar  wrote:
> >
> > sed data were integrated with the data reduction tool AIMLESS in the
> CCP4i2 suite.
>
> 
>
> 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
>



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] Multi-lattice crystal data processing strategy for structure solution

[Devbrat Kumar]

Dear Prof Gerard

Thank you for your in-depth analysis of the possibilities I am facing. The
image slicing was at 0.1 degrees.

*Warm Regards-*
*Devbrat Kumar*





On Wed, Nov 15, 2023 at 5:21 PM Gerard Bricogne 
wrote:

> Dear Devbrat,
>
>  With the unit-cell geometry you have, namely a very long axis and two
> much shorter ones, you have to be exceedingly careful about how the data
> are
> collected, and in particular about the crystal orientation and image width.
>
>  If the long axis can be brought close to being parallel to the
> rotation
> axis - either because the crystal morphology tends to make this happen or
> (better) because you can use a multi-axis goniometer to orient it that way
> -
> the images will have rows of closely-spaced spots that should be resolvable
> if the detector is placed far enough. If on the other hand the long axis is
> at a large angle to the rotation axis, there will be image ranges where
> that
> axis gets close to being parallel to the beam, so that the separation of
> reflections along that long axis (in fact, along the short reciprocal axis)
> will depends on their angular distance. Unless the image width is small
> enough, there will be overlap of the spots for consecutive reflections
> along
> that short reciprocal axis. Indexing diagnostics may then give an
> impression
> that there is more than one lattice, but most of all, because two distinct
> reflexions may overlap into a single spot, many integrated intensities will
> be corrupted.
>
>  Perhaps this is not the case, but out of curiosity: what is the
> angular
> width of your images?
>
>
>  With best wishes,
>
>   Gerard.
>
> --
> On Tue, Nov 14, 2023 at 09:25:30PM +0530, Devbrat Kumar wrote:
> > Hello everyone,
> >
> > The issue with the crystal is its multi-lattice nature; even the
> truncated
> > protein, which has been crystallized, exhibits multi-lattice
> > characteristics (detectable only after XRD).
> >
> > I have multiple native and selenium datasets with similar unit cell
> > parameters. (One axis is excessively long.) The XRD images were processed
> > using XDS in the P2 spacegroup, with unit cell parameters as follows: a =
> > 27.75 Å, b = 293.9 Å, c = 34.6 Å, and β = 113°. The XDS-processed data
> were
> > integrated with the data reduction tool AIMLESS in the CCP4i2 suite. In
> > CRANK2, the Estimation of Matthews coefficient (Program used: GCX)
> > suggested the presence of monomer NCS with a solvent content of 63.6%.
> The
> > FA estimation and substructure detection were performed by SHELXC, which
> > detected a very weak signal below 3.4 Å. Substructure determination was
> > carried out using SHELXD, yielding a maximum figure of merit of 27.8
> after
> > 640 trials and suggesting 11 atoms in the substructure with an occupancy
> of
> > at least 25%. Phasing and substructure refinement were conducted using
> the
> > BP3 program, resulting in an FOM of 0.2. During hand determination, the
> > programs suggested combined DM (density modification) FOM and phasing CLD
> > score for hand one as 6.0 and for hand two as 4.783375. The tool didn't
> > choose the hand because the value is less than the threshold. Density
> > modification with Fourier recycling suggests that the final FOM for hand
> > one and hand two is 0.428 and 0.482, respectively, while REFMAC5 gives
> the
> > R factor and Rfree factor as 0.4262 and 0.4912.
> >
> > One of the MR templates (model with balbes) works(For MR, Identity with
> the
> > PDB template is 21%), but R & Rfree are stuck at 33 & 37 for the 2.7
> > Angstrom cut-off (the total resolution in the dataset is 2 Angstrom).
> The R
> > & Rfree is not decreasing for the dataset. I have played with detector
> > distances for spot resolution, but at one pHi the spots have merged as a
> > single spot, while at 90 degrees will give us the streak of spots.
> >
> > Looking forward to hearing from you regarding dataset processing ideas
> for
> > multi-lattice crystals(Native & Se dataset) and structure solution
> strategy.
> >
> > Thank you.
> > Regards
> > Devbrat
> >
> > 
> >
> > 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] Multi-lattice crystal data processing strategy for structure solution

[Devbrat Kumar]

Dear Professor,

Thank you for your valuable suggestions.

I have attempted to investigate the NCS operator, but it seems to be a
monomer. Nevertheless, I will reconsider your suggestion regarding this
matter.

I employed the concept of using high-resolution data (2 Angstroms) without
cutoff for model building and refinement cycling, along with phases from MR
(2.7 Angstroms). I will revisit this process.

I will follow your advice for the peak search. If I cannot find any leads,
I hope to seek your guidance.

Thank you again for your support.
*Warm Regards-*
*Devbrat Kumar*





On Wed, Nov 15, 2023 at 4:38 PM Eleanor Dodson <
176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:

> You seem pretty near to having solved your structure!
> Ignoring the data problems..
>
> Extra steps I might have used.
> 1) Self rotation function. (in CCP4I2 the task is under data analysis ..)
> Does it suggest a NCS operator?
> If so is this a two fold? which might mean you have a dimer..
>
>
> 2) Now you have a reasonable R factor try extending the resolution. the
> refinement program will weight down the high resolution less reliable data
> but the extra information might marginally improve the maps.
>
> 3) Use the calculated phases to run an anomalous map - REFMAC will produce
> the appropriate coefficients and you can display the map in COOT.
> Do a peaksearch and you should of course see the sites SHELX found..
> But sometimes you get peaks over S atoms and that makes you pretty
> confident that the sequence there is correct.
>
>
>
> On Wed, 15 Nov 2023 at 10:42, Kay Diederichs <
> kay.diederi...@uni-konstanz.de> wrote:
>
>> Hello Devbrat,
>>
>> your project is difficult and there is no magic bullet to solve its
>> problems. Your approach is good because it always pays off to process
>> the data carefully.
>> In this respect, let me make a few comments.
>> 1) you don't say why you call the diffraction patterns "multi-lattice".
>> What exactly do you mean by that? Non-merohedral
>> twinning? How many lattices superimposed and visible on all frames? Can
>> they be separately indexed by XDS
>> (see https://wiki.uni-konstanz.de/xds/index.php/Indexing)?
>> 2) "XDS processing" _is_ integrating; what AIMLESS does is called scaling.
>> 3) what do you mean by "monomer NCS"? NCS implies two or more copies of
>> the same molecule in the asymmetric unit.
>> These copies often form dimers, trimers, tetramers, ... by making
>> more-or-less strong and specific interactions.
>> 4) you've advanced amazingly far and it sounds to me that with a
>> combination of your refined MR model with the SAD data you
>> should be able to improve your solution. Look up the MR-SAD pipeline (for
>> SAD after MR and for model rebuilding using anomalous
>> data) of Crank2.
>>
>> If you want me to take a look at your raw data, upload the best datasets
>> (native and SeMet) to a cloud service and send me the link.
>>
>> Good luck,
>> Kay
>>
>> On Tue, 14 Nov 2023 21:25:30 +0530, Devbrat Kumar 
>> wrote:
>>
>> >Hello everyone,
>> >
>> >The issue with the crystal is its multi-lattice nature; even the
>> truncated
>> >protein, which has been crystallized, exhibits multi-lattice
>> >characteristics (detectable only after XRD).
>> >
>> >I have multiple native and selenium datasets with similar unit cell
>> >parameters. (One axis is excessively long.) The XRD images were processed
>> >using XDS in the P2 spacegroup, with unit cell parameters as follows: a =
>> >27.75 Å, b = 293.9 Å, c = 34.6 Å, and β = 113°. The XDS-processed data
>> were
>> >integrated with the data reduction tool AIMLESS in the CCP4i2 suite. In
>> >CRANK2, the Estimation of Matthews coefficient (Program used: GCX)
>> >suggested the presence of monomer NCS with a solvent content of 63.6%.
>> The
>> >FA estimation and substructure detection were performed by SHELXC, which
>> >detected a very weak signal below 3.4 Å. Substructure determination was
>> >carried out using SHELXD, yielding a maximum figure of merit of 27.8
>> after
>> >640 trials and suggesting 11 atoms in the substructure with an occupancy
>> of
>> >at least 25%. Phasing and substructure refinement were conducted using
>> the
>> >BP3 program, resulting in an FOM of 0.2. During hand determination, the
>> >programs suggested combined DM (density modification) FOM and phasing CLD
>> >score for hand one as 6.0 and for hand two as 4.783375. The tool didn't
>> >choose the hand because the value is less than the threshold. Density
>> >modification with Fourier recycling suggests that the final FOM for hand
>> >one and hand two is 0.428 and 0.482, respectively, while REFMAC5 gives
>> the
>> >R factor and Rfree factor as 0.4262 and 0.4912.
>> >
>> >One of the MR templates (model with balbes) works(For MR, Identity with
>> the
>> >PDB template is 21%), but R & Rfree are stuck at 33 & 37 for the 2.7
>> >Angstrom cut-off (the total resolution in the dataset is 2 Angstrom).
>> The R
>> >& Rfree is not decreasing for the dataset. I have played 

Re: [ccp4bb] Multi-lattice crystal data processing strategy for structure solution

[Devbrat Kumar]

Hello Professor Kay

Thank you for your in-depth and valuable suggestions.

1. I used to observe multiple spots at one Phi, and at 90 degrees, I would
find a streak of spots. The crystal used to be spade-shaped. The spots are
visible in a count of 8 (matching the total number of crystal lattices).
2. I tried MR-SAD, but couldn't find the selenium substructure at
the Sulpher position of the sequence.
3. I will follow your suggestion to read about XDS  and conduct the
integration using XDS.
4. If I do not find any leads, I will email you the dataset.

*Warm Regards-*
*Devbrat *



On Wed, Nov 15, 2023 at 4:12 PM Kay Diederichs <
kay.diederi...@uni-konstanz.de> wrote:

> Hello Devbrat,
>
> your project is difficult and there is no magic bullet to solve its
> problems. Your approach is good because it always pays off to process
> the data carefully.
> In this respect, let me make a few comments.
> 1) you don't say why you call the diffraction patterns "multi-lattice".
> What exactly do you mean by that? Non-merohedral
> twinning? How many lattices superimposed and visible on all frames? Can
> they be separately indexed by XDS
> (see https://wiki.uni-konstanz.de/xds/index.php/Indexing)?
> 2) "XDS processing" _is_ integrating; what AIMLESS does is called scaling.
> 3) what do you mean by "monomer NCS"? NCS implies two or more copies of
> the same molecule in the asymmetric unit.
> These copies often form dimers, trimers, tetramers, ... by making
> more-or-less strong and specific interactions.
> 4) you've advanced amazingly far and it sounds to me that with a
> combination of your refined MR model with the SAD data you
> should be able to improve your solution. Look up the MR-SAD pipeline (for
> SAD after MR and for model rebuilding using anomalous
> data) of Crank2.
>
> If you want me to take a look at your raw data, upload the best datasets
> (native and SeMet) to a cloud service and send me the link.
>
> Good luck,
> Kay
>
> On Tue, 14 Nov 2023 21:25:30 +0530, Devbrat Kumar 
> wrote:
>
> >Hello everyone,
> >
> >The issue with the crystal is its multi-lattice nature; even the truncated
> >protein, which has been crystallized, exhibits multi-lattice
> >characteristics (detectable only after XRD).
> >
> >I have multiple native and selenium datasets with similar unit cell
> >parameters. (One axis is excessively long.) The XRD images were processed
> >using XDS in the P2 spacegroup, with unit cell parameters as follows: a =
> >27.75 Å, b = 293.9 Å, c = 34.6 Å, and β = 113°. The XDS-processed data
> were
> >integrated with the data reduction tool AIMLESS in the CCP4i2 suite. In
> >CRANK2, the Estimation of Matthews coefficient (Program used: GCX)
> >suggested the presence of monomer NCS with a solvent content of 63.6%. The
> >FA estimation and substructure detection were performed by SHELXC, which
> >detected a very weak signal below 3.4 Å. Substructure determination was
> >carried out using SHELXD, yielding a maximum figure of merit of 27.8 after
> >640 trials and suggesting 11 atoms in the substructure with an occupancy
> of
> >at least 25%. Phasing and substructure refinement were conducted using the
> >BP3 program, resulting in an FOM of 0.2. During hand determination, the
> >programs suggested combined DM (density modification) FOM and phasing CLD
> >score for hand one as 6.0 and for hand two as 4.783375. The tool didn't
> >choose the hand because the value is less than the threshold. Density
> >modification with Fourier recycling suggests that the final FOM for hand
> >one and hand two is 0.428 and 0.482, respectively, while REFMAC5 gives the
> >R factor and Rfree factor as 0.4262 and 0.4912.
> >
> >One of the MR templates (model with balbes) works(For MR, Identity with
> the
> >PDB template is 21%), but R & Rfree are stuck at 33 & 37 for the 2.7
> >Angstrom cut-off (the total resolution in the dataset is 2 Angstrom). The
> R
> >& Rfree is not decreasing for the dataset. I have played with detector
> >distances for spot resolution, but at one pHi the spots have merged as a
> >single spot, while at 90 degrees will give us the streak of spots.
> >
> >Looking forward to hearing from you regarding dataset processing ideas for
> >multi-lattice crystals(Native & Se dataset) and structure solution
> strategy.
> >
> >Thank you.
> >Regards
> >Devbrat
> >
> >
> >
> >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 

Re: [ccp4bb] Multi-lattice crystal data processing strategy for structure solution

[Devbrat Kumar]

Dear Professor Phil

Thank you for highlighting and providing input. Yes, The space group is P21.

*Warm Regards-*
*Devbrat Kumar*





On Wed, Nov 15, 2023 at 5:39 PM Phil Evans  wrote:

> Is the space group really P2? P21 is MUCH more common
> Phil
>
> > On 14 Nov 2023, at 15:55, Devbrat Kumar  wrote:
> >
> > sed data were integrated with the data reduction tool AIMLESS in the
> CCP4i2 suite.
>
> 
>
> 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] Multi-lattice crystal data processing strategy for structure solution

[John Bacik]

 
Hi Devbrat, here are a couple of other things to try:
- When screening crystals use rastering to find regions of the sample where 
multiple lattices may be less problematic. If multiple lattices are observed, 
often regions on the crystal(s) close to the edge will not be as affected by 
twinning/multiple lattices. Also try using fine slicing if you are not already.
- Try using AlphaFold to generate a model for the MR template.
All the best,John
On Wednesday, November 15, 2023 at 06:10:06 AM CST, Phil Evans 
 wrote:  
 
 Is the space group really P2? P21 is MUCH more common
Phil

> On 14 Nov 2023, at 15:55, Devbrat Kumar  wrote:
> 
> sed data were integrated with the data reduction tool AIMLESS in the CCP4i2 
> suite.



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] Multi-lattice crystal data processing strategy for structure solution

[Phil Evans]

Is the space group really P2? P21 is MUCH more common
Phil

> On 14 Nov 2023, at 15:55, Devbrat Kumar  wrote:
> 
> sed data were integrated with the data reduction tool AIMLESS in the CCP4i2 
> suite.



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] Multi-lattice crystal data processing strategy for structure solution

[Gerard Bricogne]

Dear Devbrat,

 With the unit-cell geometry you have, namely a very long axis and two
much shorter ones, you have to be exceedingly careful about how the data are
collected, and in particular about the crystal orientation and image width.

 If the long axis can be brought close to being parallel to the rotation
axis - either because the crystal morphology tends to make this happen or
(better) because you can use a multi-axis goniometer to orient it that way -
the images will have rows of closely-spaced spots that should be resolvable
if the detector is placed far enough. If on the other hand the long axis is
at a large angle to the rotation axis, there will be image ranges where that
axis gets close to being parallel to the beam, so that the separation of
reflections along that long axis (in fact, along the short reciprocal axis)
will depends on their angular distance. Unless the image width is small
enough, there will be overlap of the spots for consecutive reflections along
that short reciprocal axis. Indexing diagnostics may then give an impression
that there is more than one lattice, but most of all, because two distinct
reflexions may overlap into a single spot, many integrated intensities will
be corrupted. 

 Perhaps this is not the case, but out of curiosity: what is the angular
width of your images?


 With best wishes,

  Gerard.

--
On Tue, Nov 14, 2023 at 09:25:30PM +0530, Devbrat Kumar wrote:
> Hello everyone,
> 
> The issue with the crystal is its multi-lattice nature; even the truncated
> protein, which has been crystallized, exhibits multi-lattice
> characteristics (detectable only after XRD).
> 
> I have multiple native and selenium datasets with similar unit cell
> parameters. (One axis is excessively long.) The XRD images were processed
> using XDS in the P2 spacegroup, with unit cell parameters as follows: a =
> 27.75 Å, b = 293.9 Å, c = 34.6 Å, and β = 113°. The XDS-processed data were
> integrated with the data reduction tool AIMLESS in the CCP4i2 suite. In
> CRANK2, the Estimation of Matthews coefficient (Program used: GCX)
> suggested the presence of monomer NCS with a solvent content of 63.6%. The
> FA estimation and substructure detection were performed by SHELXC, which
> detected a very weak signal below 3.4 Å. Substructure determination was
> carried out using SHELXD, yielding a maximum figure of merit of 27.8 after
> 640 trials and suggesting 11 atoms in the substructure with an occupancy of
> at least 25%. Phasing and substructure refinement were conducted using the
> BP3 program, resulting in an FOM of 0.2. During hand determination, the
> programs suggested combined DM (density modification) FOM and phasing CLD
> score for hand one as 6.0 and for hand two as 4.783375. The tool didn't
> choose the hand because the value is less than the threshold. Density
> modification with Fourier recycling suggests that the final FOM for hand
> one and hand two is 0.428 and 0.482, respectively, while REFMAC5 gives the
> R factor and Rfree factor as 0.4262 and 0.4912.
> 
> One of the MR templates (model with balbes) works(For MR, Identity with the
> PDB template is 21%), but R & Rfree are stuck at 33 & 37 for the 2.7
> Angstrom cut-off (the total resolution in the dataset is 2 Angstrom). The R
> & Rfree is not decreasing for the dataset. I have played with detector
> distances for spot resolution, but at one pHi the spots have merged as a
> single spot, while at 90 degrees will give us the streak of spots.
> 
> Looking forward to hearing from you regarding dataset processing ideas for
> multi-lattice crystals(Native & Se dataset) and structure solution strategy.
> 
> Thank you.
> Regards
> Devbrat
> 
> 
> 
> 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] Multi-lattice crystal data processing strategy for structure solution

[Eleanor Dodson]

You seem pretty near to having solved your structure!
Ignoring the data problems..

Extra steps I might have used.
1) Self rotation function. (in CCP4I2 the task is under data analysis ..)
Does it suggest a NCS operator?
If so is this a two fold? which might mean you have a dimer..


2) Now you have a reasonable R factor try extending the resolution. the
refinement program will weight down the high resolution less reliable data
but the extra information might marginally improve the maps.

3) Use the calculated phases to run an anomalous map - REFMAC will produce
the appropriate coefficients and you can display the map in COOT.
Do a peaksearch and you should of course see the sites SHELX found..
But sometimes you get peaks over S atoms and that makes you pretty
confident that the sequence there is correct.



On Wed, 15 Nov 2023 at 10:42, Kay Diederichs 
wrote:

> Hello Devbrat,
>
> your project is difficult and there is no magic bullet to solve its
> problems. Your approach is good because it always pays off to process
> the data carefully.
> In this respect, let me make a few comments.
> 1) you don't say why you call the diffraction patterns "multi-lattice".
> What exactly do you mean by that? Non-merohedral
> twinning? How many lattices superimposed and visible on all frames? Can
> they be separately indexed by XDS
> (see https://wiki.uni-konstanz.de/xds/index.php/Indexing)?
> 2) "XDS processing" _is_ integrating; what AIMLESS does is called scaling.
> 3) what do you mean by "monomer NCS"? NCS implies two or more copies of
> the same molecule in the asymmetric unit.
> These copies often form dimers, trimers, tetramers, ... by making
> more-or-less strong and specific interactions.
> 4) you've advanced amazingly far and it sounds to me that with a
> combination of your refined MR model with the SAD data you
> should be able to improve your solution. Look up the MR-SAD pipeline (for
> SAD after MR and for model rebuilding using anomalous
> data) of Crank2.
>
> If you want me to take a look at your raw data, upload the best datasets
> (native and SeMet) to a cloud service and send me the link.
>
> Good luck,
> Kay
>
> On Tue, 14 Nov 2023 21:25:30 +0530, Devbrat Kumar 
> wrote:
>
> >Hello everyone,
> >
> >The issue with the crystal is its multi-lattice nature; even the truncated
> >protein, which has been crystallized, exhibits multi-lattice
> >characteristics (detectable only after XRD).
> >
> >I have multiple native and selenium datasets with similar unit cell
> >parameters. (One axis is excessively long.) The XRD images were processed
> >using XDS in the P2 spacegroup, with unit cell parameters as follows: a =
> >27.75 Å, b = 293.9 Å, c = 34.6 Å, and β = 113°. The XDS-processed data
> were
> >integrated with the data reduction tool AIMLESS in the CCP4i2 suite. In
> >CRANK2, the Estimation of Matthews coefficient (Program used: GCX)
> >suggested the presence of monomer NCS with a solvent content of 63.6%. The
> >FA estimation and substructure detection were performed by SHELXC, which
> >detected a very weak signal below 3.4 Å. Substructure determination was
> >carried out using SHELXD, yielding a maximum figure of merit of 27.8 after
> >640 trials and suggesting 11 atoms in the substructure with an occupancy
> of
> >at least 25%. Phasing and substructure refinement were conducted using the
> >BP3 program, resulting in an FOM of 0.2. During hand determination, the
> >programs suggested combined DM (density modification) FOM and phasing CLD
> >score for hand one as 6.0 and for hand two as 4.783375. The tool didn't
> >choose the hand because the value is less than the threshold. Density
> >modification with Fourier recycling suggests that the final FOM for hand
> >one and hand two is 0.428 and 0.482, respectively, while REFMAC5 gives the
> >R factor and Rfree factor as 0.4262 and 0.4912.
> >
> >One of the MR templates (model with balbes) works(For MR, Identity with
> the
> >PDB template is 21%), but R & Rfree are stuck at 33 & 37 for the 2.7
> >Angstrom cut-off (the total resolution in the dataset is 2 Angstrom). The
> R
> >& Rfree is not decreasing for the dataset. I have played with detector
> >distances for spot resolution, but at one pHi the spots have merged as a
> >single spot, while at 90 degrees will give us the streak of spots.
> >
> >Looking forward to hearing from you regarding dataset processing ideas for
> >multi-lattice crystals(Native & Se dataset) and structure solution
> strategy.
> >
> >Thank you.
> >Regards
> >Devbrat
> >
> >
> >
> >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] Multi-lattice crystal data processing strategy for structure solution

[Kay Diederichs]

Hello Devbrat,

your project is difficult and there is no magic bullet to solve its problems. 
Your approach is good because it always pays off to process 
the data carefully. 
In this respect, let me make a few comments.
1) you don't say why you call the diffraction patterns "multi-lattice". What 
exactly do you mean by that? Non-merohedral 
twinning? How many lattices superimposed and visible on all frames? Can they be 
separately indexed by XDS 
(see https://wiki.uni-konstanz.de/xds/index.php/Indexing)?
2) "XDS processing" _is_ integrating; what AIMLESS does is called scaling.
3) what do you mean by "monomer NCS"? NCS implies two or more copies of the 
same molecule in the asymmetric unit. 
These copies often form dimers, trimers, tetramers, ... by making more-or-less 
strong and specific interactions.
4) you've advanced amazingly far and it sounds to me that with a combination of 
your refined MR model with the SAD data you
should be able to improve your solution. Look up the MR-SAD pipeline (for SAD 
after MR and for model rebuilding using anomalous 
data) of Crank2.

If you want me to take a look at your raw data, upload the best datasets 
(native and SeMet) to a cloud service and send me the link.

Good luck,
Kay

On Tue, 14 Nov 2023 21:25:30 +0530, Devbrat Kumar  
wrote:

>Hello everyone,
>
>The issue with the crystal is its multi-lattice nature; even the truncated
>protein, which has been crystallized, exhibits multi-lattice
>characteristics (detectable only after XRD).
>
>I have multiple native and selenium datasets with similar unit cell
>parameters. (One axis is excessively long.) The XRD images were processed
>using XDS in the P2 spacegroup, with unit cell parameters as follows: a =
>27.75 Å, b = 293.9 Å, c = 34.6 Å, and β = 113°. The XDS-processed data were
>integrated with the data reduction tool AIMLESS in the CCP4i2 suite. In
>CRANK2, the Estimation of Matthews coefficient (Program used: GCX)
>suggested the presence of monomer NCS with a solvent content of 63.6%. The
>FA estimation and substructure detection were performed by SHELXC, which
>detected a very weak signal below 3.4 Å. Substructure determination was
>carried out using SHELXD, yielding a maximum figure of merit of 27.8 after
>640 trials and suggesting 11 atoms in the substructure with an occupancy of
>at least 25%. Phasing and substructure refinement were conducted using the
>BP3 program, resulting in an FOM of 0.2. During hand determination, the
>programs suggested combined DM (density modification) FOM and phasing CLD
>score for hand one as 6.0 and for hand two as 4.783375. The tool didn't
>choose the hand because the value is less than the threshold. Density
>modification with Fourier recycling suggests that the final FOM for hand
>one and hand two is 0.428 and 0.482, respectively, while REFMAC5 gives the
>R factor and Rfree factor as 0.4262 and 0.4912.
>
>One of the MR templates (model with balbes) works(For MR, Identity with the
>PDB template is 21%), but R & Rfree are stuck at 33 & 37 for the 2.7
>Angstrom cut-off (the total resolution in the dataset is 2 Angstrom). The R
>& Rfree is not decreasing for the dataset. I have played with detector
>distances for spot resolution, but at one pHi the spots have merged as a
>single spot, while at 90 degrees will give us the streak of spots.
>
>Looking forward to hearing from you regarding dataset processing ideas for
>multi-lattice crystals(Native & Se dataset) and structure solution strategy.
>
>Thank you.
>Regards
>Devbrat
>
>
>
>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/

[ccp4bb] Multi-lattice crystal data processing strategy for structure solution

[Devbrat Kumar]

Hello everyone,

The issue with the crystal is its multi-lattice nature; even the truncated
protein, which has been crystallized, exhibits multi-lattice
characteristics (detectable only after XRD).

I have multiple native and selenium datasets with similar unit cell
parameters. (One axis is excessively long.) The XRD images were processed
using XDS in the P2 spacegroup, with unit cell parameters as follows: a =
27.75 Å, b = 293.9 Å, c = 34.6 Å, and β = 113°. The XDS-processed data were
integrated with the data reduction tool AIMLESS in the CCP4i2 suite. In
CRANK2, the Estimation of Matthews coefficient (Program used: GCX)
suggested the presence of monomer NCS with a solvent content of 63.6%. The
FA estimation and substructure detection were performed by SHELXC, which
detected a very weak signal below 3.4 Å. Substructure determination was
carried out using SHELXD, yielding a maximum figure of merit of 27.8 after
640 trials and suggesting 11 atoms in the substructure with an occupancy of
at least 25%. Phasing and substructure refinement were conducted using the
BP3 program, resulting in an FOM of 0.2. During hand determination, the
programs suggested combined DM (density modification) FOM and phasing CLD
score for hand one as 6.0 and for hand two as 4.783375. The tool didn't
choose the hand because the value is less than the threshold. Density
modification with Fourier recycling suggests that the final FOM for hand
one and hand two is 0.428 and 0.482, respectively, while REFMAC5 gives the
R factor and Rfree factor as 0.4262 and 0.4912.

One of the MR templates (model with balbes) works(For MR, Identity with the
PDB template is 21%), but R & Rfree are stuck at 33 & 37 for the 2.7
Angstrom cut-off (the total resolution in the dataset is 2 Angstrom). The R
& Rfree is not decreasing for the dataset. I have played with detector
distances for spot resolution, but at one pHi the spots have merged as a
single spot, while at 90 degrees will give us the streak of spots.

Looking forward to hearing from you regarding dataset processing ideas for
multi-lattice crystals(Native & Se dataset) and structure solution strategy.

Thank you.
Regards
Devbrat



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/