subject:"\[COOT\] Conflicting opinions on what constitutes an alpha1\-6 L\-fucose between Coot and Privateer"

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer: Building L-fucose fails in coot-0.9.8.94

2024-04-26 Thread Markus Meier

Hi Paul,

I just rebuilt coot-0.9.8.94, coot-0.9.8.93 and coot-0.9.8.92 from scratch,
deleting the previous installs.

The glyco validator in version coot-0.9.8.94 is indeed fixed. Thank you so much!

Also, coot-0.9.8.93 and coot-0.9.8.92 now build an alpha1-6 fucose correctly
into the electron density (not sure why I got beta1-6 before), producing
identical coordinates.

However, building the same fucose in coot-0.9.8.94 fails. First, the carbohydrate moiety
is inserted into a different orientation compared to coot-0.9.8.93. During the fit, it is
being moved to the correct density blob, however it is detached completely from the
carbohydrate tree. Distance between C1 (FUC) and O6 (NAG) results into 3.56 Å! After it
is built, pressing "Refine tree" does not include the fucose and also exporting
the tree does not contain it.

I could trace the problem partially down to the library file
share/coot/data/cho-acedrg/FUC-acedrg.cif

If I remove this file from the installation so that coot-0.9.8.94 cannot load
it, the fucose builds correctly, similar (but not identical) position to that
from coot-0.9.8.93.

I can get the same result (correct build) by importing
share/coot/lib/data/monomers/f/FUC.cif *after* FUC-acedrg.cif had been loaded.
I can re-induce broken behaviour (failed build) by re-importing FUC-acedrg.cif

I believe the two libraries overwrite each other. Curiously enough,
coot-0.9.8.93 contains the same libraries (identical versions), but I don't
experience this behaviour, i.e. the fucose is built correctly no matter which
library was loaded last.

I have attached the carbohydrate trees, one built by coot-0.9.8.93, the other
by coot-0.9.8.94 (detached fucose included).

Btw. one more funny things I noticed was that the Refmac monomer libraries are
loaded from $cleaned_dir:
coot-0.9.8.94/coot-Linux-GenuineIntel---gtk2-python/share/coot/lib/data

but the acedrg libraries are loaded from $fat_dir:
coot-0.9.8.94/coot-gtk2-python/share/coot/data

(not that it matters - except for debugging...)

Thanks again!
With best regards,

Markus

On 2024-04-23 20:23, Engin Özkan wrote:

Caution! This message was sent from outside the University of Manitoba.

I see. Thanks, Paul, for the explanation. Of course you are right, and I
see that too despite having the correct ALPHA1-6:

find_glycosidic_linkage_type() for A 911 NAG,A 913 FUC returns "BETA1-6"

This settles it for me.

(I am not sure how Markus got beta-fucoses built by Coot the in the
first place, though.)

Engin

On 4/23/24 8:03 PM, Paul Emsley wrote:

We are talking about different things, I think.

Coot's glyco builder builds alpha 1-6 linked FUC correctly.

Coot's glyco validator used to (until this evening) identify the alpha
1-6 linked FUC as a beta anomer (it no longer does).

See for example the G8 FUC in 8q5u: Coot 0.9.8.94 now prints:

ASN [spec: 1 "A" 297 ""]
pyr-ASN NAG [spec: 1 "G" 1 ""]
BETA1-4 NAG [spec: 1 "G" 2 ""]
BETA1-4 BMA [spec: 1 "G" 3 ""]
ALPHA1-3 MAN [spec: 1 "G" 4 ""]
BETA1-2 NAG [spec: 1 "G" 5 ""]
BETA1-6 MAN [spec: 1 "G" 6 ""]
BETA1-2 NAG [spec: 1 "G" 7 ""]
ALPHA1-6 FUC [spec: 1 "G" 8 ""]

Paul.

On 24/04/2024 01:38, Engin Özkan wrote:

CAUTION: This email originated from outside of the LMB:
.-owner-c...@jiscmail.ac.uk-.
Do not click links or open attachments unless you recognize the
sender and know the content is safe.
If you think this is a phishing email, please forward it to
phish...@mrc-lmb.cam.ac.uk
--

Hi Markus,

I am using Coot that comes with CCP4 8, which has FUC
(alpha-L-Fucose) with an oxygen on C1 in the axial position. Which is
the correct form, I believe.

Using Coot 0.9.8.93, I get my alpha1-6 and alpha1-3 (insect) Fucoses
inserted correctly. I had trouble understanding your description:
You mean you get Coot to insert a FUC with the glycosidic bond going
equatorial, instead of axial, to Fucose?

This is how it works for me:

https://www.dropbox.com/scl/fi/yjbzstojjivmfpqdlz2os/Screen-Recording-2024-04-23-at-7.32.14-PM.mov?rlkey=zvvk7apd19wk61kt21qyc42l5=0

I'll see what happens when I update to the new Coot version where
Paul has switched the logic.

By the way, privateer tells me I have zero errors and the correct
alpha anomer for Fucose.

Fuc913-a6-GlcNAc911-b-ASN258
WURCS=2.0/2,2,1/[a2122h-1b_1-5_2*NCC/3=O][a1221m-1a_1-5]/1-2/a6-b1
...

PDB Sugar Q Phi Theta Detected type
Cnf Ctx Ok?
- -- -- ---
--- -- --- -
test NAG-A-911 0.570 280.48 5.29 beta-D-aldopyranose
4c1 47.88 (n) yes
test FUC-A-913 0.535 92.53 177.32 alpha-L-aldopyranose
1c4 70.76 (n) yes
...
Wrong anomer: 0

Best,

Engin

On 4/18/24 4:35 PM, Markus Meier wrote:

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-24 Thread Markus Meier

Hi Engin,

thank you for all the trouble you went through to make this presentation! The
glycosidic bond in your presentation is indeed correct (alpha1-6).

Not sure how to interpret the terms axial or equatorial; I followed the
definition from here:
https://www.masterorganicchemistry.com/2018/02/19/the-big-damn-post-of-sugar-nomenclature

However, judging from the drawings here
(https://chem.libretexts.org/Bookshelves/Organic_Chemistry/Supplemental_Modules_(Organic_Chemistry)/Alkanes/Properties_of_Alkanes/Cycloalkanes/Rings%3A_cis_trans_and_axial_equatorial_relationships),
yes, the glycosidic bond in my case is equatorial and privateer reports it as
beta1-6.

Its not urgent at all, but if you find a moment, could you please send me the terminal output of
Coot it reports regarding your glycosidic linkages? To do that, start Coot from a terminal and load
your carbohydrate tree. Centre on one of the atoms in your carbohydrate tree, open the "Add
N-linked Glycan" panel in the Glyco module and press "Update for Current Residue".
Coot will print debug info on the terminal.
In the unlike case that terminal output is disabled, you can enable it in
Edit->Preferences->Other->Console->Display state commands in console?->Yes

Could you also send me the coordinates of your carbohydrate tree? It would like
to load it in my Coot version(s) and check the terminal output.

Thanks again!
With best regards,
Markus

On 2024-04-23 19:38, Engin Özkan wrote:

Caution! This message was sent from outside the University of Manitoba.

Hi Markus,

I am using Coot that comes with CCP4 8, which has FUC (alpha-L-Fucose)
with an oxygen on C1 in the axial position. Which is the correct form, I
believe.

Using Coot 0.9.8.93, I get my alpha1-6 and alpha1-3 (insect) Fucoses
inserted correctly. I had trouble understanding your description: You
mean you get Coot to insert a FUC with the glycosidic bond going
equatorial, instead of axial, to Fucose?

This is how it works for me:

https://www.dropbox.com/scl/fi/yjbzstojjivmfpqdlz2os/Screen-Recording-2024-04-23-at-7.32.14-PM.mov?rlkey=zvvk7apd19wk61kt21qyc42l5=0

I'll see what happens when I update to the new Coot version where Paul
has switched the logic.

By the way, privateer tells me I have zero errors and the correct alpha
anomer for Fucose.

Fuc913-a6-GlcNAc911-b-ASN258
WURCS=2.0/2,2,1/[a2122h-1b_1-5_2*NCC/3=O][a1221m-1a_1-5]/1-2/a6-b1
...

Best,

Engin

On 4/18/24 4:35 PM, Markus Meier wrote:

Hi Engin,

hmm, it appears I was using the monomer library from
https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/dependencies/monomers-2023-01-02-23:57:29.tar.gz
(it gets installed automatically with Coot's build-it script).

However, the files in question (FUC.cif, NAG.cif, BMA.cif and MAN.cif)
are identical to current CCP4 8.0.018 version of the library
(installed by CCP4 Update Manager) as well as version ccp4-8.0.018 and
version ccp4-8.0.016 on GitHub
(https://github.com/MonomerLibrary/monomers)

MD5 checksums on these files are
9cfba4ac24f056d648df8227316f788b monomers/f/FUC.cif
958ac5b19a92ceec604d233457944b2a monomers/n/NAG.cif
697dd5242334dd99dad9ce43133c6382 monomers/b/BMA.cif
9d68ff54296ef2f4622005efb4eb080b monomers/m/MAN.cif

If they are different on your machine, please let me know.

I will switch monomer libraries anyway and do some testing.

I also was switching back and forth between my locally compiled coot
(0.9.8.92) and the the CCP4 distributed version (0.9.8.93 EL). I will
check if there is a difference between the two.

Thank you very much for your suggestion!

With best regards,
Markus

On 2024-04-18 13:48, Engin Özkan wrote:

Caution! This message was sent from outside the University of Manitoba.

Hi Markus,

There were reports to the ccp4bb and phenixbb back in 2011 about
something similar. The monomer library files for FUC and/or the linkage
definitions for refmac/phenix were not right. I think. My recollection
is that this was corrected, and I could not reproduce your observation.

I also had private discussions with Garib Murshudov, who were correcting
the glycan monomer problems back in ~2011.

I could find old links to some of the discussions on bulletin board
archives:

https://phenix-online.org/pipermail/phenixbb/2011-November/017958.html

https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg13968.html

I am curious if something got reverted or if you have some ancient
library definitions somewhere. But I hope this gives some ideas.

Best,

Engin

On 4/17/24 3:07 PM, Markus Meier wrote:

Hi Paul,

thank you for implementing the Glyco module in Coot!

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-23 Thread Engin Özkan

I see. Thanks, Paul, for the explanation. Of course you are right, and I
see that too despite having the correct ALPHA1-6:

find_glycosidic_linkage_type() for A 911 NAG,A 913 FUC returns "BETA1-6"

This settles it for me.

(I am not sure how Markus got beta-fucoses built by Coot the in the
first place, though.)

Engin

On 4/23/24 8:03 PM, Paul Emsley wrote:

We are talking about different things, I think.

Coot's glyco builder builds alpha 1-6 linked FUC correctly.

Coot's glyco validator used to (until this evening) identify the alpha
1-6 linked FUC as a beta anomer (it no longer does).

See for example the G8 FUC in 8q5u: Coot 0.9.8.94 now prints:

Paul.

On 24/04/2024 01:38, Engin Özkan wrote:

Hi Markus,

I am using Coot that comes with CCP4 8, which has FUC
(alpha-L-Fucose) with an oxygen on C1 in the axial position. Which is
the correct form, I believe.

This is how it works for me:

https://www.dropbox.com/scl/fi/yjbzstojjivmfpqdlz2os/Screen-Recording-2024-04-23-at-7.32.14-PM.mov?rlkey=zvvk7apd19wk61kt21qyc42l5=0

I'll see what happens when I update to the new Coot version where
Paul has switched the logic.

By the way, privateer tells me I have zero errors and the correct
alpha anomer for Fucose.

Fuc913-a6-GlcNAc911-b-ASN258
WURCS=2.0/2,2,1/[a2122h-1b_1-5_2*NCC/3=O][a1221m-1a_1-5]/1-2/a6-b1
...

...
Wrong anomer: 0

Best,

Engin

On 4/18/24 4:35 PM, Markus Meier wrote:

Hi Engin,

However, the files in question (FUC.cif, NAG.cif, BMA.cif and
MAN.cif) are identical to current CCP4 8.0.018 version of the
library (installed by CCP4 Update Manager) as well as version
ccp4-8.0.018 and version ccp4-8.0.016 on GitHub
(https://github.com/MonomerLibrary/monomers)

If they are different on your machine, please let me know.

I will switch monomer libraries anyway and do some testing.

I also was switching back and forth between my locally compiled coot
(0.9.8.92) and the the CCP4 distributed version (0.9.8.93 EL). I
will check if there is a difference between the two.

Thank you very much for your suggestion!

With best regards,
Markus

On 2024-04-18 13:48, Engin Özkan wrote:
Caution! This message was sent from outside the University of
Manitoba.

Hi Markus,

There were reports to the ccp4bb and phenixbb back in 2011 about
something similar. The monomer library files for FUC and/or the
linkage

definitions for refmac/phenix were not right. I think. My recollection
is that this was corrected, and I could not reproduce your
observation.

I also had private discussions with Garib Murshudov, who were
correcting

the glycan monomer problems back in ~2011.

I could find old links to some of the discussions on bulletin board
archives:

https://phenix-online.org/pipermail/phenixbb/2011-November/017958.html

https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg13968.html

I am curious if something got reverted or if you have some ancient
library definitions somewhere. But I hope this gives some ideas.

Best,

Engin

On 4/17/24 3:07 PM, Markus Meier wrote:

Hi Paul,

thank you for implementing the Glyco module in Coot!

I used its "Add

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-23 Thread Paul Emsley

We are talking about different things, I think.

Coot's glyco builder builds alpha 1-6 linked FUC correctly.

Coot's glyco validator used to (until this evening) identify the alpha
1-6 linked FUC as a beta anomer (it no longer does).

See for example the G8 FUC in 8q5u: Coot 0.9.8.94 now prints:

Paul.

On 24/04/2024 01:38, Engin Özkan wrote:

CAUTION: This email originated from outside of the LMB:
.-owner-c...@jiscmail.ac.uk-.
Do not click links or open attachments unless you recognize the sender
and know the content is safe.
If you think this is a phishing email, please forward it to
phish...@mrc-lmb.cam.ac.uk

Hi Markus,

I am using Coot that comes with CCP4 8, which has FUC (alpha-L-Fucose)
with an oxygen on C1 in the axial position. Which is the correct form,
I believe.

Using Coot 0.9.8.93, I get my alpha1-6 and alpha1-3 (insect) Fucoses
inserted correctly. I had trouble understanding your description: You
mean you get Coot to insert a FUC with the glycosidic bond going
equatorial, instead of axial, to Fucose?

This is how it works for me:

https://www.dropbox.com/scl/fi/yjbzstojjivmfpqdlz2os/Screen-Recording-2024-04-23-at-7.32.14-PM.mov?rlkey=zvvk7apd19wk61kt21qyc42l5=0

I'll see what happens when I update to the new Coot version where Paul
has switched the logic.

By the way, privateer tells me I have zero errors and the correct
alpha anomer for Fucose.

Fuc913-a6-GlcNAc911-b-ASN258
WURCS=2.0/2,2,1/[a2122h-1b_1-5_2*NCC/3=O][a1221m-1a_1-5]/1-2/a6-b1
...

...
Wrong anomer: 0

Best,

Engin

On 4/18/24 4:35 PM, Markus Meier wrote:

Hi Engin,

However, the files in question (FUC.cif, NAG.cif, BMA.cif and
MAN.cif) are identical to current CCP4 8.0.018 version of the library
(installed by CCP4 Update Manager) as well as version ccp4-8.0.018
and version ccp4-8.0.016 on GitHub
(https://github.com/MonomerLibrary/monomers)

If they are different on your machine, please let me know.

I will switch monomer libraries anyway and do some testing.

I also was switching back and forth between my locally compiled coot
(0.9.8.92) and the the CCP4 distributed version (0.9.8.93 EL). I will
check if there is a difference between the two.

Thank you very much for your suggestion!

With best regards,
Markus

On 2024-04-18 13:48, Engin Özkan wrote:

Caution! This message was sent from outside the University of Manitoba.

Hi Markus,

I also had private discussions with Garib Murshudov, who were
correcting

the glycan monomer problems back in ~2011.

I could find old links to some of the discussions on bulletin board
archives:

https://phenix-online.org/pipermail/phenixbb/2011-November/017958.html

https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg13968.html

I am curious if something got reverted or if you have some ancient
library definitions somewhere. But I hope this gives some ideas.

Best,

Engin

On 4/17/24 3:07 PM, Markus Meier wrote:

Hi Paul,

thank you for implementing the Glyco module in Coot!

I used its "Add N-linked Glycan" function to build an N-linked glycan
which contains L-fucose with an alpha1-6 glycosidic linkage on the
first N-acetyl glucosamine of the tree.

However, after exporting the tree, I checked it with privateer and
unfortunately privateer flags the L-fucose as having the wrong
anomeric linkage - beta 1-6 :-(
Coot reports the same as being in the alpha 1-6

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-23 Thread Engin Özkan

Hi Markus,

I am using Coot that comes with CCP4 8, which has FUC (alpha-L-Fucose)
with an oxygen on C1 in the axial position. Which is the correct form, I
believe.

Using Coot 0.9.8.93, I get my alpha1-6 and alpha1-3 (insect) Fucoses
inserted correctly. I had trouble understanding your description: You
mean you get Coot to insert a FUC with the glycosidic bond going
equatorial, instead of axial, to Fucose?

This is how it works for me:

https://www.dropbox.com/scl/fi/yjbzstojjivmfpqdlz2os/Screen-Recording-2024-04-23-at-7.32.14-PM.mov?rlkey=zvvk7apd19wk61kt21qyc42l5=0

I'll see what happens when I update to the new Coot version where Paul
has switched the logic.

By the way, privateer tells me I have zero errors and the correct alpha
anomer for Fucose.

Fuc913-a6-GlcNAc911-b-ASN258
WURCS=2.0/2,2,1/[a2122h-1b_1-5_2*NCC/3=O][a1221m-1a_1-5]/1-2/a6-b1
...

...
Wrong anomer: 0

Best,

Engin

On 4/18/24 4:35 PM, Markus Meier wrote:

Hi Engin,

If they are different on your machine, please let me know.

I will switch monomer libraries anyway and do some testing.

I also was switching back and forth between my locally compiled coot
(0.9.8.92) and the the CCP4 distributed version (0.9.8.93 EL). I will
check if there is a difference between the two.

Thank you very much for your suggestion!

With best regards,
Markus

On 2024-04-18 13:48, Engin Özkan wrote:

Caution! This message was sent from outside the University of Manitoba.

Hi Markus,

I also had private discussions with Garib Murshudov, who were correcting
the glycan monomer problems back in ~2011.

I could find old links to some of the discussions on bulletin board
archives:

https://phenix-online.org/pipermail/phenixbb/2011-November/017958.html

https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg13968.html

I am curious if something got reverted or if you have some ancient
library definitions somewhere. But I hope this gives some ideas.

Best,

Engin

On 4/17/24 3:07 PM, Markus Meier wrote:

Hi Paul,

thank you for implementing the Glyco module in Coot!

I used its "Add N-linked Glycan" function to build an N-linked glycan
which contains L-fucose with an alpha1-6 glycosidic linkage on the
first N-acetyl glucosamine of the tree.

I then manually built the opposite anomeric form which privateer
happily accepts as being alpha1-6. However, Coot reports it as beta
1-6.

In my understanding, the alpha configuration of a glycosidic linkage
should have the oxygen at the anomeric centre and the -CH3 group at
opposite faces of the sugar ring, so I have to agree with privateer on
this one.

Please check you code.

I have attached the two glycan trees and here is the output of
privateer and Coot for both. Coot and privateer versions are given at
the bottom.

Btw. Neither privateer nor refmac5 can read the mmCIF files that Coot
produces with the "Extract this Tree" functions. All the
_atom_site.label_*_id entries are empty!

N133_tree_privateer_alpha_anomer_FUC.cif:

Privateer:
Detailed validation data

PDB Sugar Q Phi Theta Detected type
Cnf Ctx Ok?
- -- -- ---
--- -- --- -
N133 NAG-A-459 0.597 63.27 2.85 beta-D-aldopyranose
4c1 28.09 (n) yes
N133 NAG-A-460

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-22 Thread Paul Emsley


On 17/04/2024 21:07, Markus Meier wrote:


Hi Paul,

thank you for implementing the Glyco module in Coot!

It is not in Coot 1 (yet). It needs to be rewritten.


I used its "Add N-linked Glycan" function to build an N-linked glycan 
which contains L-fucose with an alpha1-6 glycosidic linkage on the 
first N-acetyl glucosamine of the tree.


However, after exporting the tree, I checked it with privateer and 
unfortunately privateer flags the L-fucose as having the wrong 
anomeric linkage - beta 1-6 :-(

Coot reports the same as being in the alpha 1-6 anomeric form.

How was the apparent? Did you use print_glyco_tree()?


I then manually built the opposite anomeric form which privateer 
happily accepts as being alpha1-6. However, Coot reports it as beta 1-6.


In my understanding, the alpha configuration of a glycosidic linkage 
should have the oxygen at the anomeric centre and the -CH3 group at 
opposite faces of the sugar ring, so I have to agree with privateer on 
this one.


Please check you code.


OK, will do.


I have attached the two glycan trees and here is the output of 
privateer and Coot for both. Coot and privateer versions are given at 
the bottom.


Btw. Neither privateer nor refmac5 can read the mmCIF files that Coot 
produces with the "Extract this Tree" functions. All the 
_atom_site.label_*_id entries are empty!



Oof. Cif.

OK. I'll take a look after the release.

Paul.



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

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

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-18 Thread Markus Meier


Hi Engin,

hmm, it appears I was using the monomer library from 
https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/dependencies/monomers-2023-01-02-23:57:29.tar.gz
 (it gets installed automatically with Coot's build-it script).

However, the files in question (FUC.cif, NAG.cif, BMA.cif and MAN.cif) are 
identical to current CCP4 8.0.018 version of the library (installed by CCP4 
Update Manager) as well as version ccp4-8.0.018 and version ccp4-8.0.016 on 
GitHub (https://github.com/MonomerLibrary/monomers)

MD5 checksums on these files are
9cfba4ac24f056d648df8227316f788b  monomers/f/FUC.cif
958ac5b19a92ceec604d233457944b2a  monomers/n/NAG.cif
697dd5242334dd99dad9ce43133c6382  monomers/b/BMA.cif
9d68ff54296ef2f4622005efb4eb080b  monomers/m/MAN.cif

If they are different on your machine, please let me know.

I will switch monomer libraries anyway and do some testing.

I also was switching back and forth between my locally compiled coot (0.9.8.92) 
and the the CCP4 distributed version (0.9.8.93 EL). I will check if there is a 
difference between the two.

Thank you very much for your suggestion!

With best regards,
Markus

On 2024-04-18 13:48, Engin Özkan wrote:

Caution! This message was sent from outside the University of Manitoba.


Hi Markus,

There were reports to the ccp4bb and phenixbb back in 2011 about
something similar. The monomer library files for FUC and/or the linkage
definitions for refmac/phenix were not right. I think. My recollection
is that this was corrected, and I could not reproduce your observation.

I also had private discussions with Garib Murshudov, who were correcting
the glycan monomer problems back in ~2011.

I could find old links to some of the discussions on bulletin board
archives:

https://phenix-online.org/pipermail/phenixbb/2011-November/017958.html

https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg13968.html

I am curious if something got reverted or if you have some ancient
library definitions somewhere. But I hope this gives some ideas.

Best,

Engin


On 4/17/24 3:07 PM, Markus Meier wrote:

Hi Paul,

thank you for implementing the Glyco module in Coot!

I used its "Add N-linked Glycan" function to build an N-linked glycan
which contains L-fucose with an alpha1-6 glycosidic linkage on the
first N-acetyl glucosamine of the tree.

However, after exporting the tree, I checked it with privateer and
unfortunately privateer flags the L-fucose as having the wrong
anomeric linkage - beta 1-6 :-(
Coot reports the same as being in the alpha 1-6 anomeric form.

I then manually built the opposite anomeric form which privateer
happily accepts as being alpha1-6. However, Coot reports it as beta 1-6.

In my understanding, the alpha configuration of a glycosidic linkage
should have the oxygen at the anomeric centre and the -CH3 group at
opposite faces of the sugar ring, so I have to agree with privateer on
this one.

Please check you code.


I have attached the two glycan trees and here is the output of
privateer and Coot for both. Coot and privateer versions are given at
the bottom.

Btw. Neither privateer nor refmac5 can read the mmCIF files that Coot
produces with the "Extract this Tree" functions. All the
_atom_site.label_*_id entries are empty!

N133_tree_privateer_alpha_anomer_FUC.cif:

  Privateer:
    Detailed validation data
    

    PDB   Sugar   Q Phi    Theta  Detected type
Cnf    Ctx   Ok?
        -  --  -- ---
---  --  ---  -
    N133  NAG-A-459 0.597   63.27    2.85 beta-D-aldopyranose
4c1  28.09   (n)  yes
    N133  NAG-A-460 0.603  145.39    2.15 beta-D-aldopyranose
4c1  36.34   (n)  yes
    N133  BMA-A-461 0.550  333.80    4.39 beta-D-aldopyranose
4c1  45.25   (n)  yes
    N133  MAN-A-462 0.628    6.26    5.03 alpha-D-aldopyranose
4c1  46.22   (n)  yes
    N133  MAN-A-463 0.628    4.87    5.87 alpha-D-aldopyranose
4c1  50.37   (n)  yes
    N133  FUC-A-464 0.615  162.25  174.51 alpha-L-aldopyranose
1c4  31.24   (n)  yes <-

    Partially occupied monosaccharides, if any, are marked with an
asterisk (*)

    SUMMARY:

   Wrong anomer: 0
   Wrong configuration: 0
   Unphysical puckering amplitude: 0
   In higher-energy conformations: 0

   Privateer has identified 0 issues, with 0 of 6 sugars affected.
    Coot output:
    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 1
     glyco close: 1.486  A 459  C1   to A 133  ND2
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 133 ASN
returns "pyr-ASN"
    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 4
     glyco close: 2.397  A 459  C4   to A 460  C1
     glyco close: 2.339  A 459  O4   to A 460  C2
     glyco close: 2.318  A 459  O4   to A 460  O5
     glyco close: 1.387  A 459  O4   to A 460  C1
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 460 NAG
returns "BETA1-4"
    DEBUG::

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-18 Thread Engin Özkan


Hi Markus,

There were reports to the ccp4bb and phenixbb back in 2011 about 
something similar. The monomer library files for FUC and/or the linkage 
definitions for refmac/phenix were not right. I think. My recollection 
is that this was corrected, and I could not reproduce your observation.


I also had private discussions with Garib Murshudov, who were correcting 
the glycan monomer problems back in ~2011.


I could find old links to some of the discussions on bulletin board 
archives:


https://phenix-online.org/pipermail/phenixbb/2011-November/017958.html

https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg13968.html

I am curious if something got reverted or if you have some ancient 
library definitions somewhere. But I hope this gives some ideas.


Best,

Engin


On 4/17/24 3:07 PM, Markus Meier wrote:

Hi Paul,

thank you for implementing the Glyco module in Coot!

I used its "Add N-linked Glycan" function to build an N-linked glycan 
which contains L-fucose with an alpha1-6 glycosidic linkage on the 
first N-acetyl glucosamine of the tree.


However, after exporting the tree, I checked it with privateer and 
unfortunately privateer flags the L-fucose as having the wrong 
anomeric linkage - beta 1-6 :-(

Coot reports the same as being in the alpha 1-6 anomeric form.

I then manually built the opposite anomeric form which privateer 
happily accepts as being alpha1-6. However, Coot reports it as beta 1-6.


In my understanding, the alpha configuration of a glycosidic linkage 
should have the oxygen at the anomeric centre and the -CH3 group at 
opposite faces of the sugar ring, so I have to agree with privateer on 
this one.


Please check you code.


I have attached the two glycan trees and here is the output of 
privateer and Coot for both. Coot and privateer versions are given at 
the bottom.


Btw. Neither privateer nor refmac5 can read the mmCIF files that Coot 
produces with the "Extract this Tree" functions. All the 
_atom_site.label_*_id entries are empty!


N133_tree_privateer_alpha_anomer_FUC.cif:

  Privateer:
    Detailed validation data
    

    PDB   Sugar   Q Phi    Theta  Detected type 
Cnf    Ctx   Ok?
        -  --  -- ---  
---  --  ---  -
    N133  NAG-A-459 0.597   63.27    2.85 beta-D-aldopyranose  
4c1  28.09   (n)  yes
    N133  NAG-A-460 0.603  145.39    2.15 beta-D-aldopyranose  
4c1  36.34   (n)  yes
    N133  BMA-A-461 0.550  333.80    4.39 beta-D-aldopyranose  
4c1  45.25   (n)  yes
    N133  MAN-A-462 0.628    6.26    5.03 alpha-D-aldopyranose 
4c1  46.22   (n)  yes
    N133  MAN-A-463 0.628    4.87    5.87 alpha-D-aldopyranose 
4c1  50.37   (n)  yes
    N133  FUC-A-464 0.615  162.25  174.51 alpha-L-aldopyranose 
1c4  31.24   (n)  yes <-


    Partially occupied monosaccharides, if any, are marked with an 
asterisk (*)


    SUMMARY:

   Wrong anomer: 0
   Wrong configuration: 0
   Unphysical puckering amplitude: 0
   In higher-energy conformations: 0

   Privateer has identified 0 issues, with 0 of 6 sugars affected.
    Coot output:
    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 1
     glyco close: 1.486  A 459  C1   to A 133  ND2
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 133 ASN 
returns "pyr-ASN"

    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 4
     glyco close: 2.397  A 459  C4   to A 460  C1
     glyco close: 2.339  A 459  O4   to A 460  C2
     glyco close: 2.318  A 459  O4   to A 460  O5
     glyco close: 1.387  A 459  O4   to A 460  C1
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 460 NAG 
returns "BETA1-4"

    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 3
     glyco close: 2.347  A 460  O4   to A 461  C2
     glyco close: 2.346  A 460  O4   to A 461  O5
     glyco close: 1.398  A 460  O4   to A 461  C1
   debug:: find_glycosidic_linkage_type() for A 460 NAG,A 461 BMA 
returns "BETA1-4"

    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 1
     glyco close: 1.407  A 461  O6   to A 462  C1
   debug:: find_glycosidic_linkage_type() for A 461 BMA,A 462 MAN 
returns "ALPHA1-6"

    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 1
     glyco close: 1.403  A 461  O3   to A 463  C1
   debug:: find_glycosidic_linkage_type() for A 461 BMA,A 463 MAN 
returns "ALPHA1-3"

    DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 2
     glyco close: 2.385  A 459  O6   to A 464  O5
     glyco close: 1.399  A 459  O6   to A 464  C1
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 464 FUC 
returns "BETA1-6" <-


## 


N133_tree_privateer_beta_anomer_FUC.cif:

  Privateer:
    Detailed validation data

[COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

2024-04-17 Thread Markus Meier


Hi Paul,

thank you for implementing the Glyco module in Coot!

I used its "Add N-linked Glycan" function to build an N-linked glycan which 
contains L-fucose with an alpha1-6 glycosidic linkage on the first N-acetyl glucosamine 
of the tree.

However, after exporting the tree, I checked it with privateer and 
unfortunately privateer flags the L-fucose as having the wrong anomeric linkage 
- beta 1-6 :-(
Coot reports the same as being in the alpha 1-6 anomeric form.

I then manually built the opposite anomeric form which privateer happily 
accepts as being alpha1-6. However, Coot reports it as beta 1-6.

In my understanding, the alpha configuration of a glycosidic linkage should 
have the oxygen at the anomeric centre and the -CH3 group at opposite faces of 
the sugar ring, so I have to agree with privateer on this one.

Please check you code.


I have attached the two glycan trees and here is the output of privateer and 
Coot for both. Coot and privateer versions are given at the bottom.

Btw. Neither privateer nor refmac5 can read the mmCIF files that Coot produces with the 
"Extract this Tree" functions. All the _atom_site.label_*_id entries are empty!

N133_tree_privateer_alpha_anomer_FUC.cif:

  Privateer:
Detailed validation data


PDB   Sugar   Q PhiTheta  Detected type Cnf   
 Ctx   Ok?
    -  --  --  ---  ---  -- 
 ---  -
N133  NAG-A-459 0.597   63.272.85  beta-D-aldopyranose  4c1  28.09  
 (n)  yes
N133  NAG-A-460 0.603  145.392.15  beta-D-aldopyranose  4c1  36.34  
 (n)  yes
N133  BMA-A-461 0.550  333.804.39  beta-D-aldopyranose  4c1  45.25  
 (n)  yes
N133  MAN-A-462 0.6286.265.03  alpha-D-aldopyranose 4c1  46.22  
 (n)  yes
N133  MAN-A-463 0.6284.875.87  alpha-D-aldopyranose 4c1  50.37  
 (n)  yes
N133  FUC-A-464 0.615  162.25  174.51  alpha-L-aldopyranose 1c4  31.24   
(n)  yes <-

Partially occupied monosaccharides, if any, are marked with an asterisk (*)

SUMMARY:

   Wrong anomer: 0
   Wrong configuration: 0
   Unphysical puckering amplitude: 0
   In higher-energy conformations: 0

   Privateer has identified 0 issues, with 0 of 6 sugars affected.
  
  Coot output:

DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 1
 glyco close: 1.486  A 459  C1   to A 133  ND2
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 133 ASN returns 
"pyr-ASN"
DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 4
 glyco close: 2.397  A 459  C4   to A 460  C1
 glyco close: 2.339  A 459  O4   to A 460  C2
 glyco close: 2.318  A 459  O4   to A 460  O5
 glyco close: 1.387  A 459  O4   to A 460  C1
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 460 NAG returns 
"BETA1-4"
DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 3
 glyco close: 2.347  A 460  O4   to A 461  C2
 glyco close: 2.346  A 460  O4   to A 461  O5
 glyco close: 1.398  A 460  O4   to A 461  C1
   debug:: find_glycosidic_linkage_type() for A 460 NAG,A 461 BMA returns 
"BETA1-4"
DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 1
 glyco close: 1.407  A 461  O6   to A 462  C1
   debug:: find_glycosidic_linkage_type() for A 461 BMA,A 462 MAN returns 
"ALPHA1-6"
DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 1
 glyco close: 1.403  A 461  O3   to A 463  C1
   debug:: find_glycosidic_linkage_type() for A 461 BMA,A 463 MAN returns 
"ALPHA1-3"
DEBUG:: find_glycosidic_linkage_type() number of sorted distances: 2
 glyco close: 2.385  A 459  O6   to A 464  O5
 glyco close: 1.399  A 459  O6   to A 464  C1
   debug:: find_glycosidic_linkage_type() for A 459 NAG,A 464 FUC returns 
"BETA1-6" <-

##
N133_tree_privateer_beta_anomer_FUC.cif:

  Privateer:
Detailed validation data


PDB   Sugar   Q PhiTheta  Detected type Cnf   
 Ctx   Ok?
    -  --  --  ---  ---  -- 
 ---  -
N133  NAG-A-459 0.539   93.90   14.28  beta-D-aldopyranose  4c1  17.31  
(n)   yes
N133  NAG-A-460 0.5763.573.48  beta-D-aldopyranose  4c1  28.29  
(n)   yes
N133  BMA-A-461 0.534  312.399.79  beta-D-aldopyranose  4c1  42.24  
(n)   yes
N133  MAN-A-462 0.580  219.674.19  alpha-D-aldopyranose 4c1  42.02  
(n)   yes
N133  MAN-A-463 0.563  294.234.86  alpha-D-aldopyranose 4c1  47.39  
(n)   yes
N133  FUC-A-464 0.542  120.60  172.19  beta-L-aldopyranose  1c4  18.87  
(n)   no <-

Partially occupied monosaccharides, if any, are marked with an asterisk (*)

SUMMARY:

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer: Building L-fucose fails in coot-0.9.8.94

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

Re: [COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

[COOT] Conflicting opinions on what constitutes an alpha1-6 L-fucose between Coot and Privateer

9 matches

Site Navigation

Mail list logo

Footer information