Resonance Assignment/CARA/Backbone assignment: Difference between revisions

From NESG Wiki
Jump to navigation Jump to search
No edit summary
No edit summary
 
(10 intermediate revisions by the same user not shown)
Line 1: Line 1:
== '''Initializing spin systems'''  ==
Here we speak of backbone assignment in a broad sense, including assignment of CB and HB spins. Generally the approach outlined below can be applied to various set of 3D triple-resonance experiments.


Normally you would start picking new spins systems from a 2D [15N,1H] HSQC. The quickest way is to initialize them automatically.
== Initializing spin systems ==


#Open 2D [15N,1H]-HSQC in SynchroScope or PolysScope. Adjust the base contour level to estimate the optimal peak amplitude threshold, and note the ppm ranges where true backbone peaks are observed.
Normally you would start picking new spins systems from a 2D [<sup>15</sup>N,<sup>1</sup>H] HSQC. The quickest way is to initialize them automatically.  
#Run LUA script '''Pick_2D_Peaks''' (it is equivalent to the in-phase peak picking command <tt>in</tt> in XEASY, but is not interactive). The script will create new spin systems containing H and N spins. XEASY users please note that in CARA spin systems are distinct from residues and can have numbers starting with 1.
#Delete unwanted spin systems (usually erroneously picked spectral artifacts or side-chain peaks) in SynchroScope or PolyScope by selecting them, right-clicking and choosing '''Delete Systems''' from the context menu. The corresponding spins will become 'orphaned' and can be deleted as well.
#Pick additional spin systems by placing the mouse cursor on a spectral peak, right-clicking and selecting '''Pick System''' from the context menu. To help resolve overlaps in 2D HSQC you can load a 3D triple-resonance spectrum (like HNCO) or a 3D 15N-resolved NOESY into the strip panels of SynchroScope or PolyScope by clicking '''Strips''' -> '''Select Spectrum'''


After initializing spin systems you can proceed to pick 3D spectra either manually or automatically.
#Open 2D [<sup>15</sup>N,<sup>1</sup>H]-HSQC in SynchroScope or PolyScope. Adjust the base contour level to estimate the optimal peak amplitude threshold, and note the ppm ranges where true backbone peaks are observed.
#Run LUA script '''Pick_2D_Peaks''' (it is equivalent to the in-phase peak picking command <tt>in</tt> in XEASY, but is not interactive). The script will create new spin systems containing H and N spins. XEASY users please note that in CARA spin systems are distinct from residues and can have numbers starting with 1.
#Delete unwanted spin systems (usually erroneously picked spectral artifacts or side-chain peaks) in SynchroScope or PolyScope by selecting them, right-clicking and choosing '''Delete Systems''' from the context menu. The corresponding spins will become 'orphaned' and can be deleted as well.
#Pick additional spin systems by placing the mouse cursor on a spectral peak, right-clicking and selecting '''Pick System''' from the context menu. To help resolve overlaps in 2D HSQC you can load a 3D triple-resonance spectrum (like HNCO) or a 3D <sup>15</sup>N-resolved NOESY into the strip panels of SynchroScope or PolyScope by clicking '''Strips''' -&gt; '''Select Spectrum'''


== '''Picking spin systems manually'''  ==
After initializing spin systems you can proceed to pick 3D spectra either manually or automatically.
 
#Select a spin system by left-clicking in HSQC plane or using '''View -&gt; Goto System''' - this updates the strip display.
#In the strip panel set the cursor on a peak and use '''right click -&gt; Pick Label...''' or '''right click -&gt; Pick Spin''' and '''right click -&gt; Label Spin''' to pick C-1. Repeat last two steps for all spin systems.


Also see instructions on how to use SynchroScope here: http://www.cara.ethz.ch/Wiki/SynchroScope  
Also see instructions on how to use SynchroScope here: http://www.cara.ethz.ch/Wiki/SynchroScope  


== Manual resonance assignment ==


=== Picking spin systems ===


#In PolyScope or SynchroScope select a spin system by left-clicking in the 2D [15N,1H] HSQC plane, or use '''View -&gt; Goto System'''. This will update the displayed strips.
#In the strip panel set the cursor on a peak, right-click, and select '''Pick Label...''', or '''Pick Spin''' and '''Label Spin...''' to pick the appropriate spin (e.g., C-1 in HNCO).
#Repeat for all spin systems.


=== '''Picking Spins in (4,3)D GFT HNNCABCA and CABCA(CO)NHN Spectra''' ===
For example, if you are working with 3D HNCO, HN(CA)CO, CBCA(CO)NH and HNCACB spectra, you would first pick all C-1 spins in HNCO. Then you can switch the spectrum to HN(CA)CO and all C-1 spins will be displayed, allowing you to assign the remaining spins to C spins. You can follow the same procedure by picking CA-1 and CB-1 spins in CBCA(CO)NH first, then picking CA and CB in HNCACB.
 
Open (4,3)D GFT HNNCABCA and CABCA(CO)NHN in SynchroScope as you did with 3D HNCO. You can either open four separate SynchroScope windows, or open only one and switch between (4,3)D subspectra there.
 
Pick CABCA spins just like you did with C-1 spins in HNCO. It is recommended to start with (4,3)D CABCA(CO)NHN to pick CAmCA-1 CApCA-1, CAmCB-1 and CApCB-1 spins first. Then pick CAmCA CApCA, CAmCB and CApCB spins in (4,3)D GFT HNNCABCA.


Picking new spins is not as slow as in XEASY. However, it would make sense to add a script, which would create initial CABCA spins to speed up the process.
Also see instructions on how to use SynchroScope here: http://www.cara.ethz.ch/Wiki/SynchroScope


See the this page for details and examples on how to use SynchroScope: http://www.cara.ethz.ch/Wiki/SynchroScope
=== Linking and assigning ===


=== '''Calculating CA and CB Chemical Shifts''' ===
Once you have picked all the spin systems you can assign them to specific residues. This is achieved by building fragments of linked spin-systems and mapping them to the protein sequence. The procedure is outlined quite well in http://wiki.cara.nmr.ch/StripScope. Additional issues to consider:


Once the GFT spins have been picked as completely as possible, run the '''GFT_CABCA2CACB''' Lua script.
*Use StripScope to display appropriate spectra that show both sequential and intra-residue correlations, such as HNCA, HNCACB, HN(CA)CO, 15N-resolved NOESY.
*Before linking spin systems, make sure that peak shapes in 1D slices match for all such spectra. You may need to zoom in on individual peaks, since 1D slices are automatically scaled to the maximum amplitude in the zoomed region.
*The '''Show Alignment...''' window may display no match if you have highly unusual chemical shifts in your spin system. This is more likely if you include proton spins, such as HA-1 and HB-1 from 3D HBHA(CO)NH.
*To unassign spin systems right-click on a spin system tab and choose '''Unassign''' from the context menu.
*Special attention is needed when assigning and unassigning spin systems. Unassigning a spin system that is part of a linked fragment will remove assignments of all spin systems in that fragment. Conversely, assigning a spin system to a particular residue will automatically assign other spin systems in the same fragment; if the newly assigned fragment overlaps with some previously assigned fragments, they will be unassigned.
*It is thus safer to use '''Show Free Successors''' and '''Show Free Predecessors''' for routine assignments. Invoke '''Show All Successors''' and '''Show All Predecessors''' only if you cannot find matches or have reasons to suspect that existing assignments may be incorrect.
*To unlink spin systems right-click on a spin system tab and select '''Unlink Successor''' or '''Unlink Predecessor''' from the context menu.
*Display each linked fragment of spin systems by right-clicking on a spin system and selecting '''Show Fragment''' from the context menu.  


'''IMPORTANT!''' Back up the repository before running the script. Missing CA and CB spins will be created, and the chemical shifts of the existing CA and CB spins will be updated.
== Semi-automated resonance assignment ==


You will need to select the project (most likely it will be the only one) and a spectrum. The spectrum is needed to use the correct carrier offset (~ 43 ppm) of the projected dimension, saved as an attribute. See also the page on loading GFT spectra.
Semi-automated resonance assignment is usually faster than purely manual approach. Automated methods almost never yield absolutely complete and accurate results and thus require manual inspection an finalizing by an expert.


* GFT_CABCA2CACB pop-up: <br />
=== Backbone Assignment with PINE ===
    <img src="%ATTACHURLPATH%/gftcacbcalc.png" alt="gftcacbcalc.png" width='217' height='184' />


The script will produce a log in the terminal window, reporting large deviations (default threshold: 0.5 ppm), missing spins and possible glycines.
First, you would need to prepare simulated peaklists in XEASY format for a set of 3D triple-resonance experiments (for example, CBCA(CO)NH, HNCACB, HNCO, HN(CA)CO and HBHA(CO)NH ). You can use the '''Pick_3D_Peaks''' LUA script, however, can only pick spins and does not generate peaklists directly.


#Run '''Pick_3D_Peaks''' LUA script to pick spins in a 3D spectrum, such as CBCA(CO)NH. Use an appropriate generic spin label (for example, ?CA-1 would work well for CBCA(CO)NH).
#Use SynchroScope or PolyScope to walk through all spin systems and prune incorrectly picked spins and add those that were omitted.
#In PolyScope click '''File''' -> '''Export strip peaks to MonoScope...''' - this will open a MonoScope window with a simulated peaklist.
#In the MonoScope window, add peaklist to repository and update the amplitudes. Run CopyAmpToVol LUA script to set peak volumes. In the MonoScope window export the peaklist as XEASY format.
#Delete the newly picked ?CA-1 spins by executing the RemoveSpins LUA script (otherwise they will automatically show up in HNCACB, for example).
#Repeat for other 3D triple-resonance spectra that you want to use.
#Export 2D [15N,1H] HSQC peaklist useing '''File''' -> '''Export plane peaks to MonoScope...''' in PolyScope.
#Peaklists exported from CARA in XEASY format will carry spin system numbers in the form of comments on separate lines that cannot be properly read by PINE. Remove these lines with <tt>grep -v "# " hnco.peaks > hnco2.peaks</tt> command in a UNIX shell.
#Submit the data to PINE server and,
#If satisfied with the results then delete all spins and spin systems and import the assignments from the <tt>.prot</tt> in PINE output by right-clicking on your project '''Import''' -> '''Atom List''' in the main window of CARA. Don't forget to back up your repository under a different filename!
#Correct HB assignments in alanines. Due to a bug in PINE they are imported as HB1, HB2 and HB3 instead of HB.
#Verify assignments made by PINE and complete the assignment manually as described above.


At present it seems difficult to reconcile this approach with UBNMR due to:
Please note, that spin systems not assigned by PINE will not be imported back and will have to picked again. This is not a big issue as long as the automated assignments are fairly complete.
* Nomenclature differences - H vs HN
* Two sets of SRDs SRD-I and SRD-II in XEASY/UBNMR approach
* Inconvenience - using UBNMR would require reading/writing to disk; Lua scripts operate on the data in memory.


=== '''Automated Backbone Assignment with AutoAssign''' ===
=== Backbone Assignment with AutoAssign ===


Once CA, CA-1, CB, and CB-1 spins have been created you can run AutoAssign.  
Once CA, CA-1, CB, and CB-1 spins have been created you can run AutoAssign.  


To generate AutoAssign input files run Lua script '''ExportToAutoAssign'''. In a dialog window specify the directory to save the files. You will have to give a dummy filename, which will be ignored, and only the directory information will be used. This script will create four files:
To generate AutoAssign input files run Lua script '''ExportToAutoAssign'''. In a dialog window specify the directory to save the files. You will have to give a dummy filename, which will be ignored, and only the directory information will be used. This script will create four files:  
* <tt>projectname.aat</tt>
* <tt>hsqc.pks</tt>
* <tt>hncacb.pks</tt>
* <tt>hncocacb.pks</tt>
 
Here <tt>projectname</tt> is the name of you project in CARA. Peak files <tt>hncacb.pks</tt> and <tt>hncocacb.pks</tt> will contain dummy positive and negative intensities to distinguish CA and CB peaks, respectively. These peaklist will have a perfect "registration" matching in the H-N plane.
 
Modify these files, if necessary and run AutoAssign by selecting '''Default Execution'''. Save results in a file by clicking on '''Examine -> All GSs''' or '''Examine -> Assigned GSs'''.
 
Run the '''ImportFromAutoAssign''' Lua script to read the AutoAssign result file. This script will properly link spin systems and assign them to respective residues.
 
'''IMPORTANT!''' *ImportFromAutoAssign* will clear all existing of spin system assignments as well as unlink all spin system! Backup the repository, if you don't want to discard any previously made assignments.
 
The newer version of AutoAssign should be able to handle GFT shifts directly, and this protocol may need to be modified to take advantage of it.
 
[http://www.autolink.nmr-software.org/ AutoLink] is an alternative automated backbone assignment package written in LUA specifically for CARA by Jim Masse.
 
 
=== '''Verification and Manual Backbone Assignment''' ===
 
Experience tells that AutoAssign seldom yields an absolutely complete and accurate result. Typical inaccuracies occur when AutoAssign assumes that, for example, the user has mistaken the intra peak CA for sequential CA-1.
 
To verify assignments
# Open (4,3)D GFT HNNCABCA, and 15N-resolved NOESY in StripScope. You can have several StripScope sessions simultaneously.
# Sort the spin systems on the left panel by assignment column.
# Display each linked fragment of spin systems by right-clicking on a spin system and selecting '''Show Fragment''' from the context menu.
# Inspect sequential connectives between strips.
# Check how well the fragment fits the sequence by right-clicking on a spin system and selecting '''Show Alignment''' from the context menu.


To unlink spin systems right-click on a spin system tab and select '''Unlink Successor''' or '''Unlink Predecessor''' from the context menu.
*<tt>projectname.aat</tt>
*<tt>hsqc.pks</tt>
*<tt>hncacb.pks</tt>
*<tt>hncocacb.pks</tt>


To unassign spin systems right-click on a spin system tab and choose '''Unassign''' from the context menu.
Here <tt>projectname</tt> is the name of you project in CARA. Peak files <tt>hncacb.pks</tt> and <tt>hncocacb.pks</tt> will contain dummy positive and negative intensities to distinguish CA and CB peaks, respectively. These peaklist will have a perfect "registration" matching in the H-N plane.
'''IMPORTANT!''' Unassigning a spin system from a linked fragment of spin systems, will unassign the entire fragment!


See this page on how to do manual assignment in StripScope: http://www.cara.ethz.ch/Wiki/StripScope
Modify these files, if necessary and run AutoAssign by selecting '''Default Execution'''. Save results in a file by clicking on '''Examine -&gt; All GSs''' or '''Examine -&gt; Assigned GSs'''.  


To verify connectivities in 15N-resolved NOESY it may be useful to calculate HA-1 and HB-1 spins first.
Run the '''ImportFromAutoAssign''' Lua script to read the AutoAssign result file. This script will properly link spin systems and assign them to respective residues.  


%COMMENT%
'''IMPORTANT!''' *ImportFromAutoAssign* will clear all existing of spin system assignments as well as unlink all spin system! Backup the repository, if you don't want to discard any previously made assignments.


The newer version of AutoAssign should be able to handle GFT shifts directly, and this protocol may need to be modified to take advantage of it.


[http://www.autolink.nmr-software.org/AutoLink] is an alternative automated backbone assignment package written in LUA specifically for CARA by Jim Masse.




-- Main.AlexEletski - 06 Jul 2007
<br> -- Main.AlexEletski - 06 Jul 2007

Latest revision as of 02:17, 9 December 2009

Here we speak of backbone assignment in a broad sense, including assignment of CB and HB spins. Generally the approach outlined below can be applied to various set of 3D triple-resonance experiments.

Initializing spin systems

Normally you would start picking new spins systems from a 2D [15N,1H] HSQC. The quickest way is to initialize them automatically.

  1. Open 2D [15N,1H]-HSQC in SynchroScope or PolyScope. Adjust the base contour level to estimate the optimal peak amplitude threshold, and note the ppm ranges where true backbone peaks are observed.
  2. Run LUA script Pick_2D_Peaks (it is equivalent to the in-phase peak picking command in in XEASY, but is not interactive). The script will create new spin systems containing H and N spins. XEASY users please note that in CARA spin systems are distinct from residues and can have numbers starting with 1.
  3. Delete unwanted spin systems (usually erroneously picked spectral artifacts or side-chain peaks) in SynchroScope or PolyScope by selecting them, right-clicking and choosing Delete Systems from the context menu. The corresponding spins will become 'orphaned' and can be deleted as well.
  4. Pick additional spin systems by placing the mouse cursor on a spectral peak, right-clicking and selecting Pick System from the context menu. To help resolve overlaps in 2D HSQC you can load a 3D triple-resonance spectrum (like HNCO) or a 3D 15N-resolved NOESY into the strip panels of SynchroScope or PolyScope by clicking Strips -> Select Spectrum

After initializing spin systems you can proceed to pick 3D spectra either manually or automatically.

Also see instructions on how to use SynchroScope here: http://www.cara.ethz.ch/Wiki/SynchroScope

Manual resonance assignment

Picking spin systems

  1. In PolyScope or SynchroScope select a spin system by left-clicking in the 2D [15N,1H] HSQC plane, or use View -> Goto System. This will update the displayed strips.
  2. In the strip panel set the cursor on a peak, right-click, and select Pick Label..., or Pick Spin and Label Spin... to pick the appropriate spin (e.g., C-1 in HNCO).
  3. Repeat for all spin systems.

For example, if you are working with 3D HNCO, HN(CA)CO, CBCA(CO)NH and HNCACB spectra, you would first pick all C-1 spins in HNCO. Then you can switch the spectrum to HN(CA)CO and all C-1 spins will be displayed, allowing you to assign the remaining spins to C spins. You can follow the same procedure by picking CA-1 and CB-1 spins in CBCA(CO)NH first, then picking CA and CB in HNCACB.

Also see instructions on how to use SynchroScope here: http://www.cara.ethz.ch/Wiki/SynchroScope

Linking and assigning

Once you have picked all the spin systems you can assign them to specific residues. This is achieved by building fragments of linked spin-systems and mapping them to the protein sequence. The procedure is outlined quite well in http://wiki.cara.nmr.ch/StripScope. Additional issues to consider:

  • Use StripScope to display appropriate spectra that show both sequential and intra-residue correlations, such as HNCA, HNCACB, HN(CA)CO, 15N-resolved NOESY.
  • Before linking spin systems, make sure that peak shapes in 1D slices match for all such spectra. You may need to zoom in on individual peaks, since 1D slices are automatically scaled to the maximum amplitude in the zoomed region.
  • The Show Alignment... window may display no match if you have highly unusual chemical shifts in your spin system. This is more likely if you include proton spins, such as HA-1 and HB-1 from 3D HBHA(CO)NH.
  • To unassign spin systems right-click on a spin system tab and choose Unassign from the context menu.
  • Special attention is needed when assigning and unassigning spin systems. Unassigning a spin system that is part of a linked fragment will remove assignments of all spin systems in that fragment. Conversely, assigning a spin system to a particular residue will automatically assign other spin systems in the same fragment; if the newly assigned fragment overlaps with some previously assigned fragments, they will be unassigned.
  • It is thus safer to use Show Free Successors and Show Free Predecessors for routine assignments. Invoke Show All Successors and Show All Predecessors only if you cannot find matches or have reasons to suspect that existing assignments may be incorrect.
  • To unlink spin systems right-click on a spin system tab and select Unlink Successor or Unlink Predecessor from the context menu.
  • Display each linked fragment of spin systems by right-clicking on a spin system and selecting Show Fragment from the context menu.

Semi-automated resonance assignment

Semi-automated resonance assignment is usually faster than purely manual approach. Automated methods almost never yield absolutely complete and accurate results and thus require manual inspection an finalizing by an expert.

Backbone Assignment with PINE

First, you would need to prepare simulated peaklists in XEASY format for a set of 3D triple-resonance experiments (for example, CBCA(CO)NH, HNCACB, HNCO, HN(CA)CO and HBHA(CO)NH ). You can use the Pick_3D_Peaks LUA script, however, can only pick spins and does not generate peaklists directly.

  1. Run Pick_3D_Peaks LUA script to pick spins in a 3D spectrum, such as CBCA(CO)NH. Use an appropriate generic spin label (for example, ?CA-1 would work well for CBCA(CO)NH).
  2. Use SynchroScope or PolyScope to walk through all spin systems and prune incorrectly picked spins and add those that were omitted.
  3. In PolyScope click File -> Export strip peaks to MonoScope... - this will open a MonoScope window with a simulated peaklist.
  4. In the MonoScope window, add peaklist to repository and update the amplitudes. Run CopyAmpToVol LUA script to set peak volumes. In the MonoScope window export the peaklist as XEASY format.
  5. Delete the newly picked ?CA-1 spins by executing the RemoveSpins LUA script (otherwise they will automatically show up in HNCACB, for example).
  6. Repeat for other 3D triple-resonance spectra that you want to use.
  7. Export 2D [15N,1H] HSQC peaklist useing File -> Export plane peaks to MonoScope... in PolyScope.
  8. Peaklists exported from CARA in XEASY format will carry spin system numbers in the form of comments on separate lines that cannot be properly read by PINE. Remove these lines with grep -v "# " hnco.peaks > hnco2.peaks command in a UNIX shell.
  9. Submit the data to PINE server and,
  10. If satisfied with the results then delete all spins and spin systems and import the assignments from the .prot in PINE output by right-clicking on your project Import -> Atom List in the main window of CARA. Don't forget to back up your repository under a different filename!
  11. Correct HB assignments in alanines. Due to a bug in PINE they are imported as HB1, HB2 and HB3 instead of HB.
  12. Verify assignments made by PINE and complete the assignment manually as described above.

Please note, that spin systems not assigned by PINE will not be imported back and will have to picked again. This is not a big issue as long as the automated assignments are fairly complete.

Backbone Assignment with AutoAssign

Once CA, CA-1, CB, and CB-1 spins have been created you can run AutoAssign.

To generate AutoAssign input files run Lua script ExportToAutoAssign. In a dialog window specify the directory to save the files. You will have to give a dummy filename, which will be ignored, and only the directory information will be used. This script will create four files:

  • projectname.aat
  • hsqc.pks
  • hncacb.pks
  • hncocacb.pks

Here projectname is the name of you project in CARA. Peak files hncacb.pks and hncocacb.pks will contain dummy positive and negative intensities to distinguish CA and CB peaks, respectively. These peaklist will have a perfect "registration" matching in the H-N plane.

Modify these files, if necessary and run AutoAssign by selecting Default Execution. Save results in a file by clicking on Examine -> All GSs or Examine -> Assigned GSs.

Run the ImportFromAutoAssign Lua script to read the AutoAssign result file. This script will properly link spin systems and assign them to respective residues.

IMPORTANT! *ImportFromAutoAssign* will clear all existing of spin system assignments as well as unlink all spin system! Backup the repository, if you don't want to discard any previously made assignments.

The newer version of AutoAssign should be able to handle GFT shifts directly, and this protocol may need to be modified to take advantage of it.

[1] is an alternative automated backbone assignment package written in LUA specifically for CARA by Jim Masse.



-- Main.AlexEletski - 06 Jul 2007