Resonance Assignment/Abacus/Sequence specific assignment of PB fragments: Difference between revisions

From NESG Wiki
Jump to navigation Jump to search
No edit summary
 
(19 intermediate revisions by one other user not shown)
Line 1: Line 1:
<div><span>&nbsp;&nbsp; Step 1. Peak picking of NOESY spectra.</span></div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; To facilitate peak-picking of NOESY spectra, you can&nbsp;first generate expected tocsy peaks of PB-fragments using FMCGUI:&nbsp;&nbsp; </span></div><div><span>-<span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></span>open project PRJ2 {[[FMCGUI commands#Project.3ELoad|Project&gt;Load]]}</div><div>-<span>&nbsp;&nbsp; generate tocsy peaks of N15 NOESY spectrum {[[FMCGUI commands#Fragment.3EExpected_Peaks.3E|Fragment&gt;Expected Peaks&gt;N15NOESY]]}</span></div><div><span>-<span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></span>generate tocsy peaks of C13NOESY spectrum [<span>Fragment&gt;Expected Peaks&gt;C13NOESY</span>]</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Then, using SPARKY, read in the expected peaks into corresponding spectra and complete peak picking manually.</span></div><div>&nbsp;</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp; Step 2.&nbsp;Probabilistic assignment of PB-fragments to protein sequence.</span></div><div>&nbsp;</div><div>According to FMC procedure (see Introduction) you have to do the following:</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; - Probabilistic typing of PB-fragments.</span></div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; - </span>To calculate &nbsp;two fragments “contact” maps&nbsp;C<sub>NOE&nbsp;</sub>andC<sub>HNCA . </sub>C<sub>NOE</sub> is a contact map based on 15N_ and 13C_NOESY data (it could be calculated by 2 methods, “abacus” and “fawn”) and&nbsp;C<sub>HNCA</sub> is a map based on HNCA data. </div><div><span>&nbsp;&nbsp; &nbsp;&nbsp;&nbsp;- </span>To calculate assignment probabilities <span>by Simulated Annealing (SA) or Replica Exchange Method (REM) Monte Carlo simulations</span></div><div>&nbsp;</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp; This can be done by the following commands:</span><span>&nbsp;&nbsp;&nbsp;&nbsp; </span></div><div>-&nbsp;open project PRJ2 ['''<span>Project&gt;load</span>''']</div><div>-&nbsp;load 15N NOESY peak list ['''<span>DATA&gt;N15 NOESY&gt;load</span>''']</div><div>-&nbsp;load 13C NOESY peak list&nbsp;'''<span>[DATA&gt;C13NOESY H2O&gt;load</span>''']</div><div>-&nbsp;set tolerances&nbsp;['''<span>Data&gt;Tolerances</span>''']</div><div>-&nbsp;calculate typing probabilities for all fragments ['''<span>Fragment&gt;Type&gt;calculate&gt;abacus</span>''']</div><div><span>&nbsp;&nbsp;&nbsp; You have to consider the warning messages shown in the project main window and to analyze/modify typing probabilities manually by using “Fragment Property Modification” (FPM) window (see Figures 2.7 and 2.8).&nbsp;To open FPM window click on ['''Fragment&gt;Type&gt;fix''']. This window has 3 sections. Top section allows you to select a fragment (by user ID) and modify its typing probabilities. The middle section shows typing probabilities that correspond to the selected amino acid type for all fragments. Here you can fix for any fragment it’s &nbsp;typing probability corresponding to the selected amino acid type to the values of 1 or 0. </span></div><div>&nbsp;</div><div>-&nbsp;calculate fragment contact map from HNCA data ['''<span>Assignment&gt;Contacts&gt;HNCA</span>map is calculated and loaded in the memory. It is strongly recommended to check the messages in the project main window regarding HNCA peak list. In case there are inconsistencies present in the input HNCA peak list, go back to spectra and fix the list.'''].&nbsp;In the result the </div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; - calculate fragment contact map from NOESY data ['''Assignment&gt;Contacts&gt;NOE&gt;abacus'''].&nbsp;In the result, two contact maps, </span>and C<sub>NOE_F </sub>are calculated and loaded in the memory. Calculation of &nbsp;involves use of BACUS procedure for NOESY data interpretation, while <sub>&nbsp;</sub>&nbsp;does not.</div><div><span>-<span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></span>Calculate assignment probabilities. It can be done using two different Monte Carlo simulation methods. Namely, Simulating Annealing (SA) ['''<span>Assignment&gt;Calculate Probability&gt;SA</span>'''] and Replica Exchange Method (REM) ['''Assignment&gt;Calculate Probability&gt;REM''']. Before starting calculations you have to specify control parameters. The main parameters to consider are: ‘Name of the SA/REM run’, ‘Size of the pool for unassigned fragments’, ‘number of SA runs’, ‘Final Temperature’ (SA), ‘Low Temperature’ (REM), and ‘NOE contact map’.&nbsp;The results of the calculations will be stored in the directory under specified name which is created inside PRJ2/assign/ directory. The main result consists of optimal and sub-optimal fragments assignments and assignment probability map. The last one will be automatically loaded in the memory as well.</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Calculation of assignment probabilities could be repeated a few times using different methods and parameters. Result of each calculation is stored in a separate directory. Therefore, there could be a few different directories (for example, sa_run1, sa_run2, rem_run0, rem_run1, rem_run3… ) within PRJ2/assign/ directory that contain different assignment probability maps.</span></div><div>&nbsp;</div><div>Step 3.&nbsp;Sequence-specific assignment of PB-fragments by analyzing probabilities <span>.</span></div><div>&nbsp;</div><div><span>&nbsp;&nbsp; A fragment assignment to a sequence position using FMCGUI could be done in two ways, manually and using assignment probability map, respectively. </span></div><div>&nbsp;</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; - manual assignment ['''Assignment&gt;Fix Assignment&gt;manually''']. </span></div><div>This command pops up ‘Fragment Property Modification’ window. You can change the assignment ID of any selected fragment using the bottom section of the window. </div><div>&nbsp;</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; - assignment using probability map&nbsp;['''Assignment&gt;Fix Assignment&gt;using probability map'''].&nbsp;You have to select a calculation directory (sa_run# or rem_run#) that contains assignment probability map, </span><span>, </span>and to specify the probability threshold . A fragments ''k'' will be assigned to position ''s'' if the condition&nbsp;<span>&nbsp;&nbsp;&nbsp;is satisfied (see Figure 1.4).</span></div><div>&nbsp;</div><div>Step 4. Assignment analysis.</div><div>&nbsp;</div><div>In the case of poor data, only a part of the fragments get assigned unambiguously. The uncertainty in fragments assignment could be resolved manually with the help of FMCGUI command ['''<span>View&gt;Assignment</span>''']. This command pops up “Assignment Graph” window that provides you with graphical tool to visualize the current assignment and to analyze sub-optimal fragment assignments (see Figures 2.21-2.25).</div><div>&nbsp;</div><div>Step 5. Final resonance assignment.</div><div>&nbsp;</div><div>When sequence specific PB-fragment assignment is done you have to put in order assigned fragments and to assignCO resonances using command&nbsp;['''<span>Fragment&gt;Modify assigned</span>'''].</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></div><div>&lt;span /&gt;</div><div>&lt;span /&gt;</div>
=== <span>&nbsp;Step 1. Peak picking of NOESY spectra</span>  ===
<div><br></div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; To facilitate peak-picking of NOESY spectra, you can&nbsp;first generate expected tocsy peaks of PB-fragments using FMCGUI:&nbsp;&nbsp; </span></div>  
*<span><span>&nbsp;&nbsp;&nbsp;&nbsp; </span></span>open project PRJ2 { [[FMCGUI commands#Project.3ELoad|Project&gt;Load]] }  
*<span>&nbsp; &nbsp;&nbsp; generate tocsy peaks of N15 NOESY spectrum { [[FMCGUI commands#Fragment.3EExpected_Peaks.3E|Fragment&gt;Expected Peaks&gt;N15NOESY]] }</span>  
*<span><span>&nbsp;&nbsp;&nbsp;&nbsp; </span></span>generate tocsy peaks of C13NOESY spectrum {[[FMCGUI commands#Fragment.3EExpected_Peaks.3E|<span> Fragment&gt;Expected Peaks&gt;C13NOESY</span>]] }
<div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Then, using SPARKY, read the expected peaks into corresponding spectra and complete peak picking manually.</span></div><div>&nbsp;</div>
 
=== <span>Step 2.&nbsp;Probabilistic assignment of PB fragments to protein sequence</span> ===
<div></div><div>According to [[Introduction to ABACUS#Figure_1.3|FMC procedure]] you have to do the following:</div>  
#<span>&nbsp;&nbsp;&nbsp; Probabilistic typing of PB fragments</span>  
#<span>&nbsp;&nbsp;&nbsp; </span>Calculate &nbsp;two fragments [[FMCGUI objects#Contact_map|contact maps]]&nbsp;C<sub>NOE&nbsp;</sub>and C<sub>HNCA . </sub>C<sub>NOE</sub> is a contact map based on 15N_ and 13C_NOESY data (it could be calculated by 2 different methods, “abacus” and “fawn”, respectively) and&nbsp;C<sub>HNCA</sub> is a map based on HNCA data.  
#<span>&nbsp;&nbsp; </span>Calculate [[FMCGUI objects#Assignment_probabilities|assignment probabilities]] <span>by Simulated Annealing (SA) or Replica Exchange Method (REM) Monte Carlo simulations</span>
<div>&nbsp;</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp; This can be done by the following commands:</span><span>&nbsp;&nbsp;&nbsp;&nbsp; </span></div>  
*&nbsp;&nbsp; open project PRJ2 { [[FMCGUI commands#Project.3ELoad|Project&gt;load]] }
*&nbsp;&nbsp; load 15N NOESY peak list { [[FMCGUI commands#Data.3EN15_NOESY.3E|DATA&gt;N15 NOESY&gt;load]] }
*&nbsp;&nbsp; load 13C NOESY peak list { [[FMCGUI commands#Data.3EC13_NOESY.3E|DATA&gt;C13NOESY H2O&gt;load]] }
*&nbsp;&nbsp; set tolerances&nbsp; { [[FMCGUI commands#Data.3ETolerances|Data&gt;Tolerances]] }
*&nbsp;&nbsp; calculate typing probabilities for all fragments { [[FMCGUI commands#Fragment.3EType.3ECalculate.3E|Fragment&gt;Type&gt;calculate&gt;abacus]] }
<div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; You have to consider the warning messages shown in the project [[FMCGUI objects#Main_window|main window]] and to analyze/modify typing probabilities manually by using “Fragment Property Modification” (FPM) window.&nbsp;To open FPM window use command { [[FMCGUI commands#Fragment.3EType.3Efix|Fragment&gt;Type&gt;fix]] }. This window has 3 sections. Top section allows you to select a fragment (by user ID) and modify its typing probabilities. The middle section shows typing probabilities that correspond to the selected amino acid type for all fragments. Here you can fix for any fragment it’s &nbsp;typing probability corresponding to the selected amino acid type to the values of 1 or 0. </span></div><div>&nbsp;</div>  
*&nbsp; calculate fragment contact map from HNCA data { [[FMCGUI commands#Assignment.3EContacts.3EHNCA|Assignment&gt;Contacts&gt;HNCA]] }&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
 
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Contact map C<sub>HNCA</sub> is calculated and loaded in the memory. It is strongly recommended to check the messages in the project main window regarding HNCA peak list. In case there are inconsistencies present in the input HNCA peak list, go back to spectra and fix the list.&nbsp;  
 
*<span>&nbsp; calculate fragment contact map from NOESY data { [[FMCGUI commands#Assignment.3EContacts.3ENOE.3Eabacus|Assignment&gt;Contacts&gt;NOE&gt;abacus]] }.&nbsp;In the result, two contact maps, </span>C<sub>NOE_B</sub> and C<sub>NOE_F </sub>are calculated and loaded in the memory. Calculation of <span> </span>C<sub>NOE_B</sub>&nbsp; makes use of BACUS procedure for NOESY data interpretation, while calculation of&nbsp;<span> </span>C<sub>NOE_F</sub> <sub>&nbsp;</sub> does not.<span><span>&nbsp; </span></span>
*<span>&lt;span /&gt;</span>Calculate assignment probabilities.&nbsp;&nbsp;&nbsp; It can be done using two different Monte Carlo simulation methods. Namely, Simulating Annealing (SA) { [[FMCGUI commands#Assignment.3ECalculate_Probabilities.3ESA|Assignment&gt;Calculate Probabilities&gt;SA]] } and Replica Exchange Method (REM) { [[FMCGUI commands#Assignment.3ECalculate_Probabilities.3ESA|Assignment&gt;Calculate Probability&gt;REM]] }. Before starting calculations you have to specify control parameters. The main parameters to consider are: ‘''Name of the SA/REM run''’, ‘''Size of the pool for unassigned fragments''’, ‘''number of SA runs''’, ‘''Final Temperature''’ (SA), ‘''Low Temperature''’ (REM), and ‘''NOE contact map''’.&nbsp;The results of the calculations will be stored in the directory under specified name which is created inside PRJ2/assign/ directory. The main result consists of optimal and sub-optimal fragments assignments and assignment probability map. The last one will be automatically loaded in the memory as well.
<div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Calculation of assignment probabilities could be repeated a few times using different methods and parameters. Result of each calculation is stored in a separate directory. Therefore, there could be a few different directories (for example, sa_run1, sa_run2, rem_run0, rem_run1, rem_run3… ) within PRJ2/assign/ directory that contain different assignment probability maps.</span></div><div>&nbsp;</div>
 
=== Step 3.&nbsp;Sequence-specific assignment of PB fragments by analyzing probabilities ===
<div>&nbsp;</div><div><span>&nbsp;&nbsp; A fragment assignment to a sequence position using FMCGUI could be done in two ways, manually and using assignment probability map, respectively. </span></div><div>&nbsp;</div>  
*<span>&nbsp;&nbsp; manual assignment { [[FMCGUI commands#Assignment.3EFix_Assignment.3EManually|Assignment&gt;Fix Assignment&gt;manually]] }. </span>
<div>This command pops up ‘Fragment Property Modification’ window. You can change the assignment ID of any selected fragment using the bottom section of the window. </div><div>&nbsp;</div>
*<span>&nbsp; assignment using probability map&nbsp; { [[FMCGUI commands#Assignment.3EFix_Assignment.3EUsing_Probability_map|Assignment&gt;Fix Assignment&gt;using probability map]] }.&nbsp;You have to select a calculation directory (sa_run# or rem_run#) that contains [[FMCGUI objects#Assignment_probabilities|assignment probability]]  map, P<sub>SA</sub> or P<sub>REM</sub> </span><span>, </span>and to specify the probability threshold&nbsp; P<sub>min</sub>. A fragments ''f'' will be assigned to position ''s'' if the condition <span>P<sup>s</sup>(f)&lt;P<sub>min</sub> &nbsp; is satisfied (see [[Introduction to ABACUS#Figure_1.4|Figure 1.4]]).</span>
<div>&nbsp;</div>
 
=== Step 4. Assignment analysis  ===
<div>&nbsp;</div><div>In the case of poor data, only a part of the fragments get assigned unambiguously. The uncertainty in fragments assignment could be resolved manually with the help of FMCGUI command { [[FMCGUI_commands#View.3EAssignment|View&gt;Assignment]] }. This command pops up “Assignment Graph” window that provides you with graphical tool to visualize the current assignment and to analyze sub-optimal fragment assignments.</div><div>&nbsp;</div>
 
=== Step 5. Final resonance assignment  ===
<div>&nbsp;</div><div>When sequence specific PB fragment assignment is done you have to put in order assigned fragments and to assign CO resonances using command { [[FMCGUI commands#Fragment.3EModify_assigned|Fragment&gt;Modify assigned]] }.</div><div><span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></div><div><br></div>

Latest revision as of 23:48, 5 January 2010

 Step 1. Peak picking of NOESY spectra


         To facilitate peak-picking of NOESY spectra, you can first generate expected tocsy peaks of PB-fragments using FMCGUI:  
       Then, using SPARKY, read the expected peaks into corresponding spectra and complete peak picking manually.
 

Step 2. Probabilistic assignment of PB fragments to protein sequence

According to FMC procedure you have to do the following:
  1.     Probabilistic typing of PB fragments
  2.     Calculate  two fragments contact maps CNOE and CHNCA . CNOE is a contact map based on 15N_ and 13C_NOESY data (it could be calculated by 2 different methods, “abacus” and “fawn”, respectively) and CHNCA is a map based on HNCA data.
  3.    Calculate assignment probabilities by Simulated Annealing (SA) or Replica Exchange Method (REM) Monte Carlo simulations
 
     This can be done by the following commands:    
        You have to consider the warning messages shown in the project main window and to analyze/modify typing probabilities manually by using “Fragment Property Modification” (FPM) window. To open FPM window use command { Fragment>Type>fix }. This window has 3 sections. Top section allows you to select a fragment (by user ID) and modify its typing probabilities. The middle section shows typing probabilities that correspond to the selected amino acid type for all fragments. Here you can fix for any fragment it’s  typing probability corresponding to the selected amino acid type to the values of 1 or 0.
 

      Contact map CHNCA is calculated and loaded in the memory. It is strongly recommended to check the messages in the project main window regarding HNCA peak list. In case there are inconsistencies present in the input HNCA peak list, go back to spectra and fix the list. 

  •   calculate fragment contact map from NOESY data { Assignment>Contacts>NOE>abacus }. In the result, two contact maps, CNOE_B and CNOE_F are calculated and loaded in the memory. Calculation of CNOE_B  makes use of BACUS procedure for NOESY data interpretation, while calculation of  CNOE_F   does not. 
  • <span />Calculate assignment probabilities.    It can be done using two different Monte Carlo simulation methods. Namely, Simulating Annealing (SA) { Assignment>Calculate Probabilities>SA } and Replica Exchange Method (REM) { Assignment>Calculate Probability>REM }. Before starting calculations you have to specify control parameters. The main parameters to consider are: ‘Name of the SA/REM run’, ‘Size of the pool for unassigned fragments’, ‘number of SA runs’, ‘Final Temperature’ (SA), ‘Low Temperature’ (REM), and ‘NOE contact map’. The results of the calculations will be stored in the directory under specified name which is created inside PRJ2/assign/ directory. The main result consists of optimal and sub-optimal fragments assignments and assignment probability map. The last one will be automatically loaded in the memory as well.
         Calculation of assignment probabilities could be repeated a few times using different methods and parameters. Result of each calculation is stored in a separate directory. Therefore, there could be a few different directories (for example, sa_run1, sa_run2, rem_run0, rem_run1, rem_run3… ) within PRJ2/assign/ directory that contain different assignment probability maps.
 

Step 3. Sequence-specific assignment of PB fragments by analyzing probabilities

 
   A fragment assignment to a sequence position using FMCGUI could be done in two ways, manually and using assignment probability map, respectively.
 
This command pops up ‘Fragment Property Modification’ window. You can change the assignment ID of any selected fragment using the bottom section of the window.
 
 

Step 4. Assignment analysis

 
In the case of poor data, only a part of the fragments get assigned unambiguously. The uncertainty in fragments assignment could be resolved manually with the help of FMCGUI command { View>Assignment }. This command pops up “Assignment Graph” window that provides you with graphical tool to visualize the current assignment and to analyze sub-optimal fragment assignments.
 

Step 5. Final resonance assignment

 
When sequence specific PB fragment assignment is done you have to put in order assigned fragments and to assign CO resonances using command { Fragment>Modify assigned }.