Using Autoscreen

Lesson 1: Analyzing 2D 15N-HSQC spectra of calcyclin

This lesson presents the basic steps of SAR by NMR analysis using the Autoscreen module of FELIX. In this lesson you use a set of three ^¹⁵N-enriched HSQC ser files of calcyclin acquired on a Bruker spectrometer and some PDB, CAR, and MDL files.

The topics covered in this lesson are:

Setting up and adding experiments to a project.

Processing a control experiment and setting up a spectral display.

Processing and scoring test experiments.

Analyzing and editing scoring results.

Printing results.

1. Setting up for the lesson

All the files in the $BIOSYM/tutorial/felix/sar directory are required for this lesson.

To copy all files in the sar directory to your working directory, go to your working directory and enter at the UNIX prompt:

     > cp -r $BIOSYM/tutorial/felix/sar .

It is important to use this command so that you retain the relative paths for these files.

The files are briefly described below:

Table 1 Files in the sar directory
File Purpose

sar/1/*

ser and parameter files of control experiment.

sar/2/*

ser and parameter files of first test experiment.

sar/3/*

ser and parameter files of second test experiment.

sar/analysis/exps.txt

A list of experiments, used to add experiments to the Autoscreen project.

sar/ana/ysis/bmrb_assign.tbl

Resonance assignment of some of the HSQC peaks.

sar/str/mols.txt

A list of structures, used to input structural files to the Autoscreen project.

sar/str/demo.pdb

PDB file of the demo molecule.

sar/str/test.car

CAR file of a peptide.

sar/str/test2.mol

MDL MOL file of a small molecule.

Table 1 Files in the sar directory
File	Purpose
sar/1/*	ser and parameter files of control experiment.
sar/2/*	ser and parameter files of first test experiment.
sar/3/*	ser and parameter files of second test experiment.
sar/analysis/exps.txt	A list of experiments, used to add experiments to the Autoscreen project.
sar/ana/ysis/bmrb_assign.tbl	Resonance assignment of some of the HSQC peaks.
sar/str/mols.txt	A list of structures, used to input structural files to the Autoscreen project.
sar/str/demo.pdb	PDB file of the demo molecule.
sar/str/test.car	CAR file of a peptide.
sar/str/test2.mol	MDL MOL file of a small molecule.

The mols.txt file is a list of structural files, with each line specifying the structural file for one molecule, as follows:

     demo.pdb

test.car

test2.mol

The exps.txt file is a list of input experiments and their associated structural files, with each line specifying the experiment ID, ser file (with path relative to /usr/people/cpeng/sarnmr/test/scripps/; see Step 3 for project paths), file type, structural filename (optional), and comments (optional):

     calcyclin	1/ser	ser	demo	control

annexinXI_34	2/ser	ser	test.car	test-1

annexinXI_48	3/ser	ser	mols.txt

2. Starting FELIX

In your working directory, enter this UNIX command to go to the directory for saving the analysis results:

     > cd sar/analysis

Then enter this UNIX command to start the program:

     > felix

If you get the RESTORE LAST SESSION dialog box, select CANCEL.

Select File/Open and set the File Type to DBA. Specify example for the file name and select OK to build a new database.

Note: If you set up an Autoscreen project in the previous session and want to use the same spectrum display and scoring parameters as in that project, select OK in the RESTORE LAST SESSION dialog box. Spectrum-processing parameters are not inherited from session to session.

3. Setting up the project

Select the Autoscreen/Project menu item from the FELIX menu bar. When you see the AUTOSCREEN PROJECT control panel, leave the default Project Name unchanged (sar) and select OK.

You should see this in the text window:

     Created new Autoscreen project 'sar'.

Note: Currently a project name is limited to less than nine lower-case alphanumeric characters. You can create only one Autoscreen project in a database. Once a project is finished, you can select the File/New menu item to open a new database after saving the current one, and then repeat this step to create a new project.

The VERIFY DIRECTORIES control panel, which appears next, allows you to verify some important paths used to access or save the following files:

Raw data files.

Processed matrices.

Molecular files or coordinates.

ASCII files, including file list and results.

For this example, you are using relative paths for all experiments and molecular files, so it is important to verify the project paths.

Make sure that the paths in the VERIFY DIRECTORIES control panel are similar to the following:

Data: /usr2/people/cpeng/sar/
Matrices: /usr2/people/cpeng/sar/analysis/
Molecules: /usr2/people/cpeng/sar/str/
ASCII Files: /usr2/people/cpeng/sar/analysis/

You can click Browse next to any of the paths to select a directory interactively or you can directly enter a directory name.

Select OK.

This displays an empty Autoscreen Experiments Table.

Table 2 Description of items in the Autoscreen Experiments Table
Column Description

id

ID of the experiment.

score

Total score of the experiment.

thresh

Threshold used for peak picking if scored.

status

Status of the experiment, with 0 standing for nonprocessed, 1 for processed, 2 for scored, and 9 for control spectrum.

fid

File name of the FID file, if any.

type

File type of the FID file, with ser for Bruker serial file, fid for Varian FID file, mat for FELIX matrix, 2rr for Bruker processed file, var for Varian processed file, spc for NMR Compass file, ft2 for NMRPipe file, and nmr for TRIAD file.

struc

Filename of a molecule, with extension .pdb standing for PDB file, .car for CAR file, and .mol for MDL file. It can also be filename of a list of molecules if another extension is used.

comment

Comments.

Table 2 Description of items in the Autoscreen Experiments Table
	Column		Description
id	ID of the experiment.
score	Total score of the experiment.
thresh	Threshold used for peak picking if scored.
status	Status of the experiment, with 0 standing for nonprocessed, 1 for processed, 2 for scored, and 9 for control spectrum.
fid	File name of the FID file, if any.
type	File type of the FID file, with ser for Bruker serial file, fid for Varian FID file, mat for FELIX matrix, 2rr for Bruker processed file, var for Varian processed file, spc for NMR Compass file, ft2 for NMRPipe file, and nmr for TRIAD file.
struc	Filename of a molecule, with extension .pdb standing for PDB file, .car for CAR file, and .mol for MDL file. It can also be filename of a list of molecules if another extension is used.
comment	Comments.

Note: If you want to change the project paths again, use the Edit/Verify Directories menu item in the table or the Autoscreen/Experiment/Verify Directories item on the main menu bar.

4. Adding experiments to a project

The following steps demonstrate three ways of adding experiments to an Autoscreen project.

Option 1 - Add one experiment at a time

Select the Autoscreen/Experiment/Add One menu item.

In the ADD ONE EXPERIMENT control panel, select Bruker (ser) as the Spectrum File Type, enter Control as the Experiment ID, and fill in the Comment with This is the control experiment.

If you want, you may turn on the Molecule toggle and click the Browse button. When the SELECT MOLECULAR FILE OR FILE LIST file browser appears, select demo.pdb and select OK to return to the ADD ONE EXPERIMENT control panel.

Finally, select the ser file under the directory sar/1/, and select OK to add this experiment to the project.

The Autoscreen Experiments Table is updated with the newly added experiment.

The ADD ONE EXPERIMENT control panel is displayed again for you to add another experiment. Select Cancel if you are finished.

Note: If you select a molecule file interactively, be sure to do so before selecting the ser file. Otherwise, FELIX does not "remember" that you selected the ser file and you will have to do it again.

Option 2 - Add experiments from all files in a directory

First clean up from your test of the previous option:

Highlight the row of interest (or all rows if you added more than one experiment) in the Autoscreen Experiments Table and select the Edit/Delete Experiments menu item to remove the experiment.

Next, add experiments from all files:

Select Autoscreen/ Experiment/Add All Files. In the ADD ALL EXPERIMENTS control panel, change To Dir to 3, the highest experiment number. Then select OK.

All three experiments are added to the project, with the first experiment taken as the control spectrum.

Note: This function expects numbered Bruker experiments. If the experiments are not consecutively numbered, you can instruct the program to skip one or more between every two experiments, or it will automatically ignore nonexistent experiments.

Option 3 - Add all experiments listed in a file

First clean up from your test of the previous option:

Highlight all rows in the Autoscreen Experiments Table. From the table, select the Edit/Delete Experiments menu item to remove all the experiments.

Next, add experiments by reading a list in a file:

Select the Autoscreen/ Experiment/Add From File List menu item and, in the ADD EXPERIMENTS FROM FILE LIST control panel, select the file exps.txt and select OK.

This adds three experiments to the project, using the experiment list in the exps.txt file. These experiments are the ones that we will use in subsequent processing and scoring in this lesson.

5. Processing a control spectrum

Unless you are using processed data, for example, FELIX matrices or Bruker 2rr files, one of the most important steps in using Autoscreen is processing the control spectrum. The processing parameters used during this procedure are used for the subsequent processing of all other test experiments.

Highlight the calcyclin experiment in the Autoscreen Experiments Table and select Action/Process Control Spectrum from the menu bar of the table.

Next you are guided through the processing of this 2D experiment. The procedure is similar to standard 2D processing in FELIX.

Keep the default parameters unchanged and select OK in the 2D HEADER INFORMATION and 2D ACQUISITION INFORMATION control panels.

In the 2D DATA PROCESSING control panel, which appears next, select Automatic for Phasing Mode and Facelift for Baseline Correction. Select OK.

You will use the automatic phasing function (called PAMPAS) and baseline-correction function (called FACELIFT) after the Fourier transform.

Note: PAMPAS automatically determines phasing parameters for a processed matrix. You will be prompted to set parameters for it later. If you want to phase interactively, select Interactive as the Phase Mode and enter 3 for Fid to Phase.

Select OK in the SINEBELL PARAMETERS control panel. If it warns you about overwriting an existing file, click Overwrite.

In the 2D DATA PROCESSING control panel, check Linear Prediction, select Automatic as the Phasing Mode, and Facelift as the Baseline Correction. Select OK.

In the GENERAL LINEAR PREDICATION control panel, change the Number of Coefficients to 8 and select OK.

Select OK in the SINEBELL PARAMETERS control panel.

In the AUTOPHASING (PAMPAS) PARAMETERS control panel, check Correct D1 and Correct D2. Under Excluded Areas, check #1. Click the Cursor button on the same line to set the excluded area interactively.

The purpose of this action is to exclude the water signals while determining the phase parameters. The cursor changes to a cross, allowing you to define a range to exclude for D1

Click at a point between the real peaks and the water signals, keep the left mouse button depressed, and drag the cursor to the right limit of the spectrum (the Y coordinates do not matter), then release the mouse button.

The same control panel appears again with the excluded range displayed in points (for example from 374 to 512). If necessary, you can edit these numbers in the entry boxes.

Select OK.

The spectrum is automatically phased in both dimensions. In the text window, the determined phase parameters and other information are displayed.

Note:

If you choose to apply PAMPAS to the control spectrum, it is applied to all test spectra later. Since it determines phase parameters for each individual spectrum (instead of applying the same phase parameters as for the control spectrum), spectra with different phase errors are not a problem. For more reliable phasing results, it is important to exclude noise when determining phase parameters.

If you choose to extract a portion of the spectrum during processing (e.g., to extract the left half) and the truncation happens on the water ridge, autophasing restores the truncated water spectrum on the left side of the spectrum. To avoid this, use interactive phasing instead of autophasing.

Finally, the BASELINE CORRECTION (FACELIFT) PARAMETERS control panel appears. Make sure D1 and D2 are checked for Correction Dimension and select OK.

The processed HSQC spectrum is now displayed as contours.

6. Setting display reference and display limits

The reference, display limits, and threshold are set using the general FELIX menu items or icons.

Select Preference/Reference. In the REFERENCE MATRIX control panel, set these parameters:

Reference Point
D1: 462
D2: 64

Reference Shift
D1: 4.7
D2: 117.99

Axis Text
D1: D1_H1
D2: D2_N15

Select OK.

Use the Zoom icon on the FELIX tool bar to zoom in on the fingerprint area.

Select Preference/Plot Parameters. In the PLOT PARAMETERS-BASIC control panel, enter 0.025 as the Contour Threshold. Select OK.

Finally, select Autoscreen/Save Limits and Reference to save the reference, limits, and threshold.

These parameters will be used for display, hardcopy, and scoring of all experiments in the project.

Note:

Note: If you make any changes to these parameters, be sure to use Autoscreen/Save Limits and Reference to save them - otherwise the changes are lost. You can change the display limits and threshold at any time (for example, after processing and scoring some test spectra), but the reference must be set before you select Autoscreen/Setup Scoring to define scoring parameters.

7. Setting other display parameters

Many other display parameters can be changed and saved along with the project by selecting Autoscreen/Setup Display. These include the display parameters for the control spectrum and test spectra in contour mode and overlay mode and those for display of control peaks. For this lesson the default values are used.

Note:

If you have processed any test spectra, the overlay of the first available test spectrum over the control spectrum is displayed after you select Autoscreen/Setup Display. Otherwise, only the control spectrum is displayed.

When you display a single spectrum, the parameters in the Control Spectrum group are used if it is the control spectrum, and the those in the Test Spectra group are used if it is a test spectrum.

If you display an overlay of two or more spectra, the control spectrum is always used as the base spectrum and the control peaks are displayed if you select this option.

The parameters in the Cross Peaks group are used only for displaying the control peaks. Test peaks are never displayed. If you check the Draw Cross peak on Control option, control peaks are displayed on the control spectrum - either in single spectrum or overlay mode.

These parameters are saved along with the project. The next time you open the project they are automatically loaded.

8. Picking peaks and importing an assignment for the control spectrum

In the Autoscreen Experiments Table, double-click the Control spectrum to display it.

Select Peaks/Pick Region from the main menu and use the default parameters to pick all the fingerprint peaks.

About 82 peaks are picked and displayed in the Peaks-xpk:peaks table.

Note:

You should pick all peaks as control peaks, even if you are only interested in some of them. This guarantees better matching of control peaks to test peaks in the subsequent scoring process. You can define those interesting peaks as ROI peaks (see Step 15 for further details).

You should also use identical peak-picking parameters for both control peaks and test peaks to avoid artificial peak displacements. For this purpose, always select Regular as the Peak Picker in the ND PEAKPICK PARAMETERS control panel, since that is always used during peak picking of a test spectrum.

Select the Autoscreen/Import Assignments menu item. In the IMPORT ASSIGNMENTS control panel, select BMBR Assignment Table as the Assignment File Type and select the file bmrb_assign.tbl from the browser.

Select OK.

The text window reports that 17 peaks have been assigned. The assignments are also updated in the Peaks-xpk:peaks table.

To display the assignments on the spectrum, select Autoscreen/Setup Display, select Residue for Peak Labels, and select OK.

Note: The bmrb_assign.tbl file is not a real or complete file and should only be used for demonstration purposes.

9. Setup of scoring parameters

In this step you set up parameters for peak picking in and scoring of test spectra.

Select Autoscreen/Setup Scoring. In the 2D SCORING PARAMETERS control panel, click the Advanced button to review parameters in the ADVANCED PARAMETERS FOR 2D SCORING control panel. Leave the default values unchanged and click Cancel to return to the 2D SCORING PARAMETERS control panel. Leave its default values unchanged and select OK.

Note: The following briefly explains the parameters for scoring:

Basic parameters (in 2D SCORING PARAMETERS control panel):

Control Peak Table: The name of the peak table to be used for control peaks.

Peak Displacement Limits: D1 (or D2) Minimum is the minimum chemical shift difference in D1 (or D2) required for a peak to be considered as displaced by protein-ligand interaction. D1 (or D2) Maximum is the maximum possible peak displacement expected in D1 (or D2). All values are in ppm.

Scale Factors: The coefficients used to calibrate the chemical shift displacements in D1 (or D2). Default values are 1.0 and 0.2 for D1 and D2, respectively.

Advanced parameters in ADVANCED PARAMETERS FOR 2D SCORING control panel:

Peak Picking Parameters for Test Spectra: In this section, the Threshold Method determines how to set the threshold for peak picking in the test spectra. If the default Automatic is selected, FELIX automatically adjusts the threshold so that peaks are picked with a reasonable quantity and quality. If Control is selected, the same threshold as used for the control spectrum is used for all test spectra. If Define is selected, the value you enter in the associated entry box is used.

All other parameters in this section are the same as for standard 2D peak picking. The same values as for the control spectrum are recommended for them.

Peak Matching: Parameters in this section determine how test peaks are matched to control peaks during automatic scoring.

� The toggles Use Peak Widths and Use Peak Heights, if checked, allow the shape of peaks to be considered.

� If peak shape is used, Minimum Shape Similarity is a threshold for two peaks to match.

� Search Methods provides two alternative algorithms to search a best match between the test and control peaks. If Tree Search (the default) is selected, a heuristic depth-first search method is applied. Since this can be time-consuming, you can limit the CPU time spent on each test spectrum by defining a value for CPU Time Limits (default = 10 s). If Simulated Annealing is selected, the stochastic method is applied. The latter is usually fast (so CPU Time Limits is not used), yet does not guarantee a best match. The latter method is recommended when the spectra are so complicated that tree searching does not give satisfactory results in a reasonable amount of CPU time.

Unmatched Peaks: Parameters in this section determine how to deal with peaks that do not have a match in either the control or test spectrum.

� For each unmatched control peak, if Fit to Test is selected, it will be fitted to the test spectrum using the peak optimization function if the percentage of unmatched control peaks has not exceeded the Maximum (%). (For details see Peaks/Optimize (page -200) in Chapter 4, Processing, visualization, and analysis interface (1D/2D/ND) in the FELIX User Guide.) If the fitting is successful, the optimized peak is taken as its matched test peak. Otherwise it remains unmatched, and a Penalty (default 0.60) contributes to the score of the experiment.

� For unmatched test peaks, Selection allows you to select the method to define them. If None is selected, such peaks are ignored. If Close to Control Peaks is selected, only those that are close to at least one control peak, namely with displacements not bigger than the Maximum D1 (or D2) Peak Displacement Limits, are included. Otherwise, if All is selected, all peaks are included. When determining if it is a legitimate test peak, FELIX compares the peak widths and height of a test peak with the statistics of all the matching test peaks. The parameter Num.of RMSD (default 2.0) then allows you to define a tolerable deviation from the average peak widths and height. Finally, the Penalty is the contribution of each unmatched test peak to the score of the experiment. The default (0.2) is smaller than for an unmatched control peak (0.6), because automatic picking of test peaks is usually less reliable than picking of control peaks, which is normally done manually where refinement is possible.

A Peak Displacement Table is displayed. The table contains the following columns:

Table 3 Columns in the Peak Displacement Table
Column Description

id

Numbering of the control peak in Peaks-xpk:peaks table. It is always used for identifying a control peak.

asg1

The assignment of the control peak in D1. If not assigned, value is "null".

asg2

The assignment of the control peak in D2. If not assigned, value is "null".

cntrib

The contribution of the peak to the total score of the experiment. It is usually calculated based on the shift1, shift2, shape, and weight.

shift1

The absolute chemical shift displacement between the matched peak pair in ppm along D1. If no matching test peak, it remains zero.

shift2

The absolute chemical shift displacement between the matched peak pair in ppm along D2. If no matching test peak, it remains zero.

shape

The similarity of the shapes of the matched peak pair; 0 = least similar, 1 = identical. If the peak shape is not used for scoring, value = 1.

weight

Weight of contribution of the peak to the total score of the experiment. An ROI (region of interest) peak has weight greater than 0. By default all peaks in Peaks-xpk:peaks table are taken as ROI peaks when setting up scoring. You can change the weight of a peak manually from the table.

tstcen1

The chemical shift of the matching test peak in ppm along D1. If no matching test peak, it remains zero.

tstcen2

The chemical shift of the matching test peak in ppm along D2. If no matching test peak, it remains zero.

Table 3 Columns in the Peak Displacement Table
	Column		Description
id	Numbering of the control peak in Peaks-xpk:peaks table. It is always used for identifying a control peak.
asg1	The assignment of the control peak in D1. If not assigned, value is "null".
asg2	The assignment of the control peak in D2. If not assigned, value is "null".
cntrib	The contribution of the peak to the total score of the experiment. It is usually calculated based on the shift1, shift2, shape, and weight.
shift1	The absolute chemical shift displacement between the matched peak pair in ppm along D1. If no matching test peak, it remains zero.
shift2	The absolute chemical shift displacement between the matched peak pair in ppm along D2. If no matching test peak, it remains zero.
shape	The similarity of the shapes of the matched peak pair; 0 = least similar, 1 = identical. If the peak shape is not used for scoring, value = 1.
weight	Weight of contribution of the peak to the total score of the experiment. An ROI (region of interest) peak has weight greater than 0. By default all peaks in Peaks-xpk:peaks table are taken as ROI peaks when setting up scoring. You can change the weight of a peak manually from the table.
tstcen1	The chemical shift of the matching test peak in ppm along D1. If no matching test peak, it remains zero.
tstcen2	The chemical shift of the matching test peak in ppm along D2. If no matching test peak, it remains zero.

Note:

The Peak Displacement Table is updated when you score a spectrum, when you display a histogram, or when you display the overlay of a test spectrum over a control spectrum. The ID of the current test spectrum is displayed in the title of the table.

All peaks are taken as ROI peaks by default (i.e., weight = 1). If you are interested in only a subset of these peaks, see Step 15 for more information.

10. Processing and scoring test spectra

Once you have set up the scoring parameters, you can process and score all the test spectra.

Select Autoscreen/Go from the Autoscreen Experiments Table.

The two test spectra are processed and scored against the control spectrum in turn, then a histogram of scores vs. experiments is displayed. The Autoscreen Experiments Table is updated with the scores and status of the test experiments.

Note:

For each test spectrum, a summary of the scoring results is displayed in the text window. Note the fitting of unmatched control peaks to the test spectrum and the identification of unmatched test peaks. This information is also saved in an ASCII file named as TEST_CONTROL.sco in the directory defined by the project path for ASCII files (see Step 3 for more about project paths), where TEST and CONTROL are the IDs of the test and control experiments, respectively. The contents of this file are automatically displayed in the text window when you double-click the test spectrum in the Autoscreen Experiments Table.

Although unmatched test peaks contribute to the score of the experiments, they are not saved in the Peak Displacements Table. So double-clicking the test spectrum in the Autoscreen Experiments Table is the only way to view them in the text window.

Experiment annexinXI_34 shows a higher score than the other experiment, which usually indicates a stronger binding of the ligand to the protein.

There are five methods for processing and/or scoring test spectra on the Action menu in the Autoscreen Experiments Table, which are used for different purposes:

Method Purpose

Process Selected

Process the highlighted spectrum or spectra. If the control spectrum is selected, it is ignored.

Score Selected

Process (if not processed) and score the highlighted spectrum or spectra. If the control spectrum is selected, it is ignored.

Go

For each of the test spectra, process it if not processed and score it if not yet scored.

Rescore All

Re-score all test spectra.

Reprocess/Rescore All

Re-process and re-score all test spectra.

Method	Purpose
Process Selected	Process the highlighted spectrum or spectra. If the control spectrum is selected, it is ignored.
Score Selected	Process (if not processed) and score the highlighted spectrum or spectra. If the control spectrum is selected, it is ignored.
Go	For each of the test spectra, process it if not processed and score it if not yet scored.
Rescore All	Re-score all test spectra.
Reprocess/Rescore All	Re-process and re-score all test spectra.

Note: The Peak Displacement Table is not updated at this moment. To display and update it for a certain spectrum, double-click it in the Autoscreen Experiments Table.

11. Viewing clusters

Using the Autoscreen/View Clusters menu item groups experiments that share common displaced peaks, providing a way to locate the residues of the protein whose chemical shifts were affected by the close contacts of the ligand in different experiments.

Select Autoscreen/View Cluster, leave the default value of Cluster Threshold unchanged, and select OK.

The score matrix is displayed showing one cluster in green.

Move the crosshair cursor over the green area to display the peak number, experiment name, and contribution of that peak to that experiment. Press <Esc> when you are done.

Note:

The experiment numbers and peak numbers are reshuffled, so you must use the crosshair cursor (automatically displayed after selecting Autoscreen/View Cluster) to identify the peaks and experiments in the clusters. If you want to return to the crosshair cursor after pressing <Esc>, select Autoscreen/View Clusters again.

If you want to ignore peaks that have small displacements, select Autoscreen/View Cluster and increase the value of Cluster Threshold in the VIEW CLUSTER control panel. Peaks with a contribution smaller than the Cluster Threshold are ignored.

12. Analyzing the scoring results

Once you have an overview of all experiments, you can investigate the interesting experiments and interesting peaks.

First highlight Experiment annexinXI_34 (the one with the highest score) in the Autoscreen Experiments Table and click the Peak Contribution Histogram icon. Leave the default values in the PEAK CONTRIBUTION HISTOGRAM OPTIONS control panel and select OK.

A histogram of contributions vs. peaks for this experiment is displayed. The Peak Displacement Table is updated with the scoring data for Experiment annexinXI_34.

To view the peak displacements, double-click Experiment annexinXI_34 in the Autoscreen Experiments Table.

The overlay contours of Experiment annexinXI_34 over the control spectrum are displayed, together with the displacement arrows and control peak labels. The scoring results are also summarized in the text window.

To get a clearer view of the displacement arrows, you can:

Select Autoscreen/Setup Display to change the Peak Symbol, Peak Labels, and other settings.

Double-click the row of an interesting peak in the Peak Displacement Table to zoom in on that peak, or highlight several peaks in the table and click the Zoom on Peaks icon to zoom in on them. You can also use the Zoom icon on the FELIX tool bar or the <+> and <-> keys on your key pad to change the zoom ratio.

In the Autoscreen Experiments Table, click the Peak Contribution Histogram icon again to display the histogram of contributions vs. peaks of Experiment annexinXI_34.

Peak 73 has the largest contribution and seems to be an interesting peak.

On the Peak Displacement Table, click the Sort Contributions icon.

The peaks are now listed in descending order of their contributions to the score.

Highlight the first row, Peak 73, and click the Zoom on Peaks icon or simply double-click the row of peak #73.

The display zooms in on the displacement between control peak 73 and its matching test peak.

To display the titration of Peak 73, that is, its contributions in different experiments, highlight this peak in the Peak Displacement Table and click the Titration icon.

A histogram of contribution vs. experiments is displayed. This also shows that this peak has a much greater displacement in Experiment annexinXI_34 than in the other experiment.

13. Manually editing scoring results

In the Autoscreen Experiments Table, highlight Experiment annexinXI_34 and click the Overlay icon.

Experiment annexinXI_34 is displayed over the Control spectrum together with the displacement arrows.

Click the Undo Sort Contributions icon in the Peak Displacement Table, then double-click peak 7 in the Peak Displacement Table to zoom in on the spectral area around it.

All control peaks appear to be correctly matched to the test peaks, so manually editing is not needed in this experiment. For demonstration purposes, the following operations assume that you do not like the currently matched test peads for control peaks 7 and 9 and want to change them.

To remove the current matching, select Edit/Remove Displacement from the Peak Displacement Table and click control peaks 7 and 8.

This erases the displacement arrows.

Press <Esc> to exit this mode.

Now select Edit/Change Displacement from the Peak Displacement Table and click control peak 7.

Keep the mouse button depressed and drag the cursor to the center of the test peak at (7.30,124.23) and release the button.

Repeat this for control peak 8 so that it is matched with the shoulder peak centered around (7.27, 124.46).

Press <Esc> to exit this mode.

This matches peaks 7 and 8 to the desired test peaks. All the changes you've made are reported in the text window and updated in the Peak Displacement Table.

Note:

A control peak can be matched to only one test peak. This means that changing the displacement of a control peak automatically erases the original displacement.

A displacement you define is scored the same way as an automatically determined one, except that there are no minimum and maximum limits and the destination is not checked.

14. Exporting scoring results

Select Autoscreen/Export Score. In the EXPORT SCORES control panel, set Contents to All Scores and Delimiter to Tab or Space. Enter the filename Test2 under Selection (a .dat suffix will be added automatically).

Select OK.

All the experiments and their scores are listed in the Test2.dat file.

Repeat the previous box with Contents set to All Scores Sorted.

This lists all experiments and scores in descending order.

Repeat the first box of this step with Contents set to Scores and Titration, Number of Experiments set to 2, and Number of Peaks set to 10.

The scores of the top two experiments, in descending order of scores, and the contributions of the top 10 peaks that have the greatest sum of contributions to the two experiments are reported.

Note: This function is intended to give a summary of the "interesting peaks in the interesting experiments." You can choose the numbers of experiments and peaks to report.

<Shift>-click to select the two test experiments in the Autoscreen Experiments Table and <Ctrl>-click to select peaks 60, 63, and 73 in the Peak Displacements Table. Then repeat the first box of this step with Contents set to Titration Selected, Delimiter set to Tab, and Use Comments as Concentration toggled off.

The contributions of the selected peaks in the selected experiments are reported. Such a report is intended for calculation of K_{_d} based on titration. If you have specified the concentration of the experiments in the comment column in the Autoscreen Experiments Table, you can toggle on Use Comments as Concentration to include that information in the report.

Repeat the first box of this step with Contents set to C2 QSAR Table.

All experiments and scores are listed in a format suited for QSAR study with the Cerius^² program.

To import Autoscreen results into Cerius^² for QSAR study, take the following steps:

1. Start Cerius². Select the QSAR deck and click the Show Study Table item on the QSAR card. This brings up a new, empty QSAR Study Table.

2. In the QSAR Study Table, select File/Import... In the Import/Table control panel, uncheck File Contains Row labels, check File Contains Column Labels, select the filename from the list box, and click Import. The experiments and score are displayed in the Table Manager.

15. Scoring again with ROIs (regions of interest)

After getting an overview of all peaks in all experiments, you may want to focus on some interesting peaks in some interesting spectra instead of looking at all of them. When setting up the scoring parameters (see Step 9), all peaks in the Peaks-xpk:peaks table are taken as ROI peaks with weight equal to 1.0 by default. The following boxes demonstrate some of the methods for defining a subset of the peaks as ROI peaks.

Highlight the calcyclin spectrum in the Autoscreen Experiments Table and click the Draw icon to display it.

If the peak labels and peak numbers are not displayed, select Autoscreen/Setup Display and choose Small Cross for Crosspeak Symbol and Number # for Peak Labels.

Select OK.

Select Autoscreen/Define Region of Interest and click the tear-off line on the pullright menu. Move the Define Region of Interest menu to a convenient place on your screen.

Select Remove All.

This sets all peaks as nonROI peaks with weight equal to 0. Note the change of their color in the spectrum window.

Select Add Region. Drag out a rectangle around the peaks with H^¹ chemical shift greater than 9.0 ppm.

Note the change of color of these peaks and the report in the text window:

     Displaying 9 ROI peaks. Total 82 peaks.

In the Peak Displacement Table only these ROI peaks have non-zero weights.

Select Add One Peak. Click Peaks 1-5 and 73. Then press <Esc> to quit.

You have now about 15 ROI peaks. See the text window again for the number of ROI peaks.

After defining ROI, you can rescore the spectra you are interested in. Highlight Experiment annexinXI_34 in the Autoscreen Experiment Table and select Action/Score Selected Spectra.

Its score is reduced to 2.285. Only the ROI peaks, displayed in yellow by default, show displacement arrows in the spectrum window. Both the Peak Displacement Table and the text window show the scoring contributions of the ROI peaks only.

Click the Peak Contribution Histogram icon with Experiment annexinXI_34 still highlighted, to see the histogram of the contributions of the ROI peaks. You can select either Peak IDs or Residue Numbers as the x coordinates.

To score all experiments based on the newly defined ROI peaks, select Action/Re-score All Spectra and select OK in the control panel.

Note:

Only peaks from the original Peaks-xpk:peaks table can be defined as ROI or non-ROI peaks. If you want to add new peaks, you have to do peak picking with the Peaks menu and set up scoring parameters again with Autoscreen/Setup Scoring.

When you set up scoring again, you lose all scoring results if you have changed the Peaks Table.

If you zoom the spectra or resize the spectra window, the color of ROI peaks is lost. You can select Autoscreen/Define Region of Interest/Draw ROI to restore their color. You can define the color of ROI peaks with the Autoscreen/Setup Display menu item.

16. Printing spectra and histograms

To set up for printing, select Autoscreen/Setup Print. In the HARDCOPY PARAMETERS control panel, select Redraw Overlay as Plot Selection and PostScript as the Plot Device. Check Print Parameters and Send to Printer. Enter a UNIX command as the Print Command (for example, lpr -Pprintername).

Select OK.

In the Autoscreen Experiments Table, highlight Experiment annexinXI_34 and click the Print icon.

The text window reports the printing command it issues. You should receive a hardcopy of the overlay of annexinXI_34 over the Control spectrum, along with the displacement arrows and spectrum title and parameters.

Note:

Autoscreen tries to search the Bruker file title (in pdata/1/ subdirectory) or the Varian file text (in the same directory) and use the first four lines as the spectrum title. If the file is not found, Autoscreen uses the experiment ID as the title.

By default, Autoscreen redraws the highlighted spectrum as a contour or overlay within the limits you saved using Autoscreen/Save Limits and Reference. If you want to print whatever is displayed in the spectrum window (for example, a zoomed peak or a histogram), select Autoscreen/Setup Print and select No Redraw for Plot Selection. In this way, Print doesn't redraw the screen before printing. You can also use File/Print from the main menu.

17. Displaying molecules and scores in Insight II

As described in Step 1, some experiments are associated with a molecule file for demo purposes.

To display experiments associated with a molecule, first start Insight II from another UNIX shell window and go to the NMR_Refine module.

In the Autoscreen Experiments Table, highlight the calcyclin experiment and then click the Display Molecule icon to display the demo molecule.

Highlight the annexinXI_34 experiment in the Autoscreen Experiments Table. Select Action/Color Scores from the table.

In the COLOR RESIDUES BASED ON SCORE control panel, select InsightII as the Format and enter color as the filename.

Four assigned peaks with non-zero contributions are exported into file color.tab.

In Insight II, select Session/Change_Directory to move to the directory sar/analysis.

Select Query/Color_By_SAR_Score. In the control panel, set these parameters:

Sar_Molecular_Name DEMO
SAR_Score_Filename color.tab
Neutral_Color white
Num_Intervals 10
Low_Score 0.02
High_Score 0.2

Select Execute.

Note: If you are using Insight II version 980 or older, in which the Query/Color_by_Sar_Score command is not available, you can select File/Source_File to open the color_by_score.bcl script file to set up this command. The script file resides in the sar directory.

Wait until the rendering of the Connolly solid surface is complete. The residues that contribute to scores (between 0.02 and 0.2) are displayed in red.

18. Exiting FELIX

To exit FELIX, select File/Exit.

Note:

In real-world practice, you may want to start a new project at this point. To do so, select File/Save to save the current database, then select File/New to start a new database. Next follow Step 3 to start a new project.

If you exit FELIX before finishing a project, be sure to save the database when you exit (it is not important to save the session). In the new session you can open the saved database and then select Autoscreen/Project to load the project.