Newton is a multi-platform, Java-based framework for spectral analysis of multidimensional NMR data. Newton implements fast maximum likelihood reconstruction (FMLR), a form of spectral deconvolution that constructs the simplest time-domain model whose fourier processed spectrum most closely matches the spectrum of the identically processed FID.  FMLR, as implemented in Newton, is currently being applied to numerous areas of NMR spectral analysis.

 

BACKGROUND

We have developed an algorithm called fast maximum likelihood reconstruction (FMLR) that performs spectral deconvolution of 1D–2D NMR spectra for the purpose of accurate signal quantification. FMLR constructs the simplest time-domain model (e.g., the model with the fewest number of signals and parameters) whose frequency spectrum matches the visible regions of the spectrum obtained from identical Fourier processing of the acquired data. We describe the application of FMLR to quantitative metabolomics and demonstrate the accuracy of the method by analysis of complex, synthetic mixtures of metabolites and liver extracts. The algorithm demonstrates greater accuracy (0.5–5.0% error) than peak height analysis and peak integral analysis with greatly reduced operator intervention. FMLR has been implemented in a Java-based framework that is available for download on multiple platforms and is interoperable with popular NMR display and processing software. Two-dimensional 1H–13C spectra of mixtures can be acquired with acquisition times of 15 min and analyzed by FMLR in the range of 2–5 min per spectrum to identify and quantify constituents present at concentrations of 0.2 mM or greater. 

DOWNLOAD

Newton Installation Instructions

Newton requires that a Java version 1.6 or greater be installed on the host machine. The program runs on multiple platforms (Windows/Mac/Linux) and uses the native widgets of the operating system for all of its graphics.

Windows Instructions

  1. Download the zip archive containing all required contents.
  2. Unpack the archive and move the directory to your system application folder (C:\Program Files)
  3. Open the newton-1.4.10 directory and create a desktop shortcut to newton.bat
  4. Double clicking on the shortcut should launch a shell which then launches the application

Linux/Mac Instructions

  1. Download the zip archive containing all required contents.
  2. Unpack the archive and move the directory to your Desktop
  3. Make sure the newton.sh script is executable by invoking the command chmod +x newton.sh from a command shell in the newton-1.4.10 directory.
  4. The command cd ~/Desktop/newton-1.4.10 appears in your script. Modify this command if the path containing newton.sh is not ~/Desktop/newton-1.4.10
  5. On Mac platforms, you may need to create an association between the newton.sh file and the Application named "Terminal"
  6. Double clicking on the newton.sh file should launch a shell which then launches the application

The script will always launch the most current version of the application registered with this website. If you are offline and have previously run the script, the system will launch the version last downloaded when your machine was online.


MANUAL

Analysis of 2D NMR Data: Getting Started Guide

Purpose

This guide provides step-by-step instructions for performing quantitative analysis of a series of 2D NMR data files by using Newton software for Fast Maximum Likelihood Reconstruction (http://newton.nmrfam.wisc.edu). This guide assumes that the data files have been processed using NMR pipe software (data processed with Topspin is also supported by Newton).

Preconditions

This guide assumes the following setup.

1.       Installation and Launch: The Newton software package Is downloaded and installed from the “Install and Launch” link http://newton.nmrfam.wisc.edu .  If this does not launch properly from your browser, execute the command

javaws http://newton.nmrfam.wisc.edu/app/newton.jnlp

from a command shell.

2.       Example Data Set: Make sure that the example data set directory (Iscu_25Cto70C) for this guide is available on your local computer in a directory with write privileges  (do not install in a directory that has security restrictions such as read-only).

3.       General Requirements for NMR Pipe Data: The following are general requirements that should be met to analyze 2D NMR data with Newton

a.       The data needs to be processed using Topspin or NMR-Pipe

b.      There needs to be one directory per spectrum (e.g. keep the acquired directory format of Bruker and Agilent). Do not put 2 different spectra in the same directory!

 

 

Recommended Directory Structure

 

c.       If the data are processed using NMR pipe, the software will assume that the data conversion and processing scripts and the processed spectrum are located in the same directory (this is not absolutely necessary but is highly recommended!).

d.      Zero filling bug to be fixed: Currently, the software has some problems when you use multiple zero fills along with data extraction (EXT). Until this bug is fixed, you might want to just use the (ZF –auto) command no extra zero fills (slows down analysis anyway with no resolution enhancement).

4.       Configure Data Folder: The data folder should be added to the list of directories within Newton as follows

a.       Launch Newton to access the Newton Browser

b.      Select the “New Data Directory…” menu option from the File menu

c.       Use the “Browse for Folder” dialog to locate the example data directory and click on “OK”. The directory should be added to your browser so that when it is selected, its contents are displayed.

 

Overview

The following steps provide an overview of the process

Step

Description

Load NMR Pipe Scripts

Load the NMR-pipe scripts used for data conversion and processing (This step can be skipped if the data were processed using Topspin).

Create a Data Ensemble

Create an ensemble of all spectra in the study. The ensemble allows all related spectra to be viewed and analyzed in the same session.

Optional: Normalize the spectra

If the sample concentration in the study is not controlled, it is recommended practice to normalize the spectra. A simple normalization technique is to divide each spectrum by the sum of its integral.

Configure Model Parameters

Configure model parameters that control the linewidth starting values and constraints.

Configure Noise Thresholds

Configure noise thresholds for optimization. Noise thresholds are used to decide the weakest signals that are modeled.

Analyze by Deconvolution

Perform deconvolution on the data. This creates a model of all signals contained within the regions of interest. This step may take a while for large data sets (e.g. 10-30 minutes)

Define Regions of Interest

Regions of interest are footprints in the frequency spectrum that can be used to select and annotate peaks. They can be defined automatically, manually, or by importing files from related data sets.

Assign Signals To Region of Interest

Assign the fitting signals to the regions of interest.

Create the “Feature Matrix”

Output the desired “feature matrix” which is a profile of the amplitude for each region of interest per spectrum. The file is displayed in the system editor associated with the file type ‘CSV’ (usually a spreadsheet program like Excel).

Optional: Creating Vector-based Plots

Optional step to create vector-based plots (encapsulated post script) of spectra that are useful for publication.

 

Main Procedure

Load NMR Pipe Scripts

Goal

Import the data conversion and processing information required for analysis by loading the NMR pipe scripts.

Procedure

1.       Select all of the data steps that were processed in the same manner. In this example, select the 10 Grt data sets.

2.       Use the right-click menu to select the “Load NMR Pipe Scripts…” option.

 

3.       Use the file dialog to select the script or scripts required for data conversion and processing.  Do this by clicking on the “Script…” button.

 

4.       In this example, you will want to first select the “fid.com” script

 

5.       Next select the “nmrproc.com” script. The Load NMR Pipe… dialog should display both scripts.

      

 

6.       Click on the “Apply” button and the software should launch a confirmation dialog.

 

 

Create A Data Ensemble

Goal

Create an ensemble of all spectra in the study. The ensemble allows all related spectra to be viewed and analyzed in the same session.

Procedure

1.       Select all data sets in the content window of the Browser Window. This can be done by a click scrolling action or by pressing a “<Cntrl> a” key combination.

2.       Select the option to “Create Ensemble…”

3.       A new ensemble entry should be created that contains 10 data sets.

 

4.       Verify that the entry is valid by launching a viewer session. This can be done with either a double-click action, pressing the <Enter> key with the entry highlighted, or selecting the “View…” option from the menu. Note: It may take a while for spectra to be displayed because the software must create an ensemble matrix of the data sets.

5.       Edit Title (optional): The name chosen in the “DisplayName” field is the title of the spectrum. If the spectrum is derived from Bruker acquired data, this is parsed from the “Title” field. This value can be adjusted manually for each spectrum using the Edit… option (or double click on the entry in the Browser).

6.        Edit All Titles (optional): The titles can also be adjusted manually after creating a DataEnsemble by editing the TitleList xml fields from “Project Info…”

<TitleList>

            <Title>25C</Title>

            <Title>30C</Title>

            <Title>35C</Title>

            <Title>40C</Title>

            <Title>45C</Title>

            <Title>50C-after45C</Title>

            <Title>55C-after45C</Title>

            <Title>60C-after45C</Title>

            <Title>65C-after45C</Title>

            <Title>70C-after45C</Title>

     </TitleList>

 

Optional: Normalize the spectra

Goal

If the sample concentration in the study is not controlled, it is recommended practice to normalize the spectra. A simple normalization technique is to divide each spectrum by the sum of its integral.

Procedure

1.       Select the Reference->Normalize by spectrum sum… menu option

 

2.       After a few seconds, the software will redisplay the spectra showing the contours corresponding to the new normalized values.

3.       Save the session with File->Save.

 

Configure Model Parameters

Goal

Configure model parameters for optimization. These parameters are specific to the biological system and experiment type. This can be done in by either defining a prototype signal or by importing project information from a previous related study.

Define Prototype Signal

1.       If you have not done so, launch the viewer on the ensemble to be analyzed

2.       Go to the “Analyze” tab and click on the “Prototype…” button.

       

 

3.       Draw a box around an isolated signal in the spectrum. This signal should be “typical” of signals within the data set.

     

 

4.       Release the mouse button and the software should launch an analysis dialog. When the software finishes modeling the peak, the dialog will be dismissed.

5.       Go to the “Signal” tab and the spreadsheet should display the modeled signal.

     

 

6.       Use the File->Save option to save the state of the project.

     

 

7.       The parameters of this modeling can be manually adjusted by exiting the viewer and selecting the “Edit Project Info…” option on the ensemble.

     

 

8.       The modeling information is present in the Model element of the xml file.

<Model>

<Parameter fixed="false" shared="true" dim="0" type="Frequency"/>    <Parameter initValue="40.7" fixed="false" shared="true" dim="0" type="DecayRate"/>

<Parameter initValue="1.2" fixed="true" shared="true" dim="0" type="DecayPower"/>

    <Parameter fixed="false" shared="true" dim="1" type="Frequency"/>

    <Parameter initValue="17.5" fixed="false" shared="true" dim="1" type="DecayRate"/>

    <Parameter initValue="1.2" fixed="true" shared="true" dim="1" type="DecayPower"/>

    <Constraint enabled="true" mode="relative" hlim="4.0" llim="-4.0" dim="0" parmType="Frequency"/>

    <Constraint enabled="true" mode="relativeFactor" hlim="3.0" llim="0.25" dim="0" parmType="DecayRate"/>

    <Constraint enabled="true" mode="relative" hlim="4.0" llim="-4.0" dim="1" parmType="Frequency"/>

    <Constraint enabled="true" mode="relativeFactor" hlim="3.0" llim="0.25" dim="1" parmType="DecayRate"/>

</Model>

Import Related Project Procedure

For this alternative procedure, we will copy model parameters from a related procedure.

1.       Exit the viewer of the study to be analyzed.

2.       Select Source: Go to the Browser Window and locate and select the ensemble directory that already exists.

3.       Select the “Copy Project Info…” option.

    

 

4.       Select Target: Locate and select the new study to be analyzed.

5.       Select the “Paste Project Info…” option.

     

 

6.       A dialog should confirm that the project info was transferred.

7.       Launch the viewer of the study and verify that the “Analyze…”  button is enabled on the Analyze tab.

 

 

Configure Noise Thresholds

Goal

This step manually sets the noise threshold. The software starts modelling at the most intense peak and gradually adds less intense peaks until it reaches the noise threshold. This parameter thus determines the weakest peaks that are recognized by the software. Common values are between 4.0 to 8.0 in SD.

Manual Procedure

1.       Exit any viewer that is open.

2.       Select the “Edit Project Info…” option of the ensemble to configure.

3.       Edit the <peakLocator phase="1" negSN="1000.0" posSN="4.0"> negSN and posSN attributes.  A high value for the negSN value essentially means that negative peaks will be ignored. A typical value for SN is between 4.0 and 10.0 depending upon the experiment type.

4.       Save the file.

5.       Relaunch the viewer

 

Import Procedure

The procedure for importing project information described in the previous step will use the peak information from the imported data set.

 

Analyze by Deconvolution

Goal

Perform deconvolution on the data. This creates a model of all signals above the noise threshold. This step may take a while for large data sets.

Procedure

1.       Launch the viewer upon the study of interest (if it is not already launched)

2.       Start Deconvolution: To initiate deconvolution, select the “Analyze…” button on the Analyze tab of the viewer. Alternatively, you can use the “ROI Analyze…” method if regions of interest are defined. This will result in only those peaks located within regions of interest to be modeled.

3.       Analysis will take about half a minute on the example ensemble. A dialogue shows the progress of the analysis.

4.       When the analysis is complete, the dialog will close and peaks will appear on the spectrum where signals were modeled.

 

5.       Viewing Model Spectra: To view the reconstructed model spectra, select the “MODEL_GRID” option from the Matrix menu (context menu on the canvas).

     

 

6.       Viewing Residual Spectra: To view the reconstructed residual spectra (difference between the data and the model), select the “RESID_GRID” option from the Matrix menu.

Define Regions of Interest

Goal

Regions of interest are footprints in the frequency spectrum that can be used to select and annotate peaks. They can be defined automatically, manually, or by importing files from related data sets.

Automated Procedure

1.       If the spectra have been modeled (previous step), the “Auto ROI…” option should be available from the main menu when the ROI tab is selected.

     

 

2.       Selecting this option will result in footprints being constructed that fit the envelope of all peaks. When there is appreciable overlap, ROI’s will be merged.

3.       Click on File->Save to save this information.

Manual  Procedure

1.       ROI’s can be manually drawn by first clicking on the center of the desired ROI on the canvas. This will move the cursor to that position.

2.       With the ROI tab selected, select the ROI->New ROI option. This will create an ROI at the cursor position.

3.       The ROI can be translated using the cursor keys. The corners can be moved with click-drag options.

Import Procedure

1.       If the spectra have been modeled (previous step), the “Import ROI…” option should be available from the main menu when the ROI tab is selected. This option is useful for importing ROI’s constructed from similar experiments on the same biological system.

2.       Delete all of the ROI’s using the ROI->Clear All ROI… option.

3.       Select the File->Import ROI… option.

4.       You can select the “ROIRecords” file that was previously saved to read back all of the ROI’s that were constructed.

 

Assign Signals To Regions of Interest

Goal

Assign the modeled signals to the regions of interest.

Procedure

1.       Make sure the ROI tab is selected.

2.       Select the option ROI->Assign ROI….

3.       Click OK on the confirmation dialog.

4.       After assignment is complete, a dialog will display indicating how many ROI’s were used in the assignment process.

5.       Click on the “Assign” tab and verify that assigned signals are present in the spreadsheet. Clicking on these will show display on the canvas all of the signals associated with a given ROI. When switching to “MODEL_GRID”, the contours will be colored according to the corresponding ROI color of the assigned signal.

6.       Manual Assignment: To change an assignment manually, navigate to the Signal tab and click on an entry. This entry can be assigned to a different ROI (or unassigned from the current) using the ROI Assign button.

 

Create the “Feature Matrix”

Goal

Output the desired “feature matrix” which is a profile of the amplitude for each region of interest per spectrum.

Procedure

1.       Select the Report->Amplitude By ROI… option from the top menu

     

 

2.       In the resulting dialog, you can configure output options. Selection the options shown.

 

3.       Click on the “Run Report” button. The file is displayed in the system editor associated with the file type ‘CSV’ (usually a spreadsheet program like Excel).

 

Optional: Creating Vector-based Plots

Goal

Creates vector-based plots (encapsulated post script) of spectra that are useful for publication.

Procedure

1.       Navigate to any spectrum and region you want to plot.

2.       Select File->Postscript plots… from the top-level menu

3.       A resulting dialog allows you to configure the *.eps file that is generated so that it can have an aspect ratio different from what is displayed on the screen. The file “plot.eps” was generated from the dialog below.

       

 

Plot width

Width of the plot  in pixels

Plot height

Height of the plot in pixels

Plot resolution

Number of pixels/inch

PlotInches

Actual inch width and height of the plot (read-only)

File

Name of the output file (written to the project directory)

Contour width

Pen width of drawn contours

Contour intervals

Number of contours drawn in the plot

Marker width

Pixel width of peak markers

Marker height

Pixel height of peak markers

Show peak markers

Whether peak markers are displayed

Show annotations

Whether annotations are display

Show axes

Whether the axes are labeled.

Use ROI mask

Whether the contours are colored according to assigned ROI

 

 

EXAMPLES

 

Cell Wall Profiling: Getting Started Guide

Purpose

This guide provides step-by-step instructions for performing NMR cell wall profiling using Newton software for Fast Maximum Likelihood Reconstruction (http://newton.nmrfam.wisc.edu).

Preconditions

This guide assumes the following setup.

1.       Installation and Launch: The Newton software package Is downloaded and installed from the “Install and Launch” link http://newton.nmrfam.wisc.edu .  If this does not launch properly from your browser, execute the command

javaws http://newton.nmrfam.wisc.edu/app/newton_cwp.jnlp

from a command shell.

2.       Example Data Set: Make sure that the example data set directory for this guide is available on your local computer in a directory with write privileges. The key directories are:

a.       ArabidopsisDataSet

b.      CellWallPaperMetaData

3.       Configure Data Folder: The data folder should be added to the list of directories within Newton as follows

a.       Launch Newton to access the Newton Browser

b.      Select the “New Data Directory…” menu option from the File menu

 

 

c.       Use the “Browse for Folder” dialog to locate the ArabidopsisDataSet directory and click on “OK”. The directory should be added to your browser so that when it is selected, its contents are displayed.

 

 

Overview

The following steps provide an overview of the process

Step

Description

Create a Data Ensemble

Create an ensemble of all spectra in the study. The ensemble allows all related spectra to be viewed and analyzed in the same session.

Import Grouping Information

Import coding information that identifies the sample group associated with each spectrum.

Configure Regions of Interest

Import and adjust regions of interest so that the footprints of each region align with the corresponding features of the matrix.

Optional: Normalize the spectra

If the sample concentration in the study is not controlled, it is recommended practice to normalize the spectra. A simple normalization technique is to divide each spectrum by the sum of its integral.

Configure Model and Noise Thresholds

Configure model parameters and noise thresholds for optimization.

Analyze by Deconvolution

Perform deconvolution on the data. This creates a model of all signals contained within the regions of interest. This step may take a while for large data sets (e.g. 10-30 minutes)

Assign Signals To Region of Interest

Assign the fitting signals to the regions of interest.

Create the “Feature Matrix”

Output the desired “feature matrix” which is a profile of the amplitude for each region of interest per spectrum. The file is displayed in the system editor associated with the file type ‘CSV’ (usually a spreadsheet program like Excel).

Optional: Creating Vector-based Plots

Optional step to create vector-based plots (encapsulated post script) of spectra that are useful for publication.

 

Main Procedure

Create A Data Ensemble

Goal

Create an ensemble of all spectra in the study. The ensemble allows all related spectra to be viewed and analyzed in the same session.

Procedure

1.    1. Select all 98 data sets in the content window of the Browser Window. This can be done by a click scrolling action or by pressing a “<Cntrl> a” key combination.

2.     2. Select the option to “Create Ensemble…”

        3. A new ensemble entry should be created that contains 98 data sets.

 

4.     4. Verify that the entry is valid by launching a viewer session. This can be done with either a double-click action, pressing the <Enter> key with the entry highlighted, or selecting the “View…” option from the menu. Note: It may take a while for spectra to be displayed because the software must create an ensemble matrix of the 98 data sets. 

5.     5. Edit Title (optional): The name chosen in the “Display Name” field is the title of the spectrum. If the spectrum is derived from Bruker acquired data, this is parsed from the “Title” field. This value can be adjusted manually for each spectrum using the Edit… option (or double click on the entry in the Browser).

 

6.     6. Edit All Titles (optional): The titles can also be adjusted manually after creating a Data Ensemble by editing the TitleList xml fields from “Project Info…”

<TitleList>

            <Title>PMa-36-G</Title>

            <Title>PMa-35-R</Title>

            <Title>PMa-35-Q</Title>

            <Title>PMa-35-P</Title>

            <Title>PMa-35-O</Title>

 

 

Import Grouping Information

Goal

Import coding information that identifies the sample group associated with each spectrum.

Procedure

      1. Use a program like Excel to construct a spreadsheet where each row is a spectrum and two columns exist for each row. The columns contain the sample group and title of the spectrum. Note: The software will assign the group name to the spectrum based on its title so the title must be unique for each spectrum!

 

22 2. Save the spreadsheet as a file in CSV format (not native format like *.xls). For this example, the file “grouping.csv” already exists.

3.   3. If the viewer is not launched on the study, launch a viewing session.

4.    4. Select the menu option Reference->Import Grouping…

 

 5.    5.Use the dialog to locate the file “grouping1.csv” and then click on the “Assign Groups” button.

 

 6.    6. Close and reopen the viewing session (This is temporarily necessary because of a defect in the software. When the defect is fixed, this step can be skipped).

7.      7. Observe that the name of the group now appears on each spectrum.

8.     8. Verify that you can switch modes between “SINGLE”, “FILTER”, “AVERAGE” , etc. When in “FILTER” mode, you will only see spectra belonging to the selected group.

 

Configure Regions of Interest (ROI)

Goal

Import regions of interest (ROI) into the study. Regions of interest both identify the regions of the spectrum to be analyzed and provide a basis for quantification. When the spectra are modeled by deconvolution, the software will be able to report the amplitude of a region of interest based on all of the signals that appear within each ROI footprint.

Procedure

1. 11. Click on the ROI tab in the viewer

2. 22. Select the ROI->Import ROI… menu option either from the top-level menu or a context sensitive menu.

3.    3. Use the dialog to locate the file “ROIRecords.csv” in the file system.

4.   4. After exiting this dialog, a dialog will be launched informing you that 91 records were imported.

5.   5. Verify the following

a.       The ROI tab now contains the information for all regions of interest

 b.      The footprints for each ROI should appear superimposed on the spectra

c.       ROI Adjustment: The footprint of each ROI can be adjusted by clicking on it and using the arrow keys to adjust its position. You can also click and drag the corners.

 d.      Selecting Multiple ROI’s: You can use shift click to select multiple ROI’s.  The key combination of “cntrl-A” will also select all visible ROI’s.

 e.       Creating an ROI: To create a new ROI, click anywhere on the spectrum to move the cursor. Selecting the option ROI->New ROI… will result in an ROI being created at the position of the cursor.

f.        Editing ROI Info: To edit the name and other fields of the ROI, double click on the entry in the ROI spreadsheet or select the “Edit…” option from the context menu.

g.       A dialog will be launched that will allow you to edit the fields manually.

6.       6. Saving ROI Info: The ROI information you create will not be saved until you select the File->Save option.

 

 

 

Optional: Normalize the spectra

Goal

If the sample concentration in the study is not controlled, it is recommended practice to normalize the spectra. A simple normalization technique is to divide each spectrum by the sum of its integral.

Procedure

1.11. Select the Reference->Normalize by spectrum sum… menu option

 

22 2. After a few seconds, the software will redisplay the spectra showing the contours corresponding to the new normalized values.

 333. Save the session with File->Save.

 

Configure Model and Noise Thresholds

Goal

Configure model parameters and noise thresholds for optimization. These parameters are specific to the analysis of NMR data acquired on cell wall material which have much broader line widths than most samples analyzed by NMR spectroscopy.

Procedure

Configuring these parameters manually is beyond the scope of this simple quick start guide. We are going to simply copy parameters from an analogous study of plant cell wall material.

1.1 1. Exit the viewer of the study to be analyzed.

2 2 2. Select Source: Go to the Browser Window and locate and select the study that already exists for Arabidopsis.

3. 33. Select the “Copy Project Info…” option.

 4S 4. Select Target: Locate and select the new study to be analyzed.

5.5 5.Select the “Paste Project Info…” option.

 

 66 6. A dialog should confirm that the project info was transferred.

77. Launch the viewer of the study and verify that the “ROI Analyze…”  button is enabled on the Analyze tab.

 

Analyze by Deconvolution

Goal

Perform deconvolution on the data. This creates a model of all signals contained within the regions of interest. This step may take a while for large data sets (e.g. 10-30 minutes)

Procedure

1.       1. Optional: To speed up deconvolution, you can import a file containing a much smaller number of ROI’s to analyze (e.g. ROIRecords-AromaticOnly).  First clear the study of ROI’s using the ROI->Clear All… option. Import the “ROIRecords-AromaticOnly” file and you will have only 12 regions of interest.

2.      2. Start Deconvolution: To initiate deconvolution, select the “ROI Analyze…” button on the Analyze tab of the viewer.

3.     3. Analysis of the 12 ROI’s will take about several minutes on the Arabidopsis data set. Analysis of all 92 ROI’s takes about 30 minutes. A dialogue shows the progress of the analysis.

4.     4. When the analysis is complete, the dialog will close and peaks will appear on the spectrum where signals were modeled.

5 . 5. Viewing Model Spectra: To view the reconstructed model spectra, select the “MODEL_GRID” option from the Matrix menu (context menu on the canvas).

66.6. Viewing Residual Spectra: To view the reconstructed residual spectra (difference between the data and the model), select the “RESID_GRID” option from the Matrix menu.

Assign Signals to Regions of Interest

Goal

Assign the modeled signals to the regions of interest.

Procedure

1.       1. Make sure the ROI tab is selected.

2.       2. Select the option ROI->Assign ROI….

3.       3.Click OK on the confirmation dialog.

4.       4. After assignment is complete, a dialog will display indicating how many ROI’s were used in the assignment process.

5.       5. Click on the “Assign” tab and verify that assigned signals are present in the spreadsheet. Clicking on these will show display on the canvas all of the signals associated with a given ROI. When switching to “MODEL_GRID”, the contours will be colored according to the corresponding ROI color of the assigned signal.

6.       6. Manual Assignment: To change an assignment manually, navigate to the Signal tab and click on an entry. This entry can be assigned to a different ROI (or unassigned from the current) using the ROI Assign button.

 

 

 

Create the “Feature Matrix”

Goal

Output the desired “feature matrix” which is a profile of the amplitude for each region of interest per spectrum.

Procedure

1.       1. Select the Report->Amplitude By ROI… option from the top menu

 

 2.       2. In the resulting dialog, you can configure output options. Selection the options shown.

 

3.       3. Click on the “Run Report” button. The file is displayed in the system editor associated with the file type ‘CSV’ (usually a spreadsheet program like Excel).

 

 

Optional: Creating Vector-based Plots

Goal

Creates vector-based plots (encapsulated post script) of spectra that are useful for publication.

Procedure

1.       1. Navigate to any spectrum and region you want to plot.

2.       2. Select File->Postscript plots… from the top-level menu

3.       3. A resulting dialog allows you to configure the *.eps file that is generated so that it can have an aspect ratio different from what is displayed on the screen. The file “plot.eps” was generated from the dialog below.

 

Plot width

Width of the plot  in pixels

Plot height

Height of the plot in pixels

Plot resolution

Number of pixels/inch

Plot Inches

Actual inch width and height of the plot (read-only)

File

Name of the output file (written to the project directory)

Contour width

Pen width of drawn contours

Contour intervals

Number of contours drawn in the plot

Marker width

Pixel width of peak markers

Marker height

Pixel height of peak markers

Show peak markers

Whether peak markers are displayed

Show annotations

Whether annotations are display

Show axes

Whether the axes are labeled.

Use ROI mask

Whether the contours are colored according to assigned ROI

 

 

CITATION

R. A. Chylla, K. Hu, J. J. Ellinger, and J. L. Markley, “Deconvolution of Two-Dimensional NMR Spectra by Fast Maximum Likelihood Reconstruction (FMLR): Application to Quantitative Metabolomics,” Anal. Chem. 83, 4871-4880 (2011). PMCID:PMC3114465