TRENDset – Batch Fitting of Binding Affinities (or Time Courses)

Tutorial

Description

  • TRENDset works on a batch of titrations to determine the binding isotherm and KD of each titration. The set of results are then sorted by KD
  • Similar to TREND NMR, TRENDset has two menus, spectra and `fid, which function similarly.
  • TRENDset is launched using trendset.app or trendset.exe. After the batch run finishes, an HTML report with suffix of -batch_run_result.html will be opened in a browser window automatically.

    Arguments – Read from an index file

  • index
    Unlike Trendmain or TREND NMR, TRENDset can read most options from index file, which is read by index File Chooser Widget.
    trendset_1
    The index file should be a CSV format spreadsheet. Each column in the spreadsheet contains a parameter for the calculation. The title appears in the first row of the speadsheet. Three arguments are required to be be given in a index file. Other arguments are optional.

    • Required Arguments:
      • FILE_LOC: File location
        FILE_LOC specifies the location of each file or directory to be used for PCA calculations. It can be given as the absolute path from the root directory, such as /Users/username/data/titration1/1.ucsf. FILE_LOC can also be a relative path to the index file. For example, if an index file index.csv is saved as /Users/username/data/index.csv, then the previous 1.ucsf can be given as a relative path from a current directory by listing it as: titration1/1.ucsf.
        For Bruker Topspin directories, FILE_LOC should specify the <expno> level of the directory tree because <expono>/ contains a processed spectrum as the 1r or 2rr file in its /pdata/1 subdirectory. Since the raw data is present in the fid or ser file in the same <expno> directory, the fid menu can read this raw data by listing the <expno>/ directory in the same way. See brukerspectra and brukerfid in Trendmain manual for details.
        TRENDset also reads data from Agilent or Varian VnmrJ directories containing an fid file or datdir/phasefile. See ‘agilentfidandagilentspectra` in Trendmain manual for details.
      • EXP_NAME: Experiment name
        TRENDset works on a batch of NMR spectra of titrations or time-dependent measurements. EXP_NAME names each titration (or time series) for your convenience.
      • XAXIS: X axis
        In the XAXIS column, the X axis (e.g. concentration of the ligand) is defined for each file. It will be used as the x-axis values for data fitting. It is the same as the xaxis options of Trendmain and TREND NMR.
        An example of an index file directing the fittin of two titrations is:
        index
    • Optional Arguments for available in TRENDset:
      These arguments are optional and most of them can be set in the GUI of TRENDset as “global” arguments. However these arguments can also be defined in the index file to overwrite the “global” arguments. Each experiment (EXP_NAME) may have its own optional arguments.
      The below arguments are used to do PCA calculation and their details can be found in the Trendmain GUI manual and Trendmain CLI manual.

      • FILETYPE: type option in trendmain or trendnmr for choosing file format. Although TRENDset can determine the file type automatically. the user can also specify FILETYPE argument. You can use this argument to analyze FID data in the spectra menu of TRENDset.
      • SCALING: This indicates the caling method to apply to rows of the data matrix. Autoscaling is usually preferred. However, in titrations with intermediate exchange broadening, Pareto scaling is preferred because it corrects sigmoidal distortion of binding isotherms. See futhur description of scaling in the documentation of (Trendmain GUI manual and Trendmain CLI manual.
      • COLUMNSCALING: Applies the chosen scaling method to the columns of the data matrix.
      • XUNIT: specifies the units of the X-axis (given by XAXIS column in the index file).
      • THRESHOLD and THRESHOLDNUMBER: These are used as described for trendmain and trendnmr, in order to filter noise from NMR spectra.
      • BINTIMES: Specify the number of points to be merged together into a single point for the purpose of compressing the spectra for a faster calculation. See the documentation of trendmain and TRENDNMR.
      • SOLVENTFILTER: As described for handling FID signals in the documentation of trendmain and TREND NMR, solventfilter specifies a funciton to use to subtract the solvent filter signal from the FID.
      • XLIM and YLIM: indices for selecting the region of interest (ROI). These are coordinates in the NMR spectra given as a point index number, in the same fashion as done in Bruker Topspin. See more details in the next section.
    • Optional Arguments for data fitting:
      • FITTING_FUNC: Provides the location files configuring the fitting function. Its format is described below.
      • NORMALMODE: selects the type of normalization, as in trendplot.

Arguments – Read from TRENDset GUI

There are two menus of TRENDset, spectra and fid (shown below). These are designed to analyze NMR data in the frequency or time domain, respectively. As described above and in Trendmain, preprocessing can be tailored using scaling, xunit, threshold, thresholdnumber, xlim, and ylim. If these arguments are not defined in the index file, they are read by the GUI of TRENDset.
trendset_fid

  • file text box reads the configuration file for curve fitting. It can be generated by running TRENDanalysis, which generates a fitting.config file in the results folder. Or it can be created by a text editor in the following format (case-sensitive):
Parameters: KD=0.05, y_end=1.0
Constants: Pt=0.05
Function: y_end * 1.0 / (2.0 * Pt) * (KD + x + Pt - sqrt((KD + x + Pt) ** 2
- 4.0 * x * Pt))

This is an example of configuration file used to fit 1:1 protein-ligand binding isotherm.
The line starting with Function is the function for the binding isotherm in this example. (Note that it is a function, not an equation): \frac{(K_D+x+Pt-\sqrt{(K_D+x+Pt)^2-4xPt)}}{2Pt} \times y_{\_end}. The parameters are defined in the manual for TRENDanalysis.
The line starts with Parameters defines the parameters to fit, while the values specified here with initial estimates. In this example, KD, y_end are parameters to fit, while their initial values are set as 0.05 and 1.0, respectivley. If initial values are not given, then it will be set as 1.0.
The line starting with Constants defines the constant values that are fixed in the fitting. Here Pt is fixed to 0.05. The Pt can also be defined in the Parameters line as a parameter instead of a constant.
The units of KD and PT must be identical to the units of x-axis, which is defined by the xaxis argument. For example, if the xaxis option is set to mM then the units of KD and PT are also mM.
Another example is a configuration file of Exponential function:

Exponential time
Parameters: A, C, T 
Constants: None
Function: C + A * exp(-x / T)

The line of Exponential time is a flag that turns on initialization of parameters defined in the Parameters line. Usually the auto parameter initialization works fine. Hence, there is no need to list initial values in the Parameters line. TREND also supports auto parameter initialization for exponential function rate constant: C + A * exp( -k * x). To turn it on just add a line of Exponential rate. The Constants: None line specifies the absence of constants in the fitting.

  • sort argument defines which parameter to be sorted in the batch of results. By default KD will be sorted. Note If sort is not set, or the parameters set by sort is not found in one or more fitting results, TRENDset will generate a table without sorting selected parameters.
  • order defines the order of sorting. By default sorted values will be outputed in the ascending order.
  • cutoff defines which results are color-coded as a “hit”. Values larger than cutoff in a descending order, or smaller than cutoff in an ascending order will be considered as “hits” and colored in green in the generated table.
  • title is the title of output report. TRENDset will create a result folder named title (with spaces removed if there are any). PCA results, preview of selected ROIs, and fitting results of each experiment will be saved to the subdirectories named as TITLE/EXP_NAME/.
  • XLIM defines the indices of x-axis points used in the PCA analysis, e.g.1-max means using all points in x-axis. 450-500 means using points from 450 to 500. The definition of “points” is identical to that used by index in Bruker Topspin.
  • YLIM defines indices of y-axis points used for PCA analysis, provided the input data has two dimensions.
  • ROI allows manual selection of ROI. When it is checked, it will turn off the XLIM and YLIM option. A preview of the ROI will pop up the first spectrum, allowing selection of the ROI manually by dragging a rectangle across the spectrum. See TREND NMR for more details.