Some of the included features...
If you're working with local directories of data (instead of the dataChord Spectrum Server) you can scan the directories to locate any Bruker or Varian NMR data. Search patterns can be set up to extract any desired data out of text files or parameter files. The extracted data is then loaded into a dataChord table that can be used to control the selection and display of the processed data.
Screenshot
Spectra will often need to be adjusted for variation in frequency position and scaling. Any peak in the spectrum can be chosen and all spectra will be aligned. One or more regions can be selected and then the integral or median value of the region used to scale each spectrum relative to a selected spectrum. Spectra can also be explicitly scaled by values in the data table, for example, to scale by known urine volume.
Many operations may require specific regions to be identified in the spectra. Tools are available for adding, adjusting, splitting and clearing spectral regions.
An extensive library of mathematical and statistical procedures are included within dataChord Spectrum Miner. These libraries can be used to conveniently do various analyses on the spectral data. Methods included in our Graphical User Interface include Principle Component Analysis, sparse Non-Negative Matrix Factorization, and Covariance analysis. The results of these analysis, as well as simpler calculations such as integrals, are all available within interactive tables and graphs.
When a table of descriptive values are read in a selection GUI is automatically set up. The selectors can be used to interactively control which spectra will be displayed. One might for example, choose to display spectra collected on day four, from female animals dosed with the experimental compound.
Graphical selectors are automatically chosen from the column headers of input tabular data. Groupings may be used for statistical analysis and plotting.
Any items in the grouper that are turned on will be used as criteria to separate spectra into groups. All members of the chosen Spectral groups can be displayed with a common color and vertical offset so it is easy to visualize the differences between the different groups. Extracted data values (for example, integrals across a certain region) can be plotted with plot symbols and colors common to each group.
Any columns of values in the data table can be plotted in an XY plot. The XY plot is interactive, so that clicking on a plot symbol will result in the corresponding spectrum being drawn in the spectrum window. Any column in the table (typically the Grouper's group name) can be used to choose the color and symbol for the plotted items.
Box Plots are a technique of exploratory statistics that graphically render a distribution of data values. They are particularly useful for comparing the statistical distribution of different groups of data. Any quantitative data extracted from spectra or imported into the data table within dataChord can be displayed as a Box Plot. Groups of data are selected with the Grouper tool. Different groupings of data can be quickly made and the Box Plot updated.
Spectra of reference compounds can be imported in several different ways. In client-server mode reference spectra can be fetched from the server. Alternatively spectral data can be read from text files generated by dataChord or loaded from Bruker SBase files.
Reference spectra can be loaded into memory and aligned with the current experimental spectrum. Reference spectra are typically separated into signal containing regions. Each region can be independently aligned with the experimental data. Our automated alignment procedure will scan each signal region aover a specified range of points. At each position the baseline and scale is optimized to minimize the absolute value of the deviation of the reference and experimental data points. If a reference and experimental standard (DSS) signal are available and of known concentration then the absolute concentration of the compound can be calculated. Scanning can be done automatically across a whole set of spectra.
Tools are available to convert any experimental spectrum of a reference compound into the library format, including automated segmentation into signal containing regions.
Mathematical calculations can be done on values in the columns of the data table. For example, extracted integral values in one column could be scaled by urine volumes in another column to give molar quantities of a compound.
Simultaneous fitting of reference compounds can be done. A selected set of reference data is fit to a specific experimental spectrum using a Singular Value Decomposition technique. The fit can be repeated across all members of a set of experimental spectra.