v1.4.0

Installation note

• If you do not have MATLAB Runtime 2020a installed, then download SpectralAnalysis-1.4.0-win-installer-web.exe. This will automatically download and install the necessary MATLAB Runtime files and install SpectralAnalysis for you.
• If you already have MATLAB Runtime 2020a installed and don't want to install SpectralAnalysis, then you can download the standalone SpectralAnalysis-1.4.0-win.exe.

Features

• New peak thresholding algorithm. Median intensity is used as an estimate of the noise level, with thresholding applied at a user defined multiple of this level. Thanks @adamjtaylor
• Default preprocessing workflow is automatically selected based on the style of data loaded. This allows viewing spectra and generation of mean and basepeak spectra immediately, rather than requiring repetitive selection of preprocessing workflow The workflow can be changed in the same way as before.
• When performing Datacube Reduction and extracting a ROI, a new option has been included to reduce the image area down to only valid pixels in the ROI, or for all pixels to retain their original coordinates.

Interface changes

• Interface now includes icons for most buttons to make interaction more intuitive.
• x and y labels are now included on the spectrum display, based on the style of data loaded.
• Included peak list in the interface (in the same location as user selected images). If data is loaded in memory, then all channels will be displayed here.

API changes

• Parser now decides the default action for the interface when data is loaded. This allows quicker loading of data (especially for small datasets, where the data can be automatically loaded into memory and visualised). The old option of selecting which DataRepresentation is now included in the 'Tools' menu.
• MOOGL is now included within SpectralAnalysis and must no longer be downloaded separately.
• PeakDetection methods now produce arrays of instances of Peak.

Speed improvements

• InterpolationPPMRebinning is now significantly faster.

Bug fixes

• PCAInfo interface now works with all shapes of data, not just rectangular image regions.
• Interpolation rebinning occasionally produced NaN, resulting in shorter spectra than expected. The correct length spectra is now produced.

v1.3.0

Features

• More options for k-means clustering are now available (thanks @adamjtaylor)
• PImMS data is now supported (thanks @Turin508)
• Line profiles can now be visualised from ion images (thanks @Alex-Dexter)
• Mean spectra are automatically generated when visualising a dataset loaded in memory (e.g. when generating a datacube)
• Datacube reduction can now be performed with a fixed ppm (or m/z) window per peak

Interface changes

• DatacubeReduction option interface has been updated to include new functionality
• Removed empty interface when performing any PostProcessing method without additional options

API changes

• displayCallback can now be used in each Processing class to determine how the results are handled
• DataInMemory now includes MATLAB functions saveobj and loadobj for serialisation of whole object
• UsingFastMethods notification is now included in DatacubeReduction

Bug fixes

• When viewing multivariate analysis results with a diverging colour scheme, the limits were incorrectly applied such that the middle of the colour scheme was not 0.
• Multiple ROIs can now be generated from a single datacube at once
• Bruker datasets are now automatically converted to relative coordinates to avoid large empty image areas and out of memory errors

v1.2.1

Features

• Better region of interest (ROI) tools
  • Can invert an ROI
  • Can save ROIs to workspace in source code version
• Can save generated images as *.png files

Bug fixes

• Fixed deploying for R2019a
• Correct handling of ROIs when trying to perform MVA (such as PCA) on ROIs or data that has been loaded into memory

v1.2.0

Features

• Better region of interest (ROI) tools
  • Copy / clone ROI - new ROI created with same size and shape as selected ROI
  • Move ROI - freely move ROI to a new location
  • Load/save ROI lists (or individual ROIs)
• Image generation can now be performed on PPM of Da fixed values for the entire image list
• Improved interface for multivariate analysis (MVA)
• Latent plot visualised when performing PCA
• New rebinning methods for better handling of centroided data
• New peak filter for selecting decimal fraction of peak (for isolating molecular species)

Bug fixes

• Fixed export to imzML (now using latest version of jimzMLParser library)
• Can now save image lists

SpectalAnalysis v1.1.0

Features

• Exporting to imzML now has multiple options for data type and whether the data is 'processed' or 'continuous' for more flexibility over the resulting data size.
• Region of interest statistics interface.

Bug Fixes

• ROI overlay causing display of underlying data to darken (occasionally to the point of not being visible)
• With certain preprocessing parameters, mean/total spectrum and image generation produced no data to be displayed
• Rebinning in deployed version now works. Updated deploy script to compile all mex files.

SpectralAnalysis v1.0.1-alpha

Bug fixes:

• Spectral overlay not working
• Zooming when spectrum opened in new window (MATLAB graphics update induced bug)

Compatibility:

• Added strsplit function to enable use in MATLAB R2009a

SpectralAnalysis v1.0.0-alpha

Initial alpha version of SpectralAnalysis.