Software for automated unmixing of spectral micrographs

Spectral microscopy has the ability to generate a large volume of data in a short period of time. Unfortunately, commercial software has been limited in its ability to handle batch data. I developed a suite of MATLAB scripts called SpectralMagic to address these problems. SpectralMagic leverages the Bio-Formats library to gain access to microscope metadata. Using their unique metadata attributes, images are automagically processed with parallelized spectral unmixing algorithms. The batch functionality of the software allows users to set up an analysis job and get back to the wet bench.

I released SpectralMagic during the 2014 Physiology course at the Marine Biological Laboratory. The students used the software to analyze hundreds of gigabytes of spectral images, including the 100GB spectral z-stacks shown below. The image displays a focus-through of a spectral z-stack where each color represents a different bacterial taxon. As of this writing, SpectralMagic is still used for routine analysis by the Borisy Lab at The Forsyth Institute and the Phillips Lab at Caltech.

Spectral z-stack processed using SpectralMagic. Image credits: B. Weinstein, L. Mohapatra and M. Gralka