Compressive sensing of jointly sparse signals as a method for dimensionality reduction of mass spectrometry data

Mass spectrometry (MS) is a technique that is applied in chemical and biomedical applications for molecular analysis. MS data has extremely high dimensionality that can be represented as a three dimension(3D) dataset. In this paper, we exploit 3D data structure and propose an effective model of compressive sensing (CS) for dimensionality reduction of MS data. A large set of MS data can be reduced significantly with high quality of data recovery. Our recovery approach is based on clustering the MS data as joint sparse signals in a Multiple Measurement Vector (MMV) where the sparsity of the signal occurs in common locations. The MS data is considered during the acquisition process. The results showed that the MS data with high dimensionality could be recovered with no quality degradation from a low dimensional data set.