Unsupervised learning for exploring MALDI imaging mass spectrometry ‘omics’ data

Matrix Assisted Laser Desorption Ionization-Imaging Mass Spectrometry (MALDI-IMS) is an emerging data acquisition technology in biological research. It has gained its popularity in `omics´ sciences because of its ability to explore the spatial distributions of various bio-molecules in detail. The sheer volume of data generated through this technology and the often limited a priori knowledge about the molecular compositions of biological samples, call for efficient data analysis methods. In this paper, first we review the available computational methods for analyzing the high-dimensional imaging datasets highlighting their advantages and limitations. Then, we propose a more recent unsupervised method as a means of exploring MALDI-IMS data and demonstrate its competency by extracting hidden significant spatial distribution patterns of a rat brain imaging dataset. Finally, we explain the potential future advances of `omics´ research associated with MALDI-IMS and the foreseeable challenges in analyzing the resultant data.