Mass spectrometry imaging enables label-free, high-resolution spatial mapping of the chemical composition of complex, biological samples. Typical experiments require selecting ions and/or positions from the images: ions for fragmentation studies to identify keystone compounds and positions for follow up validation measurements using microdissection or other orthogonal techniques. Unfortunately, with modern imaging machines, these must be selected from an overwhelming amount of raw data. Existing techniques to reduce the volume of data, the most popular of which are principle component analysis and non-negative matrix factorization, have the disadvantage that they return difficult-to-interpret linear combinations of actual data elements. In this work, we show that CX and CUR matrix decompositions can be used directly to address this selection need. CX and CUR matrix decompositions use empirical statistical leverage scores of the input data to provide provably good low-rank approximations of the measured data that are expressed in terms of actual ions and actual positions, as opposed to difficult-to-interpret eigenions and eigenpositions. We show that this leads to effective prioritization of information for both ions and positions. In particular, important ions can be found either by using the leverage scores as a ranking function and using a deterministic greedy selection algorithm or by using the leverage scores as an importance sampling distribution and using a random sampling algorithm; however, selection of important positions from the original matrix performed significantly better when they were chosen with the random sampling algorithm. Also, we show that 20 ions or 40 locations can be used to reconstruct the original matrix to a tolerance of 17% error for a widely studied image of brain lipids; and we provide a scalable implementation of this method that is applicable for analysis of the raw data where there are often more than a million rows and/or columns, which is larger than SVD-based low-rank approximation methods can handle. These results introduce the concept of CX/CUR matrix factorizations to mass spectrometry imaging, describing their utility and illustrating principled algorithmic approaches to deal with the overwhelming amount of data generated by modern mass spectrometry imaging.

Mass spectrometry imaging (MSI) has primarily been applied in localizing biomolecules within biological matrices. Although well-suited, the application of MSI for comparing thousands of spatially defined spotted samples has been limited. One reason for this is a lack of suitable and accessible data processing tools for the analysis of large arrayed MSI sample sets. The OpenMSI Arrayed Analysis Toolkit (OMAAT) is a software package that addresses the challenges of analyzing spatially defined samples in MSI data sets. OMAAT is written in Python and is integrated with OpenMSI ( http://openmsi.nersc.gov ), a platform for storing, sharing, and analyzing MSI data. By using a web-based python notebook (Jupyter), OMAAT is accessible to anyone without programming experience yet allows experienced users to leverage all features. OMAAT was evaluated by analyzing an MSI data set of a high-throughput glycoside hydrolase activity screen comprising 384 samples arrayed onto a NIMS surface at a 450 μm spacing, decreasing analysis time >100-fold while maintaining robust spot-finding. The utility of OMAAT was demonstrated for screening metabolic activities of different sized soil particles, including hydrolysis of sugars, revealing a pattern of size dependent activities. These results introduce OMAAT as an effective toolkit for analyzing spatially defined samples in MSI. OMAAT runs on all major operating systems, and the source code can be obtained from the following GitHub repository: https://github.com/biorack/omaat .

Metabolomics is a widely used technology for obtaining direct measures of metabolic activities from diverse biological systems. However, ambiguous metabolite identifications are a common challenge and biochemical interpretation is often limited by incomplete and inaccurate genome-based predictions of enzyme activities (that is, gene annotations). Metabolite Annotation and Gene Integration (MAGI) generates a metabolite-gene association score using a biochemical reaction network. This is calculated by a method that emphasizes consensus between metabolites and genes via biochemical reactions. To demonstrate the potential of this method, we applied MAGI to integrate sequence data and metabolomics data collected from Streptomyces coelicolor A3(2), an extensively characterized bacterium that produces diverse secondary metabolites. Our findings suggest that coupling metabolomics and genomics data by scoring consensus between the two increases the quality of both metabolite identifications and gene annotations in this organism. MAGI also made biochemical predictions for poorly annotated genes that were consistent with the extensive literature on this important organism. This limited analysis suggests that using metabolomics data has the potential to improve annotations in sequenced organisms and also provides testable hypotheses for specific biochemical functions. MAGI is freely available for academic use both as an online tool at https://magi.nersc.gov and with source code available at https://github.com/biorack/magi .

An integrated omics approach using genomics, transcriptomics, metabolomics (MALDI mass spectrometry imaging, MSI), and bioinformatics was employed to study spatiotemporal formation and deposition of health-protecting polymeric lignans and plant defense cyanogenic glucosides. Intact flax (Linum usitatissimum) capsules and seed tissues at different development stages were analyzed. Transcriptome analyses indicated distinct expression patterns of dirigent protein (DP) gene family members encoding (-)- and (+)-pinoresinol-forming DPs and their associated downstream metabolic processes, respectively, with the former expressed at early seed coat development stages. Genes encoding (+)-pinoresinol-forming DPs were, in contrast, expressed at later development stages. Recombinant DP expression and DP assays also unequivocally established their distinct stereoselective biochemical functions. Using MALDI MSI and ion mobility separation analyses, the pinoresinol downstream derivatives, secoisolariciresinol diglucoside (SDG) and SDG hydroxymethylglutaryl ester, were localized and detectable only in early seed coat development stages. SDG derivatives were then converted into higher molecular weight phenolics during seed coat maturation. By contrast, the plant defense cyanogenic glucosides, the monoglucosides linamarin/lotaustralin, were detected throughout the flax capsule, whereas diglucosides linustatin/neolinustatin only accumulated in endosperm and embryo tissues. A putative biosynthetic pathway to the cyanogens is proposed on the basis of transcriptome coexpression data. Localization of all metabolites was at ca. 20 μm resolution, with the web based tool OpenMSI enabling not only resolution enhancement but also an interactive system for real-time searching for any ion in the tissue under analysis.