Evaluation of Digital Image Recognition Methods for Mass Spectrometry Imaging Data Analysis

Incorporation of Three Different Optical Trains into the IR-MALDESI Mass Spectrometry Imaging Platform to Characterize Artemisia annua

Artemisinin is the leading medication for the treatment of malaria and is only produced naturally in Artemisia annua. The localization of artemisinin in both the glandular and non-glandular trichomes of the plant makes it an ideal candidate for mass spectrometry imaging (MSI) as a model system for method development. Infrared matrix-assisted laser desorption electrospray ionization MSI (IR-MALDESI-MSI) has the capability to detect hundreds to thousands of analytes simultaneously, providing abundance information in conjunction with species localization throughout a sample. The development of several new optical trains and their application to the IR-MALDESI-MSI platform has improved data quality in previous proof-of-concept experiments but has not yet been applied to analysis of native biological samples, especially the MSI analysis of plants. This study aimed to develop a workflow and optimize MSI parameters, specifically the laser optical train, for the analysis of Artemisia annua with the NextGen IR-MALDESI platform coupled to an Orbitrap Exploris 240 mass spectrometer. Two laser optics were compared to the conventional set up, of which include a Schwarzschild-like reflective objective and a diffractive optical element (DOE). These optics, respectively, enhance the spatial resolution of imaging experiments or create a square spot shape for top-hat imaging. Ultimately, we incorporated and characterized three different optical trains into our analysis of Artemisia annua to study metabolites in the artemisinin pathway. These improvements in our workflow, resulted in high spatial resolution and improved ion abundance from previous work, which will allow us to address many different questions in plant biology beyond this model system.

DeepION: A Deep Learning-Based Low-Dimensional Representation Model of Ion Images for Mass Spectrometry Imaging

Mass spectrometry imaging (MSI) is a high-throughput imaging technique capable of the qualitative and quantitative in situ detection of thousands of ions in biological samples. Ion image representation is a technique that produces a low-dimensional vector embedded with significant spectral and spatial information on an ion image, which further facilitates the distance-based similarity measurement for the identification of colocalized ions. However, given the low signal-to-noise ratios inherent in MSI data coupled with the scarcity of annotated data sets, achieving an effective ion image representation for each ion image remains a challenge. In this study, we propose DeepION, a novel deep learning-based method designed specifically for ion image representation, which is applied to the identification of colocalized ions and isotope ions. In DeepION, contrastive learning is introduced to ensure that the model can generate the ion image representation in a self-supervised manner without manual annotation. Since data augmentation is a crucial step in contrastive learning, a unique data augmentation strategy is designed by considering the characteristics of MSI data, such as the Poisson distribution of ion abundance and a random pattern of missing values, to generate plentiful ion image pairs for DeepION model training. Experimental results of rat brain tissue MSI show that DeepION outperforms other methods for both colocalized ion and isotope ion identification, demonstrating the effectiveness of ion image representation. The proposed model could serve as a crucial tool in the biomarker discovery and drug development of the MSI technique.

Metabolite Annotation Confidence Score (MACS): A Novel MSI Identification Scoring Tool

Mass spectrometry imaging (MSI) is an analytical technique capable of measuring and visualizing the spatial distribution of thousands of ions across a sample. Measured ions can be putatively identified and annotated by comparing their mass-to-charge ratio (m/z) to a database of known compounds. For high-resolution, accurate mass (HRAM) imaging data sets, this is commonly performed by the annotation platform METASPACE. Annotations are reported with a metabolite-signal-match (MSM) score as a measure of the annotation's confidence level. However, the MSM scores reported by METASPACE often do not reflect a reasonable confidence level of an annotation and are not assigned consistently. The metabolite annotation confidence score (MACS) is an alternative scoring system based on fundamental mass spectrometry imaging metrics (mass measurement accuracy, spectral accuracy, and spatial distribution) to generate values that reflect the confidence of a specific annotation in HRAM-MSI data sets. Herein, the MACS system is characterized and compared to MSM scores from ions annotated by METASPACE.

The development and application of matrix assisted laser desorption electrospray ionization: The teenage years

In the past 15 years, ambient ionization techniques have witnessed a significant incursion into the field of mass spectrometry imaging, demonstrating their ability to provide complementary information to matrix-assisted laser desorption ionization. Matrix-assisted laser desorption electrospray ionization is one such technique that has evolved since its first demonstrations with ultraviolet lasers coupled to Fourier transform-ion cyclotron resonance mass spectrometers to extensive use with infrared lasers coupled to orbitrap-based mass spectrometers. Concurrently, there have been transformative developments of this imaging platform due to the high level of control the principal group has retained over the laser technology, data acquisition software (RastirX), instrument communication, and image processing software (MSiReader). This review will discuss the developments of MALDESI since its first laboratory demonstration in 2005 to the most recent advances in 2021.

Emerging Computational Methods in Mass Spectrometry Imaging

Mass spectrometry imaging (MSI) is a powerful analytical technique that generates maps of hundreds of molecules in biological samples with high sensitivity and molecular specificity. Advanced MSI platforms with capability of high-spatial resolution and high-throughput acquisition generate vast amount of data, which necessitates the development of computational tools for MSI data analysis. In addition, computation-driven MSI experiments have recently emerged as enabling technologies for further improving the MSI capabilities with little or no hardware modification. This review provides a critical summary of computational methods and resources developed for MSI data analysis and interpretation along with computational approaches for improving throughput and molecular coverage in MSI experiments. This review is focused on the recently developed artificial intelligence methods and provides an outlook for a future paradigm shift in MSI with transformative computational methods.

What are we imaging? Software tools and experimental strategies for annotation and identification of small molecules in mass spectrometry imaging

Mass spectrometry imaging (MSI) has become a widespread analytical technique to perform nonlabeled spatial molecular identification. The Achilles' heel of MSI is the annotation and identification of molecular species due to intrinsic limitations of the technique (lack of chromatographic separation and the difficulty to apply tandem MS). Successful strategies to perform annotation and identification combine extra analytical steps, like using orthogonal analytical techniques to identify compounds; with algorithms that integrate the spectral and spatial information. In this review, we discuss different experimental strategies and bioinformatics tools to annotate and identify compounds in MSI experiments. We target strategies and tools for small molecule applications, such as lipidomics and metabolomics. First, we explain how sample preparation and the acquisition process influences annotation and identification, from sample preservation to the use of orthogonal techniques. Then, we review twelve software tools for annotation and identification in MSI. Finally, we offer perspectives on two current needs of the MSI community: the adaptation of guidelines for communicating confidence levels in identifications; and the creation of a standard format to store and exchange annotations and identifications in MSI.

Antiretroviral drug exposure in lymph nodes is heterogeneous and drug dependent

Introduction

HIV reservoirs and infected cells may persist in tissues with low concentrations of antiretrovirals (ARVs). Traditional pharmacology methods cannot assess variability in ARV concentrations within morphologically complex tissues, such as lymph nodes (LNs). We evaluated the distribution of six ARVs into LNs and the proximity of these ARVs to CD4⁺ T cells and cell‐associated RT‐SHIV viral RNA.

Methods

Between December 2014 and April 2017, RT‐SHIV infected (SHIV+; N = 6) and healthy (SHIV–; N = 6) male rhesus macaques received two selected four‐drug combinations of six ARVs over 10 days to attain steady‐state conditions. Serial cryosections of axillary LN were analysed by a multimodal imaging approach that combined mass spectrometry imaging (MSI) for ARV disposition, RNAscope in situ hybridization for viral RNA (vRNA) and immunohistochemistry for CD4⁺ T cell and collagen expression. Spatial relationships across these four imaging domains were investigated by nearest neighbour search on co‐registered images using MATLAB.

Results

Through MSI, ARV‐dependent, heterogeneous concentrations were observed in different morphological LN regions, such as the follicles and medullary sinuses. After 5–6 weeks of infection, more limited ARV penetration into LN tissue relative to the blood marker heme was found in SHIV+ animals (SHIV+: 0.7 [0.2–1.4] mm; SHIV–: 1.3 [0.5–1.7] mm), suggesting alterations in the microcirculation. However, we found no detectable increase in collagen deposition. Regimen‐wide maps of composite ARV distribution indicated that up to 27% of SHIV+ LN tissue area was not exposed to detectable ARVs. Regions associated with B cell follicles had median 1.15 [0.94–2.69] ‐fold reduction in areas with measurable drug, though differences were only statistically significant for tenofovir (p = 0.03). Median co‐localization of drug with CD4⁺ target cells and vRNA varied widely by ARV (5.1–100%), but nearest neighbour analysis indicated that up to 10% of target cells and cell‐associated vRNA were not directly contiguous to at least one drug at concentrations greater than the IC50 value.

Conclusions

Our investigation of the spatial distributions of drug, virus and target cells underscores the influence of location and microenvironment within LN, where a small population of T cells may remain vulnerable to infection and low‐level viral replication during suppressive ART.

ADVANCES IN HIGH-RESOLUTION MALDI MASS SPECTROMETRY FOR NEUROBIOLOGY

Research in the field of neurobiology and neurochemistry has seen a rapid expansion in the last several years due to advances in technologies and instrumentation, facilitating the detection of biomolecules critical to the complex signaling of neurons. Part of this growth has been due to the development and implementation of high-resolution Fourier transform (FT) mass spectrometry (MS), as is offered by FT ion cyclotron resonance (FTICR) and Orbitrap mass analyzers, which improves the accuracy of measurements and helps resolve the complex biological mixtures often analyzed in the nervous system. The coupling of matrix-assisted laser desorption/ionization (MALDI) with high-resolution MS has drastically expanded the information that can be obtained with these complex samples. This review discusses notable technical developments in MALDI-FTICR and MALDI-Orbitrap platforms and their applications toward molecules in the nervous system, including sequence elucidation and profiling with de novo sequencing, analysis of post-translational modifications, in situ analysis, key advances in sample preparation and handling, quantitation, and imaging. Notable novel applications are also discussed to highlight key developments critical to advancing our understanding of neurobiology and providing insight into the exciting future of this field. © 2020 John Wiley & Sons Ltd. Mass Spec Rev.

MS imaging of multicellular tumor spheroids and organoids as an emerging tool for personalized medicine and drug discovery

MS imaging (MSI) is a powerful tool in drug discovery because of its ability to interrogate a wide range of endogenous and exogenous molecules in a broad variety of samples. The impressive versatility of the approach, where almost any ionizable biomolecule can be analyzed, including peptides, proteins, lipids, carbohydrates, and nucleic acids, has been applied to numerous types of complex biological samples. While originally demonstrated with harvested organs from animal models and biopsies from humans, these models are time consuming and expensive, which makes it necessary to extend the approach to 3D cell culture systems. These systems, which include spheroid models, prepared from immortalized cell lines, and organoid cultures, grown from patient biopsies, can provide insight on the intersection of molecular information on a spatial scale. In particular, the investigation of drug compounds, their metabolism, and the subsequent distribution of their metabolites in 3D cell culture systems by MSI has been a promising area of study. This review summarizes the different ionization methods, sample preparation steps, and data analysis methods of MSI and focuses on several of the latest applications of MALDI-MSI for drug studies in spheroids and organoids. Finally, the application of this approach in patient-derived organoids to evaluate personalized medicine options is discussed.

Investigations of β-carotene radical cation formation in infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI)

RATIONALE

Radical cationization of endogenous hydrocarbons in cherry tomatoes was previously reported using infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI), a mass spectrometry imaging technique that operates at ambient conditions and requires no sample derivatization. Due to the surprising nature of this odd-electron ionization, subsequent experiments were performed on β-carotene to determine the amount of radical cationization across different sampling conditions.

METHODS

β-Carotene was analyzed across a variety of sample states using IR-MALDESI followed by Orbitrap mass spectrometric analysis: first, as a standard in ethanol in a well plate; second, as particulates on printer paper; and third, as particulates covered by an ice matrix. These techniques were also performed with a β-carotene standard either in solution with a reducing agent (ascorbic acid) or with ascorbic acid in the electrospray solution.

RESULTS

Tandem mass spectrometry confirmed the presence of the radical cation of β-carotene by comparing fragments against NIST and METLIN databases. It was always analyzed as a radical cation when sampled from solution, where ascorbic acid increased radical cation abundance when in solution with β-carotene. Mixed-mode ionization between radical cationization and proton adduction was observed from dried particulates using IR-MALDESI.

CONCLUSIONS

There are several potential mechanisms for β-carotene radical cationization prior to IR-MALDESI analysis, with multiphoton ionization, thermal degradation, and/or reaction with oxygen appearing to be the most logical explanations. Furthermore, although not the primary cause, changing certain aspects of sample conditions can result in significant mixed-mode ionization with competing protonation.

Infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI) mass spectrometry imaging analysis of endogenous metabolites in cherry tomatoes

We report the spatially resolved metabolic profiling of cherry tomatoes using infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI), a mass spectrometry imaging (MSI) technique that operates at ambient conditions and requires no sample derivatization. Tomatoes were flash frozen, cryosectioned and imaged with adequate spatial resolution to distinguish between the major tissue structures of a tomato including the skin, mesocarp, endocarp, locular tissue, septum, placenta, seed and seed coating. Metabolites were imaged from 100-1200 m/z, enabling significant coverage of a diverse array of metabolites including amino acids and lipids along with the major secondary metabolite classes: terpenes, phenolics, glycosides, and alkaloids. During the metabolic profiling, we found endogenous carotenoid hydrocarbons, namely lycopene or its structural isomer β-carotene, ionized as radical cations. To our knowledge, this is the first demonstration of ionizing hydrocarbons in the MSI field.

Structure-Preserving and Perceptually Consistent Approach for Visualization of Mass Spectrometry Imaging Datasets

Mass spectrometry imaging (MSI) has become an important tool for 2D profiling of biological tissues, allowing for the visualization of individual compound distributions in the sample. Based on this information, it is possible to investigate the molecular organization within any particular tissue and detect abnormal regions (such as tumor regions) and many other biologically relevant phenomena. However, the large number of compounds present in the spectra hinders the productive analysis of large MSI datasets when utilizing standard tools. The heterogeneity of samples makes exploratory visualization (a presentation of the general idea of the molecular and structural organization of the inspected tissues) challenging. Here, we explore the application of various dimensionality reduction techniques that have been used extensively in the visualization of hyperspectral images and the MSI data specifically, such as principal component analysis, independent component analysis, non-negative matrix factorization, t-distributed stochastic neighbor embedding, and uniform manifold approximation and projection. Further, we propose a new approach based on a combination of structure preserving visualization with nonlinear manifold embedding of normalized spectral data. This way, we aim to preserve as much spatially overlapping signals as possible while augmenting them with information on compositional (spectral) variation. The proposed approach can be used for exploratory visualization of MSI datasets without prior deep chemical or histological knowledge of the sample. Thus, different datasets can be visually compared employing the proposed method. The proposed approach allowed for the clear visualization of the molecular layer, granular layer, and white matter in chimpanzee and macaque cerebellum slices.

Single-Cell Metabolomics by Mass Spectrometry Imaging

Multicellular organisms achieve their complex living activities through the highly organized metabolic interplay of individual cells and tissues. This complexity has driven the need to spatially resolve metabolomics down to the cellular and subcellular level. Recent technological advances have enabled mass spectrometry imaging (MSI), especially matrix-assisted laser desorption/ionization (MALDI), to become a powerful tool for the visualization of molecular species down to subcellular spatial resolution. In the present chapter, we summarize recent advances in the field of MALDI-MSI, with respect to single-cell level resolution metabolomics directly on tissue. In more detail, we focus on advancements in instrumentation for MSI at single-cell resolution, and the applications towards metabolomic scale imaging. Finally, we discuss new computational tools to aid in metabolite identification, future perspective, and the overall direction that the field of single-cell metabolomics directly on tissue may take in the years to come.

Lipidomic profiling of single mammalian cells by infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI)

To better understand cell-to-cell heterogeneity, advanced analytical tools are in a growing demand for elucidating chemical compositions of each cell within a population. However, the progress of single-cell chemical analysis has been restrained by the limitations of small cell volumes and minute cellular concentrations. Here, we present a rapid and sensitive method for investigating the lipid profiles of isolated single cells using infrared matrix-assisted laser desorption electrospray ionization mass spectrometry (IR-MALDESI-MS). In this work, HeLa cells were dispersed onto a glass slide, and the cellular contents were ionized by IR-MALDESI and measured using a Q-Exactive HF-X mass spectrometer. Importantly, this approach does not require extraction and/or enrichment of analytes prior to MS analysis. Using this approach, 45 distinct lipid species, predominantly phospholipids, were detected and putatively annotated from the single HeLa cells. The proof-of-concept study demonstrates the feasibility and efficacy of IR-MALDESI-MS for rapid lipidomic profiling of single cells, which provides an important basis for future work on differentiation between normal and diseased cells at various developmental states, which can offer new insights into cellular metabolic pathways and pathological processes. Although not yet accomplished, we believe this approach can be readily used as an assessment tool to compare the number of identified species during source evolution and method optimization (intra-laboratory), and also disclose the complementary nature of different direct analytical approaches for the coverage of different types of endogenous analytes (inter-laboratory).Graphical abstract.

ColocML: machine learning quantifies co-localization between mass spectrometry images

Motivation

Imaging mass spectrometry (imaging MS) is a prominent technique for capturing distributions of molecules in tissue sections. Various computational methods for imaging MS rely on quantifying spatial correlations between ion images, referred to as co-localization. However, no comprehensive evaluation of co-localization measures has ever been performed; this leads to arbitrary choices and hinders method development.

Results

We present ColocML, a machine learning approach addressing this gap. With the help of 42 imaging MS experts from nine laboratories, we created a gold standard of 2210 pairs of ion images ranked by their co-localization. We evaluated existing co-localization measures and developed novel measures using term frequency–inverse document frequency and deep neural networks. The semi-supervised deep learning Pi model and the cosine score applied after median thresholding performed the best (Spearman 0.797 and 0.794 with expert rankings, respectively). We illustrate these measures by inferring co-localization properties of 10 273 molecules from 3685 public METASPACE datasets.

Availability and implementation

https://github.com/metaspace2020/coloc.

Supplementary information

Supplementary data are available at Bioinformatics online.

Structural Similarity Image Analysis for Detection of Adenosine and Dopamine in Fast-Scan Cyclic Voltammetry Color Plots

Fast-scan cyclic voltammetry (FSCV) is widely used for in vivo detection of neurotransmitters, but identifying analytes, particularly mixtures, is difficult. Data analysis has focused on identifying dopamine from cyclic voltammograms, but it would be better to analyze all the data in the three-dimensional FSCV color plot. Here, the goal was to use image analysis-based analysis of FSCV color plots for the first time, specifically the structural similarity index (SSIM), to identify rapid neurochemical events. Initially, we focused on identifying spontaneous adenosine events, as adenosine cyclic voltammograms have a primary oxidation at 1.3 V and a secondary oxidation peak that grows in over time. Using SSIM, sample FSCV color plots were compared with reference color plots, and the SSIM cutoff score was optimized to distinguish adenosine. High-pass digital filtering was also applied to remove the background drift and lower the noise, which produced a better LOD. The SSIM algorithm detected more adenosine events than a previous algorithm based on current versus time traces, with 99.5 ± 0.6% precision, 95 ± 3% recall, and 97 ± 2% F₁ score (n = 15 experiments from three researchers). For selectivity, it successfully rejected signals from pH changes, histamine, and H₂O₂. To prove it is a broad strategy useful beyond adenosine, SSIM analysis was optimized for dopamine detection and is able to detect simultaneous events with dopamine and adenosine. Thus, SSIM is a general strategy for FSCV data analysis that uses three-dimensional data to detect multiple analytes in an efficient, automated analysis.

Spatial Metabolomics and Imaging Mass Spectrometry in the Age of Artificial Intelligence

Spatial metabolomics is an emerging field of omics research that has enabled localizing metabolites, lipids, and drugs in tissue sections, a feat considered impossible just two decades ago. Spatial metabolomics and its enabling technology-imaging mass spectrometry-generate big hyper-spectral imaging data that have motivated the development of tailored computational methods at the intersection of computational metabolomics and image analysis. Experimental and computational developments have recently opened doors to applications of spatial metabolomics in life sciences and biomedicine. At the same time, these advances have coincided with a rapid evolution in machine learning, deep learning, and artificial intelligence, which are transforming our everyday life and promise to revolutionize biology and healthcare. Here, we introduce spatial metabolomics through the eyes of a computational scientist, review the outstanding challenges, provide a look into the future, and discuss opportunities granted by the ongoing convergence of human and artificial intelligence.

Investigating host-pathogen meta-metabolic interactions of Magnaporthe oryzae infected barley using infrared matrix-assisted laser desorption electrospray ionization mass spectrometry

Infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI) mass spectrometry imaging is a useful tool for identifying important meta-metabolomic features pertinent for enhancing our understanding of biological systems. Magnaporthe oryzae (M. oryzae) is a filamentous fungus that is the primary cause of rice blast disease. True to its name, M. oryzae primarily destroys rice crops and can also destroy other cereal crops as well. In a previous study, the F-box E3 ligase protein in M. oryzae was noted to be crucial for its growth and pathogenicity. In this study, we inoculated three separate sets of barley with wild-type M. oryzae, an F-box E3 ligase protein knock out of M. oryzae, and a control solution. Over the course of the infection (8 days), we imaged each treatment after development of an advanced polarity switching method, which allowed for the detection of low and high molecular weight compounds that ionize in positive or negative polarities. A set of features from initial experiments were chosen for another analysis using tandem mass spectrometry. Serotonin, a barley defense metabolite, was a compound identified in both positive and negative modes. Serotonin was putatively identified using MS1 data including carbon estimation and sulfur counting then confirmed based on tandem mass spectrometry fragmentation patterns. Metabolites in the melanin pathway, important for infection development of M. oryzae, were also identified using MS1 data but were unable to be confirmed with MS/MS due to their low abundances.