Imaging mass spectrometry is a powerful technology enabling spatial metabolomics, yet metabolites can be assigned only to a fraction of the data generated. METASPACE-ML is a machine learning-based approach addressing this challenge which incorporates new scores and computationally-efficient False Discovery Rate estimation. For training and evaluation, we use a comprehensive set of 1,710 datasets from 159 researchers from 47 labs encompassing both animal and plant-based datasets representing multiple spatial metabolomics contexts derived from the METASPACE knowledge base. Here we show that, METASPACE-ML outperforms its rule-based predecessor, exhibiting higher precision, increased throughput, and enhanced capability in identifying low-intensity and biologically-relevant metabolites.

Summary: Imaging mass spectrometry (imaging MS) has advanced spatial and single-cell metabolomics, but the reliance on MS1 data complicates the accurate identification of molecular structures, not being able to resolve isomeric and isobar molecules. This prevents application of conventional methods for overrepresentation analysis (ORA) and metabolite set enrichment analysis (MSEA). To address this, we introduce S2IsoMEr R package and a web app for METASPACE, which uses bootstrapping to propagate isomeric/isobaric ambiguities into the enrichment analysis. We demonstrate S2IsoMEr for single-cell metabolomics and the METASPACE web app for spatial metabolomics. Availability and implementation: METASPACE web app can be used on existing and new datasets submitted to METASPACE (https://metaspace2020.org). The source code for the S2IsoMEr R package is available on GitHub (https://github.com/alexandrovteam/S2IsoMEr).

Understanding dormancy in prostate cancer is challenging because of model availability. In this study, using murine and human prostate cancer cell lines, we generated a stress-induced model of dormancy in vitro and demonstrated that the phenotype could be sustained upon intrailiac artery delivery into the bone marrow microenvironment. RNA sequencing analysis revealed that the transcription factor positive regulatory domain–containing 16 (PRDM16) was commonly upregulated in dormant prostate cancer cells compared with controls. Furthermore, bone marrow–disseminated prostate cancer cells from primary orthotopic tumors were largely positive for PRDM16. Genetic ablation and forced ectopic expression supported a role for PRDM16 in maintaining prostate cancer dormancy in vitro and in vivo. Clinically, PRDM16 negatively correlated with disease recurrence and with the E2F cell-cycle program in disseminated tumor cells derived from the bone marrow of patients with prostate cancer. Gene enrichment and characterization studies implicated PRDM16 as a regulator of metabolic and cell-cycle pathways. Chromatin immunoprecipitation–qPCR further revealed that PRDM16 binds upstream of the promoter of RB1, a potent repressor of E2F activity. Overall, this study developed a straightforward method for inducing cancer cell dormancy and applied this approach to find that PRDM16 governs an intrinsic dormancy program in prostate cancer. Significance: PRDM16 initiates a dormancy program in prostate cancer cells that is sustained in the bone marrow microenvironment, highlighting PRDM16 as a potential biomarker for relapse and target for eliminating dormant cancer cells.

Campylobacter spp. represents the most common cause of gastroenteritis worldwide with the potential to cause serious sequelae. The ability of Campylobacter to survive stressful environmental conditions has been directly linked with food-borne illness. Toxin-antitoxin (TA) modules play an important role as defense systems against antimicrobial agents and are considered an invaluable strategy harnessed by bacterial pathogens to survive in stressful environments. Although TA modules have been extensively studied in model organisms such as Escherichia coli K12, the TA landscape in Campylobacter remains largely unexplored. Therefore, in this study, a comprehensive in silico screen of 111 Campylobacter (90 C.jejuni and 21 C.coli) isolates recovered from different food and clinical sources was performed. We identified 10 type II TA systems belonging to four TA families predicted in Campylobacter genomes. Furthermore, there was a significant association between the clonal population structure and distribution of TA modules; more specifically, most (12/13) of the Campylobacter isolates belonging to ST-21 isolates possess HicB-HicA TA modules. Finally, we observed a high degree of shared synteny among isolates bearing certain TA systems or even coexisting pairs of TA systems. Collectively, these findings provide useful insights about the distribution of TA modules in a heterogeneous pool of Campylobacter isolates from different sources, thus developing a better understanding regarding the mechanisms by which these pathogens survive stressful environmental conditions, which will further aid in the future designing of more targeted antimicrobials.

Summary: Single-cell metabolomics promises to resolve metabolic cellular heterogeneity, yet current methods struggle with detecting small molecules, throughput, and reproducibility. Addressing these gaps, we developed HT SpaceM, a high-throughput single-cell metabolomics method with novel cell preparation, custom glass slides, small-molecule MALDI imaging mass spectrometry protocol, and batch processing. We propose a unified framework covering essential data analysis steps including quality control, characterization, differential analysis, structural validation and functional analysis. Interrogating human HeLa and mouse NIH3T3 cells, we detected 73 diverse small-molecule metabolites validated by bulk LC-MS/MS, achieving high reproducibility across wells and slides. Interrogating nine NCI-60 cancer cells and HeLa, we identified cell-type markers in small subpopulations. Functional analysis revealed overrepresented metabolic pathways, co-abundant metabolites, and metabolic hubs. We demonstrate the ability of SCM to analyze over 120,000 cells from over 112 samples, and provide guidance to interpret single-cell metabolic heterogeneity, revealing metabolic insights beyond population averages.

Imaging mass spectrometry (imaging MS) has advanced spatial and single-cell metabolomics, but the reliance on MS1 data complicates the accurate identification of molecular structures, not being able to resolve isomeric and isobar molecules. This prevents application of conventional methods for overrepresentation analysis (ORA) and metabolite set enrichment analysis (MSEA). To address this, we introduce S2IsoMEr R package and a web app for METASPACE, which uses bootstrapping to propagate isomeric/isobaric ambiguities into the enrichment analysis. We demonstrate S2IsoMEr for single-cell metabolomics and the METASPACE web app for spatial metabolomics.